Disease Control Priorities in Developing Countries SECOND EDITION Editors Dean T. Jamison Joel G. Breman Anthony R. Measham George Alleyne Mariam Claeson David B. Evans Prabhat Jha Anne Mills Philip Musgrove Disease Control Priorities Project Disease Control Priorities in Developing Countries SECOND EDITION Disease Control Priorities in Developing Countries SECOND EDITION Editors Dean T. Jamison Joel G. Breman Anthony R. Measham George Alleyne Mariam Claeson David B. Evans Prabhat Jha Anne Mills Philip Musgrove A copublication of Oxford University Press and The World Bank ©2006 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 09 08 07 06 A copublication of The World Bank and Oxford University Press. Oxford University Press 165 Madison Avenue New York NY 10016 This volume was funded in part by a grant from the Bill & Melinda Gates Foundation and is a product of the staff of the International Bank for Reconstruction and Development / the World Bank, the World Health Organization, and the Fogarty International Center of the National Institutes of Health. The findings, inter- pretations, and conclusions expressed in this volume do not necessarily reflect the views of the executive direc- tors of the World Bank or the governments they represent, the World Health Organization, or the Fogarty International Center of the National Institutes of Health. The World Bank, the World Health Organization, and the Fogarty International Center of the National Institutes of Health do not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of the World Bank, the World Health Organization, or the Fogarty International Center of the National Institutes of Health concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete infor- mation to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN-10: 0-8213-0821361791 ISBN-13: 978-0-821-36179-5 eISBN: 0-8213-6180-5 Library of Congress Cataloguing-in-Publication Data has been applied for. Dedication This book is dedicated to Bill and Melinda Gates, whose vision, leadership, and financing over the past decade have catalyzed global support for transforming the lives of the world's poor through inexpensive but powerful health interventions. Contents Foreword by Jaime Sepúlveda xiii Preface xvii Editors xix Advisory Committee to the Editors xxiii Contributors xxv Disease Control Priorities Project Partners xxxv Acknowledgments xxxvii Abbreviations and Acronyms xxxix Part One Summary and Cross-Cutting Themes 1 Summary Chapter 1 Investing in Health 3 Dean T. Jamison Chapter 2 Intervention Cost-Effectiveness: Overview of Main Messages 35 Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles Chapter 3 Strengthening Health Systems 87 Anne Mills, Fawzia Rasheed, and Stephen Tollman Cross-Cutting Themes Chapter 4 Priorities for Global Research and Development of Interventions 103 Barry R. Bloom, Catherine M. Michaud, John R. La Montagne, and Lone Simonsen Chapter 5 Science and Technology for Disease Control: Past, Present, and Future 119 David Weatherall, Brian Greenwood, Heng Leng Chee, and Prawase Wasi Chapter 6 Product Development Priorities 139 Adel Mahmoud, Patricia M. Danzon, John H. Barton, and Roy D. Mugerwa Chapter 7 Economic Approaches to Valuing Global Health Research 157 David Meltzer Chapter 8 Improving the Health of Populations: Lessons of Experience 165 Carol Ann Medlin, Mushtaque Chowdhury, Dean T. Jamison, and Anthony R. Measham vii Chapter 9 Millennium Development Goals for Health: What Will It Take to Accelerate Progress? 181 Adam Wagstaff, Mariam Claeson, Robert M. Hecht, Pablo Gottret, and Qiu Fang Chapter 10 Gender Differentials in Health 195 Mayra Buvini´c, André Médici, Elisa Fernández, and Ana Cristina Torres Chapter 11 Fiscal Policies for Health Promotion and Disease Prevention 211 Rachel Nugent and Felicia Knaul Chapter 12 Financing Health Systems in the 21st Century 225 George Schieber, Cristian Baeza, Daniel Kress, and Margaret Maier Chapter 13 Recent Trends and Innovations in Development Assistance for Health 243 Robert M. Hecht and Raj Shah Chapter 14 Ethical Issues in Resource Allocation, Research, and New Product Development 259 Dan W. Brock and Daniel Wikler Chapter 15 Cost-Effectiveness Analysis for Priority Setting 271 Philip Musgrove and Julia Fox-Rushby Part Two Selecting Interventions 287 Infectious Disease, Reproductive Health, and Undernutrition Chapter 16 Tuberculosis 289 Christopher Dye and Katherine Floyd Chapter 17 Sexually Transmitted Infections 311 Sevgi O. Aral and Mead Over, with Lisa Manhart and King K. Holmes Chapter 18 HIV/AIDS Prevention and Treatment 331 Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, Lisa M. DeMaria, Becca Feldman, Helene Gayle, Julian Gold, Robert Grant, and Michael T. Isbell Chapter 19 Diarrheal Diseases 371 Gerald T. Keusch, Olivier Fontaine, Alok Bhargava, Cynthia Boschi-Pinto, Zulfiqar A. Bhutta, Eduardo Gotuzzo, Juan A. Rivera, Jeffrey Chow, Sonbol A. Shahid-Salles, and Ramanan Laxminarayan Chapter 20 Vaccine-Preventable Diseases 389 Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, Mark Miller, and Neal A. Halsey Chapter 21 Conquering Malaria 413 Joel G. Breman, Anne Mills, Robert W. Snow, Jo-Ann Mulligan, Christian Lengeler, Kamini Mendis, Brian Sharp, Chantal Morel, Paola Marchesini, Nicholas J. White, Richard W. Steketee, and Ogobara K. Doumbo Chapter 22 Tropical Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Onchocerciasis, and Leprosy 433 Jan H. F. Remme, Piet Feenstra, P. R. Lever, André Médici, Carlos Morel, Mounkaila Noma, K. D. Ramaiah, Frank Richards, A. Seketeli, Gabriel Schmunis, W. H. van Brakel, and Anna Vassall Chapter 23 Tropical Diseases Lacking Adequate Control Measures: Dengue, Leishmaniasis, and African Trypanosomiasis 451 Pierre Cattand, Phillippe Desjeux, M. G. Guzmán, Jean Jannin, A. Kroeger, André Médici, Philip Musgrove, Mike B. Nathan, Alexandra Shaw, and C. J. Schofield viii | Contents Chapter 24 Helminth Infections: Soil-Transmitted Helminth Infections and Schistosomiasis 467 Peter J. Hotez, Donald A. P. Bundy, Kathleen Beegle, Simon Brooker, Lesley Drake, Nilanthi de Silva, Antonio Montresor, Dirk Engels, Matthew Jukes, Lester Chitsulo, Jeffrey Chow, Ramanan Laxminarayan, Catherine M. Michaud, Jeff Bethony, Rodrigo Correa-Oliveira, Xiao Shu-Hua, Alan Fenwick, and Lorenzo Savioli Chapter 25 Acute Respiratory Infections in Children 483 Eric A. F. Simoes, Thomas Cherian, Jeffrey Chow, Sonbol A. Shahid-Salles, Ramanan Laxminarayan, and T. Jacob John Chapter 26 Maternal and Perinatal Conditions 499 Wendy J. Graham, John Cairns, Sohinee Bhattacharya, Colin H. W. Bullough, Zahidul Quayyum, and Khama Rogo Chapter 27 Newborn Survival 531 Joy E. Lawn, Jelka Zupan, Geneviève Begkoyian, and Rudolf Knippenberg Chapter 28 Stunting, Wasting, and Micronutrient Deficiency Disorders 551 Laura E. Caulfield, Stephanie A. Richard, Juan A. Rivera, Philip Musgrove, and Robert E. Black Chapter 29 Health Service Interventions for Cancer Control in Developing Countries 569 Martin L. Brown, Sue J. Goldie, Gerrit Draisma, Joe Harford, and Joseph Lipscomb Noncommunicable Disease and Injury Chapter 30 Diabetes: The Pandemic and Potential Solutions 591 K. M. Venkat Narayan, Ping Zhang, Alka M. Kanaya, Desmond E. Williams, Michael M. Engelgau, Giuseppina Imperatore, and Ambady Ramachandran Chapter 31 Mental Disorders 605 Steven Hyman, Dan Chisholm, Ronald Kessler, Vikram Patel, and Harvey Whiteford Chapter 32 Neurological Disorders 627 Vijay Chandra, Rajesh Pandav, Ramanan Laxminarayan, Caroline Tanner, Bala Manyam, Sadanand Rajkumar, Donald Silberberg, Carol Brayne, Jeffrey Chow, Susan Herman, Fleur Hourihan, Scott Kasner, Luis Morillo, Adesola Ogunniyi, William Theodore, and Zhen-Xin Zhang Chapter 33 Cardiovascular Disease 645 Thomas A. Gaziano, K. Srinath Reddy, Fred Paccaud, Susan Horton, and Vivek Chaturvedi Chapter 34 Inherited Disorders of Hemoglobin 663 David Weatherall, Olu Akinyanju, Suthat Fucharoen, Nancy Olivieri, and Philip Musgrove Chapter 35 Respiratory Diseases of Adults 681 Frank E. Speizer, Susan Horton, Jane Batt, and Arthur S. Slutsky Chapter 36 Diseases of the Kidney and the Urinary System 695 John Dirks, Giuseppe Remuzzi, Susan Horton, Arrigo Schieppati, and S. Adibul Hasan Rizvi Chapter 37 Skin Diseases 707 Roderick Hay, Sandra E. Bendeck, Suephy Chen, Roberto Estrada, Anne Haddix, Tonya McLeod, and Antoine Mahé Chapter 38 Oral and Craniofacial Diseases and Disorders 723 Douglas Bratthall, Poul Erik Petersen, Jayanthi Ramanathan Stjernswärd, and L. Jackson Brown Chapter 39 Unintentional Injuries 737 Robyn Norton, Adnan A. Hyder, David Bishai, and Margie Peden Contents | ix Chapter 40 Interpersonal Violence 755 Mark L. Rosenberg, Alexander Butchart, James Mercy, Vasant Narasimhan, Hugh Waters, and Maureen S. Marshall Risk Factors Chapter 41 Water Supply, Sanitation, and Hygiene Promotion 771 Sandy Cairncross and Vivian Valdmanis Chapter 42 Indoor Air Pollution 793 Nigel Bruce, Eva Rehfuess, Sumi Mehta, Guy Hutton, and Kirk Smith Chapter 43 Air and Water Pollution: Burden and Strategies for Control 817 Tord Kjellström, Madhumita Lodh, Tony McMichael, Geetha Ranmuthugala, Rupendra Shrestha, and Sally Kingsland Chapter 44 Prevention of Chronic Disease by Means of Diet and Lifestyle Changes 833 Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, Courtenay Dusenbury, Pekka Puska, and Thomas A. Gaziano Chapter 45 The Growing Burden of Risk from High Blood Pressure, Cholesterol, and Bodyweight 851 Anthony Rodgers, Carlene M. M. Lawes, Thomas A. Gaziano, and Theo Vos Chapter 46 Tobacco Addiction 869 Prabhat Jha, Frank J. Chaloupka, James Moore, Vendhan Gajalakshmi, Prakash C. Gupta, Richard Peck, Samira Asma, and Witold Zatonski Chapter 47 Alcohol 887 Jürgen Rehm, Dan Chisholm, Robin Room, and Alan D. Lopez Chapter 48 Illicit Opiate Abuse 907 Wayne Hall, Chris Doran, Louisa Degenhardt, and Donald Shepard Consequences of Disease and Injury Chapter 49 Learning and Developmental Disabilities 933 Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, Karin B. Nelson, Geoffrey C. Solarsh, Nicole Bellows, Richard M. Scheffler, and Karen J. Hofman Chapter 50 Loss of Vision and Hearing 953 Joseph Cook, Kevin D. Frick, Rob Baltussen, Serge Resnikoff, Andrew Smith, Jeffrey Mecaskey, and Peter Kilima Chapter 51 Cost-Effectiveness of Interventions for Musculoskeletal Conditions 963 Luke B. Connelly, Anthony Woolf, and Peter Brooks Chapter 52 Pain Control for People with Cancer and AIDS 981 Kathleen M. Foley, Judith L. Wagner, David E. Joranson, and Hellen Gelband PartThree Strengthening Health Systems 995 Strengthening Public Health Services Chapter 53 Public Health Surveillance: A Tool for Targeting and Monitoring Interventions 997 Peter Nsubuga, Mark E. White, Stephen B. Thacker, Mark A. Anderson, Stephen B. Blount, Claire V. Broome, Tom M. Chiller, Victoria Espitia, Rubina Imtiaz, Dan Sosin, Donna F. Stroup, Robert V. Tauxe, Maya Vijayaraghavan, and Murray Trostle x | Contents Chapter 54 Information to Improve Decision Making for Health 1017 Sally K. Stansfield, Julia Walsh, Ndola Prata, and Timothy Evans Chapter 55 Drug Resistance 1031 Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, Philip Jenkins, Thomas O'Brien, Iruka N. Okeke, Ariel Pablo-Mendez, and Keith P. Klugman Chapter 56 Community Health and Nutrition Programs 1053 John B. Mason, David Sanders, Philip Musgrove, Soekirman, and Rae Galloway Chapter 57 Contraception 1075 Ruth Levine, Ana Langer, Nancy Birdsall, Gaverick Matheny, Merrick Wright, and Angela Bayer Chapter 58 School-Based Health and Nutrition Programs 1091 Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, Kathleen Beegle, Amaya Gillespie, Lesley Drake, Seung-hee Frances Lee, Anna-Maria Hoffman, Jack Jones, Arlene Mitchell, Delia Barcelona, Balla Camara, Chuck Golmar, Lorenzo Savioli, Malick Sembene, Tsutomu Takeuchi, and Cream Wright Chapter 59 Adolescent Health Programs 1109 Elizabeth Lule, James E. Rosen, Susheela Singh, James C. Knowles, and Jere R. Behrman Chapter 60 Occupational Health 1127 Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut Chapter 61 Natural Disaster Mitigation and Relief 1147 Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio Chapter 62 Control and Eradication 1163 Mark Miller, Scott Barrett, and D. A. Henderson Strengthening Personal Health Services Chapter 63 Integrated Management of the Sick Child 1177 Cesar G. Victora, Taghreed Adam, Jennifer Bryce, and David B. Evans Chapter 64 General Primary Care 1193 Stephen Tollman, Jane Doherty, and Jo-Ann Mulligan Chapter 65 The District Hospital 1211 Mike English, Claudio F. Lanata, Isaac Ngugi, and Peter C. Smith Chapter 66 Referral Hospitals 1229 Martin Hensher, Max Price, and Sarah Adomakoh Chapter 67 Surgery 1245 Haile T. Debas, Richard Gosselin, Colin McCord, and Amardeep Thind Chapter 68 Emergency Medical Services 1261 Olive C. Kobusingye, Adnan A. Hyder, David Bishai, Manjul Joshipura, Eduardo Romero Hicks, and Charles Mock Chapter 69 Complementary and Alternative Medicine 1281 Haile T. Debas, Ramanan Laxminarayan, and Stephen E. Straus Capacity Strengthening and Management Reform Chapter 70 Improving the Quality of Care in Developing Countries 1293 John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and Julio Frenk Chapter 71 Health Workers: Building and Motivating the Workforce 1309 Charles Hongoro and Charles Normand Contents | xi Chapter 72 Ensuring Supplies of Appropriate Drugs and Vaccines 1323 Susan Foster, Richard Laing, Bjørn Melgaard, and Michel Zaffran Chapter 73 Strategic Management of Clinical Services 1339 Alexander S. Preker, Martin McKee, Andrew Mitchell, and Suwit Wilbulpolprasert Glossary 1353 Index 1357 Credits 1401 xii | Contents Foreword The 1993 publication of the now classic book, Disease Control particular circumstances in each country, as well as the indi- Priorities in Developing Countries, by Oxford University Press vidual institutional capacities to deliver goods and services and and of its companion document, the World Development Report to implement policies and processes. 1993: Investing in Health, published by the World Bank that Context-specific strategies and responses are essential, same year, constitute a landmark in the public health literature. because application of the Disease Control Priorities Project's For the first time, decision makers and public health practi- findings will vary according to each country's circumstances: tioners had a comprehensive review of the cost-effectiveness of one size does not fit all. Understanding that most health inter- available interventions to address the most common health ventions require a minimum level of institutional capacity to problems in the developing world. They were also provided deliver goods and services is equally important, and such capac- with the useful metric known as disability-adjusted life years ity may have to be built up before money or physical inputs can to calculate the burden of disease and the cost-effectiveness of yield any benefits. Accordingly, goals and priorities should be interventions more accurately than in the past. established and tailored to each country's context. As was the case with the first edition, this second edition of Disease Control Priorities in Developing Countries will serve an array of audiences. One primary audience consists of people TRANSITION IN HEALTH working in the health sector, ranging from those who are responsible for making evidence-based decisions to those who Every developing region is facing a transition in its epidemio- practice medicine and public health under often suboptimal logical profile from an environment with high fertility rates field conditions. A second audience consists of people working and high mortality from preventable causes to one in which a in finance and planning ministries, who will benefit from the combination of lower fertility rates and changing lifestyles has solid recommendations for improving the health of popula- led to aging populations and epidemics of tobacco addiction, tions through sound resource reallocation and cost-effective obesity, cardiovascular disease, cancers, diabetes, and other practices. chronic ailments. The 20th century will be remembered for, among other things, witnessing the largest universal increase in life expectancy in history. While life expectancy is PURPOSE highest in the richest countries, the upward trend is apparent in almost every society. Moreover, in the past 50 years, variations The purpose of this book is to provide information about what in this health indicator across and within countries have works--specifically, the cost-effectiveness of health interven- decreased. This convergence of improved life expectancy and tions in a variety of settings. Such information should influence reduced variations, which has occurred even in the presence of the redesign of programs and the reallocation of resources, widening income gaps in many regions, can be explained solely thereby helping to achieve the ultimate goal of reducing mor- by the impact of knowledge expansion and direct public health bidity and mortality. interventions. The increase in life expectancy worldwide will, however, FUNDAMENTAL POLICY CONSIDERATIONS soon reach a plateau, and a retraction has occurred in many countries. HIV/AIDS and civil unrest in Africa, vaccine- Although economic and budgetary constraints are clearly preventable diseases and alcoholism in Eastern Europe, and important considerations, money is not the only limitation. obesity in the United States have reduced--or will soon do Additional factors fundamental to improving outcomes are the so--the years of life their populations can expect. xiii SCALING UP EFFECTIVE INTERVENTIONS Ideally, this should become a tide that lifts all boats to yield global benefits. The challenge is to harness the information The late Jim Grant, former executive director of the United technology revolution to foster the growth of economies. One Nations Children's Fund, was one of the first leaders with a step in the right direction is the open access movement, which vision for setting specific health goals and priorities within a promotes and permits free and immediate access to research time frame and on a global scale. He recognized the need to results and other components of knowledge transfer. raise awareness of the dramatic disparities in children's health and to mobilize political will accordingly. His missionary zeal SPENDING MORE AND SPENDING BETTER for universal child immunization and for organizing the first summit of world leaders for children's health and rights in 1990 It is indeed a paradox to observe that even though the money permitted the scaling up of interventions of proven efficacy. spent on health worldwide has reached 10 percent of overall The Millennium Development Goals are a natural consequence global income, that amount is both insufficient and poorly of that vision and an extremely useful instrument for main- allocated. The World Health Organization's Commission on taining both focus and social pressure. Achieving these ambi- Macroeconomics and Health and several other global initia- tious goals will require not only the universal implementation tives make a persuasive plea for a larger investment in health. At of effective interventions that are currently available, but also the same time, this book is dedicated to making the case for the development of new interventions. better spending--that is, deriving more health benefits from every dollar spent. The aim should be to reduce inequalities in NEED FOR ONGOING RESEARCH health investment between and within countries: a 100-fold difference between the rich and the poor in money spent on Today, most vaccines, medical devices, diagnostic tools, and health services still persists in many places. Despite a lack of drugs have been subjected to careful investigation in the labo- clarity about what constitutes the optimum balance of health ratory, at the bedside, and in the field. However, not enough spending, a larger share should go to prevention. This book investment has gone into research to increase well-being and looks at several prevention options and clinical interventions development globally. We need more epidemiological and that are not being fully implemented. health systems research to improve the efficiency of available interventions, technological research to reduce their costs, and SELECTING INTERVENTIONS biomedical research to develop new tools for dealing with as yet unsolved and emerging health problems. This book persuasively makes the case that both clinical and public health interventions depend on the capacity of a given OPPORTUNITIES AND CHALLENGES country's health system to deliver, noting that some interven- OF GLOBALIZATION tions are more demanding than others in terms of infrastruc- ture and human resources. Therefore, both the costs and the One of the greatest opportunities and challenges for interna- likelihood of success of the more complex interventions are a tional public health is globalization. We live in an era when the function of the health capacity in place. In addition, decisions explosion of trade, travel, and communications is spreading about which interventions should be given priority will new cultural influences and lifestyles faster than ever before, depend on assessments of the local burden of disease, local and the division between domestic and international health health infrastructure, and other social factors as well as on problems is becoming increasingly obsolete. At the same time, cost-effectiveness analyses. The following chapters identify the globalization also permits the spread of risks, pathogens, and health system capacity needed for scaling up a given interven- other threats. The ever-increasing movement of people every- tion. Even middle-income countries with relatively better where increases the potential for epidemics. Travelers, refugees, health infrastructure often pursue sophisticated approaches to and displaced people are more vulnerable to infectious dis- medical care that result in fewer health gains per amount of eases, and their movement contributes to spreading pathogens money invested. Every country, regardless of level of develop- into new areas. Overall, however, the positive consequences ment, could benefit from the recommendations presented outweigh the negative ones, and cautious optimism about this here. irreversible trend is justified. Certainly, one of the most valu- able contributions of globalization is the rapid accrual and DIAGONAL APPROACH spread of knowledge about useful tools for controlling disease and ways to implement those tools on a large scale. The medical literature has long debated which approach to In recent years, the huge advances in information techno- delivering health interventions is more effective: vertical pro- logy have greatly boosted the globalization of knowledge. grams or horizontal programs. Vertical programs refer to xiv | Foreword focused, proactive, disease-specific interventions on a massive anything but the most efficient methods for organizing and scale, whereas horizontal programs refer to more integrated, implementing health care. This book is a fundamental compo- demand-driven, resource-sharing health services. This is a false nent for fostering equitable outcomes in health and develop- dilemma, because both need to coexist in what could be called ment. It will inspire all those who seek the highly complex but a diagonal approach--that is, the proactive, supply-driven pro- attainable goal of universal good health for all members of the vision of a set of highly cost-effective interventions on a large global community. scale that bridges health clinics and homes. This approach often starts vertically (polio vaccination, for instance) but FACILITATING PROGRESS moves toward an increasing number of interventions (for example, oral rehydration, other vaccines, residual spraying We all share global responsibility: governments and interna- and bednets for malaria control, micronutrient supplementa- tional agencies, public and private sectors, and society and tion, and supervised tuberculosis treatment), making full use individuals all have specific tasks. We must all strive toward of field health workers and existing infrastructure. This could more equitable distribution of the benefits of new knowledge well be the equivalent of a public health polypill. to reduce health and development gaps between rich and poor, between countries, and within countries. The second edition of MULTIDISCIPLINARY ORIENTATION Disease Control Priorities in Developing Countries is a new step in precisely the right direction. If we succeed in conveying the What makes this book unique, in addition to its comprehensive main lessons and messages of this book, public health in devel- scope, is its truly multidisciplinary approach to disease control, oping countries will progress farther and faster. which merges the best of the medical and economic sciences. Every recommendation has been carefully researched and doc- Jaime Sepúlveda, Director, National Institutes of Health of umented. Evidence-based approaches must be the foundation Mexico, Mexico City, Mexico for allocating scarce resources. The poor cannot afford Chair, Advisory Committee to the Editors Foreword | xv Preface In the late 1980s, the World Bank initiated a review of priorities World Bank 1993). Closely related efforts in collaboration with for the control of specific diseases and used this information as the World Health Organization led to the first global and input for comparative cost-effectiveness estimates of interven- regional estimates of numbers of deaths by age, sex, and cause tions addressing most conditions important in developing and of the burden (including the disability burden) from more countries. The purpose of the comparative cost-effectiveness than 100 specific diseases and conditions (Murray, Lopez, and work was to inform decision making within the health sectors Jamison 1994; World Bank 1993). of highly resource-constrained low- and middle-income coun- This second edition of Disease Control Priorities in tries. This process resulted in the 1993 publication of the first Developing Countries (DCP2) seeks to update and improve edition of Disease Control Priorities in Developing Countries guidance on the "what to do" questions in DCP1 and to address (DCP1) (Jamison and others 1993). That volume's preface the institutional, organizational, financial, and research capac- stated its purpose as follows: ities essential for health systems to deliver the right interven- tions. DCP2 is the principal product of the Disease Control Between 1950 and 1990, life expectancy in developing Priorities Project, an alliance of organizations designed to countries increased from forty to sixty-three years with a review, generate, and disseminate information on how to concomitant rise in the incidence of the noncommuni- improve population health in developing countries. In addition cable diseases of adults and the elderly. Yet there remains to DCP2, the project produced numerous background papers, a huge unfinished agenda for dealing with undernutri- an extensive range of interactive consultations held around the tion and the communicable childhood diseases. These world, and several additional major publications. The other trends lead to increasingly diverse and complicated epi- major publications are as follows: demiological profiles in developing countries. At the same time, new epidemic diseases like AIDS are emerg- · Global Burden of Disease and Risk Factors (Lopez and others ing; and the health of the poor during economic crisis is 2006), undertaken in collaboration with the World Health a source of growing concern. These developments have Organization intensified the need for better information on the effec- · Millions Saved: Proven Successes in Global Health (Levine tiveness and cost of health interventions. To assist coun- and the What Works Working Group 2004), undertaken in tries to define essential health service packages, this book collaboration with the Center for Global Development provides information on disease control interventions · "The Intolerable Burden of Malaria: II. What's New, What's for the commonest diseases and injuries in developing Needed" (Breman, Alilio, and Mills 2004), undertaken in countries. collaboration with the Multilateral Initiative on Malaria To this end, DCP1 aimed to provide systematic guidance on the · Priorities in Health (Jamison and others 2006), a brief and selection of interventions to achieve rapid health improve- nontechnical companion to this volume. ments in an environment of highly constrained public sector budgets through the use of cost-effectiveness analysis. Each product of the Disease Control Priorities Project marries DCP1 provided limited discussion of investments in health economic approaches with those of epidemiology, public system development. Other major efforts undertaken at the health, and clinical medicine. World Bank at about the same time, including the World While general lessons emerge from the Disease Control Development Report 1993: Investing in Health, used the findings Priorities Project, they result from careful consideration of of DCP1 and dealt more explicitly with the financial and health individual cases. The diversity of health conditions necessitates systems aspects of implementation (Feachem and others 1992; specificity of analysis. Arrow clearly stated the need for xvii technical analyses to underpin health economics: "Another development priorities, but also applies to control priorities. lesson of medical economics is the importance of recognizing In this regard, DCP2 continues in the spirit of DCP1 in the specific character of the disease under consideration. The assessing cost-effectiveness analyses of major changes, but it policy challenges that arise in treating malaria are simply very does so more systematically for each of the six regional different from those attached to other major infectious groupings of low- and middle-income countries used scourges (Arrow,Panosian,and Gelband 2004, xi­xii)." Chapters throughout this volume (see map 1, inside the front cover). in this volume address this need for specificity, yet use cost- effectiveness analysis in a way that makes findings on the What was becoming clear in 1990 is clearer today: focusing relative attractiveness of interventions comparable. health system attention on delivering efficacious and often rel- DCP2 goes beyond DCP1 in a number of important ways as atively inexpensive health interventions can lead to dramatic follows: reductions in mortality and disability at modest cost. A valu- able dimension of globalization has been the diffusion of · While virtually all chapters of DCP1 were structured around knowledge about what these interventions are and how to clusters of conditions, DCP2 provides integrative deliver them. The pace of this diffusion into a country deter- chapters--for example, on school health systems, surgery, mines the pace of health improvement in that country much and integrated management of childhood illness--that more than its level of income. Our purpose is to help speed this draw together the implementation-related responses to a diffusion of life-saving knowledge. number of conditions. These and other chapters reflect DCP2's inclusion of implementation and system issues. The Editors · DCP2 includes explicit discussions of research and product development opportunities. · Although DCP1 dealt with policy mechanisms to change behavior (or the environment), DCP2 attempts to do so in a REFERENCES more systematic way. In particular, a number of chapters assess in depth the public sector instruments for influencing Arrow, K. J., C. Panosian, and H. Gelband, eds. 2004. Saving Lives, Buying behavior change that were described briefly in DCP1: infor- Time: Economics of Malaria Drugs in an Age of Resistance. Washington, DC: National Academies Press. mation, education, and communication; laws and regula- Breman, J. G., M. S. Alilio, and A. Mills, eds. 2004. "The Intolerable Burden tions; taxes and subsidies; engineering design, such as speed of Malaria: II. What's New, What's Needed." American Journal of bumps; and facility location and characteristics. Hygiene and Tropical Medicine 71 (2 Suppl): 1­282. · Different interventions place different levels of demand on a Feachem, R. G. A., T. Kjellstrom, C. J. L. Murray, M. Over, and M. Phillips, country's health system capacity. DCP2 builds on earlier eds. 1992. Health of Adults in the Developing World. New York: Oxford work (Gericke and others 2005) in attempting, in some University Press. chapters, to identify which interventions require relatively Gericke, C. A., C. Kurowski, M. K. Ranson, and A. Mills. 2005. "Intervention Complexity: A Conceptual Framework to Inform less system capacity for scaling up and which require more. Priority-Setting in Health." Bulletin of the World Health Organization · Although DCP1 briefly discussed the nonhealth outcomes 83 (4): 285­93. of interventions, DCP2 does so in a more systematic way, Jamison, D. T., J. G. Breman, A. R. Measham, G. Alleyne, M. Claeson, D. B. including looking at the consequences of interventions (and Evans, P. Jha, A. Mills and P. Musgrove, eds. 2006. Priorities in Health. Washington, DC: World Bank. intervention financing) for reducing financial risks at the Jamison, D. T., W. H. Mosley, A. R. Measham, and J. L. Bobadilla, eds. 1993. household level. Other important nonhealth outcomes Disease Control Priorities in Developing Countries. New York: Oxford include, for example, the time-saving value of having piped University Press. water close to the home, the increased labor productivity of Levine, R., and the What Works Working Group. 2004. Millions Saved: healthy workers, and the amenity value of clean air. Proven Successes in Global Health. Washington, DC: Center for Global Development. · An important element of DCP1 was its assumption that to Lopez A. D., C. D. Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray, inform broad policy, major changes from the status quo eds. 2006. Global Burden of Disease and Risk Factors. New York: Oxford need to be considered, not just marginal ones. For cost- University Press. effectiveness analysis, any major change needs to be Murray, C. J. L., A. D. Lopez, and D. T. Jamison. 1994. "The Global Burden informed by burden of disease assessments in a way not of Disease in 1990: Summary Results, Sensitivity Analysis, and Future required for judging the attractiveness of marginal change, Directions." In Global Comparative Assessments in the Health Sector: Disease Burden, Expenditures, and Intervention Packages, ed. C. J. L. because the size of the burden affects total costs and the fea- Murray, and A. D. Lopez, 97­138. Geneva: World Health Organization. sibility of extending the intervention to all who would ben- World Bank. 1993. World Development Report 1993: Investing in Health. efit. This is particularly true when considering research and New York: Oxford University Press. xviii | Preface Editors Dean T. Jamison is a professor of health economics in the eradication. In 1976, in the Democratic Republic of Congo School of Medicine at the University of California, San (formerly Zaire), Dr. Breman investigated the first outbreak of Francisco (UCSF), and an affiliate of UCSF Global Health Ebola hemorrhagic fever. Sciences. Dr. Jamison concurrently serves as an adjunct profes- Following the confirmation of smallpox eradication in sor in both the Peking University Guanghua School of 1980, Dr. Breman returned to the U.S. Centers for Disease Management and in the University of Queensland School of Control, where he began work on the epidemiology and con- Population Health. trol of malaria. Dr. Breman joined the Fogarty International Before joining UCSF, Dr. Jamison was on the faculty of the Center in 1995 and has been director of the International University of California, Los Angeles, and also spent many years Training and Research Program in Emerging Infectious at the World Bank, where he was a senior economist in the Diseases and senior scientific adviser. He has been a member research department; division chief for education policy; and of many advisory groups, including serving as chair of the division chief for population, health, and nutrition. In 1992­93, World Health Organization's Technical Advisory Group on he temporarily rejoined the World Bank to serve as director of Human Monkeypox and as a member of the World Health the World Development Report Office and as lead author for Organization's International Commission for the Certification the Bank's World Development Report 1993: Investing in Health. of Dracunculiasis (guinea worm) Eradication. Dr. Breman has His publications are in the areas of economic theory, public written more than 100 publications on infectious diseases and health, and education. Dr. Jamison studied at Stanford (B.A., research capacity strengthening in developing countries. He philosophy; M.S., engineering sciences) and at Harvard (Ph.D., was guest editor of two supplements to the American Journal economics, under K. J. Arrow). In 1994, he was elected to mem- of Tropical Medicine and Hygiene: "The Intolerable Burden of bership in the Institute of Medicine of the U.S. National Malaria: A New Look at the Numbers" (2001) and "The Academy of Sciences. Intolerable Burden of Malaria: What's New, What's Needed" (2004). Joel G. Breman, M.D., D.T.P.H., is senior scientific adviser, Fogarty International Center of the National Institutes of Anthony R. Measham is co-managing editor of the Disease Health, and comanaging editor of the Disease Control Control Priorities Project at the Fogarty International Center Priorities Project. He was educated at the University of of the National Institutes of Health; deputy director of the California, Los Angeles; the Keck School of Medicine, the Communicating Health Priorities Project at the Population University of Southern California; and the London School of Reference Bureau, Washington, DC; and a member of the Hygiene and Tropical Medicine. Dr. Breman trained in medi- Working Group of the Global Alliance for Vaccines and cine at the University of Southern California­Los Angeles Immunization on behalf of the World Bank. County Medical Center; in infectious diseases at the Boston Born in the United Kingdom, Dr. Measham practiced family City Hospital, Harvard Medical School; and in epidemiology at medicine in Dartmouth, Nova Scotia, before devoting the the U.S. Centers for Disease Control and Prevention. remainder of his career to date to international health. He Dr. Breman worked in Guinea on smallpox eradication spent 15 years living in developing countries on behalf of the (1967­69); in Burkina Faso at the Organization for Population Council (Colombia), the Ford Foundation Coordination and Cooperation in the Control of the Major (Bangladesh), and the World Bank (India). Early in his inter- Endemic Diseases (1972­76); and at the World Health national health career (1975­77), he was deputy director of the Organization, Geneva (1977­80), where he was responsible Center for Population and Family Health at Columbia for orthopoxvirus research and the certification of smallpox University, New York. He then served for 17 years on the staff xix of the World Bank, as health adviser from 1984 until 1988 and his services to medicine. In 2001, he was awarded the Order as chief for policy and research of the Health, Nutrition, and of the Caribbean Community, the highest honor that can be Population Division from 1988 until 1993. conferred on a Caribbean national. Dr. Measham has spent most of his career providing techni- cal assistance, carrying out research and analysis, and helping Mariam Claeson, M.D., M.P.H., is the program coordinator for to develop projects in more than 20 developing countries, pri- AIDS in the South Asia Region of the World Bank since January marily in the areas of maternal and child health, family plan- 2005. She was the lead public health specialist in the Health, ning, and nutrition. He was an editor of the first edition of Nutrition, and Population, Human Development Network, of Disease Control Priorities in Developing Countries and has the World Bank (1998­2004), managing the Health, Nutrition, authored approximately 60 monographs, book chapters, and and Population Millennium Development Goals work program journal articles. to support accelerated progress in countries. Dr. Measham graduated in medicine from Dalhousie Dr. Claeson coauthored the call for action by the Bellagio University, Halifax, Nova Scotia. He received a master of science study group on child survival in 2003, Knowledge into Action and a doctorate in public health from the University of North for Child Survival, and the World Bank's 2005 report on Carolina in Chapel Hill and is a diplomat of the American Board The Millennium Development Goals for Health: Rising to the of Preventive Medicine and Public Health. His honors include Challenges. She was a member of the What Works Working being elected to the Alpha Omega Alpha Honor Medical group hosted by the Center for Global Development that re- Society; being appointed as special professor of International sulted in the report Millions Saved: Proven Successes in Global Health, University of Nottingham Medical School, Nottingham, Health (2005). Dr. Claeson coauthored the health chapter of the United Kingdom; and being named Dalhousie University Poverty Reduction Strategy source book, promoting a life-cycle Medical Alumnus of the Year in 2000­1. approach to maternal and child health and nutrition. As a coor- dinator of the public health thematic group (1998­2002), she George Alleyne, M.D., F.R.C.P., F.A.C.P. (Hon.), D.Sc. (Hon.), led the development of the strategy note Public Health and World is director emeritus of the Pan American Health Organization, Bank Operations and promoted multisectoral approaches to where he served as director from 1995 to 2003. Dr. Alleyne is a child health within the World Bank and in Bank-supported native of Barbados and graduated from the University of the country operations, analytical work, and lending. West Indies in medicine in 1957. He completed his postgradu- Prior to joining the World Bank, Dr. Claeson worked with ate training in internal medicine in the United Kingdom and the World Health Organization from 1987 until 1995, in later did further postgraduate work in that country and in the years as program manager for the Global Program for the United States. He entered academic medicine at the University Control of Diarrheal Diseases. She has several years of field of the West Indies in 1962, and his career included research in experience working in developing countries; in clinical practice the Tropical Metabolism Research Unit for his doctorate in at the rural district level in Bangladesh, Bhutan, and Tanzania; medicine. He was appointed professor of medicine at the in national program management of immunization and University of the West Indies in 1972, and four years later he diarrheal disease control programs in Ethiopia; and in health became chair of the Department of Medicine. He is an emeri- sector development projects in middle- and low-income tus professor of the University of the West Indies. Dr. Alleyne countries. joined the Pan American Health Organization in 1981, in 1983 he was appointed director of the Area of Health Programs, and David B. Evans, Ph.D., is an economist by training. Between in 1990 he was appointed assistant director. 1980 and 1990, he was an academic, first in economics depart- Dr. Alleyne's scientific publications have dealt with his ments and then in a medical school, during which time he research in renal physiology and biochemistry and various undertook consultancies for the World Bank, the World Health aspects of clinical medicine. During his term as director of the Organization, and governments. From 1990 until 1998, he Pan American Health Organization, he dealt with and pub- sponsored and conducted research into social and economic lished on issues such as equity in health, health and develop- aspects of tropical diseases and their control in the United ment, and international cooperation in health. He has also Nations Children's Fund, United Nations Development addressed several aspects of health in the Caribbean and the Programme, World Bank, and World Health Organization problems the area faces. He is a member of the Institute of Special Programme on Research and Training in Tropical Medicine and chancellor of the University of the West Indies. Diseases. He subsequently became director of the Global Dr. Alleyne has received numerous awards in recognition Programme on Evidence for Health Policy and then the of his work, including prestigious decorations and national Department of Health Systems Financing of the World Health honors from many countries of the Americas. In 1990, he was Organization, where he is now responsible for a range of made Knight Bachelor by Her Majesty Queen Elizabeth II for activities relating to the development of appropriate health xx | Editors financing strategies and policies. These activities include the including books on the role of government in health in World Health Organization's CHOICE project, which has developing countries, health planning in the United Kingdom, assessed and reported the costs and effectiveness of more than decentralization, health economics research in developing 700 health interventions, the costs of scaling up interventions, countries, and the public-private mix. Her most recent research the levels of health expenditures and accounts, and the extent interests have been in the organization and financing of health of financial catastrophe and impoverishment caused by out-of- systems, including the evaluation of contractual relationships pocket payments for health and which has assessed the impact between the public and private sectors and the application of of different ways to raise funds for health, pool them, and use economic evaluation techniques to improve the efficiency of them to provide or purchase services and interventions. He has disease control programs. published widely in these areas. Dr. Mills has had extensive involvement in supporting the health economics research activities of the World Health Prabhat Jha is Canada research chair of health and develop- Organization's Tropical Disease Research Programme. She ment at the University of Toronto. He is also the founding founded, and is head of, the Health Economics and Financing director of the Centre for Global Health Research, St. Michael's Programme, which has become one of the world's leading Hospital; associate professor in the Department of Public groups in developing and applying theories and techniques Health Sciences, University of Toronto; research scholar at the of health economics to increase knowledge on how best to McLaughlin Centre for Molecular Medicine; and professeur improve the equity and efficiency of developing countries' extraordinaire at the Université de Lausanne, Switzerland. health systems. She has acted as adviser to a number of multi- Dr. Jha is lead author of Curbing the Epidemic: Governments lateral and bilateral agencies--notably, the United Kingdom and the Economics of Tobacco Control and coeditor of Tobacco Department for International Development and the World Control in Developing Countries. Both are among the most influ- Health Organization. She guided the creation of the Alliance ential books on tobacco control. He is the principal investigator for Health Policy and Systems Research and chairs its board. of a prospective study of 1 million deaths in India, researching Most recently, she has been a member of the Commission for mortality from smoking, alcohol use, fertility patterns, indoor Macroeconomics and Health and cochair of its working group air pollution, and other risk factors among 2.3 million homes on improving the health outcomes of the poor. and 15 million people. This work is currently the world's largest prospective study of health. He also conducts studies of HIV Philip Musgrove is deputy editor--global health for Health transmission in various countries, focusing on documenting Affairs, which is published by Project HOPE in Bethesda, the risk factors for the spread of HIV and interventions to pre- Maryland. He worked for the World Bank (1990­2002), includ- vent the spread of the HIV/AIDS epidemic. His studies have ing two years on secondment to the World Health Organization received more than $5 million in peer-reviewed grants. (1999­2001), retiring as a principal economist. He was previ- Dr. Jha has published widely on tobacco, HIV/AIDS, and ously an adviser in health economics at the Pan American health of the global poor. His awards include a Gold medal Health Organization (1982­90) and a research associate at the from the Poland Health Promotion Foundation (2000), the Brookings Institution and at Resources for the Future Top 40 Canadians under Age 40 Award (2004), and the Ontario (1964­81). Premier's Research Excellence Award (2004). Dr. Jha was a Dr. Musgrove is an adjunct professor at the School of research scholar at the University of Toronto and McMaster Advanced International Studies, Johns Hopkins University, University in Canada. He holds an M.D. from the University of and has taught at George Washington University, American Manitoba and a D. Phil. in epidemiology and public health University, and the University of Florida. He holds degrees from Oxford University, where he studied as a Rhodes Scholar from Haverford College (B.A., 1962, summa cum laude); at Magdalen College. Princeton University (M.P.A., 1964); and Massachusetts Institute of Technology (Ph.D., 1974). Anne Mills, Ph.D., is professor of health economics and policy Dr. Musgrove has worked on health reform projects in at the London School of Hygiene and Tropical Medicine. She Argentina, Brazil, Chile, and Colombia and has dealt with a has more than 20 years of experience in research pertaining to variety of issues in health economics, financing, equity, and health economics in developing countries and has published nutrition. His publications include more than 50 articles in widely in the fields of health economics and health planning, economics and health journals and chapters in 20 books. Editors | xxi Advisory Committee to the Editors J. R. Aluoch Peter Lachmann Professor, Nairobi Women's Hospital, Nairobi, Kenya Past President, U.K. Academy of Medical Sciences, Cambridge, United Kingdom Jacques Baudouy Director, Health, Nutrition, and Population, World Bank, Mary Ann Lansang Washington, DC, United States Executive Director, INCLEN Trust International Inc., Manila, Philippines Fred Binka Executive Director, INDEPTH Network, Accra, Ghana Christopher Lovelace Director, Kyrgyz Republic Country Office and Central Asia Mayra Buvinic´ Human Development, World Bank, Bishkek, Kyrgyz Republic Director, Gender and Development, World Bank, Anthony Mbewu Washington, DC, United States Executive Director, Medical Research Council of South Africa, David Challoner, Co-Chair Tygerberg, South Africa Foreign Secretary, Institute of Medicine, Rajiv Misra U.S. National Academies, Gainesville, Florida, Former Secretary of Health, Government of India, United States Haryana, India Guy de Thé, Co-Chair Perla Santos Ocampo Research Director and Professor Emeritus, Institut Pasteur, President, National Academy of Science and Technology, Paris, France San Juan, Philippines Timothy Evans G. B. A. Okelo Assistant Director General, Evidence and Information for Secretary General and Executive Director, African Academy of Policy, World Health Organization, Geneva, Switzerland Sciences, Nairobi, Kenya Richard Horton Sevket Ruacan Editor, The Lancet, London, United Kingdom General Director, MESA Hospital Ankara, Turkey Sharon Hrynkow Pramilla Senanayake Acting Director, Fogarty International Center, National Chairman, Foundation Council of the Global Forum for Institutes of Health, Bethesda, Maryland, Health Research, Colombo, Sri Lanka United States Jaime Sepúlveda, Chair Gerald Keusch Director, National Institutes of Health of Mexico, Mexico City, Provost and Dean for Global Health, Boston University Mexico School of Public Health, Boston, Massachusetts, United States Chitr Sitthi-amorn Director, Institute of Health Research, and Dean, Kiyoshi Kurokawa Chulalongkorn University, College of Public Health, Bangkok, President, Science Council of Japan, Kanawaga, Japan Thailand xxiii Sally Stansfield Zhengguo Wang Associate Director, Global Health Strategies, Bill & Melinda Professor, Chinese Academy of Engineering, Daping, China Gates Foundation, Seattle, Washington, United States Witold Zatonski Misael Uribe Professor, Health Promotion Foundation, Warsaw, Poland President, National Academy of Medicine of Mexico, Mexico City, Mexico xxiv | Advisory Committee to the Editors Contributors Taghreed Adam Geneviève Begkoyian World Health Organization United Nations Children's Fund Sarah Adomakoh Jere R. Behrman Chronic Disease Research Centre, University of the West University of Pennsylvania Indies Associates for International Development Nicole Bellows Olu Akinyanju University of California, Berkeley Sickle Cell Foundation, Nigeria Sandra E. Bendeck University Teaching Hospital, Nigeria University of Texas at Southwestern Mark A. Anderson Stefano Bertozzi U.S. Centers for Disease Control and Prevention Instituto Nacional de Salud Pública Sevgi O. Aral Jeff Bethony U.S. Centers for Disease Control and Prevention George Washington University Samira Asma Alok Bhargava U.S. Centers for Disease Control and Prevention University of Houston Cristian Baeza Sohinee Bhattacharya World Bank University of Aberdeen Rob Baltussen Zulfiqar A. Bhutta Erasmus MC Aga Khan University Delia Barcelona Nancy Birdsall United Nations Population Fund Center for Global Development Scott Barrett David Bishai Johns Hopkins University Johns Hopkins University John H. Barton Robert E. Black Stanford University Johns Hopkins Bloomberg School of Public Health Jane Batt University of Toronto Barry R. Bloom St. Michaels Hospital Harvard School of Public Health Angela Bayer Stephen B. Blount Johns Hopkins University Bloomberg School of Public Health U.S. Centers for Disease Control and Prevention Kathleen Beegle Cynthia Boschi-Pinto World Bank World Health Organization xxv Douglas Bratthall Laura E. Caulfield World Health Organization Collaborating Centre Johns Hopkins University Bloomberg School of Public Health Centre for Oral Health Sciences, Malmo University Frank J. Chaloupka Carol Brayne University of Illinois at Chicago University of Cambridge Vijay Chandra Joel G. Breman World Health Organization, Regional Office for Fogarty International Center, National Institutes of Health South-East Asia Disease Control Priorities Project Heng Leng Chee Logan Brenzel National University of Singapore World Bank Suephy Chen Dan W. Brock Emory University Harvard Medical School Atlanta Veterans Administration Medical Center Simon Brooker Thomas Cherian London School of Hygiene and Tropical Medicine World Health Organization Peter Brooks Tom M. Chiller University of Queensland U.S. Centers for Disease Control and Prevention Claire V. Broome Dan Chisholm U.S. Centers for Disease Control and Prevention World Health Organization L. Jackson Brown Lester Chitsulo American Dental Association World Health Organization Martin L. Brown Jeffrey Chow National Cancer Institute, National Institutes of Health Resources for the Future Nigel Bruce Mushtaque Chowdhury University of Liverpool Bangladesh Rural Advancement Committee Columbia University Jennifer Bryce Independent consultant Mariam Claeson World Bank Colin H. W. Bullough University of Aberdeen Luke B. Connelly University of Queensland Donald A. P. Bundy World Bank Joseph Cook International Trachoma Initiative Alexander Butchart World Health Organization Rodrigo Correa-Oliveira Centro de Pesquisas Rene Rachou--FIOCRUZ Mayra Buvinic´ World Bank Mark Cullen Yale University Sandy Cairncross London School of Hygiene and Tropical Medicine Patricia M. Danzon Wharton School, University of Pennsylvania John Cairns London School of Hygiene and Tropical Medicine Haile T. Debas University of California, San Francisco Balla Camara Ministry of Public Health, Guinea Louisa Degenhardt Ministry of Education, Guinea University of New South Wales Pierre Cattand Lisa M. DeMaria Association against Trypanosomiasis Instituto Nacional de Salud Pública xxvi | Contributors Nilanthi de Silva Timothy Evans University of Kelaniya, Sri Lanka World Health Organization Phillippe Desjeux Qiu Fang Institute for OneWorld Health World Bank Claude de Ville de Goyet Piet Feenstra Independent consultant Royal Tropical Institute, Netherlands John Dirks Becca Feldman International Society of Nephrology Harvard School of Public Health University of Toronto Instituto Nacional de Salud Pública Jane Doherty Alan Fenwick University of the Witwatersrand, South Africa Imperial College Elisa Fernández Chris Doran Inter-American Development Bank University of Queensland Marilyn Fingerhut Ogobara K. Doumbo National Institute for Occupational Safety University of Bamako, Mali and Health, United States Gerrit Draisma Katherine Floyd Erasmus MC World Health Organization Lesley Drake Kathleen M. Foley St. Mary's Medical School Memorial Sloan-Kettering Cancer Center Maureen S. Durkin Weill Medical College of Cornell University University of Wisconsin Medical School Olivier Fontaine University of Wisconsin-Madison World Health Organization Adriano Duse Susan Foster University of the Witwatersrand, South Africa Boston University School of Public Health National Health Laboratory Service Julia Fox-Rushby Courtenay Dusenbury Brunel University, United Kingdom Emory University Julio Frenk Christopher Dye Secretaria de Salud de Mexico World Health Organization Kevin D. Frick Michael M. Engelgau Johns Hopkins Bloomberg School of Public Health U.S. Centers for Disease Control and Prevention Suthat Fucharoen Dirk Engels Mahidol University, Thailand World Health Organization Vendhan Gajalakshmi Mike English Epidemiological Research Center, India Kenya Medical Research Institute Rae Galloway University of Oxford World Bank Victoria Espitia-Hardeman Helene Gayle U.S. Centers for Disease Control and Bill & Melinda Gates Foundation Prevention Thomas A. Gaziano Roberto Estrada Brigham and Women's Hospital Universidad Autónoma de Guerrero, Mexico Harvard Medical School David B. Evans Hellen Gelband World Health Organization Institute of Medicine, National Academies Contributors | xxvii Amaya Gillespie Martin Hensher United Nations Study on Violence against Children Department of Health, United Kingdom Julian Gold Susan Herman Prince of Wales Hospital, Australia University of Pennsylvania Sue J. Goldie Eduardo Romero Hicks Harvard School of Public Health Secretaria de Salud Guanajuato, Mexico University of Guanajuato, Mexico Chuck Golmar World Health Organization Anna-Maria Hoffman United Nations Educational Scientific, and Cultural Richard Gosselin Organization University of California, Berkeley Karen J. Hofman Pablo Gottret Fogarty International Center, National Institutes of Health World Bank King K. Holmes Eduardo Gotuzzo University of Washington Universidad Peruana Cayetano Heredia Harborview Medical Center Hospital Nacional Cayetano Heredia, Peru Charles Hongoro Wendy J. Graham London School of Hygiene and Tropical Medicine University of Aberdeen Aurum Health Research Institute London School of Hygiene and Tropical Medicine Susan Horton Robert Grant Wilfrid Laurier University, Canada J. David Gladstone Institutes, United States Peter J. Hotez Brian Greenwood George Washington University London School of Hygiene and Tropical Medicine Fleur Hourihan Prakash C. Gupta University of Newcastle, Australia Healis-Sekhsaria Institute of Public Health, India Arnold School of Public Health, United States Guy Hutton Swiss Tropical Institute M.G. Guzmán Pedro Kouri Tropical Medicine Institute, Cuba Adnan A. Hyder Johns Hopkins Bloomberg School of Public Health Anne Haddix Rollins School of Public Health, Emory University Steven Hyman Harvard University Wayne Hall Harvard Medical School University of Queensland Giuseppina Imperatore Neal A. Halsey U.S. Centers for Disease Control and Prevention Johns Hopkins University Bloomberg School of Public Health Rubina Imtiaz Joe Harford U.S. Centers for Disease Control and Prevention National Cancer Institute, National Institutes of Health Michael T. Isbell Roderick Hay Independent consultant Queen's University Belfast Dean T. Jamison Robert M. Hecht University of California, San Francisco International AIDS Vaccine Initiative Disease Control Priorities Project D. A. Henderson Jean Jannin University of Pittsburgh Medical Center World Health Organization xxviii | Contributors Philip Jenkins Olive C. Kobusingye World Health Organization World Health Organization Regional Office for Africa Prabhat Jha Jeffrey P. Koplan University of Toronto Emory University Centre for Global Health Research Daniel Kress T. Jacob John Bill & Melinda Gates Foundation Christian Medical College (retired) A. Kroeger National HIV/AIDS Reference Center, India (retired) World Health Organization Jack Jones Richard Laing World Health Organization Boston University School of Public Health, now World Health David E. Joranson Organization University of Wisconsin Comprehensive Cancer Center John R. La Montagne (Deceased) World Health Organization National Institute of Allergy and Infectious Diseases, Manjul Joshipura National Institutes of Health Academy of Traumatology, India Claudio F. Lanata Apollo Hospitals, India Instituto de Investigación Nutricional, Peru Matthew Jukes Imperial College London Ana Langer EngenderHealth Alka M. Kanaya University of California, San Francisco Carlene M. M. Lawes University of Auckland Scott Kasner University of Pennsylvania Joy E. Lawn Save the Children­USA Ronald Kessler Institute of Child Health Harvard Medical School Ramanan Laxminarayan Gerald T. Keusch Resources for the Future Boston University Medical Campus Boston University School of Public Health Seung-Hee Frances Lee Save the Children­USA Peter Kilima International Trachoma Initiative Christian Lengeler Swiss Tropical Institute Sally Kingsland National Centre for Epidemiology and Population Health, P. R. Lever Australian National University Royal Tropical Institute, Netherlands Tord Kjellström Ruth Levine Australian National University Center for Global Development National Institute of Public Health, Sweden Joseph Lipscomb Keith P. Klugman Rollins School of Public Health, Emory University Rollins School of Public Health, Emory University Madhumita Lodh Felicia Knaul Commonwealth Department of Transport and Regional Fundación Mexicana para la Salud Services, Australia Secretaria de Educación Pública de Mexico Alan Lopez Rudolf Knippenberg University of Queensland United Nations Children's Fund Harvard School of Public Health James C. Knowles Elizabeth Lule Independent consultant World Bank Contributors | xxix Antoine Mahé David Meltzer Programme National de Lutte Contre le SIDA University of Chicago Adel Mahmoud Kamini Mendis Merck & Company Inc. World Health Organization Case Western Reserve University James Mercy Margaret Maier U.S. Centers for Disease Control and Prevention RAND Corporation Catherine Michaud Lisa Manhart Harvard School of Public Health University of Washington Mark Miller Bala Manyam Fogarty International Center, National Institutes Texas A&M University HSC School of Medicine of Health Paola Marchesini Anne Mills World Health Organization London School of Hygiene and Tropical Medicine Maureen S. Marshall Andrew Mitchell Task Force for Child Survival and Development Harvard School of Public Health John B. Mason Arlene Mitchell Tulane University School of Public Health World Food Programme and Tropical Medicine Charles Mock Gaverick Matheny University of Washington University of Maryland Harborview Medical Center Colin McCord Antonio Montresor Columbia University World Health Organization James Moore Martin McKee U.S. Centers for Disease Control and Prevention London School of Hygiene and Tropical Medicine Chantal Morel Tonya McLeod London School of Hygiene and Tropical Medicine Emory University Oxford Outcomes Tony McMichael Luis Morillo Australian National University Javeriana University Anthony R. Measham Roy D. Mugerwa World Bank (retired) Makerere University, Uganda Disease Control Priorities Project Case Western Reserve University Jeffrey Mecaskey Jo-Ann Mulligan Axios International London School of Hygiene and Tropical Medicine André Médici Philip Musgrove Inter-American Development Bank Health Affairs Carol Ann Medlin Disease Control Priorities Project University of California, San Francisco Vasant Narasimhan Sumi Mehta Novartis Pharma AG, Switzerland World Health Organization K. M. Venkat Narayan Health Effects Institute Centers for Disease Control and Prevention Bjørn Melgaard Rollins School of Public Health at Emory World Health Organization University xxx | Contributors Mike B. Nathan Rajesh Pandav World Health Organization World Health Organization, Regional Office for South-East Asia Karin B. Nelson National Institute for Neurological Disorders and Stroke, Vikram Patel National Institutes of Health London School of Hygiene and Tropical Medicine Isaac Ngugi Vikram S. Pathania KEMRI/Wellcome Trust Programme, Kenya University of California, Berkeley Mounkaila Noma John W. Peabody African Programme for Onchocerciasis Control University of California, San Francisco University of California, Los Angeles Charles Normand University of Dublin, Trinity College Richard Peck London School of Hygiene and Tropical Medicine University of Illinois at Chicago Robyn Norton Margie Peden George Institute for International Health World Health Organization University of Sydney Poul Erik Petersen Peter Nsubuga World Health Organization U.S. Centers for Disease Control and Prevention Ndola Prata Rachel Nugent University of California, Berkeley Population Reference Bureau Alexander S. Preker Fogarty International Center, National Institutes World Bank of Health Max Price Thomas O'Brien University of the Witwatersrand, South Africa Brigham and Women's Hospital Pekka Puska Adesola Ogunniyi National Public Health Institute, Finland University of Ibadan Zahidul Quayyum University College Hospital, Nigeria University of Aberdeen Iruka N. Okeke Sadanand Rajkumar Haverford College University of Newcastle Eidgenossische Technische Hochschule, Switzerland Bloomfield Hospital Nancy Olivieri Ambady Ramachandran Hemoglobinopathy Research Program, University Health M.V. Hospital for Diabetes, India Network, Canada Diabetes Research Centre, India Claudio Osorio K. D. Ramaiah Pan American Health Organization Vector Control Research Centre, India Mead Over Geetha Ranmuthugala World Bank Australian National University Ariel Pablos-Mendez Fawzia Rasheed World Health Organization Independent consultant Columbia University K. Srinath Reddy Fred Paccaud All India Institute of Medical Sciences University of Lausanne Initiative for Cardiovascular Health in the University of Montreal Developing Countries Nancy S. Padian Eva Rehfuess University of California, San Francisco World Health Organization Contributors | xxxi Jürgen Rehm Arrigo Schieppati Centre for Addiction and Mental Health, Canada Mario Negri Institute for Pharmacological ISGF/ARI, Switzerland Research, Italy Azienda Ospedaliera Ospedali Riuniti di Bergamo, Italy Jan H. F. Remme World Health Organization Gabriel Schmunis Pan American Health Organization Giuseppe Remuzzi Mario Negri Institute for Pharmacological Research, Italy Helen Schneider Azienda Ospedaliera Ospedali Riuniti di Bergamo, Italy University of the Witwatersrand, South Africa Serge Resnikoff C. J. Schofield World Health Organization London School of Hygiene and Tropical Medicine Stephanie A. Richard A. Seketeli Johns Hopkins University Bloomberg School of African Programme for Onchocerciasis Control Public Health Malick Sembene Frank Richards Ministry of National Education, Senegal U.S. Centers for Disease Control and Prevention Sheldon Shaeffer The Carter Center of Emory University United Nations Educational, Scientific, and Juan A. Rivera Cultural Organization Instituto Nacional de Salud Pública, Mexico Raj J. Shah S. Adibul Hasan Rizvi Bill & Melinda Gates Foundation Sindh Institute of Urology and Transplantation Sonbol A. Shahid-Salles David A. Robalino Disease Control Priorities Project World Bank Population Reference Bureau Anthony Rodgers Brian Sharp University of Auckland Medical Research Council Khama Rogo Alexandra Shaw World Bank AP Consultants Robin Room Donald Shepard Stockholm University Schneider Institute for Health Policy, Heller School, James E. Rosen Brandeis University World Bank Rupendra Shrestha Mark L. Rosenberg Australian National University Task Force for Child Survival and Xiao Shu-Hua Development National Institute of Parasitic Diseases, China Linda Rosenstock Donald Silberberg University of California, Los Angeles University of Pennsylvania David Sanders Eric A. F. Simoes University of the Western Cape University of Colorado Health Sciences Center Lorenzo Savioli Children's Hospital World Health Organization Lone Simonsen Richard M. Scheffler National Institute of Allergy and Infectious Diseases, University of California, Berkeley National Institutes of Health George Schieber Susheela Singh World Bank Alan Guttmacher Institute, United States xxxii | Contributors Arthur S. Slutsky Stephen B. Thacker St. Michael's Hospital U.S. Centers for Disease Control and Prevention University of Toronto William Theodore Andrew Smith National Institute for Neurological Disorders and Stroke, World Health Organization National Institutes of Health Kirk Smith Amardeep Thind School of Public Health, University of California, Berkeley University of Western Ontario University of California, Los Angeles Peter C. Smith Centre for Health Economics, University of York Stephen Tollman Medical Research Council Robert W. Snow University of the Witwatersrand, South Africa Centre for Tropical Medicine, University of Oxford Kenya Medical Research Institute Ana Cristina Torres World Bank Soekirman Institut Pertanian Bogor, Indonesia Murray Trostle U.S. Agency for International Development Geoffrey C. Solarsh Monash University, Australia Vivian Valdmanis University of the Sciences in Philadelphia Dan Sosin U.S. Centers for Disease Control and Prevention W. H. van Brakel Royal Tropical Institute, Netherlands Frank E. Speizer Harvard Medical School Anna Vassall Harvard School of Public Health Royal Tropical Institute, Netherlands Sally K. Stansfield Cesar G. Victora Bill & Melinda Gates Foundation Universidade Federal de Pelotas, Brazil University of Washington Maya Vijayaraghavan Richard W. Steketee U.S. Centers for Disease Control and Prevention PATH Theo Vos Jayanthi Ramanathan Stjernswärd University of Queensland Malmø University Judith L. Wagner Stephen E. Straus Institute of Medicine, United States National Center for Complementary and Alternative Adam Wagstaff Medicine, National Institutes of Health World Bank Donna F. Stroup Julia Walsh U.S. Centers for Disease Control and Prevention University of California, Berkeley Mario M. Taguiwalo Prawese Wasi Department of Health, Republic of the Siriraj Hospital Philippines Mahidol University Tsutomu Takeuchi Hugh Waters School of Medicine, Keio University, Japan Johns Hopkins University Bloomberg School of Public Health Caroline Tanner Parkinson's Institute David Weatherall University of Oxford Robert V. Tauxe U.S. Centers for Disease Control and Jeny Wegbreit Prevention University of California, San Francisco Contributors | xxxiii Mark E. White Cream Wright U.S. Centers for Disease Control and Prevention United Nations Children's Fund Nicholas J. White Merrick Wright Mahidol University, Thailand Independent consultant University of Oxford Michel Zaffran Harvey Whiteford World Health Organization University of Queensland Ricardo Zapata Marti Daniel Wikler United Nations Economic Commission for Latin America Harvard School of Public Health and Caribbean Suwit Wilbulpolprasert Witold Zatonski Ministry of Public Health, Thailand Cancer Center and Institute of Oncology, Poland Health Promotion Foundation Walter C. Willett Harvard School of Public Health Ping Zhang Harvard Medical School U.S. Centers for Disease Control and Prevention Desmond E. Williams Zhen-Xin Zhang U.S. Centers for Disease Control and Prevention Peking Union Medical College Hospital Chinese Academy of Medical Science Lara J. Wolfson World Health Organization Jelka Zupan World Health Organization Anthony Woolf Peninsula Medical School, United Kingdom Royal Cornwall Hospital, United Kingdom xxxiv | Contributors Disease Control Priorities Project Partners The Disease Control Priorities Project is a joint enterprise of US$18 billion to US$22 billion each year in loans to its client the Fogarty International Center of the National Institutes of countries, provided US$1.27 billion for health, nutrition, and Health, the World Health Organization, the World Bank, and the population in 2004. The World Bank is working in more than Population Reference Bureau. 100 developing economies, bringing a mix of analytical work, The Fogarty International Center is the international compo- policy dialogue, and lending to improve living standards-- nent of the U.S. National Institutes of Health. It addresses global including health and education--and reduce poverty. health challenges through innovative and collaborative research The Population Reference Bureau informs people around and training programs and supports and advances the mission of the world about health, population, and the environment and the U.S. National Institutes of Health through international empowers them to use that information to advance the well- partnerships. being of current and future generations. For 75 years, the The World Health Organization is the specialized agency for bureau has analyzed complex data and research results to health of the United Nations. Its objective, as set out in its con- provide objective and timely information in a format easily stitution, is the attainment by all peoples of the highest possi- understood by advocates, journalists, and decision makers; has ble level of health, with health defined as a state of complete conducted workshops around the world to give key audiences physical, mental, and social well-being and not merely the the tools they need to understand and communicate effectively absence of disease or infirmity. about relevant issues; and has worked to ensure that policy The World Bank Group is one of the world's largest makers in developing countries base policy decisions on sound sources of development assistance. The Bank, which provides evidence. xxxv Acknowledgments Preparation of this volume required efforts over four years by (Prabhat Jha), and Resources for the Future (Ramanan many institutions and almost 1,000 individuals: chapter coau- Laxminarayan). The Center for Global Development (Ruth thors, advisory committee members, peer reviewers, copy edi- Levine) collaborated with the chapter authors in an effort to tors, and research and staff assistants. We have many contri- identify proven successes in global health, the results of which butions to acknowledge. We particularly thank our chapter were used both in this book and in a separate publication. We authors, who worked extremely hard through a long and exact- are grateful to each of these institutions and individuals. ing process of writing, review, and revision. We also owe much We were particularly fortunate to have the strong collab- gratitude to the institutional sponsors of this effort: oration of the Inter-Academies Medical Panel (IAMP), an asso- ciation of the medical academies or medical divisions of the sci- · The Fogarty International Center (FIC) of the U.S. National entific academies of 44 countries. David Challoner and Guy de Institutes of Health. The FIC supported both the senior editor Thé cochaired the Steering Committee of the IAMP and invest- and one of the co-managing editors of this project, as well as ed much time and effort into facilitating the collaboration. In support staff. Gerald Keusch, former director of the FIC, ini- particular, the IAMP helped establish the productive Advisory tiated and facilitated this effort, and FIC's acting director, Committee to the Editors, chaired by Jaime Sepúlveda, on Sharon Hrynkow, continued to provide support and counsel. which many members of the IAMP Steering Committee served. · The World Bank. Successive directors of the World Bank's The IAMP's second global meeting hosted the launch of this Health, Nutrition, and Population Department, Christopher volume in Beijing in April 2006, and the IAMP also sponsored Lovelace and Jacques Baudouy, provided support, guidance, the peer review process for all the chapters. We are most grate- and critical reactions and facilitated the involvement of ful to David Challoner and Guy de Thé, as well as to Jaime Bank staff as coauthors and reviewers. Sepúlveda and other members of the Advisory Committee to · The World Health Organization. Successive leaders of the the Editors. The U.S. member of the IAMP, the Institute of World Health Organization's Evidence and Information Medicine of the National Academy of Sciences, played a critical for Policy Cluster, Christopher Murray and Timothy Evans, role in facilitating all aspects of the IAMP's collaboration. coordinated the involvement of the World Health Patrick Kelley, Patricia Cuff, Dianne Stare, Stacey Knobler, and Organization, which had been agreed by Gro Harlem Leslie Baer at the Institute of Medicine and Mohamed Hassan Brundtland, then the director-general. and Muthoni Fanin at the IAMP managed this effort and · The Bill & Melinda Gates Foundation. Richard Klausner, provided critical, substantive inputs. Sally Stansfield, and Beth Peterman arranged for the foun- The Office of the Publisher at the World Bank provided out- dation to provide major financial support and interacted standing assistance, enthusiastic advice, and support during closely with us throughout the past four years. Initial every phase of production of this volume and helped coordi- conversations with and encouragement from William Gates nate publicity and initial distribution. We particularly wish to Senior are gratefully acknowledged. thank Dirk H. Koehler, the publisher; Carlos Rossel; Mary Fisk; Santiago Pombo-Bejarano; Nancy Lammers; Randi Park; In undertaking the work leading to this volume, we bene- Valentina Kalk; Alice Faintich; Joanne Ainsworth; Enid Zafran; fited from the close engagement of three institutions that Deepa Menon; and Janice Tuten for their timely, high-quality helped organize and host consultations and arranged for back- professionalism. ground analyses to be undertaken. These institutions were the Donald Lindberg, director of the National Library of London School of Hygiene and Tropical Medicine (Anne Medicine (NLM) of the U.S. National Institutes of Health, and Mills), the University of Toronto's Center for Global Health Julia Royall, chief, International Programs, NLM, graciously xxxvii offered the competent services of the NLM's Information · National Cancer Institute, National Institutes of Health, Engineering Branch of the National Center for Biotechnology consultation on cancer prevention, treatment, and pain Information to convert the text into an electronic product avail- control, Bethesda, Maryland (June 2003) able to all visitors to the National Library of Medicine's PubMed · Oswaldo Cruz Foundation, World Health Organization, and Web site. We would like to extend our gratitude to the National Pan-American Health Organization, consultation on tropi- Center for Biotechnology Information team members--David cal infectious diseases, Rio de Janeiro, Brazil (April­May Lipman, Jo McEntyre, and Mohammad Al-Ubaydli, and 2003) Belinda Beck--for their technical expertise and commitment. · Université de Lausanne, consultation on cardiovascular dis- With this volume now in the dissemination phase, the ease, Lausanne, Switzerland (March 2002) Population Reference Bureau is charged to communicate its · University of California, Berkeley, consultation on learning findings in formats likely to be of use to a range of audiences. and developmental disorders, Berkeley, California (August We greatly value the work of the bureau's William P. Butz, pres- 2003) ident, and Nancy Yinger, director of international programs, in · University of California, San Francisco, consultation on rapidly initiating this effort. surgery, San Francisco, California (July 2003) Multiple institutions from around the world contributed to · University of Queensland, School of Population Health, organizing and hosting meetings that facilitated the prepara- authors' meeting on psychiatric disorders, neurology, and tion of this book and providing background for such meetings. alcohol and other substance abuse, Brisbane, Australia We greatly appreciate the contributions and hospitality of these (August 2003) institutions, including the following: · University of Washington, consultation on sexually trans- mitted infections, Seattle, Washington (July 2003) · Chinese Academy of Engineering and Chinese Academy of · University of the Witwatersrand, consultations on health Sciences, Disease Control Priorities Project Launch and systems and on capacity strengthening and management Inter-Academies Medical Panel Global Meeting, Beijing, reform, Johannesburg, South Africa (July 2004) China (April 2006) · World Health Organization, Division of Mental Health, and · Italian Ministry of Health, Veneto Region, consultation on National Institutes of Health, National Institute of Mental child health and nutrition, Venice, Italy (January 2004) Health, consultation on mental health economics, Geneva, · Instituto Nacional de Salud Pública, Advisory Committee to Switzerland (March 2004). the Editors meeting, Cuernavaca, Mexico (June 2002) · Institut Pasteur, Advisory Committee to the Editors meet- Coordination of the work leading to this publication and back- ing, Paris, France (March and December 2004) ground research were undertaken by a small secretariat. Nancy · Johns Hopkins Bloomberg School of Public Health, consul- Hancock, Pamela Maslen, and Sonbol A. Shahid-Salles provided tation on maternal and child health, Annapolis, Maryland outstanding research assistance; Andrew Marshall ably man- (May 2002) aged the budget and process; Candice Byrne provided key com- · Johns Hopkins Paul H. Nitze School of Advanced munications guidance, staff and editorial assistance; and International Studies, consultation on elimination and Mantra Singh and Cherice Holloway provided staff assistance. eradication of disease, and vaccinations, Washington, DC Richard Miller, Lauren Sikes and Tommy Freeman of the FIC (October 2004) provided excellent administrative support to the Disease · Merck & Company Inc., consultation on research and prod- Control Priorities Project. Their work was absolutely essential uct development priorities, Whitehouse Station, New Jersey in producing this book, and we are deeply grateful for their (September 2004) commitment and productivity. With so many authors and · Multilateral Initiative on Malaria, consultations on the bur- institutions involved, we are aware that many more people gave den of malaria: countless hours to this endeavor. We thank them also for their · National Institute of Medical Research, Arusha, Tanzania dedication. (November 2002) · University of Yaoundé, Cameroon (November 2005) The Editors xxxviii | Acknowledgments Abbreviations and Acronyms ACE angiotensin-converting enzyme CBR cost-benefit ratio ACER average cost-effectiveness ratio CDC U.S. Centers for Disease Control and Prevention ACT artemisinin combination therapy CDD control of diarrheal diseases AD Alzheimer's disease CEA cost-effectiveness analysis ADB Asian Development Bank CEmOC comprehensive emergency obstetric care ADHD attention deficit and hyperactivity disorder CER cost-effectiveness ratio AED antiepileptic drug CFR case-fatality rate AHEAD applied health education and development CHA community health aide AIDS acquired immunodeficiency syndrome CHD coronary heart disease AIN-C atención integral a la niñez comunitaria CHF congestive heart failure ALRI acute lower respiratory infection CHNP community-based health and nutrition program AMI acute myocardial infarction CHNW community health and nutrition worker ANW anganwadi worker CHOICE choosing interventions that are cost-effective aP acellular pertussis vaccine CI confidence interval APOC African Programme for Onchocerciasis Control CKD chronic kidney disease ARF acute rheumatic fever CL cutaneous leishmaniasis ARI acute respiratory infection CL/P cleft lip and palate ART atraumatic restorative treatment CM cerebral malaria ASD autism spectrum disorder CMH Commission on Macroeconomics and Health ATLS advanced trauma life support CML chronic myeloid leukemia AUD alcohol-use disorder CO carbon monoxide AZT Zidovudine COBRA combination therapy for rheumatoid arthritis BCC behavior-change communication COHRED Council on Health Research for Development BCG Bacillus Calmette-Guérin COM chronic otitis media BEmOC basic emergency obstetric care COPCORD Community-Oriented Program for Control of BINP Bangladesh Integrated Nutrition Program Rheumatic Disease BMI body mass index COPD chronic obstructive pulmonary disease BMT buprenorphine maintenance treatment CoV coronavirus BOD burden of disease COX cyclo-oxygenase BRAC Bangladesh Rural Advancement Committee CRA comparative risk analysis BRFSS behavioral risk factor surveillance system CT computed tomography BZA benzimidazole anthelmintic CVD cardiovascular disease CABG coronary artery bypass graft CVS chorionic villus sampling CAD coronary artery disease CYP couple-year of protection CAM complementary and alternative medicine DAH development assistance for health CAPP Country/Area Profile Programme DALY disability-adjusted life year CBA cost-benefit analysis dBHL decibel hearing level CBE clinical breast examination DCP1 Disease Control Priorities in Developing CBHI community-based health insurance Countries, first edition xxxix DCP2 Disease Control Priorities in Developing GFHR Global Forum on Health Research Countries, second edition GIS geographic information system DCPP Disease Control Priorities Project GM genetic modification DDT dichlorodiphenyltrichloroethane GMP good manufacturing practice DEET N,N-diethyl-meta-toluamide GNI gross national income DF dengue fever GNP gross national product DHF dengue hemorrhagic fever GSE glutathione S-transferase DHS demographic and health survey GUSTO global use of strategies to open occluded DMARD disease-modifying antirheumatic drug coronary arteries DMFT decayed, missing, and filled teeth HAART highly active antiretroviral therapy for the DNA deoxyribose nucleic acid treatment of HIV/AIDS DOT directly observed therapy Hb hemoglobin DOTS directly observed therapy short course HBV hepatitis B virus DRC Democratic Republic of Congo HDL high-density lipoprotein DSM-IVTR Diagnostic and Statistical Manual of Mental HepB hepatitis B Disorders HHV human herpes virus DSS dengue shock syndrome Hib Haemophilus influenzae type B DTP diphtheria-tetanus-pertussis HIC high-income country EAP economically active population HIS health information system EBM evidence-based medicine HIV human immunodeficiency virus ED emergency department HMN Health Metrics Network EFA education for all HPLC high-performance liquid chromatography EFM electronic fetal monitoring HPS health promoting school EHCAP Effective Health Care Alliance Programme HPV human papillomavirus EIR entomological inoculation rate HR human resource ELISA enzyme-linked immunosorbent assay HRT hormone replacement therapy EMR electronic medical record HSV-1 herpes simplex virus type 1 EMS emergency medical services HSV-2 herpes simplex virus type 2 EPI Expanded Program on Immunization IAEA International Atomic Energy Agency ESRD end-stage renal disease IAP indoor air pollution EUROSTAT European Statistical Office IAVI International AIDS Vaccine Initiative FA folic acid ICD-10 International Statistical Classification of Diseases FBD food-borne disease and Related Health Problems, 10th revision FCTC Framework Convention on Tobacco Control ICDS integrated child development services FDA U.S. Food and Drug Administration ICER incremental cost-effectiveness ratio FDC fixed-dose combinations ICPD international conference on population and FEFO first expiry, first out development FETP Field Epidemiology Training Program ICT information and communication FEV1 forced expiratory volume in one second technologies FGM female genital mutilation IDA International Development Association FHP family health program IDD iodine deficiency disorders FIC fully immunized child IDSR integrated disease surveillance and response FRESH focusing resources on effective school health IEC information, education, and communication FTE full-time equivalent IFF International Finance Facility G6PD glucose-6-phosphate dehydrogenase IHD ischemic heart disease G-7 Group of Seven ILO International Labour Organisation GATB Global Alliance for TB Drug Development IMCI integrated management of infant and childhood GAVI Global Alliance for Vaccines and Immunization illness GDP gross domestic product IMF International Monetary Fund GET 2020 World Health Organization Alliance for the IMR infant mortality rate Global Elimination of Trachoma INCB International Narcotics Control Board xl | Abbreviations and Acronyms INDEPTH International Network of Field Sites with MR mental retardation Continuous Demographic Evaluation of MRI magnetic resonance imaging Populations and Their Health in Developing MSF Médecins Sans Frontières (Doctors Without Countries Borders) INFECTOM information, feedback, contracting with MTCT mother-to-child transmission providers to adhere to practice guidelines, and MVA modified vaccinia virus Ankara ongoing monitoring NAFTA North American Free Trade Agreement IPT intermittent preventive treatment NAP nonaffective psychosis IPTi intermittent preventive treatment in NCCAM National Center for Complementary and infancy Alternative Medicine IPV inactivated polio vaccine NCE new chemical entity IRB institutional review board NDP national drug policy IRR internal rate of return NGO nongovernmental organization IRS indoor residual spraying NHA national health account ISDR international strategy for disaster reduction NHS national health service ISIC international standard industrial classification of NIH National Institutes of Health all economic activities NIOSH National Institute for Occupational Safety ITN insecticide-treated net and Health IUATLD International Union against Tuberculosis and NIPA national income and product accounts Lung Disease NMR neonatal mortality rate IUD intrauterine device NO2 nitrogen dioxide IUGR intrauterine growth retardation NORA national occupational research agenda JE Japanese encephalitis NOx nitrogen oxide and nitrogen dioxide LAAM levo-alpha-acetyl-methadol NRA national regulatory authority LBW low birthweight NRT nicotine replacement therapies LDD learning and developmental disability NSAID nonsteroidal anti-inflammatory drug LDL low-density lipoprotein NSO national statistics office LE 20 life expectancy at age 20 NTD neural tube defect LF lymphatic filariasis OA osteoarthritis LIC low-income country OCP Onchocerciasis Control Program LMICs low- and middle-income countries ODA official development assistance LPG liquid petroleum gas OECD Organisation for Economic Co-operation LRI lower respiratory tract infection and Development LSD lysergic acid diethylamide OEPA Onchocerciasis Elimination Program MBB marginal budgeting for bottlenecks for the Americas MCE multi-country evaluation of IMCI effectiveness, OP osteoporosis cost, and impact OPV oral polio vaccine MCH maternal child and health ORS oral rehydration solution MDA mass drug administration ORT oral rehydration therapy MDG Millennium Development Goal PAHO Pan American Health Organization MDMA methylenedioxymethamphetamine PAL practical approach to lung health MDR-TB multidrug-resistant tuberculosis PARIS21 Partnership in Statistics for Development in the MDT multidrug therapy 21st Century MEASURE monitoring and evaluation to assess and use PCBs polychlorinated biphenyls results PCD Partnership for Child Development MIC middle-income country PCP Pneumocystis carinii pneumonia MMR measles-mumps-rubella PCR polymerase chain reaction MMT methadone maintenance treatment PCV protein-conjugated polysaccharide vaccine MMV Medicines for Malaria Venture PD Parkinson's disease MNCH maternal, neonatal, and child health PDOH Philippine Department of Health MOH ministry of health PDSA plan-do-study-act Abbreviations and Acronyms | xli PFGE pulsed-field-gel-electrophoresis TB tuberculosis PHC primary health care TCA tricyclic antidepressant PHSWOW public health school without walls TDR Special Programme for Research and Training in PLACE Priorities for Local AIDS Control Effort Tropical Diseases PM particulate matter TEHIP Tanzania Essential Health Interventions PMTCT prevention of mother-to-child transmission Program PopEd population and family life education THC tetrahydrocannabinol ppm parts per million TINP Tamil Nadu Integrated Nutrition Program PPPs public-private partnerships TLTI treatment for latent tuberculosis infection PRSC poverty reduction support credit TLV threshold limit value PRSP Poverty Reduction Strategy Paper TM traditional medicine PSV polysaccharide vaccine TRIPS Agreement on Trade-Related Aspects of PTA parent-teacher association Intellectual Property Rights PTCA percutaneous transluminal coronary angioplasty UN United Nations PTSD posttraumatic stress disorder UNAIDS Joint United Nations Programme on HIV/AIDS PZQ Praziquantel UNEP United Nations Environment Programme QALY quality-adjusted life year UNESCO United Nations Education, Scientific, and RA rheumatoid arthritis Cultural Organization R&D research and development UNFPA United Nations Population Fund RCT randomized clinical trial UNICEF United Nations Children's Fund RDI recommended dietary intake UNIDO United Nations Industrial Development RESU regional epidemiology and surveillance unit Organization RHD rheumatic heart disease URI upper respiratory tract infection RNA ribonucleic acid USAID U.S. Agency for International Development ROP retinopathy of prematurity VAD vitamin A deficiency RRT renal replacement therapy VC vital capacity RSV respiratory syncytial virus VCT voluntary counseling and testing RTI road traffic injury VERC village education resource center rt-PA recombinant tissue plasminogen activator VF ventilation factor SAFE surgery, antibiotics to control the infection, VIA visual inspection after application of an acetic facial cleanliness, and environmental acid solution improvements VL visceral leishmaniasis SAR search and rescue VOI value-of-information (techniques) SARS severe acute respiratory syndrome VSL value of a statistical life SBP systolic blood pressure WFP World Food Programme SCC short-course chemotherapy WHA World Health Assembly SD standard deviation WHO World Health Organization SiC significant caries (index) WHO/TDR WHO Special Programme for Research and SMA severe malarial anemia Training in Tropical Diseases SO2 sulfur dioxide WHOCC WHO Collaborating Center SP sulfadoxine-pyrimethamine WISE work improvement in small enterprises SSO social security organization WTO World Trade Organization SSRI selective serotonin reuptake inhibitor YF yellow fever STATCAP statistical capacity building YLD year of life lived with disability STH soil-transmitted helminth YLL year of life lost STI sexually transmitted infection YLS year of life saved SWAp sectorwide approach All dollar amounts are U.S. dollars unless otherwise indicated. xlii | Abbreviations and Acronyms Part One Summary and Cross-Cutting Themes · Summary · Cross-Cutting Themes Chapter 1 Investing in Health Dean T. Jamison A girl born in Chile in 1910 could expect to live only to age 33. Although the magnitude of possible gains in health was Since then, her life expectancy has more than doubled to its clear by the early 1990s, it is even clearer today: focused atten- current level of 78 years. What has this increase meant for her? tion by health systems on delivering powerful but often inex- The probability that she will die before her fifth birthday has pensive interventions can lead to dramatic improvements in declined from 36 percent to less than 2 percent. Throughout health at modest cost. Globalization has helped diffuse knowl- middle age the likelihood that she will die is also far lower: edge about what those interventions are and how health sys- death in childbearing or from tuberculosis (TB) as a young tems can deliver them. The pace of diffusion of such knowledge adult are no longer threats, and she is less likely to die in mid- into a country--much more than its level of income-- dle age from cancer. Mirroring this mortality reduction--but determines the pace of health improvement in that country. less easily quantified--are marked improvements in health- Our purpose in Disease Control Priorities in Developing related quality of life. She will be able to choose to have fewer Countries, 2nd edition (DCP2), is to help speed the diffusion of children and thus spend less time in pregnancy and child rear- policy-relevant knowledge. ing. From an average of about 5.3 children at midcentury, This introductory chapter to DCP2 serves two purposes: Chilean women's fertility rate has dropped to its current level · First, it provides the context for the rest of the book by dis- of 2.3. She will have fewer infections, less anemia, greater cussing broad trends in health conditions, by summarizing strength and stature, and a quicker mind. Her life is not only health conditions of the world at the dawn of the 21st cen- much longer; it is much healthier as well. tury, and by pointing to recent research suggesting that the Chile's history of health improvements is unusually well economic benefits from successful investments in health are documented but typifies changes that have occurred in much of likely to be exceptionally high. the world. These dramatic improvements in health have, more- · Second, it highlights some of the main messages for policy over, been possible without major increases in income. In the that emerge from the 37 chapters that deal with conditions early 1900s, income levels in the United States were roughly and risk factors and the 21 chapters that deal with strength- the same as they are in Chile today, yet U.S. life expectancy then ening health systems. These highlights are deliberately brief was 25 years shorter. New knowledge, new vaccines, and new because chapters 2 and 3 summarize the remainder of the drugs have inexpensively enabled major gains in health that book: chapter 2 summarizes findings about intervention were not possible before, even for those whose incomes were cost-effectiveness from across the book, and chapter 3 syn- high. Although those gains are now possible, they do not occur thesizes findings on strengthening health systems. unless health systems and policies effectively realize the avail- able potential. Box 1.1 summarizes the main messages of this chapter. 3 Box 1.1 Disease Control Priorities Chapters in this volume convey compact distillations of 3. Although health improvements constituted an enor- current knowledge concerning interventions to improve mous success for human welfare in the 20th century, health and the related delivery systems. Chapter 2 sum- four critical challenges face developing countries (and marizes main messages of the chapters dealing with the world) at the beginning of the 21st century: interventions, and chapter 3 summarizes the main mes- · high levels and rapid growth (for mostly demo- sages concerning health systems. Chapter 1 provides graphic reasons) of noncommunicable conditions context and conveys examples of the range of findings in the disease profiles of developing countries from across the volume. Here, in brief, are the main · the still unchecked HIV/AIDS pandemic messages of chapter 1: · the possibility of a successor to the influenza pan- demic of 1918 1. Average life expectancy in low- and middle-income · the persistence in many countries and many popu- countries increased dramatically in the past half- lation subgroups of high but preventable levels century, while cross-country health inequalities of mortality and disability from diseases such as decreased. In the countries with the best health indi- malaria, TB, diarrhea, and pneumonia; from cators, life expectancy increased a substantial two and micronutrient malnutrition; and, for both mothers one-half years per decade since 1960; low- and and infants, from childbirth. middle-income countries on average, with life expectancy gains of about five years per decade, The main purpose of this volume is to facilitate diffu- have been converging toward the countries with sion of appropriate approaches for addressing those the longest life expectancy. Improvement in average problems. income and education levels contributed to these 4. The volume's conclusions concerning interventions worldwide gains in health. Of much greater quantita- include the following: tive significance, however, have been the generation · Although 50 percent of deaths (including still- and diffusion of new knowledge and of low-cost, births) of children under age five occur at ages appropriate technologies. Increased access to knowl- younger than 28 days, relatively little attention has edge and technology has accounted for perhaps as been paid to this age group. Cost-effective interven- much as two-thirds of the impressive 2 percent per tions exist. year rate of decline in under-five mortality rates. · Treatment of HIV-positive mothers, treatment of 2. Improved health has contributed significantly to eco- sexually transmitted infections, free distribution nomic welfare. Per capita GNP rose rapidly in devel- of condoms, and other interventions can cost- oping countries in the decades following 1960, and effectively interrupt HIV transmission. These economic research suggests that health improvements preventive interventions continue to receive inade- led to perhaps 10 percent to 15 percent of that GNP quate attention from health systems and workers. growth. Although GNP includes the costs of provid- · Controlling tobacco use, particularly through taxa- ing medical care and reflects changes in health-related tion, is feasible in developing countries and is the consumption, such as the quantity and quality of single most important intervention for reducing food, it omits altogether the value that mortality noncommunicable disease. reduction represents for countries. Recent economic · Lifelong medical management of risk factors in research has extended measurement to a broader individuals at high risk for heart attacks or strokes, indicator, known as full income, that reflects reason- using aspirin and other drugs, is cost-effective and able valuation of changes in mortality. For many would benefit tens of millions of individuals. countries, recent mortality changes exceed in value 5. This volume's findings concerning health services and the growth of GNP. More widespread use of full- systems include the following: income measures to calculate the rate of return to · Provider incentives matter. Financial or other investments in health--and health research--will recognition for timely, responsive service increases almost certainly conclude that, today, most countries the likelihood of such services. Conversely, financial substantially undervalue those investments. incentives for excessive or inappropriate use of 4 | Disease Control Priorities in Developing Countries | Dean T. Jamison drugs or diagnostic tests is an all-too-common health in the 20th century. Every reason exists to cause of high costs and poor health outcomes. believe that continued progress--meeting the chal- · Provider experience matters. Having providers do a lenges of noncommunicable disease, HIV/AIDS, few things frequently, rather than attempting to potential pandemics, and neglected populations-- provide diverse services, facilitates quality improve- will also rely heavily on new knowledge. The rapidly ment with potentially major improvements in growing commitment of high-income countries to health outcomes. providing development assistance for health would be · Strengthening surgical capacity at district hospitals more effectively used if a larger share were devoted to is likely to be cost-effective and would address research and development. Public-private partner- broad needs. ships provide a promising institutional mechanism · In low-income countries, targeting the very limited for new product development. A particularly public sector resources for health to control of important--and much neglected--type of knowl- diseases--such as TB--that particularly affect the edge results from tight evaluations of interventions poor would be efficient. and systems. · In middle-income countries, public finance--or This volume represents an attempt to learn systematically publicly mandated finance--of a substantial pack- from the enormous successes of the past half-century in age of clinical care for all would be not only equi- improving human health. Knowledge that has been table but also efficient in terms of meeting health gained--and that this volume pulls together--creates a needs, controlling costs, and providing financial platform for addressing the problems that remain. protection to populations. 6. The generation and diffusion of new knowledge and products underpinned the enormous improvements in Source: Author. THE 20TH CENTURY TAKEOFF IN HUMAN HEALTH Life expectancy in years Percent 90 90 The 20th century differed markedly from previous history in 85 85 two critical domains: 80 Japan 80 75 75 Nordic countries 75% · First, the rapid economic growth that had begun in the 19th 70 70 50% century in countries of the North Atlantic diffused widely 65 25% 65 around the globe while continuing in the countries where it 60 60 New Zealand originated (DeLong 2000; Maddison 1999). 55 55 50 50 · Second, human mortality rates plummeted, and other Norway World 45 Estimates for population 45 dimensions of health improved dramatically. These changes 40 United Kingdom interquartile 40 also began in the North Atlantic countries in the 19th cen- range 35 35 tury but remained modest until the 20th century, during 30 Sierra Leone 30 which the rate of improvement increased and spread to Lower viability limit 25 25 most of the rest of the world (Easterlin 1996, 1999; Oeppen 20 20 and Vaupel 2002). 1550 1600 1650 1700 1750 1800 1850 1900 1950 2000 2050 Source: Oeppen 1999. Improvements in Health Figure 1.1 Trends in Maximum Female Life Expectancy, 1600­2000 This section briefly documents the magnitude of health improvements and then points to the challenges that remain. For the past 160 years, life expectancy in the healthiest countries life expectancy. From about 1600 to about 1840, there is fluctu- has increased steadily. At the same time, differences in life ation but no clear trend; after 1840, the graph turns upward at expectancy between those countries and much of the rest of the a surprisingly uniform rate of improvement: maximum life world have narrowed. Figure 1.1 depicts trends in female life expectancy increased by about two and one-half years per expectancy in the country with the highest estimated level of decade for 160 years. Investing in Health | 5 Table 1.1 Levels and Changes in Life Expectancy, 1960­2002, by World Bank Region Life expectancy (years) Rate of change (years per decade) Region 1960 1990 2002 1960­90 1990­2002 Low- and middle-income countries 44 63 65 6.3 1.7 East Asia and the Pacific 39 67 70 9.3 2.5 (China) (36) (69) (71) (11) (1.7) Europe and Central Asia -- 69 69 -- 0.0 Latin America and the Caribbean 56 68 71 4.0 2.5 Middle East and North Africa 47 64 69 5.7 4.2 South Asia 44 58 63 4.7 4.2 (India) (44) (59) (64) (5) (4.6) Sub-Saharan Africa 40 50 46 3.3 3.3 High-income countries 69 76 78 2.3 1.7 World 50 65 67 5.0 1.7 Source: World Bank 2004 (CD-ROM version). -- not available. Note: Entries are the average of male and female life expectancies. Table 1.1 shows progress in life expectancy by World Bank identified many specific examples of low-cost interventions region between 1960 and 2002. (Map 1 on the inside front leading to large and carefully documented health improvements cover depicts the World Bank regions.) For the first three (Levine and others 2004). The public sector initiated and decades of this period, progress was remarkably fast--a gain of financed virtually all of these interventions. The goal of this 6.3 years in life expectancy per decade on average, albeit with book is to assist decision makers--particularly those in the pub- substantial regional variation. Progress continued between lic sector--to realize the potential for low-cost intervention to 1990 and 2002 in the low- and middle-income countries but at rapidly improve the health and welfare of their populations. a much slower pace. This slower pace is due, in great part, to mortality increases from HIV/AIDS. Sub-Saharan Africa actu- Remaining Challenges ally lost more than four years of life expectancy. Four central challenges for health policy ensue from the pace Since 1950, life expectancy in the median country has and unevenness of the progress just documented and from the steadily converged toward the maximum and cross-country evolving nature of microbial threats to human health. differences have decreased markedly. This reduction in inequal- ity in health contrasts with long-term increases in income Epidemiological Transition. First, the next two decades will inequality between and within countries. Despite the magni- see continuation of trends resulting from the dramatic mortal- tude of global improvements, many countries and populations ity declines of recent decades. The key phenomenon is that have failed to share in the overall gains or have even fallen the major noncommunicable diseases--circulatory system behind. Some countries--for example, Sierra Leone--remain diseases, cancers, and major psychiatric disorders--are fast far behind (figure 1.1). China's interior provinces lag behind replacing (or adding to) the traditional scourges--particularly the more advantaged coastal regions. Indigenous people every- infectious diseases and undernutrition in children. This phe- where probably lead far less healthy lives than do others in their nomenon results in substantial part from rapid relative respective countries, although confirmatory data are scant. population growth at the older ages, when noncommunicable Reasons for remaining health inequalities lie only partially in diseases become manifest. Additionally, injuries resulting from income inequality: the experiences of China, Costa Rica, Cuba, road traffic are replacing more traditional forms of injury. Sri Lanka, and Kerala state in India, among many others, con- Using data from Chile, figure 1.2 illustrates the huge increase in clusively show that dramatic improvements in health can occur the relative importance of injuries, cancers, and cardiovascular without high or rapidly growing incomes. The experiences of disease between 1909 and 1999. Responding to this epidemio- countries in Europe in the late 19th and early 20th centuries sim- logical transition with sharply constrained resources is a key ilarly show that health conditions can improve without prior or challenge. Tables 1.A1 and 1.A2 (see annex 1.A) provide cause- concomitant increases in income (Easterlin 1996). A recent specific summaries of death and disease burden, measured in review, undertaken in part as background for this volume, DALYs, in 2001 for the world as a whole and for low- and 6 | Disease Control Priorities in Developing Countries | Dean T. Jamison Percentage distribution and within countries. Bourguignon and Morrisson (2002) have 35 stressed that global inequalities are declining if one properly 1909 1999 accounts for convergence across countries in health conditions, 30 which more than compensates for income divergence. However, in far too many countries health conditions remain 25 unacceptably--and unnecessarily--poor. This factor is a source of grief and misery, and it is a sharp brake on economic 20 growth and poverty reduction. From 1990 to 2001, for exam- 15 ple, the under-five mortality rate remained stagnant or increased in 23 countries. In another 53 countries (including 10 China), the rate of decline in under-five mortality in this period was less than half of the 4.3 percent per year required to 5 reach the fourth Millennium Development Goal (MDG-4) (see map 2 on the inside back cover of this book). Meeting the 0 MDG for under-five mortality reduction by 2015 is not remotely possible for these countries. Yet the examples of many Injuries Cancers infectious Diarrheal Respiratory other countries, often quite poor, show that with the right poli- infections diseases diseasesTuberculosis diseases CardiovascularOther cies dramatic reductions in mortality are possible. A major goal Cause of deaths of this volume is to identify strategies for implementing inter- Source: WHO 1999, 13. ventions that are known to be highly cost-effective for dealing Note: For 1909, 35.1 percent of deaths were categorized as "other," and for 1999, the with the health problems of countries remaining behind--for corresponding percentage was 17.5. The cause-specific percentages shown in the figure are the number from the indicated cause as a percentage of the total number example, treatment for diarrhea, pneumonia, TB, and malaria; classified into a specific cause for that year. immunization; and other preventive measures against a large Figure 1.2 Distribution of Deaths by Cause in Chile, 1909 and 1999 proportion of those diseases. middle-income countries as a group as well as for high-income THE ECONOMIC BENEFITS OF BETTER HEALTH countries. Those summaries indicate that noncommunicable disease already accounts for over half of all deaths in the low- The dramatic health improvements globally during the 20th and middle-income countries, although nearly 40 percent of century arguably contributed as much or more to improve- deaths continue to be from infection, undernutrition, and ments in overall well-being as did the equally dramatic innova- maternal conditions, creating a "dual burden" that Julio Frenk tion in and expansion of the availability of material goods and and colleagues have pointed to (Bobadilla and others 1993). services. To the substantial extent that appropriate investments in health can contribute to continued reductions in morbidity HIV/AIDS Epidemic. A second key challenge is the HIV/AIDS and mortality, the economic welfare returns to health invest- epidemic. Control efforts and successes have been very real but, ments are likely to be exceptional and positive--with pre- with only a few exceptions, limited to upper-middle-income viously unrecognized implications for public sector resource and high-income countries. Poorer countries remain in the allocation. These returns go far beyond the contribution better epidemic's deadly path. health makes to per capita income, which itself appears substantial (see Bloom, Canning, and Jamison 2004; Lopez- New Pandemics. The global influenza pandemic of 1918 Casasnovas, Rivera, and Currais 2005). This section first sum- resulted in more than 40 million human deaths, exceeding the marizes the evidence concerning health's effect on per capita 20th-century toll of HIV/AIDS or of World Wars I and II. income and then turns to more recent literature concerning the Continued evolution of the influenza virus leaves the world at effect of health changes on a broader measure of economic risk of another such pandemic--as has been much discussed in well-being than per capita gross domestic product (GDP). the press as this book goes to print. If the H5N1 strain of avian influenza, for example, evolved so that (like the human flu) it Health and Income could be efficiently transmitted from human to human, a major pandemic would be likely. Preparing for such an eventu- How does health influence GDP per person? Healthy workers ality is the third great challenge to global health. are more productive than workers who are similar but not healthy. Supporting evidence for this plausible observation Unequal Progress. A fourth key challenge results from con- comes from studies that link investments in health and nutrition tinued high levels of inequality in health conditions across of the young to adult wages (Strauss and Thomas 1998). Better Investing in Health | 7 health also raises per capita income through a number of other income growth in low-income countries than in high-income channels. One involves altering decisions about expenditures ones. Although attribution of causality is never unequivocal and savings over the life cycle. The idea of planning for retire- in analyses like these, different types of evidence point consis- ment occurs only when mortality rates become low enough for tently to a likely causal effect of health on growth. retirement to be a realistic prospect. Rising longevity in devel- Health declines can precipitate downward spirals, setting off oping countries has opened a new incentive for the current impoverishment and further ill health. For example, the effect generation to save--an incentive that can dramatically affect of HIV/AIDS on per capita GDP could prove devastating in the national saving rates. Although this saving boom lasts for only long run. An enormous waste of human capital occurs as one generation and is offset by the needs of the elderly after prime-age workers die. A high-mortality environment deters population aging occurs, it can substantially boost investment the next generation from investing in education and creating and economic growth rates while it lasts. human capital. The creation of a generation of orphans means Encouraging foreign direct investment is another channel: that children may be forced to work to survive and may not get investors shun environments in which the labor force suffers a the education they need. High rates of mortality may reduce heavy disease burden. Endemic diseases can also deny humans investment. Saving rates are likely to fall, and retirement access to land or other natural resources, as occurred in much becomes less likely. A foreign company is less likely to invest in of West Africa before the successful control of river blindness. a country with a high HIV prevalence rate because of the threat Boosting education is yet another channel. Healthier chil- to the firm's own workers, the prospect of high labor turnover, dren attend school and learn more while they are there. A and the loss of workers who have gained specific skills by work- longer life span increases the returns on investment in ing for the firm. The International Monetary Fund recently education. published a collection of important studies of the multiple Demographic channels also play an important role. Lower mechanisms through which a major AIDS epidemic can be infant mortality initially creates a "baby-boom" cohort and expected to affect national economies (Haacker 2004). leads to a subsequent reduction in the birth rates as families choose to have fewer children in the new low-mortality regime. A baby-boom cohort thereby affects the economy profoundly Health and Economic Welfare as its members enter the educational system, find employment, Judging countries' economic performance by GDP per person save for retirement, and finally leave the labor market. The fails to differentiate between situations in which health condi- cohorts before and after a baby boom are much smaller; hence, tions differ: a country whose citizens enjoy long and healthy for a substantial transition period, this cohort creates a large lives clearly outperforms another with the same GDP per per- labor force relative to overall population size and the potential son but whose citizens suffer much illness and die sooner. for accelerated economic growth (Bloom, Canning, and Individual willingness to forgo income to work in safer envi- Malaney 2000). ronments and social willingness to pay for health-enhancing If better health improves the productive potential of safety and environmental regulations provide measures, albeit individuals, good health should accompany higher levels of approximate, of the value of differences in mortality rates. national income in the long run. Countries that have high levels Many such willingness-to-pay studies have been undertaken of health but low levels of income tend to experience relatively in recent decades, and their results are typically summarized faster economic growth as their income adjusts. How big an as the value of a statistical life (VSL). Chapter 7 discusses these overall contribution does better health make to economic issues in the context of assessing the economic returns to growth? Evidence from cross-country growth regressions sug- investments in health research and development. gests the contribution is consistently substantial. Indeed, the Although the national income and product accounts initial health of a population has been identified as one of the include the value of inputs into health care (such as drugs and most robust and potent drivers of economic growth--among physician time), standard procedures do not incorporate infor- such well-established influences as the initial level of income mation on the value of changes in longevity. In a seminal paper, per capita, geographic location, institutional environment, Usher (1973) first brought the value of mortality reduction economic policy, initial level of education, and investments in into national income accounting. He did this by generating education. Bloom, Canning, and Sevilla (2004) found that one estimates of the growth in what Becker, Philipson, and Soares extra year of life expectancy raises GDP per person by about (2003) have called full income--a concept that captures the 4 percent in the long run. Jamison, Lau, and Wang (2005) value of changes in life expectancy by including them in an estimated that reductions in adult mortality explain 10 to assessment of economic welfare. Estimates of changes in full 15 percent of the economic growth that occurred from 1960 to income are typically generated by adding the value of changes 1990. Not all countries benefit equally from this link. Bhargava in annual mortality rates (calculated using VSL figures) to and others (2001) found that better health matters more for changes in annual GDP per person. These estimates of change 8 | Disease Control Priorities in Developing Countries | Dean T. Jamison in full income are conservative in that they incorporate only the Annual change as percentage of initial year GDP per capita value of mortality changes and do not account for the total 10 value of changes in health status. Valuation of changes in mor- Change in GDP per capita Change in full income tality, it should be noted, is only one element--albeit a quanti- tatively important one--of potentially feasible additions to 5 national account to deal with nonmarket outcomes. The U.S. National Academy of Sciences has recently proposed broad changes for the United States that would include but go beyond 0 valuation of mortality change (Abraham and Mackie 2005). For many years, little further work was done on the effects of mortality change on full income although, as Viscusi and Aldy (2003) document, the number of carefully constructed esti- 5 mates of VSLs increased enormously. Bourguignon and Morrisson (2002) address the long-term evolution of inequality among world citizens, starting from the premise that a "com- 10 1960­70 1970­80 1980­90 1990­2000 prehensive definition of economic well-being would consider individuals over their lifetime." Their conclusion is that rapid Source: Jamison, Sachs, and Wang 2001. increases in life expectancy in poorer countries had resulted in Figure 1.3 Changes in GDP and Full Income in Kenya, 1960­2000 declines in inequality (broadly defined) beginning sometime after 1950, even though income inequality had continued to rise. In another important paper, Nordhaus (2003) assessed the Jamison, Jamison, and Sachs (2003) have adapted standard growth of full income per capita in the United States in the 20th cross-country growth regressions to model determinants of full century. He concludes that more than half of the growth in full income (rather than GDP per person). Like Becker, Philipson, income in the first half of the century--and less than half in the and Soares (2003), they conclude that inequalities have been second half of the century--had resulted from mortality decreasing. decline. In this period, real income in the United States increased The dramatic mortality declines of the past 150 years--and sixfold and life expectancy increased by more than 25 years. their reversal in Africa by AIDS subsequent to 1990--have had Three lines of more recent work extend those methods to major economic consequences. The effect of health on GDP the interpretation of the economic performance of developing is substantial. The intrinsic value of mortality changes-- countries. All reach conclusions that differ substantially from measured in terms of VSL--is even more substantial. What are analyses based on GDP alone. Two of those studies--one the implications of these findings for development strategy and undertaken for the Commission on Macroeconomics and for benefit-cost analyses of public sector investment options? Health (CMH) of the World Health Organization (WHO) Using full income in benefit-cost analyses of investments in (Jamison, Sachs, and Wang 2001) and the other at the health (and in health-related sectors such as education, water International Monetary Fund (Crafts and Haacker 2004)-- supply and sanitation, and targeted food transfers) would assessed the impact of the AIDS epidemic on full income. Both markedly increase estimates of net benefits or rates of return. studies conclude that the AIDS epidemic in the 1990s had far A careful, quantitative reassessment of competing policies for more adverse economic consequences than previous estimates improving a country's living standards would probably con- of effects on per person GDP growth would suggest. clude that development assistance and budgetary allocations to Accounting for mortality decline in Africa before the 1990s, on health deserve greater relative priority. the other hand leads to estimates of much more favorable over- all economic performance than does the trend in GDP per per- son. Figure 1.3 shows that in Kenya, for example, full income WHY HAS MORTALITY DECLINED AT SUCH grew more rapidly in GDP per person before 1990 (and far DIFFERENT RATES IN DIFFERENT COUNTRIES? more rapidly in the 1960s). After 1990 the mounting death toll from AIDS appears to have only a modest effect on GDP per This section explores some of the reasons mortality has declined person but a dramatically adverse impact on changes in full so rapidly and at such different rates in different countries. It income. Becker, Philipson, and Soares (2003) extended the ear- considers the question of whether income levels or growth rates lier work of Bourguignon and Morrisson (2002) in finding play an important role in achieving better health or whether strong absolute convergence in full income across countries good policies can potentially lead to good health for low-income over time, in contrast to the standard finding of continued populations. The section concludes with a snapshot of health divergence (increased inequality) of GDP per person. Finally, conditions in the world at the dawn of the 21st century. Investing in Health | 9 The 20th century witnessed huge and unprecedented methods of disease treatment that can be applied at reasonable declines in mortality rates at all ages and in most parts of the cost. The reduction was rapid because it did not depend on world. Easterlin (1996) and Crafts (2000) place an emphasis on general economic development or social modernization (Davis mortality transformation that is comparable to their emphasis 1956, 306­7, 314). Some strands of the literature, however, on economic growth in their retrospectives on the unprece- attribute the high correlation of income and life expectancy at dented changes in the human condition during the 20th any given time to a significant causal effect of income on health century. Understanding the sources of mortality changes is (see, for example, Pritchett and Summers 1996). important for understanding one of the defining events of Background work for this volume (Jamison, Sandbu, and world history and also for devising policies to address the needs Wang 2004) attempted to provide a better sense of the impor- of the perhaps 25 percent of the world's population whose mor- tance of income as a determinant of mortality by exploring the tality rates remain far higher than those of the rest of humanity. relationships among income, technical progress (or diffusion), Several approaches shed light on the sources of mortality and mortality decline. Previous econometric research either decline. Epidemiologists and demographers have carefully has given little emphasis to technical progress--in part simply tracked specific communities for many years to assess levels of because much of the research is cross-sectional and therefore mortality and causes of death. In rural Senegal, rapid mortality fails to address developments over time--or has assumed the decline followed introduction of interventions addressing spe- rate of technical progress or technology adoption to be con- cific conditions (Pison and others 1993). stant across countries. The background work for this volume Another approach is historical. Easterlin (1996, 1999) exam- relaxed the assumption that the rate of technology adoption is ined the interplay of economic growth, urbanization, and constant across countries. Allowing for cross-country variation mortality in 19th- and 20th-century Europe. He concluded that in the rate of adapting new methods resulted in weaker esti- although income growth in the 19th century probably did play mated effects of income on infant mortality rates than previ- a role in reducing mortality (through its influence on food ously found, although education's estimated effect was robust availability and environmental conditions), the magnitude of with respect to this change. the effect was small. Fogel (1997) stressed the importance of Much of the variation in country outcomes results from the increases in food availability during this period. Positive effects very substantial cross-country variation in the rate of technical of income growth were partially offset by increased infectious progress--from essentially no decline in infant mortality rate disease transmission resulting from urbanization. Easterlin caused by technical progress to reductions of up to 5 percent (1999) concludes that 20th-century mortality decline, which per year from that source. Deaton (2004) provides a comple- was much more rapid than that of the 19th century, had its ori- mentary and extended discussion of the importance of techno- gin in technical progress, and Powles (2001) has pointed to the logical diffusion for improvements in health. Many factors importance and nature of the institutional changes required to from outside the health sector also affect the pace of health translate technical change and economic improvements into improvement; the education levels of populations are most mortality reduction. Mosk and Johansson's (1986) assessment important. Box 1.2 briefly discusses the multisectoral nature of of the interplay between income and mortality in Japan illus- health's determinants. The importance of technical progress trates the role that adoption of public health knowledge and and diffusion should be viewed in this larger context. institutional development played in mortality decline in the However technical progress or diffusion may be manifested, country that now has the world's lowest mortality rates. the large differences in its magnitude across countries suggest Most analysts agree that advances in science and technology important effects of a country's health-related policies (Fuchs have underpinned the 20th-century transformations both of 1980; Oeppen 1999). This point bears reiterating in a slightly dif- income and of mortality levels. Models of economic growth ferent way: income growth is neither necessary nor sufficient for rely heavily on technological progress to account for economic sustained improvements in health. Today's tools for improving change (Boskin and Lau 2000; Easterly and Levine 1997; Solow health are so powerful and inexpensive that health conditions can 1957). Preston (1975, 1980) and Fuchs (1974) provided early be reasonably good even in countries with low incomes. quantitative assessments of the central importance of technical progress in accounting for 20th-century increases in life CHILD HEALTH expectancy. [Economists use the term technical progress to denote advances in knowledge that lead to new products, like A small number of conditions accounts for most of the (large) vaccines, or that can inform behavior change, like knowledge of differences in health between the poor and the not so poor. the germ theory of disease (Preston and Haines 1998).] Davis Less than 1 percent of all deaths from AIDS, TB, and malaria, (1956) had already concluded that the unprecedented reduc- for example, occur in the high-income countries. Available tion in mortality in underdeveloped areas since 1940 is the technical options--exemplified by but going well beyond result primarily of the discovery and dissemination of new immunization--can address most of the conditions that affect 10 | Disease Control Priorities in Developing Countries | Dean T. Jamison Box 1.2 The Multisectoral Determinants of Health Malnourished children easily acquire diseases, and they · risky sexual activity--5.3 percent (5.1 percent) easily die from the diseases that they acquire. Dwellings · alcohol use--3.6 percent (3.4 percent). and neighborhoods without sanitation provide fertile Underlying most proximal risks are more general environments for transmission of intestinal infections. determinants of health, such as education and, to a lesser Cooking with wood and coal results in air dense with extent, income. The effects of income and education oper- particulates and gases, which destroy lungs and lives. ate for the most part through influencing risk (and per- Hopeless life circumstances thrust young girls (and boys) mitting effective use of health services). If an important into commercial sex work with its attendant risks of vio- fraction of ill health results from poverty and low educa- lence and sexually transmitted infections, including tional levels--or from their consequences in inadequate HIV/AIDS. Manufacturers of tobacco and alcohol profit food or sanitation or other specific risks--then ought the enormously from advertising and promotion that spread task of the health professional lie principally in addressing addiction. Rapid growth in vehicular traffic--often with these underlying problems? In one sense, the answer is untrained drivers on unsafe roads--generates a rising toll surely yes: the health community should measure the of injury. Poorly designed irrigation creates breeding effects on health of actions outside the health sector. It grounds for vectors of disease. The point is clear: determi- should ensure that these findings are communicated and nants of health are truly multisectoral. are considered by those making policy choices. The mag- WHO coordinated a group of more than 100 individu- nitude of the demonstrated effect of girls' education on als to generate estimates of the percentage of deaths, by health and fertility outcomes, for example, provides one region and globally, associated with a range of 26 risk fac- powerful argument for investing in expansion of educa- tors (Ezzati and others 2004). Those estimates were tional access to girls. Millions of premature deaths, to take revised and updated for the Disease Control Priorities another example, could be averted in Africa alone in the Project. The results give a sense of the extent to which next quarter century with appropriate policies toward multisectoral factors contributed to mortality and disease supply of energy for household use (Bailis, Ezzati, and burden in low- and middle-income countries in 2001. The Kammen 2005). It is essential that the health sector docu- following, for example, are estimates of the percentage of ment and advocate opportunities such as these. disease burden (and, in parentheses, of deaths) in those The health community has limited capacity for direct countries attributable to the indicated risk factors: action outside the health sector, however. It will make more · tobacco smoking--4.7 percent (8.5 percent) of a difference if it focuses its energy, expertise, and re- · indoor air pollution--2.7 percent (3.2 percent) sources on ensuring that health systems efficiently deliver · inadequate water and sanitation--3.4 percent (2.8 per- the powerful interventions provided by modern science. cent) Source: Author. Note: The estimates reported here of DALYs and deaths that are attributable to various risk factors come from Ezzati and others (2006). children, and can do so with great efficacy and at modest cost. ambitious. Yet its implication of an average 4.3 percent per That short list of conditions, including undernutrition, relates year decline is well within recent experience. In the first directly to achieving the MDGs for health. Public expenditures half of the MDG period (1990­2002), 46 countries achieved to address those conditions have, in the past, benefited the rel- rates of decline in under-five mortality greater than 4.3 per- atively well off, albeit within poor countries (although global cent per year. Figure 1.4 displays trends in the rate inequities have decreased because many poor countries have of decline in under-five mortality relative to the requisite made much progress). 4.3 percent per year for China, India, Latin America and the Caribbean, and Sub-Saharan Africa. Africa's slowed Under-Five Health Problems and Intervention Priorities progress probably stems mostly from HIV/AIDS and the MDG-4 for under-five mortality (reducing its level in 2015 spread of resistance to previously effective and widely used by two-thirds relative to what it was in 1990) is highly antimalarial drugs. Map 2 (on the inside back cover of this Investing in Health | 11 Rate of decline of under-five mortality rates (percent) Under-five deaths per 1,000 births 7 160 151 1960s 1970s 1980s 1990s 1990 2001 6 140 122 120 5 100 4 80 3 60 2 38 40 26 1 20 12 3 0 0 China India MDG Sub-Saharan Latin America AIDS Malaria Other causes requirement Africa and the Caribbean Source: Lopez, Begg, and Bos 2006, table 2.4. Source: Calculations based on data in World Bank 2004 (CD-ROM version). Figure 1.5 Under-Five Deaths from AIDS, Malaria, and Other Note: The black bar in the center shows the 4.3 percent per year rate of decline Causes, per Thousand Births, 1990 and 2001, Sub-Saharan Africa required for the period 1990­2015 to meet MDG-4 of reducing under-five mortality by two-thirds. Figure 1.4 Rate of Progress in Reducing Under-Five Mortality, 1960­2000: China, India, Latin America and the Caribbean, and leading to an even greater rise in malaria mortality and mor- Sub-Saharan Africa bidity that could be substantial. Figure 1.5 illustrates increases in malaria death rates in under-five children in Sub-Saharan Africa in the period from 1990 to 2001. The design of instru- book) shows country-specific progress in reducing under-five ments for financing a rapid transition to effective new mortality: treatments--artemisinin combination therapies (ACTs)--is a high priority (chapter 21; Arrow, Gelband, and Jamison 2005). · Countries colored in green experienced annual rates of The other intervention priorities for addressing under-five decline greater than 4.3 percent in the first half of the MDG mortality are for the most part familiar: period (1990­2002). · Countries colored in red saw no decrease (or an increase) in · Expand immunization coverage. their under-five mortality. · Expand the use of the simple and low cost but highly effec- · Countries colored in yellow and orange depict countries in tive treatments for diarrhea and child pneumonia through between--with yellow indicating performance in the top half integrated management of childhood illness or other of the range between 0 and 4.3 percent, and orange indicat- mechanisms. ing poorer performance in the bottom half of the range. · Prevent transmission of and mortality from malaria by expanding coverage of insecticide-treated bednets, by Basic knowledge about the cost-effectiveness of interven- expanding use of intermittent preventive treatment for tions to address maternal and child health has been available pregnant women, and, particularly, by financing the adop- from the 1980s. DCP2's work provides a reassessment with few tion of ACTs to replace the now widely ineffective drugs surprises but some additions. It makes two important relatively chloroquine and SP. new points. The first results from noting that half of under-five · Ensure widespread distribution of key micronutrients. deaths occur at ages less than 28 days, when the substantial · Expand the use of a package of measures to prevent mother- but usually neglected problem of stillbirth is considered. DCP2 to-child transmission of HIV (further discussed in the next identifies some highly cost-effective approaches to interven- section on HIV/AIDS). tion against stillbirth and neonatal death (chapter 27). The second new point results from the rapid spread of resistance In addition to interventions to reduce under-five mortality, of the malaria parasite to chloroquine and sulfadoxine- one other priority is clear. The world's most prevalent infec- pyrimethamine (SP). These inexpensive, highly effective, tions are intestinal helminth (worm) infections, and children of widely available drugs provided an important partial check on all ages are among the most heavily affected. Chapter 24 the high levels of malaria mortality in Africa. Their loss is discusses these infections, which a low-cost drug (albendazole), 12 | Disease Control Priorities in Developing Countries | Dean T. Jamison taken every six months to a year, can control effectively. in buying interventions, in buying out of prevailing system Chapter 58 on school health services points to both the impor- constraints, and in investing in relevant system capacity for the tance to children's school progress of taking albendazole where future? What needs to be constantly borne in mind throughout needed and the potential efficacy of school health programs as this continued controversy is that something works: under-five a vehicle for delivery. In the long run, improved sanitation and mortality rates have plunged by more than half since 1960 in water supplies will prevent transmission. Use of albendazole is the low- and middle-income countries. only an interim solution, but it is one that may be required for decades if the experience of the currently high-income coun- HIV/AIDS tries is relevant. For dozens of countries around the world--including several Delivering Child Health Interventions of the most populous--the AIDS epidemic threatens every The list of potential interventions is far from exhaustive, and aspect of development. No other threat comes close, with the different regions, countries, and communities will face possible exceptions of use of nuclear weapons in densely pop- different mixes of the problems these interventions address. ulated areas or a devastating global pandemic similar to the However, there can be little dispute that any short list of 1917­18 influenza episode. Most governments of affected low- intervention priorities for under-five mortality in low- and and middle-income countries and most providers of develop- middle-income countries would include many on the list in the ment assistance have only recently begun to respond more preceding section. Why not, then, simply put money into scal- than minimally. Creation of the Global Fund to Fight AIDS, ing up these known interventions to a satisfactory level? Tuberculosis, and Malaria can be viewed as an attempt of the To greatly oversimplify--and these issues are discussed world's top political leaders to improve on the records of exist- more substantially in chapter 3--two schools of thought exist. ing institutions. The Global Fund's initial years have seen One line of thinking--often ascribed to macroeconomist substantial success, but that success is potentially undermined Jeffrey Sachs and his work as chair of the WHO CMH-- by sharp constraints on resource availability (Bezanson 2005). concludes that more money and focused effort are the solu- tions. Although acknowledging dual constraints--of money Tools to Control the Epidemic and of health system capacity--Sachs and his colleagues In contrast to the initially slow programmatic movement of (WHO CMH 2001; Sachs 2005) contend that money can buy most national leaders and international institutions, the (or develop, or both) relevant system capacity even over a peri- research and development community--public and private-- od as short as five years. Major gains are affordable and health has made rapid progress in developing tools to control the system capacity constraints can be overcome. Immunization HIV/AIDS epidemic, although both a vaccine and a curative provides an example of where, even in the short term, money drug remain distant objectives. Sensitive, specific, and inexpen- can substitute for system capacity. Adding antigens for sive diagnostics are available; means of prevention have been Haemophilus influenzae type B (Hib) and hepatitis B (HepB) to developed and tested; modes of transmission are well under- the immunization schedule is costly (although still cost- stood; and increasingly powerful drugs for controlling viral effective). In some environments, however, it proves less load allow radical slowing of disease progression. Tools for demanding of system capacity than expanding coverage does. dealing with HIV/AIDS are thus available: As emphasized in Money can be effectively spent by adding antigens at the same chapter 18, a number of countries show by example that those time as investing in the capacity to extend coverage. tools can be put to effective use. Most of the high-income A second school of thought acknowledges the need for more countries have done so, and Brazil and Mexico provide exam- money but asserts that health system capacity is often a bind- ples of upper-middle-income countries that have forestalled ing short- to medium-term constraint on substantial scaling potentially serious epidemics. Mexico succeeded, for example, up of interventions. Critical priorities are, therefore, system with a policy of responding both early and forcefully to the reform and strengthening while ensuring that such reforms epidemic (del Rio and Sepúlveda 2002). The major successes of focus clearly on achieving improved health outcomes and Thailand and Uganda demonstrate that countries with fewer financial protection. financial resources can also succeed--and succeed against Chapter 3, as indicated, discusses these issues further in the more established epidemics that had already penetrated deeply context of all the problems facing a health system, and into their populations. chapter 9 provides a thoughtful assessment of how to overcome the constraints facing achievement of the MDGs for health. From an individual country's perspective, however, if financial Prevention and Management resources are available, the question is very much an empirical Prevention underpins success. At the time the World Bank's one: to what extent can those resources be effectively deployed World Development Report: Investing in Health (World Bank Investing in Health | 13 1993) was being written in 1992 and 1993, the only tool for resistant mutants that undermined the efficacy of therapy. In a dealing with the epidemic was prevention. In collaboration remarkably short time, scientific advances have substantially with the then­Global Programme against AIDS at WHO, the attenuated those problems, making feasible, at least in princi- World Development Report commissioned very approximate ple, antiretroviral therapy in low-income settings. WHO's "3 by estimates of the consequence for the new infection rate of 5" program had as its objective, for example, to reach 3 million fully implementing available preventive measures (its opti- people in low- and middle-income countries with antiretrovi- mistic case scenario) or of doing very little (worst case). ral therapy by 2005. Although that goal was far from being met, Actual incidence numbers for 2000, unfortunately, fall very the global effort to make treatment widely available is well close to the worst-case projection, and chapter 18 points out under way. that even by 2003 fewer than one in five people at high risk of Despite the indicated progress against the problems infection had access to the most basic preventive services. In with antiretroviral drugs, challenges to their effective use in much of the world, little has been spent on prevention, and low-income environments remain formidable. The complexity little has been achieved. In addition, the current U.S. admin- of patient management is very real. Management requires istration may be partially responsible for discouraging con- high levels of human resources and other capacities in many dom use in some countries and in stigmatizing and alienating of the countries where those capacities need to be most commercial sex workers who are particular priorities for pre- carefully rationed. Perhaps in consequence, achieving effective vention programs. Despite those problems, the potential for implementation has been difficult on even a limited scale. prevention is very real, and a number of successful countries Chapter 18 reviews those problems and how they might be have shown the possibility of using that potential well. addressed. Chapter 17 on sexually transmitted infections (STIs) and Three points concerning widespread antiretroviral drug use chapter 18 on AIDS discuss a broad menu of preventive mea- are particularly noteworthy: sures and experiences with their implementation. Among them, treatment of STIs may be of particular salience both · Poor implementation (low adherence, development of because the diseases are well worth treating in their own right resistance, interruptions in drug supplies) is likely to lead to and because the absence of STIs greatly reduces transmission very limited health gains, even for individuals on therapy. of HIV. (This outcome is unlike that of a weak immunization pro- In addition to prevention, better management of patients gram in which health gains still exist in the fraction of the with AIDS could avert much misery, both by treating oppor- population that is immunized.) Poorly implemented anti- tunistic infections and by ameliorating the often excruciating retroviral drug delivery programs could divert substantial pain associated with many AIDS deaths. Medically inappropri- resources from prevention or from other high-payoff activ- ate restrictions on the use of inexpensive but powerful opiates ities in the health sector. Even worse, they could lead to a for pain control continue to deny dignity and comfort to mil- false sense of complacency in affected populations: evidence lions of patients with AIDS and cancer in their final days from some countries suggests that treatment availability has (chapter 52). led to riskier sexual behavior and increased HIV transmis- sion. The injunction to "do no harm" holds particular salience. Antiretroviral Treatment · Unless systematic efforts are made to acquire hard knowl- Intensive research and development efforts have led in the past edge about which approaches work and which do not, the decade to the availability of well over a dozen antiretroviral likelihood exists that unsuccessful implementation efforts drugs that can greatly reduce the quantity of HIV in an infected will be continued without the appropriate reallocation of person. This reduction in viral load slows or halts progression resources to successful approaches. Learning what works of AIDS and can return individuals from serious illness to rea- will require major variations in approach and careful evalu- sonable health. Available drugs leave a residual population of ation of effects. Failing to learn will lead to large numbers of HIV in the body, however, and this population grows if the needless deaths. Most efforts to scale up antiretroviral ther- drugs are stopped. At present the drugs must be taken for life. apy unconscionably fail to commit the substantial resources Widespread use of these drugs in high-income (and some required for evaluation of effects. Such evaluations are middle-income) countries has transformed the life prospects of essential if ineffective programs are to be halted or effective HIV-infected individuals. ones are to receive more resources. Early generation antiretroviral drugs suffered notable short- · Many programs rely exclusively on the cheapest possible comings: they were enormously costly; regimens for their use drugs, thereby risking problems with toxicity, adherence, were complicated, making adherence difficult; their use gener- and drug resistance. From the outset a broader range of ated unpleasant side effects; and rapid evolution of HIV led to drug regimens needs to be tested. 14 | Disease Control Priorities in Developing Countries | Dean T. Jamison NONCOMMUNICABLE DISEASE AND INJURY Tobacco deaths worldwide in the indicated quarter century (millions) 350 At the same time that most low- and middle-income countries Baseline If proportion of young 300 need to address health problems that are now effectively con- 300 adults taking up smoking 280 trolled in high-income countries, they are increasingly sharing halves by 2020 the high-income countries' heavy burdens of cardiovascular 250 If adult consumption halves by 2020 system disease (chapters 33, 44, and 45); cancers (chapter 29); 200 psychiatric disorders (chapter 31); and automobile-related injuries (chapter 39). The public health research and policy 150 148 150 150 community has been surprisingly silent about these epidemics 120 even though, for example, cardiovascular disease (CVD) in 100 low- and middle-income countries killed over twice as many people in 2001 as did AIDS, malaria, and TB combined 50 50 (table 1.A1). An important early exception is Feachem and oth- ers (1992), who indicated approaches to treatment and preven- 0 tion of these conditions that can be adapted to the tighter 1975­2000 2000­25 2025­50 Time interval budget constraints of developing countries. The World Health Organization provides a valuable and more up-to-date discus- Source: Chapter 46, figure 46.2. Adapted from Jha and Chaloupka 2000; Peto and Lopez 2001. sion that emphasizes prevention (WHO 2005). In addition, low-cost but effective approaches to long-term management of Figure 1.6 Increase in Tobacco-Related Deaths as Populations Age chronic conditions need to be developed and implemented. The remainder of this section briefly discusses, as examples, the prevention and management of cardiovascular diseases, starting. Reducing smoking levels is well demonstrated to be psychiatric disorders, and injuries. within the control of public policy. The principal instrument is through taxation: Complementary measures discussed in chapter 46 are important as well. Cardiovascular Disease The main risk factors for CVD account for very large frac- Cardiovascular diseases in low- and middle-income countries tions of the deaths (and even more of the burden) from those result in about 13 million deaths each year, over a quarter of all diseases. For ischemic heart disease, they collectively account deaths in those countries. Most cardiovascular deaths result for 78 percent of deaths in low- and middle-income countries; from ischemic heart disease (5.7 million) or cerebrovascular for stroke, they account for 61 percent (Ezzati and others 2006). disease (4.6 million). Because such deaths occur at older ages, Measures to reduce the levels of those risk factors--high blood they account for a substantially smaller fraction of total disease pressure, high cholesterol, smoking, obesity, excessive alcohol burden in disability-adjusted life years (DALYs)--12.9 per- use, physical inactivity, and low fruit and vegetable cent--than they do of deaths (table 1.A2). consumption--are the goals for prevention. Unlike the favor- Growing tobacco use accounts for a substantial and avoid- able experience with controlling tobacco use, attempts to able fraction of CVD and of cancers. Reasonable projections change the behaviors leading to obesity, hypertension, high show the number of tobacco-related deaths to be not only large cholesterol, or physical activity appear to have had little success but also growing, particularly in developing countries. In 2000, at a population level. However, as chapter 44 documents, many the number of tobacco-related deaths in developing countries promising approaches remain to be tried. Common sense sug- about equaled the number in high-income countries; projec- gests that they should be initiated even while more systematic tions suggest that by 2030 developing countries will have more efforts to develop and evaluate behavior-change packages are than twice as many. For those reasons, controlling smoking is ramped up. a key element of any national strategy for preventing CVD or Pharmaceutical interventions to manage two major compo- for promoting health more generally. Preventing the initiation nents of cardiovascular risk--hypertension and high choles- of smoking is important because addiction to tobacco makes terol levels--are well established and are highly cost-effective smoking cessation very difficult, even for the numerous for individuals at high risk of a stroke or heart attack. From at individuals who would like to do so. However, helping people least the time of publication of Disease Control Priorities in quit smoking is at least as important as preventing initiation. Developing Countries, 1st edition (DCP1), researchers have rec- Figure 1.6 portrays estimates showing that far more lives could ognized that the low cost and high effectiveness of drugs to be saved between now and 2050 with successful efforts to help prevent the reoccurrence of a cardiovascular event made their people stop smoking than with efforts to keep them from long-term use potentially cost-effective in low-income Investing in Health | 15 environments (Pearson, Jamison, and Trejo-Gutierrez 1993). Injuries Even if sustained behavior change proves difficult to achieve, Injuries constitute an additional major and neglected compo- medications have the potential to reduce CVD risks by 50 per- nent of disease burden in developing countries. This volume's cent or more. Chapters 33 and 45 develop the current chapters on injury (chapters 39 and 40) emphasize prevention. evidence on that point. A key problem, however, concerns the Timely treatment is also important, and chapters 67 (on sur- health care personnel and systems requirements associated gery) and 68 (on emergency medical services) point to the with the need for lifelong medication use, a problem also faced potential, at low cost, for much better treatment of injury vic- with antiretroviral therapy for AIDS and the use of medica- tims than is typical today. tions to target several major psychiatric disorders. How to The great diversity of both causes and consequences of achieve effective long-term management of lifesaving drugs injury precludes an attempt in this chapter to do more than is a key delivery and research challenge for health system highlight their importance. Chapter 2 and the injury-related reformers. chapters just mentioned provide a rich menu of practical In contrast to the lifelong requirement for drug use associ- options. It is worth pointing out here the central importance of ated with CVD risk reduction in high-risk individuals, treat- two specific categories of injury--road-traffic injuries (1.07 mil- ment of acute heart attacks with inexpensive drugs is both less lion deaths in 2001 in low- and middle-income countries) and demanding of system resources and highly cost-effective (chap- suicides (0.75 million deaths). Safer roads, safer driving, safer ter 45). Given the high incidence of these problems, systemwide vehicles, and better emergency care have sharply reduced the efforts to achieve high rates of appropriate drug use in response toll from road-traffic injuries in high-income countries, but to acute heart disease are a high priority. unless dramatic action is taken in developing countries, the toll will surely rise. Although there has been less success in reduc- Psychiatric Disorders ing suicide rates, the improved treatments now available for psychiatric disorders are proving to be one important approach Although neurological and psychiatric disorders lead to only in suicide prevention. about 1.4 percent of deaths in low- and middle-income coun- This discussion of noncommunicable diseases and injury tries (1.8 percent in high-income countries), they cause suffer- highlights the huge and growing burden from those conditions ing and disability far beyond what the mortality numbers sug- and conveys a consistent message that constructive action is gest. About 10 percent of disease burden in DALYs in low- and feasible at relatively modest cost. No attempt has been made to middle-income countries results from these conditions; three be comprehensive (chapters 29 through 38 all deal with non- major psychiatric diseases--unipolar major depression (3.1 communicable diseases); rather the discussion points to the percent of DALYs), bipolar disorder (0.6 percent), and schizo- need for health systems to systematically incorporate effective phrenia (0.8 percent)--account for much of it (table 1.A2). responses to noncommunicable diseases and injuries as their Chapter 31 provides a concise overview of advances made in capacities grow. recent years in treating these conditions (as well as panic disor- der), summarizes information on their burden, and develops estimates of the cost-effectiveness of drug-based and cognitive HEALTH SYSTEM DEVELOPMENT AND FINANCE behavioral therapies in different settings (hospital based and community based). Although the cost-effectiveness estimates DCP1 focused principally on intervention priority. What mix reported in chapter 31 suggest interventions are only moder- of public health and clinical interventions would best respond ately cost-effective, the authors suggest that a fuller analysis of to important disease conditions in highly resource constrained benefits than is captured by a health metric such as the DALY environments? Given the results of those assessments, where would justify substantial investments. They analyze a basic were the most important overall best buys? Where were package of mental health services that could provide a practical resource commitments likely to be of low value? DCP2 returns vehicle for providing these interventions in environments with to those questions but goes beyond them in assessing the steps tightly constrained financial and implementation resources. A required for strengthening of health services and systems in continuing theme in this volume--and one of particular rele- ways that will allow the appropriate mix of interventions to be vance here--is that without careful evaluations of the effects of delivered equitably and well. De Savigny and others (2004) alternative approaches to large-scale intervention against psy- describe a specific example from Tanzania that links system chiatric disorders, the world will fail to develop hard knowledge reform to intervention selection. of what does--and does not--work. Without that knowledge, Part 3 of the volume addresses strengthening of health sys- far less health and financial security will be gained than is tems. For valuable discussions of the goals of health systems, potentially possible from the inevitably limited resources see WHO's World Health Reports for 1999 and 2000 (WHO available. 1999, 32­33; 2000, 23­25) and Roberts and others (2003). 16 | Disease Control Priorities in Developing Countries | Dean T. Jamison Part 3 first reviews options for public health services with This overview of the topics on health systems provides a chapters on surveillance and information (chapters 53 and 54), sense of the breadth of the issues considered. Chapter 3 pro- drug resistance (chapter 55), community health and nutrition vides a concise and integrated statement of the main findings. programs (chapter 56), contraception (chapter 57), school- The remainder of this section deals briefly with assessing based health (chapter 58), adolescent health (chapter 59), the performance of health systems and with the key issue of occupational health (chapter 60), natural disaster relief (chap- finance. Before we turn to those topics, however, it is worth ter 61), and disease elimination and eradication (chapter 62). highlighting several particular points. A major point implied by the simple number of chapters First, in low-income countries, limited health system capac- devoted to public health is that health system strengthening and ity has sometimes led governments (and development assis- reform efforts need to commit substantial financial resources tance agencies) to focus their capacity on a few high-priority and political and managerial attention to public health. items--such as immunization or control of HIV/AIDS. The A second cluster of chapters in part 3 deals with strengthen- objective may be a reasonable one: a greater reduction in dis- ing personal health services. The first of those chapters deals ease burden in the population and more financial protection with an important facet of community-level health services, the for it are likely to be achieved by doing a few important things integrated management of the sick child (chapter 63). Chapters well than by doing many things poorly. Yet if this focused effort 64 to 66 deal with levels of care: general primary care, the dis- is undertaken by establishing vertical structures outside the trict hospital, and the referral hospital, respectively. Three chap- health system, then important opportunities for increasing ters address services offered at multiple levels of the system: sur- capacity may be missed. Chapter 3 stresses a critical point: a gery (chapter 67), emergency medical services (chapter 68), and focused program should be designed so that it contributes to, complementary and alternative medicine (chapter 69). The rather than detracts from, long-term system strengthening. final cluster of four chapters addresses capacity strengthening Second, quality of care is important; it can be measured, and and management reform: quality of care (chapter 70), the it can be improved (box 1.3). Third, providing basic surgical health workforce (chapter 71), supplies of drugs and vaccines services, particularly at the district hospital level appears to (chapter 72), and management of clinical services (chapter 73). offer major but neglected opportunities for addressing Box 1.3 Tangible Approaches to Improving Quality of Care A 2001 report from the Institute of Medicine of the U.S. · Use evidence-based criteria to link quality of care to out- National Academy of Sciences (Institute of Medicine comes. This approach can be implemented by training 2001b) highlights great variation in the quality of clinical and creating incentives for adapting clinical guidelines or care in the United States. Its publication catalyzed reform by using the collaborative improvement model. efforts. In a recent evaluation, Leape and Berwick (2005) · Improve system-level and provider incentives. found that those reform efforts had a major effect on pro- Minimally, do no harm with the structure of financial fessional attitudes and organizational culture, although incentives facing providers, for example, by establishing less effect, so far, on mortality. Chapter 70 on quality of a legal and ethical environment where care providers care documents the similarly large variation in quality in do not profit personally from sale of drugs, diagnostic low- and middle-income countries and the associated cost procedures, or referrals to expensive specialized care. in lives and money. Improving quality of care amplifies · Emphasize high-volume care for selected surgical pro- the effect of investments in health. Promising approaches cedures and prevalent medical conditions. Such an in improving the quality of care include the following: approach can lead to higher quality and lower cost even while, in some cases (for example, cataract removal), · Invest in measuring quality and feeding that informa- allowing lower-level workers to substitute for more tion back into the system. This approach has been expensive and scarcer physicians. shown to be possible (for example, clinical vignettes) and effective. Source: This box was prepared with input from John Peabody. Investing in Health | 17 significant sources of disease burden. An important substantive ance of countries is better understood even though relating component of health sector reforms should often involve country performance to performance of its health systems may strengthening surgical capacity. remain only judgmental for the moment. For example, Brazil and China had under-five mortality rates that were quite close in 2002: 37 per 1,000 for Brazil and 38 per 1,000 for China. In Health System Performance 1990, however, Brazil's rate was 60 per 1,000 and China's was 49 Since about 1940, the publication of economic performance per 1,000: the rate of improvement in Brazil was far more rapid indicators in national income and product accounts has than in China. This measure is only one dimension of out- made it possible to hold political leaders accountable for eco- come, and many explanations are possible. Yet hard numbers nomic management. Additionally, measures of economic on country perform do exist to initiate discussions of policy. performance--such as GDP growth rates and unemployment rates--have allowed economists to move toward evidence- based assessments of which policies facilitate good economic Financing Health Services performance and which do not. Chapters 12 and 13 in this volume discuss domestic and exter- In many ways, unfortunately, the assessment of health sys- nal financing of health systems. Different issues arise in low- tem performance remains where economic performance mea- income countries than in middle-income ones, and the discus- sures were before the development of national income and sion that follows is so divided. Table 1.2 provides context by product accounts in the United Kingdom in the late 1930s. conveying the level of health expenditures in 2001 in different Chapter 3 observes, for example, that "The body of knowledge income groupings of countries, the fraction of GDP spent on [on health systems] represents a largely ad hoc and disjointed health, and the extent to which those expenditures are publicly collection of facts, figures, and points of view. Making confi- financed. Almost 10 percent of the total product of the world dent recommendations relevant to strengthening health system pays for health services. In the low-income countries, about capacity is thus difficult." In its 2000 World Health Report, three-quarters of expenditures are from private, out-of-pocket WHO made an ambitious effort to provide the performance payment. In the high-income European countries, only about measures for health systems that would enable progress toward one-quarter of expenditures is private. Middle-income coun- more systematic knowledge of the policies to improve health tries spend about 5 times as much per capita on health services systems (WHO 2000). Such knowledge could replace what is as do low-income ones and over 10 times as much through the now frequently simply ideology and opinion. The 2000 World public sector. Although available data sets (for example, from Health Report proved to be highly controversial, and its ranking the World Bank or WHO) provide no direct evidence on trends of health system performance may in the end be judged as over time in health expenditures (for more than very short more of a first attempt than an initial approximation (Jamison periods), current levels of expenditure are likely to substantially and Sandbu 2001). WHO set an agenda that will certainly con- exceed those of several decades ago, even as a percentage tinue to be advanced. of growing incomes. The availability of physicians provides Despite current inability to judge health system perform- one indicator: in a large sample of countries the number of ance and the consequently ad hoc character of knowledge, physicians per 100,000 population increased from 54 in the much is in fact known that bears on health policy. Chapters 2 mid 1960s to 116 in the early 1990s, an annual rate of increase and 3 of this volume summarize very specific knowledge about of 2.8 percent. intervention characteristics and system design that can inform Before we turn to questions of financing health services (or policy. Although broad prescriptions may still elude us, partic- insurance), briefly discussing related issues concerning the ular knowledge is still important. Additionally, the perform- public sector's financial role is worthwhile. Those issues address Table 1.2 Health Expenditures by Country Income Level, Public and Total, 2001 Health expenditure Health expenditure Public sector expenditures Country group per capita (2001 US$) (percentage of GDP) (percentage of total health expenditures) Low income 23 4.4 26.3 Middle income 118 6.0 51.1 High income 2,841 10.8 62.1 (Countries in the European Monetary Union) (1,856) (9.3) (73.5) World 500 9.8 59.2 Source: World Bank 2004, table 2.14. 18 | Disease Control Priorities in Developing Countries | Dean T. Jamison what chapter 11 calls "healthy fiscal policy and fiscal policy for increasingly clear that a strong government presence in finance health." An example of unhealthy fiscal policy was the Polish is the least bad way of dealing with these problems. Such a government's subsidy of fatty animal products. Elimination of presence is necessary to achieve universal access to health care that subsidy was a gain for the treasury and resulted in and makes it easier to impose the hard budget constraints that improved diets and health. Minimally, a healthy fiscal policy impose discipline in resource allocation. Additional evidence identifies and corrects such inappropriate subsidies. Fiscal pol- indicates that introducing universal mandatory health coverage icy for health is exemplified by tobacco taxation, which chapter favorably affects both wages and employment levels. Gruber 46 deals with at length and chapter 11 deals with more briefly. and Hanratty (1995) provide thorough documentation of these Fiscal policy for health involves taxes whose principal purposes effects in Canada. Some combination of these factors likely lie more in changing health-related behaviors than in generat- underpins the choices of the high-income democracies to fund ing revenue (although the latter can be important as well). a large fraction of private clinical services with public resources. Public financing of services for all does not imply that all Financing Health in Middle-Income Countries. A major services can be provided. Indeed, given their resource con- cause of poverty (and economic insecurity more generally) straints, countries face hard choices about what to include (and results from highly uneven and unpredictable needs to finance exclude) in the universal benefits package--choices that this health expenditures. In consequence, most societies have volume seeks to inform. moved toward prepaid care as income rises. The current high- Middle-income countries vary substantially in the extent to income countries, with only two exceptions, have decided in which health care providers are financed on a fee-for-service favor of universal public financing (rather than private volun- basis--that is, by direct payment for specific services. Although tary insurance) as the principal means of meeting the demand that is traditionally the chief way to pay for private care, it is for prepaid care. Taiwan (China) and the Republic of Korea, worth clarifying that government providers can also be financed several years ago, and Mexico and Thailand more recently, have (legally or illegally) on a fee-for-service basis--as is increas- also taken the path toward universal public, financing. The ingly the case in China, for example. Similarly, providers are health sector is exceptional: no one in the mature capitalist sometimes compensated through other means, such as capita- democracies would contemplate substantial public financing tion, and individual physicians in the private sector are some- for food or housing, and public subsidies and protection for times on salary or a combination of salary plus capitation. Out- agriculture result from unusually powerful interest groups. of-pocket payments to a public sector provider are usually Public financing for health, including for clinical services for called user fees, but they differ little from fee-for-service com- well-off individuals, has been the result of the democratic pensation of private providers. process in all the major capitalist countries except Switzerland What is the OECD experience with user fees? Basically, it is and the United States. (Public financing is, of course, consistent that both providers and patients respond strongly to the incen- with private provision of services, and the countries of the tive environment. Indeed, a problem exists of providers being Organisation for Economic Co-operation and Development too responsive: much low-value or useless surgery, diagnosis, (OECD) display substantial diversity in this regard.) Efficiency and drug use is, in some systems, highly profitable to the as well as equity concerns underlie this pattern. Barr (2001) provider, and often the provider must, as agent for the patient, examines in detail the efficiency rationales that have under- decide what to do. This conflict of interest has led to cost esca- pinned major public sector financial involvement in health, lation and to inappropriate care. A case may be made for education, and social protection in the high-income countries. divorcing provider compensation from the delivery of individ- Why do market economies choose public sector financing ual services, drugs, or diagnostic tests unless a need exists to (either public spending or publicly mandated social insurance) accelerate coverage of critical services by giving bonuses to for many of their personal clinical services? The case for providers for providing them, as the United Kingdom's publicly financing interventions that are shared by all (for National Health Service has done with immunization. example, antitobacco advertising or water fluoridation) or If fee-for-service financing can generate a perverse incentive where significant externalities exist (such as interruption of environment, does that imply that a system must forgo charg- transmission of TB by treatment of infections) is widely ing beneficiaries for services they receive? Not at all: other ways accepted. Providing personal clinical services, like hernia exist to ensure that funds are adequate for costs--ways that repair, has none of those attributes. Nonetheless, as Arrow may be more effective. Earmarking payroll taxes to finance (1963) articulated in a now-classic article, the pervasiveness health care for workers and their dependents (usually called of incomplete information for decision makers (patients, social insurance) is one approach for recovering costs that is providers, insurers) dominates private health insurance and consistent with provider compensation mechanisms relying delivery of clinical services. These personal clinical services principally on salaries or capitation rather than fee-for-service. account for the bulk of health expenditure. The evidence is It has been argued that cost recovery through payroll taxes Investing in Health | 19 will generate more economic distortions than do income, made on how to allocate highly limited public sector resources. consumption, or sin taxes--although recent evidence suggests Much of this volume deals with resource allocation across that may not be true (Blanchard and Katz 1997). Nonetheless, interventions. Public finance must address an additional set of when general revenue mechanisms are incapable of financing decisions. Do interventions with substantial positive externali- the nationally defined basic package of services for all, the ties have a particular claim on public resources--beyond the option of cost recovery through payroll taxes for the privileged amount of health and financial protection they buy per million workers in the formal sector is clearly desirable on equity dollars spent? Should public resources be spent only on indi- grounds. This form of taxation also links contributions to a viduals with low income? Or should health systems provide specific service, which increases its acceptability. universal public finance for the very limited range of interven- tions that can be afforded? Should public finance emphasize Financing Health in Low-Income Countries. Approximately providing interventions that maximize financial protection 2.5 billion people live in countries the World Bank classifies as or improvements in health? What patterns of public sector low-income--that is, with a per capita gross national income in resource allocation are likely to prove politically sustainable? 2002 of less than US$735 per year. These countries include Fewer tradeoffs may exist among these criteria than at first India but not China. Table 1.2 reflects that the estimated aver- seems to be the case. age per capita health expenditure for these 2.5 billion people is A starting point for thinking about these criteria is the avail- about US$23 per year, of which US$5 or US$6 comes from ability of an increasing number of good benefit incidence public sources. Chapter 12, on financing health systems, points studies--that is, studies of how the benefits of a public inter- to the severe challenges in setting priorities that these resource vention distribute across income (or asset) quintiles of the limitations imply. Not only are expenditure levels currently population. Those studies find that in a great majority of coun- very low, but also the fiscal space needed to increase them is, tries wealthier people are more likely to benefit from public in most low-income countries, sharply constrained. Fiscal programs than are the poor, at least where benefits are meas- space results from an excess of potential government revenues, ured in expenditures. The World Bank's 1993 World including reasonable projections of official development assis- Development Report pointed to that pattern some time ago tance (ODA), over public expenditures. The concept of fiscal (although noting a number of important exceptions), and space combines both short-term fiscal balance and long-term more recent studies add support to that conclusion. The caveat debt sustainability. Grant ODA can help with short-term bal- "measured in expenditures" is important and insufficiently ance, and soft loans (such as International Development noted. The value or welfare benefit to the poor of a given level Association credits from the World Bank) can reduce the repay- of transfer may well exceed the value received by the well off ment burden from a given level of incurred debt. Health financ- from the same level of transfer. A landmark benefit incidence ing policy for low-income countries must focus heavily on study of the U.S. Medicare program, a mandatory health insur- mobilizing public sector resources and concentrating resources ance for the elderly, found it to be regressive in dollar terms but on true priorities (although the broader range of issues just dis- pro-poor in welfare outcomes (McClellan and Skinner 1997). cussed that middle-income countries must address is relevant Public programs that are not universal appear to systemati- to low-income countries with large formal sectors). cally benefit the better off, and that pattern is understandable The chapters in this volume make clear that incremental from a political perspective. It follows that if an immunization resources for health, well spent, could have an enormous effect: program, for example, is differentially benefiting the well off, resource mobilization is important. Increasing public sector then making immunization universal would be pro-poor in expenditures in health by 0.5 percent or more of GDP will be terms of incremental public expenditures. Figure 1.7 uses data possible in some countries, but not in all, and even where it is from a careful benefit incidence assessment in the Philippines possible other investment priorities will also be pressing. (Gwatkin and others 2000) to illustrate this point for immu- However, increases of as much as 1 percent of GDP may possi- nization and for attended deliveries. ble where the political will exists, as is now being attempted in Making coverage universal for cost-effective interventions India. Cost estimates for meeting just the health-related MDGs, for conditions important to the poor is thus likely to prove to as reported in chapter 9, can exceed that amount, and other be an efficient way of both improving health outcomes and estimates have run higher. Development assistance for health, enhancing equity. Many of these interventions address infec- discussed in chapter 13 and here, can expand the available tious disease where control has significant externalities, and resource envelope, but even multiples of current levels of devel- implementing universal coverage is likely to prove more politi- opment assistance would likely prove insufficient to finance cally sustainable than targeting population subgroups. Lindert attainment of the MDGs in some countries. (2004) extensively discusses the experience in high-income Achieving gains for health (and frequently concomitant countries with universalization of public financing of educa- gains in financial protection) requires that critical decisions be tion, health, and old-age pensions and concludes not only that 20 | Disease Control Priorities in Developing Countries | Dean T. Jamison Percentage of incremental benefit received by each quintile answer to this question exists, an important factor has been 50 advance in scientific knowledge and its application both in Universal delivery attendance creating powerful interventions and in guiding behavior. 45 43 Universal immunization coverage Acquisition and use of health research and development or its 40 products becomes, then, an essential function of a country's 35 health system. Moreover, it is important that research extend 32 30 29 beyond development of new products to encompass knowl- edge generation on health system financing and performance. 25 22 Much knowledge is embodied in global public goods: 20 19 once a vaccine against hepatits B has been developed any- 16 15 15 where, it becomes, in some sense, available everywhere. 10 Although monopoly pricing made possible by patents may 10 9 slow the diffusion of some innovations, the temporary nature 5 4 of patent-induced monopoly pricing limits that effect. 0 However, an innovation's being cheap, powerful, and globally Poorest Second Third Fourth Richest available in no way assures its global use. The implication is Quintile clear: globally available knowledge and products offer enor- Source: Calculations based on data reprinted in Gwatkin and others (2000). mous opportunities to countries, but national policies and Notes: Universal immunization coverage means complete coverage with the standard immunization schedule. The Philippines achieved 75 percent coverage overall in 1998, national health systems determine whether that knowledge is and Gwatkin and others (2000) show the coverage level by quintile. The black bar shows put to local use. Additionally, although some information for what the percentage distribution of benefits by population quintile would be of moving from the 1998 status quo to universal coverage. Universal delivery attendance means improving outcomes is principally local and must be locally that all births would be attended by a doctor, nurse, or nurse-midwife. The Philippines achieved 63 percent coverage overall, and the white bar shows what the percentage produced, making the results available contributes to a grow- distribution of benefits by population quintile would be of moving from the current ing evidence base. Chapter 4 on health research stresses the pattern of coverage to universal coverage. value of contributing to a global evidence base and summa- Figure 1.7 Equity Implications of Providing Universal Coverage for rizes with the observation that "all health care is national" and Immunization and Attended Delivery in the Philippines, 1998 "all health research is global." What are the implications for policy? One is that if knowl- edge gains prove even partially as important for future health it is politically sustainable but also that no evidence indicates improvements as they have in the past century--and chapters that the resulting higher taxes have harmed economic growth. in this volume point to a number of reasons for expecting this Two final points are worth stating about making coverage to be so--then investments in health research and develop- universal: ment will continue to have high payoffs in health status and economic productivity. Chapter 7 points to the potential for · First, early adoption of universalization of coverage for pub- enormous economic returns. Ensuring an adequate level of licly financed interventions--even if only a few can be research and development investment, therefore, holds strong financed--sets the stage for expansion, in a middle-income claim on health budgets--a claim for more than the approxi- environment, to universal public financing of health care, mately 3 percent now committed. Equally important--or more the overwhelming choice of the democratic process in high- important--is that the investments be efficient in generating income countries. useful new knowledge and products. Fauci (2005) discusses the · Second, the implementation capacity of health systems in need for greater efficiency in conducting research and develop- low-income countries will often be highly constrained. ment in an environment of tightening budgets in U.S. agencies, Capacity is likely to grow most rapidly by building on a base and he points to a number of specific directions for doing so. of doing a few things well rather than many things poorly. In some cases, additional resources (probably from growth Universal coverage implies tight focus in highly resource- within national health budgets or health aid budgets) will be constrained environments. required to meet these research and development needs ade- quately. In many cases, institutional change will be necessary to create the information and incentives required for efficient RESEARCH AND DEVELOPMENT resource allocation. At the international level, resource alloca- tion has often lacked focus, failing to bring results to the point Why has health improved so dramatically after controlling for of application, and has neglected important conditions and income and, hence, the availability of commodities that, like issues while providing, often generously, for less important food, are essential for health? Although no unambiguous ones. Reform is needed. Successful models of competitively Investing in Health | 21 driven international funding and experience-sharing networks approach to public goods such as research and development; should be applied to currently neglected clusters of conditions. here, responsibility for catalyzing collective action lies princi- Just as the quality and productivity of research efforts vary pally in the hands of the global community. Far from over- dramatically from one institution to another within the high- shadowing action at the national level, global efforts help both income countries, they vary in the low- and middle-income to make national research and development efforts more pro- countries. Exemplary work is done in a number of institutions ductive and to lead to a global result that exceeds the sum of and countries; but in general, the obstacles to high quality are national ones. Thus, among the many competing demands on greater when countries' incomes are lower. Inadequate train- the funds allocated to international assistance for health, those ing, insufficient staff motivation, and lack of competition contributing to generation of new knowledge, products, and prevent many institutions from attaining their potential. The interventions that can be shared by all have special merit. instability of short-term funding, isolation from peers, and poor access to the research literature all compound the prob- lem and prevent researchers from responding rapidly to ever- DEVELOPMENT ASSISTANCE FOR HEALTH changing demands. Given the shortage of good researchers, an argument exists for the talent to move to countries (including Development assistance, wisely focused, has the potential for low- and middle-income countries) whose policies are likely to unusual effect. First, because health gains for the poor can be facilitate productive research (WHO 1996). Donor funding relatively inexpensive (compared to the cost of achieving sig- should reflect this possibility. nificant effect in other sectors), development assistance itself Institutions are more likely to succeed not only if they can achieve much, particularly if it serves as a channel for dif- receive stable core funding but also if a proportion of their fusion of new technologies and best practices. Second, evidence work is funded competitively. Some institutions, such as the suggests that development assistance in health can be more Oswaldo Cruz Foundation in Brazil, have already moved in effective than other development assistance in poor policy and these directions with great success--for example, by freeing up weak institutional environments. Third, the economic benefits intramural resources for competitive allocation between of investing in health can be exceptionally high. Finally, because groups and within the institution, with assessments being research and development have had high impact (chapter 7) made by an external review group. Notable successes have and are an international public good, development assistance occurred in assisting with capacity strengthening, such as the has a particular comparative advantage in ensuring their Special Programme for Research and Training in Tropical finance. Diseases collaboratively supported by WHO, the World Bank, Those conclusions point to a proactive strategy within and the United Nations Development Programme. development assistance agencies and governments for achiev- The failure of current incentive structures, essentially the ing major shifts in staffing and budgetary allocations toward patent system, to produce health products for the lowest specific high-payoff investments in health. They also point to income groups demands remedial action (chapter 5). In the need, in order to achieve the potential benefits, for a essence, either the public sector must harness the skills, energy, focused concentration of health system development on a and capacity of the private sector to develop and bring promptly limited set of priority health goals--for example, controlling to market products for the lowest income groups, or it must AIDS, controlling smoking, meeting the health-related MDGs, take responsibility for doing so itself. In reality, a combination and--for middle-income countries--implementing finance of the two is likely, as is exemplified in successful public-private reforms that lead toward universal public financing. The sec- partnerships such as the Medicines for Malaria Venture or the tion argues that although financial fungibility--the capacity to International AIDS Vaccine Initiative. Recently proposed pre- redirect government resources away from areas supported by commitments by the public sector to purchase specific new external financing--can dilute the effect of development assis- products are an additional potential instrument to generate tance in health, as in other sectors, designing development incentives for private sector investment (Kremer and assistance for health that minimizes the fungibility problem is Glennerster 2004). Developing countries that participate in possible. Performance-based budget support will be one private sector innovation will be positioned to more quickly instrument. learn of and have access to the technical progress that is critical In 2003, the world committed to ODA of almost US$100 in driving health improvements. billion, and news reports in May 2005 suggested the possibility Global challenges demand, in some sense, a global response. of substantial increases by European donors. Approximately All nations share the fruits of research and development. Even 10 percent of ODA is spent for health, a percentage that has though each country may invest a relatively modest sum grown rapidly. Table 13.1 in chapter 13 shows recent trends in toward collective goals, the aggregate effort potentially benefits external financing for health, of which ODA (that is, grant or all substantially. Collective action is the economically rational highly concessionary loans) is only a part: these numbers are 22 | Disease Control Priorities in Developing Countries | Dean T. Jamison for commitments, not actual disbursements, which are smaller highly targeted development assistance, can be as successfully and lag behind commitments. (The Global Fund to Fight implemented as immunization programs where health systems AIDS, Tuberculosis, and Malaria is one of the few providers of are weak. developmental finance that reports disbursements as well as commitments.) External financing for health has grown from about US$6.7 billion in 1998 to US$9.3 billion in 2002 Project Support versus Budget Support (Michaud 2003). For some countries, development assistance Development assistance is tending to move away from project constitutes a significant and growing fraction of health expen- support--for example, of an immunization program, an AIDS ditures. Economists have recently returned to the question of control program, or an extension of a road network--and the returns to expenditures on development assistance, and toward general budgetary support, often to be provided several recent trends have important potential implications for through pooling of donor assistance. There are many reasons health. for this tendency, some of which are good (Kanbur and Sandler 1999, 106). The usefulness (and even propriety) of budget sup- port is contingent, however, on adequacy of the policy and Aid Effectiveness institutional environments. Chapter 3 points to arguments that Recent work has been reassessing aid effectiveness and has as health systems evolve, development assistance should move focused on the following questions: Is there any evidence that from project assistance toward program assistance. The Global infusions of development assistance have affected economic Alliance for Vaccines and Immunization (GAVI) is pointing growth rates? Is there any evidence that infusions of economic to ways that support for immunization programs can be assistance have affected mortality rates or levels of poverty? advanced within the context of this tendency to move toward These questions are clearly not easy to answer. Nonetheless, general budget support. GAVI's innovation is to support some data provide insights. Burnside and Dollar (2000) con- immunization programs based on performance--US$20 for a clude, for example, that development assistance does seem to fully immunized child. The country gets the US$20 for immu- work in countries where a good policy environment and a good nizing the child in whatever way it decides; thus, GAVI provides institutional environment exist, but not in countries lacking general budget support that is conditioned on performance. those elements. Recent work focuses on aid directed to eco- GAVI's concern has been with transitional financing (rather nomic development and greatly strengthens the inclusiveness than with sustained assistance), but its approach points the way of the conclusion that aid boosts growth (Clemens, Radelet, for designing long-term budget support conditioned on meas- and Bhavnani 2004). The effect of development assistance on urable performance with respect to specific health goals. growth is quantitatively important even in countries with poor Jamison (2004) outlined design of long-term development policies and institutions, although the effect is stronger in assistance for health that could meet this objective, maintaining countries with better policies. Interestingly, aid's effect appears incentives for countries to increase coverage (or performance) larger in countries with higher life expectancy. That develop- while scaling back the volume of aid as a country's income ment assistance contributes broadly to growth does not, of increased. Adequate measurement underpins assessment of course, imply that development assistance for health will accel- performance and can be difficult even for immunization cov- erate health improvements. However, it is certainly suggestive erage. Measurement requires resources that must be planned of the potential in health to know that development assistance for and budgeted. works for growth. Even if development assistance is viewed as working better in strong institutional and policy environments, a dilemma Macroeconomic Consequences of Aid exists in that the countries that most need aid are often ones Another concern in the aid community--particularly in the that have weak policies and weak institutions (Radelet 2003, International Monetary Fund--is that development assistance 194). Experiences with ODA in health complement the recent could have adverse domestic macroeconomic consequences-- research on aid for growth in suggesting that ODA can pay off essentially inflationary consequences (see WHO 2002, despite limited institutional or absorptive capacity. Polio has chapter 8). This argument needs to be taken seriously. It is in certainly been eliminated in countries with good health sys- essence an argument about the generation of domestic infla- tems, but it has also been eliminated from most countries with tionary pressures--of projects chasing after those few good weak ones. No smallpox exists today in countries with bad engineers or doctors with an increasing amount of foreign policies and bad institutions. A number of those countries have money and creating an inflationary spiral in that way. However, immunization rates of 60 or 70 percent, or as high as in the if the principal proposed use for the money is for drugs or United States. An important question concerns the extent to vaccines--for example, the US$10 increment for adding Hib which other development assistance for health, particularly and HepB vaccines to the Expanded Program on Immunization Investing in Health | 23 schedule--that money is almost all foreign exchange, and the of uncertain duration. Jamison and Radelet (2005) point to macroeconomic arguments about inflationary consequences ways of using such aid that can be minimally disruptive. simply would not apply. Careful project design can respond to what on the whole are serious concerns from the macro- CONCLUSIONS economic part of the development assistance community. Economic analysis can provide information--such as this vol- A volume as large as this one can provide only a sampling of ume attempts to provide--on getting the maximum health and opportunities and potential pitfalls for investments in health. financial protection outcomes from the development assistance Indeed, the International Statistical Classification of Diseases available and for designing interventions (tradable and com- (WHO 2003a) takes more than 1,500 pages simply to list the modity intensive) that will minimize potentially adverse conditions that a health system must address. Yet the diseases macroeconomic consequences. accounting for most of the burden can be listed in perhaps half a dozen pages, and the diseases that account for most of the differences in outcome between high- and low-mortality coun- The Millennium Development Goals tries can be listed on a page. Chapters in this volume assess 115 An additional and significant direction in thinking about ODA population interventions and 204 personal interventions that concerns achievement of the MDGs (chapter 9). The MDGs are address most conditions of importance. The conclusions listed very specific targets for improvement in education, health, and in this final section of the chapter simply highlight important income-related poverty. Interestingly, focusing development conclusions for policy without attempting to summarize the assistance on achieving the MDGs stands in at least partial volume as a whole. Chapters 2 and 3 complement this one with opposition to the move toward budget support. a fuller summary. These considerations point to several directions for the Table 1.3 provides a sense of the nature of the findings in design of development assistance for health. Radelet (2003, much of the book, and box 1.4 provides a brief description of the 194) provides detailed quantitative examples to show that, even methods.The table shows the number of DALYs that we estimate under very favorable circumstances, in a lower-middle-income could be averted (or years of healthy life that could be bought) by country development assistance is likely to be needed for spending a million dollars on a few of the interventions address- decades. Some conclusions follow that are drawn from the pre- ing major sources of disease burden. If we think of these num- ceding discussion and from the need for predictability and long bers as the prices for buying health by different means, the price time horizons in donor behavior. ODA should move toward variation is enormous, ranging from one or two DALYs per mil- the following: lion dollars up to well over 100,000. All the standard caveats more than apply to these numbers. Nonetheless, they do convey · providing aid over long-term perspectives (10 or more information relevant for policy. Expanding coverage of the cur- years) rently used mix of vaccines,for example,appears more attractive · ensuring predictability in assistance commitments than does adding new vaccines (except when the lower demands · emphasizing demand-side support (with concomitant on system capacity of adding new vaccines are considered). country control of resources) Bypass surgery used even in the most appropriate circumstances · providing incentives for countries to maintain high cover- is an expensive way to buy a year of healthy life, but for many age for cost-effective programs common indications it is inordinately expensive. Findings from · avoiding perverse incentives table 1.3 exemplify findings of the cost-effectiveness analyses · including a transparent exit strategy (for example, reduced from throughout the book that are summarized in chapter 2. grant support with per capita GDP growth). A general conclusion follows from the preceding discussion. There are many inexpensive ways to reduce mortality rates and There is a strong analogy to within-country programs like improve health. A country that focuses on those interventions Mexico's Progresa, which provides cash transfers to poor can expect to achieve major improvements--even with very households contingent on getting children immunized or into limited resources. There are also ways to spend money on school. Gertler (2004) has reported evaluation results indicat- health that can dissipate even a substantial budget with almost ing a high degree of effectiveness. The effectiveness of coverage no return--either for better health or for the financial protec- incentives is well exemplified by the Bill & Melinda Gates tion of populations. Intervention selection matters. Many of Foundation in its work on polio with both GAVI and the World the good and bad buys are now well known. Some are not, and Bank in providing a financial incentive for enhanced coverage our main purpose in this volume has been to assemble the evi- (chapter 13). Although donors increasingly state a commit- dence based on what is known. ment to providing aid predictably and over long periods, the We now turn to more specific conclusions. Different items reality for many countries is that aid flows will be volatile and on the list are relevant in different countries. (Chapter 2 24 | Disease Control Priorities in Developing Countries | Dean T. Jamison Table 1.3 How Much Health Will a Million Dollars Buy? Estimated DALYs averted Service or intervention Cost per DALY (US$) per million US$ spent Reducing under-five mortality Improving care of children under 28 days old 10­400 2,500­100,000 (including resuscitation of newborns) Expanding immunization coverage with standard child vaccines 2­20 50,000­500,000 Adding vaccines against additional diseases to the standard 40­250 4,000­24,000 child immunization program (particularly Hib and HepB) Switching to the use of combination drugs (ACTs) against 8­20 50,000­125,000 malaria where resistance exists to current inexpensive and previously highly effective drugs (Sub-Saharan Africa) Preventing and treating HIV/AIDS Preventing mother-to-child transmission (antiretroviral-nevirapine 50­200 5,000­20,000 prophylaxis of the mother; breastfeeding substitutes) Treating STIs to interrupt HIV transmission 10­100 10,000­100,000 Using antiretroviral therapy that achieves high adherence for a large 350­500 2,000­3,000 percentage of patients Using antiretroviral therapy that achieves high adherence for only a Because of very limited gains by individual small percentage of patients patients and the potential for adverse changes in population behavior, it is possi- ble that more life years would be lost than saved. Preventing and treating noncommunicable disease Taxing tobacco products 3­50 24,000­330,000 Treating AMI (heart attacks) with 10­25 40,000­100,000 an inexpensive set of drugs Treating AMI with inexpensive drugs plus streptokinase (costs and 600­750 1,300­1,600 DALYs for this intervention are in addition to what would have occurred with inexpensive drugs only) Treating heart attack and stroke survivors for life with a daily polypill 700­1,000 1,000­1,400 combining four or five off-patent preventive medications Performing coronary artery bypass grafting (bypass surgery) in specific 25,000 40 identifiable high-risk cases--for example, disease of the left main coronary artery (incremental to treatment with polypill) Using bypass surgery for less severe coronary artery disease Very high Very small (incremental to treatment with polypill) Other Detecting and treating cervical cancer 15­50 20,000­60,000 Operating a basic surgical ward at the district hospital level that focuses 70­250 4,000­15,000 on trauma, high-risk pregnancy, and other common surgically treatable conditions Source: Authors. AMI acute myocardial infarction. suggests, for example, major differences between the priorities money or political leadership or health system capacity, it will for South Asia and those for Sub-Saharan Africa.) Many inter- often be necessary to focus the available resources on a few key ventions or policy changes that are important are not on this priorities. Chapter 12 makes it disappointingly clear that, for list, but they are included in the more extensive discussions in the low-income countries, not only are financial resources chapters 2 and 3, which synthesize messages from the rest of sharply limited now, but prospects for more than modest the book on setting intervention priorities and strengthening increases seem unlikely for many years. (Financial constraints health system capacity. Given often quite limited availability of in the middle-income countries, although real, are less Investing in Health | 25 Box 1.4 Cost-Effectiveness Analysis in This Volume A starting point for cost-effectiveness analysis is to observe substitute for different aspects of system capacity (see also that health systems have two objectives: (a) to improve the Gericke and others 2003). An important mechanism for level and distribution of health outcomes in the popula- strengthening capacity, inherent in highly outcome- tion and (b) to protect individuals from financial risks oriented programs, may simply be to use it successfully-- that are often very substantial and that are frequent causes learning by doing. Several chapters discuss capacity of poverty. Financial risk results from illness-related loss of strengthening at different levels of the clinical system, in income as well as expenditures on care; the loss can be public health, and in health research and development. ameliorated by preventing illness or its progression and by The literature on economic evaluation of health proj- using appropriate financial architecture for the system. ects typically reports the cost per unit of achieving some For the purposes of this book, we consider two classes measure of health outcome--quality-adjusted life years of resources to be available: financial resources and health (QALYs) or DALYs or deaths averted--and at times system capacity. To implement an intervention in a popu- addresses how that cost varies with the level of intervention lation, the system uses some of each resource. Just as and other factors. Pritchard (2004) provides a valuable some interventions have higher dollar costs than others, introduction to this literature. DCP1 reported such cost- some interventions are more demanding of system capac- effectiveness findings for a broad range of interventions; ity than others. In countries with limited health system DCP2 does so as well. DCP2 authors were asked to use capacity, it is clearly important to select interventions that methods described in Jamison (2003); chapter 15 discusses require relatively little of such capacity. Human resource actual implementation. Cost-effectiveness calculations capacity constitutes a particularly important aspect of sys- provide important insights into the economic attractive- tem capacity, discussed in chapter 71 and in a recent ness of an intervention, but other considerations--such as report of the Joint Learning Initiative (2004). consequences for financial protection and demands on Although in the very short run little tradeoff may exist health system capacity--are also relevant. between dollars and human resources or system capacity DCP2 also makes a preliminary attempt to accumulate more generally, investing in the development of such information about the extent to which interventions place capacity can help make more of that resource available in demands on health system capacity; this information is the future. Chapter 3 discusses different types of health qualitative. DCP2 provides only an initial effort, but quali- system capacity and intervention complexity, and it points tative information does provide helpful input to policy. to the importance of and potential for responding to low Kim (2005) develops a more quantitative approach in an capacity by selecting interventions that are less demanding analysis dealing with cervical cancer. Much less has been of capacity and by simplifying interventions. Chapter 3 done on the extent to which specific interventions provide also explores the extent to which financial resources can financial protection for patients and their families. Source: Author. binding.) Selecting priorities will be hard. This section provides treatment for diarrhea, malaria, and acute respiratory a starting point for discussing which activities should be high infections; and improved prenatal and delivery care. In priorities. The conclusions are grouped under four headings: cases of sharply limited resources--financial or health interventions; health services, systems, and financing; research system capacity--often the single highest priority will and development; and development assistance. be expansion of immunization coverage with the basic antigens: poliomyelitis, measles, diphtheria, tetanus, Interventions pertussis, and perhaps BCG. 1. Standard interventions for reducing under-five mortality 2. Cost-effective interventions exist to address the 50 percent have long been known to be highly cost-effective. The chal- of under-five deaths that occur under 28 days of age, lenge is to scale up while conserving and strengthening including stillbirths. These are underused relative to inter- scarce health system capacity. These interventions include ventions for older children, and correcting this neglect is a immunization; micronutrient supplement delivery; priority. 26 | Disease Control Priorities in Developing Countries | Dean T. Jamison 3. Standard interventions to treat TB are also known to be reduce problems of obesity, hypertension, and dyslipi- highly cost-effective, although probably more demanding demia and their consequences for vascular disease. of health system capacity than are some of the child health Successes are rare but suggestive that large-scale efforts interventions. Scaling up by using already developed mod- could be worthwhile. Careful impact evaluation will be els for strengthening relevant health system capacity is a essential to ascertain whether these investments deliver priority. value for money. 4. Many well-tested preventive interventions for AIDS are 8. Lifetime medical management--eventually using variants effective and cost-effective. Such interventions include of the polypill--of individuals at high risk for stroke or treating STIs, promoting condom use, providing voluntary ischemic heart disease is cost-effective and important for counseling and testing, promoting peer intervention, using tens of millions of individuals. The clearest indications of antiretroviral therapies to prevent mother-to-child trans- high risk are a previous vascular event or diabetes. mission, ensuring safe blood supplies, and encouraging the use of breast milk alternatives by HIV-positive mothers. Health Services, Systems, and Financing Scaling up treatment for STIs may prove particularly 9. Focused funding for particular diseases or programs--for important. Much more rapid implementation of these example, TB or immunization--is a fact of life in many low- interventions is of highest priority and needs to be accom- income countries. Using such funding to build health sys- panied by effective mechanisms of surveillance and evalu- tem capacity is feasible as well as desirable, but it is far from ation. The appropriate mix and distribution of interven- automatic. As capacity grows, the potential advantages of tions depends on the stage of the epidemic. In particular, categorical programs are likely to fade while a more inte- limited financial and institutional resources imply focus- grated (but still outcome-oriented) health system assumes ing effort on populations at high risk early in an epidemic. responsibility for dealing with the relevant conditions. 5. Antiretroviral drugs have been successful on a wide scale 10. Quality of clinical care makes an enormous difference, in high-income countries (and in some upper-middle- both to the cost of care and to health outcomes. Tangible income countries--notably Brazil and Mexico) in sharply actions can be taken to improve quality: important among reducing viral load and extending the life expectancy of them is having each provider do a few things well rather patients who are HIV positive. Health system capacity for than many things poorly. achieving durable benefits from antiretroviral drug use at 11. Strengthening capacity for surgery at the district hospital scale in resource-constrained environments, however, level is a frequently neglected priority. Major important remains to be demonstrated. Failure to achieve good uses of this capacity will be to deal with injuries and adherence in such an environment would provide minimal obstetrical emergencies. benefits to the patient, increase risks of drug resistance, 12. In low-income countries public funding for health will and incur substantial costs: the financial and human losses remain highly constrained as a percentage of GDP for the could be enormous. Multiple approaches to successful foreseeable future. Targeting these funds to provide uni- maintenance on antiretroviral drugs should be tried and versal access to a limited number of interventions that are evaluated in large-scale pilots or as part of implementation high priority for poor people is both efficient and equity scale-up. Given the magnitude of the AIDS problem, enhancing, but it will require clear setting of priorities, undertaking and evaluating variations in implementation particularly for incremental resources as they become (including possible variation in the choice of first-line available. drugs) in parallel rather than serially is important. This 13. Middle-income countries can learn from the OECD expe- approach is not now being used. Similarly important is rience that universal public financing of a substantial pack- being rigorous in dropping unsuccessful implementation age of clinical care is both efficient and equity enhancing. models before they consume substantial resources that could otherwise have greatly affected AIDS prevention or other priorities in the health sector. Research and Development 6. Control of tobacco use is the cornerstone of proven 14. Impact evaluation of interventions in many domains is an approaches to primary prevention of heart disease, stroke, essential priority and should be done around planned vari- chronic pulmonary disease, and many types of cancer. ations in implementation. One specific area of importance Instruments for control of tobacco use centering on is evaluation of effective ways to manage lifelong drug taxation and improved public information are well use--for example, for AIDS, secondary prevention of established. vascular disease, diabetes, and major psychiatric disorders. 7. A range of potential approaches to changing dietary and 15. Public-private partnerships such as the Medicines for exercise patterns of populations would, if successful, Malaria Venture and the International AIDS Vaccine Investing in Health | 27 Initiative provide promising models for developing impor- others 2006) will provide more comprehensive tables of results tant new drugs, vaccines, and diagnostic products to deal for a much finer disaggregation of conditions, a full exposition with the major diseases of poverty as well as the problem of methods and data sources, and sensitivity analyses (includ- of drug resistance and microbial evolution more generally. ing assessments of the sensitivity of results to including still- birth). All numbers in this annex are consistent with those in Development Assistance the companion volume. 16. Development assistance for health has begun to become This annex first provides a brief background on assess- performance based, and this trend should accelerate, along ments of deaths by cause and disease burden and then an with making development assistance more stable and long overview of the uses of such measures for health policy. It term (contingent on performance). This change may concludes with aggregated tables on deaths and on disability- involve at least a partial shift away from the sectorwide adjusted life years (DALYs) by selected causes or groups of approaches to development assistance that recent evidence causes. The tables present estimates both with and without suggests may lead to neglect of focus on outcomes. It will stillbirths, which constituted approximately 5.5 percent of require renewed attention to outcome measurement. deaths globally in 2001. Estimates are provided separately for 17. Resistance of the malaria parasite that is responsible for high-income countries and for the low- and middle-income most deaths to chloroquine and SP is now widespread and countries as a group. rapidly increasing. A particular challenge is overcoming financial and institutional barriers to virtually complete Background replacement of those drugs with ACTs, which minimize Many countries, including all high-income ones, maintain vital resistance and can decrease transmission. Absent such an registration systems that provide data (usually annual) on the effort, malaria mortality is likely to continue rising. A number of deaths by cause, age, sex, and sometimes race. Some centralized procurement mechanism receiving subsidies countries additionally compute years of life lost (or YLL) by from development assistance agencies and making low- cause, which assigns a number of years of life lost attributable cost ACTs available to public and private supply chains to each cause that depends on the age of death and some rele- globally would address this problem. vant measure of life expectancy. As of the early 1990s, no simi- 18. Investing in global capacity to respond effectively to a new lar estimates existed for many developing countries or for influenza pandemic, particularly within the resource con- regional groupings of them. Experts on individual conditions straints of low-income countries, is a priority for the inter- or the relevant disease program at WHO generated estimates national system. Such capacity would include effective for the diseases of interest to them. When added up across dis- surveillance, surge manufacturing capacity for drugs and eases, however, such estimates exceeded, often by a factor of 2 vaccines, stockpiles of drugs that could be used to attempt or more, any plausible estimate of the total number of deaths to contain epidemics, and mass media messages and pub- occurring in each age group. DCP1 and the World Development lic policies prepared in advance to be deployed if needed. Report 1993 (Lopez 1993 and World Bank 1993) generated esti- 19. Because research and development is so important for mates of the number of deaths by cause that were consistent health and because it is a classic international public good, with demographically determined death totals for eight a substantial fraction of incremental development assis- regional groupings of countries. WHO collaborated closely on tance for health should go to research and development. this work. The number of deaths from a disease is one measure of the magnitude of its burden, and YLL constitutes for many The content of these specific recommendations, and of rec- purposes a better measure. Neither takes account of the dis- ommendations throughout this volume, point to the enormous ability or suffering associated with a nonfatal disease. potential we now have to reduce further the human and finan- The 1993 World Development Report also developed a vari- cial burden of ill health. Scientific advance created this potential. ant of the quality-adjusted life year (QALY) from the health Its more widespread realization requires the focused attention of economics literature to add a disability dimension to YLLs in health systems to finance and deliver priority interventions. order to generate a more comprehensive measure of burden. The result, called a disability-adjusted life year, measures bur- ANNEX 1.A: THE BURDEN OF DISEASE IN 2001 den from a specific cause as the sum of years of life lost from that cause and the equivalent years of life lost (in a sense that is This annex provides estimates of the burden of different dis- made quite specific) from the disability caused by the condi- eases and injuries in 2001. Alan Lopez, Colin Mathers, tion. Original publications on disease burden included esti- Christopher Murray, and their colleagues at WHO generated mates that discounted future events at 0 percent or at 3 percent the estimates, aggregated them by World Bank regions, and per year. They also included estimates that weighted the value provided final updates. A companion volume (Lopez and of a year of life uniformly across all age groups and estimates 28 | Disease Control Priorities in Developing Countries | Dean T. Jamison that placed greater value on middle-aged groups. These are The national income and product accounts for the labeled DALYs (r,k) with the first number indicating the dis- United States (NIPAs), and kindred accounts in other count rate in percent per year and the second indicating nations, have been among the major contributions to whether uniform or nonuniform age weights were used. The economic knowledge over the past half century. . . . most widely reported variant on the DALY is the DALY (3,1)-- Several generations of economists and practitioners that is, one that uses a 3 percent discount rate and nonuniform have now been able to tie theoretical constructs of age weighting. This chapter and the companion volume report income, output, investment, consumption, and sav- DALYs (3,0)--that is, with discounting but uniform age ings to the actual numbers of these remarkable weighting. accounts with all their fine detail and soundly meshed Estimates of DALYs by cause for 1990 first appeared as interrelations. (Eisner 1989, 1) appendix B of the 1993 World Development Report and, in expanded form, in Murray, Lopez, and Jamison (1994). Disease burden measures have the potential of serving a Christopher Murray, Alan Lopez, and colleagues later produced similar purpose for health policy. updated estimates for 1990 and a fuller account of the methods 2. Generating a forum for informed debate of values and priori- used (Murray and Lopez 1996a, 1996b). The relative burden of ties. The assessment of disease burden in a country in prac- different conditions as measured by numbers of deaths corre- tice involves participation of a broad range of national lates highly with DALYs, but important exceptions exist. The disease specialists, epidemiologists, and, often, policy mak- massive burden of major psychiatric conditions, for example, is ers. Debating the appropriate values, say, for disability captured by DALYs but missed in estimates of deaths by cause weights or for years of life lost at different ages helps clarify or YLLs. Table 1.A1 summarizes the current estimates of deaths values and objectives for national health policy. Discussing by cause in 2001 from Mathers, Murray, and Lopez (2006). the interrelations among diseases and their risk factors in the Table 1.A2 shows disease burden in DALYs. Additional light of local conditions sharpens consideration of priorities. columns in these tables show the effect of including stillbirths 3. Identifying national control priorities. Many countries on the percentage distribution of burden across conditions now identify a relatively short list of interventions, the full while leaving unchanged the other numbers (see Jamison and implementation of which becomes an explicit priority for others 2006). national political and administrative attention. Examples Disease burden can be assessed by risk factor as well as by include interventions to control TB, poliomyelitis, HIV disease or condition. An initial assessment of risk factor burden infection, smoking, and specific micronutrient deficiencies. appeared in the 1993 World Development Report and later in Because political attention and high-level administrative Murray and Lopez (1997) and WHO (1996). WHO published capacity are in relatively fixed and short supply, the benefits a much fuller set of estimates in its 2003 World Health Report from using those resources will be maximized if they are (WHO 2003b). Ezzati and others (2006) provided a substantial directed to interventions that are both cost-effective and update adjusted to the same methodological assumptions as for aimed at problems associated with a high burden. Thus, deaths and DALYs, including use of DALYs (3,0). national assessments of disease burden are one input in establishing a potential short list of control priorities. In the Uses of Disease Burden Measures summary of the cost-effectiveness analyses reported in this DALYs are useful for informing health policy in at least six volume, chapter 2 pays particular attention to identifying ways. Estimates of deaths by cause or YLL serve these same cost-effective interventions capable of averting a large dis- purposes, but for some uses less well. ease burden. 4. Allocating training time for clinical and public health practi- 1. Assessing performance. A country-specific (or regional) tioners. Medical schools offer a fixed number of instruc- assessment of the burden of disease provides an outcome tional hours; training programs for other levels and types of indicator that can be used over time to judge progress or practitioners are likewise limited. A major instrument for across countries or regions to judge relative performance. implementing policy priorities is allocating this fixed-time The most natural comparison is to the development of resource well. Again that means allocation of time to train- national income and product accounts (NIPAs) by Simon ing in cost-effective interventions in which disease burden is Kuznets and others in the 1930s, which culminated in 1939 high. with a complete NIPA for the United Kingdom prepared by 5. Allocating research and development resources. Whenever a James Meade and Richard Stone at the request of the U.K. fixed effort will have a benefit proportional not only to the Treasury. NIPAs have, in the subsequent decades, trans- size of the effort but also to the size of the problem being formed the empirical underpinnings of economic policy addressed, estimates of disease burden become essential for analysis. One leading scholar has put it this way: formulating policy. Developing a vaccine for a broad range Investing in Health | 29 Table 1.A1 Causes of Deaths in Low- and Middle-Income and High-Income Countries and the World, 2001 (percent) Low- and Middle-Income High-Income World Stillbirths Stillbirths Stillbirths Stillbirths Stillbirths Stillbirths excluded included excluded included excluded included Population (thousands) 5,221,572 928,660 6,150,233 Births (thousands) 118,505 121,733 11,371 11,416 129,878 133,150 Total deaths (thousands) 48,377 51,605 7,936 7,981 56,268 59,542 Causes of death (percent) I. COMMUNICABLE DISEASES, 36.4 34.1 7.0 6.9 32.3 30.5 PREGNANCY OUTCOMES, AND NUTRITIONAL DEFICIENCIES A Infectious and parasitic diseases 22.1 20.7 1.9 1.9 19.3 18.2 1 Tuberculosis 3.3 3.1 0.2 0.2 2.9 2.7 2 STIs excluding HIV 0.4 0.3 0.0 0.0 0.3 0.3 3 HIV/AIDS 5.3 4.9 0.3 0.3 4.6 4.3 4 Diarrheal diseases 3.7 3.4 0.1 0.1 3.2 3.0 5 Childhood diseases 2.8 2.6 0.0 0.0 2.4 2.3 a Pertussis 0.6 0.6 0.0 0.0 0.5 0.5 b Poliomyelitis 0.0 0.0 0.0 0.0 0.0 0.0 c Diphtheria 0.0 0.0 0.0 0.0 0.0 0.0 d Measles 1.6 1.5 0.0 0.0 1.4 1.3 e Tetanus 0.6 0.6 0.0 0.0 0.5 0.5 6 Meningitis 0.3 0.3 0.1 0.1 0.3 0.3 8 Malaria 2.5 2.3 0.0 0.0 2.1 2.0 Other I.A. (7, 9­15) 3.8 3.6 1.3 1.3 3.5 3.3 B Respiratory infections 7.2 6.7 4.4 4.4 6.8 6.4 C Maternal conditions 1.0 1.0 0.0 0.0 0.9 0.9 D Perinatal conditions 5.1 4.8 0.4 0.4 4.5 4.2 1 Low birth weight 2.7 2.5 0.1 0.1 2.3 2.2 2 Birth asphyxia and birth trauma 1.5 1.4 0.1 0.1 1.3 1.2 3 Other perinatal conditions 1.0 0.9 0.1 0.1 0.9 0.8 E Nutritional deficiencies 0.9 0.9 0.2 0.2 0.8 0.8 II. NONCOMMUNICABLE CONDITIONS 53.8 50.5 86.5 86.0 58.5 55.3 A Malignant neoplasms 10.2 9.6 26.0 25.9 12.5 11.8 C Diabetes mellitus 1.6 1.5 2.6 2.5 1.7 1.6 E Neuropsychiatric disorders 1.4 1.4 4.8 4.7 1.9 1.8 1 Unipolar major depression 0.0 0.0 0.0 0.0 0.0 0.0 2 Bipolar disorder 0.0 0.0 0.0 0.0 0.0 0.0 3 Schizophrenia 0.0 0.0 0.0 0.0 0.0 0.0 Other II.E. (4­16) 1.4 1.3 4.7 4.7 1.9 1.8 G Cardiovascular disease 27.6 25.9 38.3 38.1 29.1 27.5 3 Ischaemic heart disease 11.8 11.0 17.2 17.1 12.6 11.9 4 Cerebrovascular disease 9.5 8.9 9.8 9.8 9.6 9.1 Other II.G. (1, 2, 5, 6) 6.3 5.9 11.3 11.2 7.0 6.6 H Respiratory diseases 6.5 6.1 6.0 6.0 6.4 6.1 I Digestive diseases 3.3 3.1 4.2 4.2 3.4 3.3 M Congenital anomalies 1.0 0.9 0.4 0.4 0.9 0.9 Other II. (B, D, F, J­L, N) 2.2 2.0 4.3 4.3 2.5 2.3 III. INJURIES 9.8 9.1 5.9 5.9 9.2 8.7 A Unintentional 6.6 6.2 4.0 4.0 6.3 5.9 1 Road traffic accidents 2.2 2.1 1.5 1.5 2.1 2.0 Other III. A. (2­6) 4.4 4.2 2.5 2.5 4.2 3.9 B Intentional 3.1 2.9 1.9 1.9 2.9 2.8 1 Self-inflicted 1.5 1.5 1.6 1.6 1.6 1.5 Other III.B. (2­4) 1.6 1.5 0.3 0.3 1.4 1.3 Sources: Estimates in the columns excluding stillbirths come from Mathers, Lopez, and Murray (2006). Estimates in the columns including stillbirths come from Jamison and others (2006), which uses the estimates from Mathers, Lopez, and Murray (2006) while adding in stillbirths. 30 | Disease Control Priorities in Developing Countries | Dean T. Jamison Table 1.A2 The Burden of Disease in Low- and Middle-Income and High-Income Countries and the World, 2001 (percent) Low- and Middle-Income High-Income World DALYsa DALYsSB b DALYsa DALYsSB b DALYsa DALYsSB b Total DALYs (thousands) 1,387,426 1,260,643 149,161 148,316 1,536,587 1,412,600 Causes of death (percent) I. COMMUNICABLE DISEASES, 39.8 33.6 5.7 5.4 36.5 30.5 PREGNANCY OUTCOMES, AND NUTRITIONAL DEFICIENCIES A Infectious and parasitic diseases 23.1 21.0 2.3 2.2 21.1 18.9 1 Tuberculosis 2.6 2.8 0.1 0.1 2.3 2.5 2 STIs excluding HIV 0.7 0.7 0.1 0.1 0.6 0.6 3 HIV/AIDS 5.1 5.3 0.4 0.4 4.7 4.8 4 Diarrheal diseases 4.2 2.6 0.3 0.3 3.9 2.6 5 Childhood diseases 3.1 2.4 0.1 0.1 2.8 2.4 a Pertussis 0.8 0.7 0.1 0.1 0.8 0.6 b Poliomyelitis 0.0 0.0 0.0 0.0 0.0 0.0 c Diphtheria 0.0 0.0 0.0 0.0 0.0 0.0 d Measles 1.7 1.5 0.0 0.0 1.5 1.4 e Tetanus 0.6 0.5 0.0 0.0 0.5 0.4 6 Meningitis 0.4 0.4 0.1 0.1 0.4 0.3 8 Malaria 2.9 2.1 0.0 0.0 2.6 1.8 Other I.A. (7, 9­15) 4.1 4.2 1.1 1.1 3.8 4.2 B Respiratory infections 6.3 4.6 1.7 1.7 5.8 4.3 C Maternal conditions 1.9 2.1 0.3 0.3 1.7 1.9 D Perinatal conditions 6.4 3.7 0.9 0.6 5.9 3.4 1 Low birth weight 3.1 1.4 0.3 0.2 2.8 1.5 2 Birth asphyxia and birth trauma 2.3 1.5 0.4 0.4 2.1 1.4 3 Other perinatal conditions 1.1 0.6 0.3 0.3 1.0 0.5 E Nutritional deficiencies 2.1 2.2 0.6 0.6 2.0 2.0 II. NONCOMMUNICABLE CONDITIONS 48.9 52.4 86.7 87.2 52.6 56.4 A Malignant neoplasms 5.4 5.9 17.4 17.4 6.6 7.1 C Diabetes mellitus 1.1 1.3 2.8 2.8 1.3 1.4 E Neuropsychiatric disorders 9.9 10.8 20.9 21.0 11.0 11.9 1 Unipolar major depression 3.1 3.4 5.6 5.7 3.4 3.7 2 Bipolar disorder 0.6 0.7 0.7 0.7 0.6 0.7 3 Schizophrenia 0.8 0.8 0.7 0.8 0.8 0.8 Other II.E. (4­16) 5.4 5.9 13.8 13.9 6.2 6.7 G Cardiovascular disease 12.9 14.2 20.0 20.1 13.6 14.7 3 Ischaemic heart disease 5.2 5.7 8.3 8.4 5.5 6.0 4 Cerebrovascular disease 4.5 5.0 6.3 6.3 4.7 5.1 Other II.G. (1, 2, 5, 6) 3.2 3.5 5.4 5.5 3.4 3.7 H Respiratory diseases 4.2 4.5 6.6 6.6 4.4 4.7 I Digestive diseases 3.8 4.0 4.4 4.4 3.8 4.1 M Congenital anomalies 1.7 1.3 1.0 0.8 1.6 1.3 Other II. (B, D, F, J­L, N) 9.9 10.9 13.7 13.8 10.3 11.2 III. INJURIES 11.2 12.1 7.5 7.5 10.9 11.6 A Unintentional 8.2 8.8 5.3 5.3 7.9 8.4 1 Road traffic accidents 2.3 2.5 2.0 2.0 2.3 2.5 Other III. A. (2­6) 5.9 6.3 3.2 3.3 5.6 6.1 B Intentional 3.1 3.4 2.3 2.3 3.0 3.2 1 Self-inflicted 1.3 1.4 1.7 1.7 1.3 1.4 Other III.B. (2­4) 1.8 2.0 0.5 0.5 1.7 1.8 Sources: Mathers, Lopez, and Murray (2006) provide the reported estimates of DALYs. Jamison and others (2006) provide the estimates for DALYsSB. a. The burden of disease is measured in DALYs. DALYs form a class of measures that aggregate years of life lost from premature mortality with years of life lost due to disability. The DALYs reported here are calculated at a 3 percent per year discount rate with no age-weights, i.e. a year of life at any age is valued the same. These are referred to as DALYs (3,0) in the accompaning volume on burden of disease and risk factors (Lopez and others, 2006). b. The DALYsSB is analagous to the DALY except that it includes stillbirths in the estimates of burden and assumes a gradual "acquisition of life potential" that allows the burden associated with a death near the time of birth to grow gradually with age rather than instantaneously increasing from 0 to a high value at birth or some earlier time. Jamison and others (2006) provide the estimates used here, which they label DALYsSB (3,0,0.54). Investing in Health | 31 of viral pneumonias, for example, would have perhaps hun- REFERENCES dreds of times the effect of a vaccine against Hanta virus Abraham, K. G., and C. Mackie, eds. 2005. Beyond the Market: Designing infection. Thus, information on disease or risk factor burden Nonmarket Accounts for the United States. Washington, DC: The is one vital input (of several) to inform research and devel- National Academies Press. opment resource allocation, as discussed in chapters 4 and 5. Arrow, K. J. 1963. "Uncertainty and the Welfare Economics of Medical 6. Allocating resources across health interventions. Here disease Care." American Economic Review 53 (5): 851­83. burden assessment often plays a minor role; the task is to Arrow, K. J., H. Gelband, and D. T. Jamison. 2005. "Making Antimalarial Agents Available in Africa." New England Journal of Medicine 353: shift resources to interventions, which, at the margin, will 333­35. generate the greatest reduction in DALY loss. When there Bailis, R., M. Ezzati, and D. M Kammen. 2005. "Mortality and Greenhouse are major fixed costs in mounting an intervention, as is the Gas Impacts of Biomass and Petroleum Energy Futures in Africa." case with political and managerial attention for national Science 308: 98­103. control priorities, burden estimates are required to improve Barr, N. 2001. The Welfare State as Piggy Bank: Information, Risk, resource allocation. Likewise, major fixed costs may be asso- Uncertainty, and the Role of the State. Oxford: Oxford University Press. ciated with making the use of an intervention universal (or Becker, G. S., T. J. Philipson, and R. R. Soares. 2003. "The Quantity and Quality of Life and the Evolution of World Inequality." NBER Working expanding it to cover a major percentage of the population), Paper 9765, National Bureau of Economic Research, Cambridge, MA. and if so, the cost-effectiveness of the expansion will depend Bezanson, K. 2005. "Replenishing the Global Fund: An Independent in part on the size of the burden. Assessment." http://www.theglobalfund.org/en/files/about/replenish- ment/assessment_report_en.pdf. Bhargava, A., D. T. Jamison, L. J. Lau, and C. J. L. Murray. 2001. "Model- Results ing the Effects of Health on Economic Growth." Journal of Health Tables 1.A1 and 1.A2 convey summaries of deaths by cause and Economics 20 (May): 423­40. burden of disease in 2001, respectively. Blanchard, O., and L. F. Katz. 1997. "What We Know and Do Not Know about the Natural Rate of Unemployment." Journal of Economic Perspectives 11 (1): 51­72. ACKNOWLEDGMENTS Bloom, D. E., D. Canning, and D. T. Jamison. 2004. "Health, Wealth and Welfare." Finance and Development 41 (1): 10­15. Sonbol Shahid-Salles provided invaluable research support and Bloom, D. E., D. Canning, and P. Malaney. 2000. "Demographic Change and Economic Growth in Asia." Supplement to Population and critical advice during preparation of this chapter. Mantra Singh Development Review 26: 257­90. expertly provided word-processing support. Candice Byrne Bloom, D. E., D. Canning, and J. Sevilla. 2004. "The Effect of Health on provided valuable comments. The other editors of Disease Economic Growth: A Production Function Approach." World Control Priorities in Developing Countries, 2nd edition, Development 32: 1­13. provided extensive critical reactions to the messages and text, Bobadilla, J. L., J. Frenk, R. Lozano, T. Frejka, and C. Stern. 1993. "The Epidemiologic Transition and Health Priorities." In Disease Control and the chapter is consequently very different than it would Priorities in Developing Countries, ed. D. T. Jamison, W. H. Mosley, have been. The Advisory Committee to the editors of this vol- A. R. Measham, and J. L. Bobadilla, 746. New York: Oxford University ume, chaired by Jaime Sepúlveda, provided invaluable com- Press. ments and reaction during a meeting at the Institut Pasteur, Boskin, M. J., and L. J. Lau. 2000. "Generalized Solow-Neutral Technical Paris, in December 2004. Progress and Postwar Economic Growth." NBER Working Paper 8023, National Bureau of Economic Research, Cambridge, MA. In the early 1990s, the World Bank initiated efforts to Bourguignon, F., and C. Morrisson. 2002. "Inequality among World understand and disseminate policies to address the remaining Citizens: 1820­1992." American Economic Review 92: 727­44. large burden of disease affecting the world's poor. The World Burnside, C., and D. Dollar. 2000. "Aid, Policies and Growth." American Bank's (1993) World Development Report: Investing in Health Economic Review 90: 847­68. reported the results of that assessment, which drew on a Clemens, M., S. Radelet, and R. Bhavnani. 2004. "Counting Chickens second publication, Disease Control Priorities in Developing When They Hatch: The Short-Term Effect of Aid on Growth." Working Paper 44, Center for Global Development, Washington, DC. Countries, 1st edition (DCP1) (Jamison and others 1993). Crafts, N. 2000. "Globalization and Growth in the Twentieth Century." Enormous changes both in the world and in our knowledge IMF Working Paper WP/00/44, International Monetary Fund, base occurred during the subsequent decade, leading to the Washington, DC. conclusion that a major revision, update, and expansion of Crafts, N., and M. Haacker. 2004. "Welfare Implications of HIV/AIDS." DCP1 would be of value. In a collaborative undertaking, the In The Macroeconomics of HIV/AIDS, ed. M. Haacker, 182­97. World Bank, the World Health Organization, and the Fogarty Washington, DC: International Monetary Fund. International Center of the U.S. National Institutes of Health Davis, K. 1956. "The Amazing Decline of Mortality in Underdeveloped Areas." American Economic Review (Papers and Proceedings) 46 (2): sponsored this new effort (DCP2) with substantial financial 305­18. support from the Bill & Melinda Gates Foundation. This book Deaton, A. 2004. "Health in an Age of Globalization." NBER Working results from that collaboration. Paper 10669, National Bureau of Economic Research, Cambridge, MA. 32 | Disease Control Priorities in Developing Countries | Dean T. Jamison DeLong, J. B. 2000."Cornucopia: The Pace of Economic Growth in the ------. 2004. "External Finance of Immunization Programs: Time for a Twentieth Century." Working Paper 7602. Cambridge, MA: National Change in Paradigm?" In Vaccines: Preventing Disease and Protecting Bureau of Economic Research. Health, ed. C. de Quadros, 325­32. Scientific and Technical Publication del Rio, C., and J. Sepúlveda. 2002. "AIDS in Mexico: Lessons Learned and 596. Washington, DC: Pan American Health Organization. Implications for Developing Countries." AIDS 16: 1445­57. Jamison, D. T., E. A. Jamison, and J. D. Sachs. 2003. "Assessing the de Savigny, D., H. Kasale, C. Mbuya, and G. Reid. 2004. Fixing Health Determinants of Growth When Health Is Explicitly Included in the Systems. Ottawa: International Development Research Centre. Measure of Economic Welfare." Paper presented at the 4th World Congress of the International Health Economics Association, Easterlin, R. A. 1996. Growth Triumphant: The Twenty-First Century in San Francisco, June. Historical Perspective. Ann Arbor: University of Michigan Press. Jamison, D. T., L. J. Lau, and J. Wang. 2005. "Health's Contribution to ------. 1999. "How Beneficent Is the Market? A Look at the Modern Economic Growth in an Environment of Partially Endogenous History of Mortality." European Review of Economic History 3: 257­94. Technical Progress." In Health and Economic Growth: Findings and Easterly, W., and R. Levine. 1997. "Africa's Growth Tragedy: Policies and Policy Implications, ed. G. Lopez-Casasnovas, B. Rivera, and L. Currais, Ethnic Divisions." Quarterly Journal of Economics 112: 1203­50. 67­91. Cambridge, MA: MIT Press. Eisner, R. 1989. The Total Incomes System of Accounts. Chicago: University Jamison, D. T., W. H. Mosley, A. R. Measham, and J. L. Bobadilla, eds. 1993. of Chicago Press. Disease Control Priorities in Developing Countries. New York: Oxford Ezzati, M., A. D. Lopez, A. Rodgers, and C. J. L. Murray, eds. 2004. University Press. Comparative Quantification of Health Risks: Global and Regional Jamison, D. T., and S. Radelet. 2005. "Making Aid Smarter." Finance and Burden of Disease Attributable to Selected Major Risk Factors. Vols. 1­2. Development 42 (2): 42­46. Geneva: World Health Organization. Jamison, D.T., J. Sachs, and J. Wang. 2001. "The Effect of the AIDS Ezzati, M., S. vander Hoorn, A. D. Lopez, G. Danaei, A. Rodgers, C. D. Epidemic on Economic Welfare in Sub-Saharan Africa." CMH Mathers, and C. J. L. Murray. 2006. "Comparative Quantification of Working Paper WG1:13, Commission on Macroeconomics and Mortality and Burden of Disease Attributable to Selected Major Risk Health, World Health Organization, Geneva. Factors." In Global Burden of Disease and Risk Factors, ed. A. D. Lopez, Jamison, D. T., and M. E. Sandbu. 2001. "The WHO Ranking of Health C. Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. New York: Systems." Science 293: 1595­96. Oxford University Press. Jamison, D. T., M. Sandbu, and J. Wang. 2004. "Why Has Infant Mortality Fauci, A. S. 2005. "The Global Challenge of Infectious Diseases: The Decreased at Such Different Rates in Different Countries?" Working Evolving Role of the National Institutes of Health in Basic and Clinical Paper 21, Disease Control Priorities Project, Bethesda, MD. Research." Nature Immunology 6 (8): 743­47. Jamison, D. T., S. Shahid-Salles, J. S. Jamison, J. Lawn, and J. Zupan. 2006. Feachem, R. G. A., T. Kjellstrom, C. J. L. Murray, M. Over, and M. Phillips "Incorporating Deaths Near the Time of Birth into Estimates of the (Eds.). 1992. Health of Adults in the Developing World. New York: Global Burden of Disease." In Global Burden of Disease and Risk Oxford University Press. Factors, ed. A. D. Lopez, C. D. Mathers, M. Ezzati, D. T. Jamison, and Fogel, R. W. 1997. "New Findings on Secular Trends in Nutrition and C. J. L. Murray. New York: Oxford University Press. Mortality: Some Implications for Population Theory." In Handbook Jha, P., and F. J. Chaloupka, eds. 2000. Tobacco Control in Developing of Population and Family Economics, Vol. 1A, ed. M. Rosenzweig and Countries. Oxford, U.K.: Oxford University Press. O. Stark, 433­81. Amsterdam: Elsevier Science. Joint Learning Initiative. 2004. Human Resources for Health: Overcoming Fuchs, V. 1974. "Some Economic Aspects of Mortality in Developed the Crisis. Washington, DC: Communications Development. Countries." In The Economics of Health and Medical Care, ed. M. Perlman, 174­93. London: Macmillan. Kanbur, R., and T. Sandler. 1999. The Future of Development Assistance: Common Pools and International Public Goods. Washington, DC: ------. 1980. "Comment." In Population and Economic Change in Overseas Development Council. Developing Countries, ed R. Easterlin, 348­51. Chicago: University of Kim, J. J. 2005. "Using Mathematical Modeling to Evaluate the Public Chicago Press. Health Impact and Cost-Effectiveness of Cervical Cancer Screening Gericke, C. A., C. Kurowski, M. K. Ranson, and A. Mills. 2003. "Feasibility Strategies in Different World Regions." Ph.D. dissertation, Program in of Scaling-up Interventions: The Role of Intervention Design." Health Policy, Harvard University, Cambridge, MA. Working Paper 13, Disease Control Priorities Project, Bethesda, MD. Kremer, M., and R. Glennerster. 2004. Strong Medicine: Creating Incentives Gertler, P. 2004. "Do Conditional Cash Transfers Improve Child Health? for Pharmaceutical Research on Neglected Diseases. Princeton, NJ: Evidence from PROGRESA's Control Randomized Experiment." Princeton University Press. Health, Health Care, and Economic Development 94 (2): 336­41. Leape, L. L., and D. M. Berwick. 2005. "Five Years after to Err Is Human: Gruber, J., and M. Hanratty. 1995. "The Labor-Market Effects of What Have We Learned?" Journal of American Medical Association 293 Introducing National Health Insurance: Evidence from Canada." (19): 2384­90. Journal of Business and Economic Statistics 13 (2): 163­73. Levine, R. and the What Works Working Group. 2004. Millions Saved: Gwatkin, D. R., S. Rustein, K. Johnson, R. P. Pande, and A. Wagstaff. 2000. Proven Successes in Global Health. Washington, DC: Center for Global "Socio-Economic Differences in Health, Nutrition, and Population in Development. the Philippines." Washington, DC: World Bank. Lindert, P. H. 2004. Growing Public: Social Spending and Economic Growth Haacker, M., ed. 2004. The Macroeconomics of HIV/AIDS. Washington, since the Eighteenth Century. Vol. 1. Cambridge, U.K.: Cambridge DC: International Monetary Fund. University Press. Institute of Medicine. 2001. Crossing the Quality Chasm. Washington, DC: Lopez, A. D. 1993. "Causes of Death in Industrial and Developing National Academies Press. Countries: Estimates for 1985­90." In Disease Control Priorities in Jamison, D. T. 1993."Investing in Health." Finance and Development 30 (2): Developing Countries, eds. D. T. Jamison, W. H. Mosley, A. R. Measham, 2­5. and J. L. Bobadilla, 35­50. New York: Oxford University Press. ------. 2003. "Cost-Effectiveness Analysis: Concepts and Applications." Lopez, A. D., S. Begg, and E. Bos. 2006. "Demographic and In The Oxford Textbook of Public Health, ed. R. Detels, J. McEwen, Epidemiological Characteristics of Major Regions of the World, 1990 R. Beaglehole, and H. Tamaka, 2: 903­19. and 2001." In Global Burden of Disease and Risk Factors, ed. A. D. Lopez, Investing in Health | 33 C. D. Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. New York: Powles, John. 2001. "Healthier Progress: Historical Perspectives on the Oxford University Press. Social and Economic Determinants of Health." In The Social Origins of Lopez, A. D., C. D. Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray, Health and Well-Being, ed. R. Eckersly, J. Dixon, and B. Douglas, 3­24. eds. 2006. "Measuring the Global Burden of Disease and Risk Factors." Cambridge, U.K.: Cambridge University Press. In Global Burden of Disease and Risk Factors, ed. A. D. Lopez, C. D. Preston, S. H. 1975. "The Changing Relation between Mortality and Level Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. New York: of Economic Development." Population Studies 29 (2): 231­48. Oxford University Press. ------. 1980. "Causes and Consequences of Mortality Declines in Less Lopez-Casasnovas, G., B. Rivera, and L. Currais, eds. 2005. Health and Developed Countries during the Twentieth Century." In Population Economic Growth: Findings and Policy Implications. Cambridge, MA: and Economic Change in Developing Countries, ed. R. Easterlin, MIT Press. 289­360. Chicago: University of Chicago Press. Maddison, A. 1999. "Poor until 1820." Wall Street Journal Europe, July 11. Preston, S. H., and M. Haines. 1991. Fatal Years: Child Mortality in Late Mathers, C. D., C. J. L. Murray, and A. D. Lopez. 2006. "The Burden of 19th Century America. Princeton, NJ: Princeton University Press. Disease and Mortality by Condition: Data, Methods and Results for Pritchard, C. 2004. "Developments in Economic Evaluation in Health the Year 2001." In Global Burden of Disease and Risk Factors, ed. A. D. Care: A Review of HEED." OHE Briefing 40, Office of Health Lopez, C. D. Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. Economics, London, March 2004. New York: Oxford University Press. Pritchett, L., and L. H. Summers. 1996. "Wealthier Is Healthier." Journal of McClellan, M., and J. Skinner. 1997. "The Incidence of Medicare." NBER Human Resources 31(4): 841­68. Working Paper 6013, National Bureau of Economic Research, Radelet, S. 2003. Challenging Foreign Aid. Washington, DC: Center for Cambridge, MA. Global Development. Michaud, C. 2003."Development Assistance for Health: Recent Trends and Roberts, M., W. Hsiao, P. Berman, and M. Reich. 2003. Getting Health Resource Allocation." Paper prepared for the Second Consultation Reform Right: A Guide to Improving Performance and Equity. New York: Commission on Macroeconomics and Health, World Health Oxford University Press. Organization, Geneva. Sachs, J. D. 2005. The End of Poverty: Economic Possibilities for Our Time. Mosk, C., and S. R. Johannson. 1986. "Income and Mortality: Evidence New York: The Penguin Press. from Modern Japan." Population and Development Review 12: 415­40. Solow, R. 1957. "Technical Change and the Aggregate Production Murray, C. J. L., and A. D. Lopez. 1996a. Global Health Statistics: A Function." Review of Economics and Statistics 39: 312­20. Compendium of Incidence, Prevalence and Mortality Estimates for Over 200 Conditions. Cambridge, MA: Harvard University Press. Strauss, J., and D. Thomas. 1998. "Health, Nutrition, and Economic Development." Journal of Economic Literature 36: 766­817. ------. 1996b. The Global Burden of Disease, Volume 1. Cambridge, MA: Harvard University Press. Usher, D. 1973. "An Imputation to the Measure of Economic Growth for Changes in Life Expectancy." In The Measurement of Economic and ------. 1997. "Global Mortality, Disability and the Contribution of Risk Social Performance, ed. M. Moss, 193­226. Chicago: Columbia Factors: Global Burden of Disease Study." Lancet 349 (9063): 1436­42. University Press for the National Bureau of Economic Research. Murray, C. J. L., A. D. Lopez, and D. T. Jamison. 1994. "The Global Burden Viscusi, W. K., and J. E. Aldy. 2003. "The Value of a Statistical Life: A of Disease in 1990: Summary Results, Sensitivity Analysis and Future Critical Review of Market Estimates from Around the World." Journal Directions." In Global Comparative Assessments in the Health Sector: of Risk and Uncertainty 27: 5­76. Disease Burden, Expenditures and Intervention Packages, ed. C. J. L. World Bank. 1993. World Development Report: Investing in Health. New Murray and A. D. Lopez, 97­138. Geneva: World Health Organization. York: Oxford University Press. (A shorter version of this paper appeared in Bulletin of the World Health Organization 72: 495­509.) ------. 2004. World Development Indicators. New York: Oxford University Press. Available annually. Nordhaus, W. 2003. "The Health of Nations: The Contributions of Improved Health to Living Standards." In Measuring the Gains from World Health Organization. 1996. "Investing in Health Research and Health Research: An Economic Approach, ed. K. M. Murphy and R. H. Development." Report of the Ad Hoc Committee on Health Research Topel, 9­40. Chicago: University of Chicago Press. Relating to Future Intervention Options (Document TDR/GEN/ 96.1). Geneva: WHO. Oeppen, J. 1999. "The Health and Wealth of Nations since 1820." Paper presented at the Social Science History Conference, Fort Worth, TX, ------. 1999. The World Health Report: Making a Difference. Geneva: November. WHO. Oeppen, J., and J. W. Vaupel. 2002. "Demography. Broken Limits to Life ------. 2000. The World Health Report: Health Systems. Geneva: WHO. Expectancy." Science 296 (5570): 1029­31. ------. 2002. The World Health Report: Reducing Risks, Promoting Healthy Life. Geneva: WHO. Pearson, T. A., D. T. Jamison, and J. Trejo-Gutierrez. 1993. "Cardiovascular Disease." In Disease Control Priorities in Developing Countries, ed. D. T. ------. 2003a. International Statistical Classification of Diseases and Jamison, W. H. Mosley, A. R. Measham, and J. L. Bobadilla, 746. New Related Health Problems, 10th revision. Geneva: WHO. York: Oxford University Press. ------. 2003b. The World Health Report 2003: Shaping the Future. Geneva: Peto, R. S., and A. D. Lopez. 2001. "The Future Worldwide Health Effects WHO. of Current Smoking Patterns." In Critical Issues in Global Health, ed. ------. 2005. Preventing Chronic Diseases: A Vital Investment. Geneva: C. E. Koop, C. E. Pearson, and M. R. Schwarz. New York: Jossey-Bass. WHO. Pison, G., J. F. Trape, M. Lefebvre, and C. Enel. 1993. "Rapid Decline in WHO CMH (World Health Organization Commission on Macro- Child Mortality in a Rural Area of Senegal." International Journal of economics and Health). 2001. Macroeconomics and Health: Investing in Epidemiology 22(1): 72­80. Health for Economic Development. Geneva: WHO. 34 | Disease Control Priorities in Developing Countries | Dean T. Jamison Chapter 2 Intervention Cost-Effectiveness: Overview of Main Messages Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles Deeper understanding of the role of human health as a critical of disease.2 This chapter provides broad conclusions on the component of economic development has stimulated interest economic efficiency of using these interventions to improve in improving the efficiency with which the modest health health. resources available in low- and middle-income countries (LMICs) are spent. In recent years, exponential growth in the number of economic evaluations of health interventions, PRIORITY SETTING spurred in part by the first edition of this volume (Jamison and others 1993), has created a wider knowledge base for evaluating Information on the costs of purchasing health in conjunction the costs and benefits of interventions to enable better targeting with regional or national realities regarding disease priorities, of financial resources in the health sector (box 2.1). Although private willingness to pay for health, and public budget con- efficient spending on health has always been a desirable goal, it straints can be used to identify widely prevalent investments is particularly critical in the face of recent threats, such as that are not cost-effective (shaded in figure 2.1) and highly HIV/AIDS and drug-resistant bacteria, as well as the problems cost-effective opportunities to improve health that policy mak- presented by increasing prevalence of chronic diseases, such ers are currently neglecting. Throughout the chapter, "not cost- as diabetes and cardiovascular disease (CVD), that threaten to effective" describes an intervention that has a relatively high roll back the significant health gains achieved in the past two ratio of costs to effectiveness. The information provided also decades. This book is an opportunity to assess anew the costs may be helpful in identifying interventions that are not cost- associated with and the health gains attainable from specific effective and are rarely used and cost-effective interventions interventions and thereby better inform the allocation of new that are justifiably widely used (unshaded in figure 2.1). The health funding. broad objective of this exercise is to help improve global popu- Drawing from the collective knowledge and analytical work lation health by improving understanding of the implications of the many experts who have contributed to this volume, of investing in different interventions. Some of the interven- this chapter provides a broader perspective on the relative effi- tions considered are widely prevalent, whereas others are less ciency and effect on health of a number of interventions than well known. Although some interventions are personal, others is possible in a single, condition-specific chapter.1 The objective are population-based (see annex 2.A for definitions). They is to provide information on the cost-effectiveness estimates encompass the spectrum of disease conditions covered in this for 319 interventions covering nearly every disease condition book but are by no means exhaustive of the universe of possi- considered in the volume, and the resulting avertable burden ble interventions. 35 Box 2.1 Use of Cost-Effectiveness to Set Policy: The Directly Observed Treatment Strategy Nearly 2 million people die from tuberculosis (TB) each and studies showing DOTS' cost-effectiveness, the DOTS year, 98 percent of whom live in developing countries and strategy was established worldwide as the most effective most of whom are 15 to 49 years old. Meanwhile, anti-TB response to TB. By 1995, DOTS had expanded to 73 coun- medicines are 95 percent effective in curing TB, even in tries, and by 2003, it had reached more than 180 countries low-income countries, and cost as little as US$10 for a worldwide (WHO 2004). six-month course of treatment or directly observed ther- As of 1999, DOTS has been implemented with the col- apy short course (DOTS). The TB chapter of the first edi- laboration of WHO in 13 provinces of China and has tion of this book (Murray, Styblo, and Rouillon 1993) and achieved a cure rate of 90 percent. The population in East studies by Joesoef, Remington, and Jiptoherijanto (1989) Asia and the Pacific with access to DOTS increased from and Kamolratanakul and others (1993) describe treatment 44 percent in 1995 to 57 percent in 1997, with the propor- of smear-positive TB with short-course chemotherapy as tion of registered TB patients who are enrolled in a DOTS an extremely cost-effective intervention for TB. program also increasing, from 30 percent in 1995 to Since 1980, the World Health Organization (WHO) has 46 percent in 1997 (WHO 1999). The progress is mainly collaborated closely with many countries in East Asia and attributable to high-prevalence countries in the region, the Pacific to introduce short-course chemotherapy and which include Cambodia, China, and the Philippines. then the DOTS strategy to achieve global targets, with a By 2002, national TB programs reported that 69 per- cure rate of 85 percent and a case-detection rate of 70 per- cent of the world's population lived in countries or parts cent. In 1990, 10 countries were using short-course of countries with DOTS coverage. DOTS programs chemotherapy. In 1993, as a result of growing TB preva- treated a total of 13.3 million TB patients and 6.8 million lence rates, WHO's declaration of a global emergency, smear-positive patients between 1995 and 2002. chapter also focuses on the total burden of disease avertable by Cost-effective expanding population coverage of an intervention. The deliv- Neglected interventions High opportunities ery of many interventions, including those that are relatively used widely cost-effective, may require a certain degree of institutional and organizational capacity on the part of a health system, and Interventions for countries will have to pay attention to this important consider- Interventions to which scaling up Cost-effectiveness Low scale back ation.3 These factors, in combination with other considerations is inefficient such as equity, social justice, medical suitability, and epidemio- Low High logical appropriateness, should guide where money may be Current coverage spent most effectively (Cookson and Dolan 1999, 2000). Source: Authors. Cost-effectiveness ratios can be used to set health priorities in two ways. One approach is to use a cutoff level of cost- Figure 2.1 Efficiency of Interventions effectiveness beyond which interventions are no longer used. This cutoff can vary from place to place depending on the availability of health resources, the disease burden, and the Cost-Effectiveness local preferences for health spending. The World Bank has The specific measure of cost-benefit analysis adopted in this described health interventions that cost less than US$100 per volume is cost-effectiveness. Effectiveness is measured in year of life saved as highly cost-effective for poor countries, but natural units (deaths averted and years of life saved) and in this benchmark is arbitrary, as chapter 15 makes clear by not- disability-adjusted life years (DALYs), a composite measure ing the interaction with income, budget levels, and the disease that combines years lived with disability and years lost to pre- burden (Jamison and others 1993). mature death in a single metric (see chapter 15 for an explana- An alternative approach to using cost-effectiveness data to tion of how DALYs are calculated). Nevertheless, dollars per set intervention priorities is to interpret the cost-effectiveness DALY averted can at best be only one consideration in the allo- ratio as the "price" of equivalent units of health using different cation of resources to different diseases and interventions. This interventions (box 2.2 explains this approach). Reinterpreted 36 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles Box 2.2 A Framework for Using Cost-Effectiveness Information to Set Health Priorities A frequent, often justified, criticism of cost-effectiveness socioeconomic status of target populations (including the analyses is that they address only one of many criteria extent to which they can obtain treatment from their own that could be used to evaluate health interventions. resources), and the ministry of health's ability to deliver Epidemiological, medical, political, ethical, and cultural the program effectively. After the tradeoffs have been made factors often also play important roles in the decision to and can be represented by an indifference curve (see the allocate resources to a specific health condition or inter- figure and explanation), the cost-effectiveness information vention; however, determining how one might weigh cost- is useful in determining how much of the policy makers' effectiveness ratios alongside these other considerations fixed budget should be allocated to each intervention-- when setting priorities for spending is difficult. Musgrove that is, at what coverage of one problem should they start (1999) shows how to take some of these connections devoting resources to the other? The indifference curve into account, including circumstances in which cost- represents health planners' willingness to trade off between effectiveness is an adequate criterion by itself. One investment in antimalarial drugs and treatment for approach is for the policy maker to think of cost- Parkinson's based on all the relevant factors and independ- effectiveness ratios as the relative "price" of purchasing a ent of the budget constraint. unit of health (a DALY, for instance) using different inter- ventions. These costs, along with the budget constraint, can DALYs gained help determine the optimal allocation of resources among from treating a given set of interventions. Parkinson's Consider, for instance, a policy maker in a country in disease Sub-Saharan Africa facing the choice between treating Y Parkinson's disease and expanding malaria treatment pro- grams while constrained by a fixed budget allocated by the ministry of finance. The cost per DALY averted of treating X Parkinson's disease using carbidopa is vastly greater than that for the malaria control program. A simplistic interpre- DALYs gained from tation of the cost-effectiveness information would be to treating malaria expand the malaria treatment program to the maximum extent possible before turning to the treatment of The solid line represents the budget, and its slope is the Parkinson's disease. This solution could be desirable in ratio of the cost-effectiveness ratios of the two interven- some situations, particularly if the budget is large enough to tions. The dashed line represents an alternative scenario in deal fully with the malaria problem and still allow for some which the cost-effectiveness of treating Parkinson's is bet- treatment of Parkinson's. Emphasizing as complete cover- ter (more DALYs can be gained) than the ratio represented age as possible of a particular problem may be especially by the solid line. The axes show how many DALYs can be appropriate in an epidemic situation, in which turning to gained from each treatment, so a large number of DALYs another disease first can mean the epidemic will be worse in corresponds to a low "price" of health or cost per DALY. the future. Devoting some resources to each problem The figure shows the simple case, in which these prices are instead of concentrating on either may be more sensible constant for either budget line--that is, expanding either when neither presents the threat of a growing epidemic. program does not raise the unit cost--although this case is Asking policy makers to make a binary choice between unlikely when part of the population is difficult to reach, is two sets of interventions on the basis of cost-effectiveness harder to treat, or has more severe disease, in which case ratios alone may be unrealistic and misleading. Rather, pol- the rise in unit cost means that an intervention becomes icy makers should first determine their willingness to trade relatively less cost-effective, giving a further reason to start off health improvements in children (malaria) versus the devoting resources to an alternative intervention. elderly (Parkinson's). Policy makers may want to avert at When the price of buying a unit of health to treat least some burden from Parkinson's even if these cases are Parkinson's is relatively high in terms of cost per DALY relatively expensive to treat for each unit of health gained averted, the relatively flat (solid) budget line applies, and the because of such considerations as the target age group, the optimal balance of investment in the two interventions is at Intervention Cost-Effectiveness: Overview of Main Messages | 37 Box 2.2 Continued point X. If the cost of buying a unit of health to treat that policy makers' willingness to trade off between buying Parkinson's is relatively low, then the steeper (dashed) health from the two approaches is just a straight line, in budget line applies and the relative allocation of resources is which case they would want to invest the maximum amount represented by point Y. Therefore, policy makers would possible in the lower-cost intervention (malaria) before allocate relatively more resources to treating Parkinson's turning to the higher-cost intervention (Parkinson's). The when the price of buying a unit of health through this inter- role of the cost-effectiveness information is to make policy vention is relatively low, and they would allocate fewer makers aware of differences in the price of improving health resources when the price of health obtained through this using different interventions. Interventions with a high intervention is relatively high. The figure shows the general price should, all else being equal, be used less, whereas those likely shape of an indifference curve, but one possibility is with a low price should be used to a greater extent. this way, there is no one-dimensional economic criterion that asymmetries--between profit-making providers and patients. interventions must attain to be declared economically fit, and Private providers may encourage unnecessary procedures and cost-effectiveness plays the more useful function of informing excessively invasive procedures that, in some instances, can be tradeoffs that policy makers are forced to make when investing more dangerous than no treatment at all. Governments have a in a portfolio of health interventions. role to play in lowering these information asymmetries, partly by providing information to populations, for instance, on the importance of childhood vaccinations. Moreover, even if gov- Target Audiences ernment expenditures are not directly influenced by the lack of efficiency in privately delivered health care, they are affected by The general notion of efficiency in how resources, both public the inefficiency of private health systems if private patients seek and private, are spent on improving health is of interest not public emergency care or require other state assistance. In their only to severely resource-constrained countries that each year role as large purchasers of health care, governments--even in spend only a few public dollars on health for each individual, largely privately financed health care systems, such as in the but also to relatively wealthier nations with many competing United States--exercise enormous influence over the choices of priorities for public and private resources. The primary audi- drugs used and interventions provided and can play an impor- ences for cost-effectiveness information are ministries of health tant role in promoting policies to facilitate greater efficiency in and finance and policy makers in other branches of govern- health care systems. ment in LMICs, both to help reallocate existing outlays in the The costs and efficacy of interventions may vary greatly, health sector and to allocate new monies efficiently. Other even within a single geographical region, depending on local audiences include aid agencies, international development health system capacity, cultural context, disease epidemiology, lending institutions, nongovernmental organizations, and pri- and a host of other factors. Greater efficiency in how countries vate health care providers. spend their health care resources can have a tremendous effect on the health of their populations. Box 2.3 discusses gains from improved priority setting found in the lifesaving study by Priority Setting in the Private Sector Harvard University (Tengs 1997; Tengs and others 1995). The use of the efficiency criterion in priority setting should not be limited to public resources. A large proportion of health care in developing countries is paid for out of pocket, and greater METHODOLOGY clarity on interventions that are efficient from an economic perspective is no less urgent when the payer is private: ineffi- This chapter compares the cost-effectiveness of interventions cient private spending on health care in developing countries is that cover a broad spectrum of health conditions prevalent in wasteful as well. Much of this inefficiency may be attributed to developing countries. All results are presented in U.S. dollars significant differences in knowledge--termed information discounted to the year 2001 using a 3 percent annual discount 38 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles Box 2.3 The Harvard Lifesaving Study A study by Harvard University in the United States 592,000 life years. The same amount of money could have showed potential gains of life years saved by choosing saved another 636,000 life years had funds been redirected interventions on the basis of their cost-effectiveness. The from less to more cost-effective interventions. Following study's authors assessed more than 500 types of lifesaving an assessment of varying factors regarding interventions-- interventions, defined as any behavioral or technological those that affect the most people and are the most effective, strategy that reduces the probability of premature death least cost, and most cost-effective--the study reports that among a specified target population. The study focused on if the goal is to save the most life years, cost-effectiveness is 185 interventions and the extent to which each interven- a useful approach that will result in the most efficient allo- tion was currently being implemented, without taking its cation of resources. The study indicates that choosing the cost-effectiveness into account. most cost-effective interventions could double the life The authors estimated that the selected 185 interven- years saved. tions would cost US$21.4 billion per year and would save Sources: Tengs 1997; Tengs and others 1995. rate. Chapter 15 summarizes the general guidelines governing effectiveness for some conditions, but data to estimate such the analysis leading to the results reported for all LMICs. differences are lacking. Caveats Regional Variations The findings in this chapter are subject to a number of caveats. Where possible and appropriate, intervention cost-effectiveness First, despite efforts to ensure the consistency of cost- ratios and other information have been disaggregated by World effectiveness numbers across chapters, the approaches taken in Bank region. In discussing the estimates, this chapter focuses arriving at these numbers vary significantly. Although some on differences in the costs of interventions rather than on chapters rely on cost-effectiveness numbers drawn from the differences in their effectiveness in specific regional settings, literature, other chapters have analyzed these numbers afresh although both contribute to differences in cost-effectiveness using the standardized resource costs described elsewhere. estimates across regions. Cost-effectiveness estimates also differ Table 2.1 contains definitions of indicators used to assess the among regions because of variations in underlying mortality, age structure, disease prevalence, and efficiency with which Table 2.1 Quality of Cost-Effectiveness Evidence for interventions are implemented. The analyses take all but the Interventions last of these considerations into account. Interregional cost differences are attributable to differences Level Source of cost-effectiveness evidence in the local costs of goods and services that are not easily trad- 1 Literature review of one cost-effectiveness study in one country able. For components that are tradable, such as patented drugs 2 Literature review of several cost-effectiveness studies for multiple and specialized medical equipment typically imported from countries industrial nations, the analyses assume uniform international 3 Literature review of several cost-effectiveness studies for a single costs for all LMICs, usually adjusted for local transportation intervention in a single region and distribution costs. By using a single composite set of 4 Original economic analysis by authors following the volume editors' resource costs for each region, the analyses mask intraregional guidelines in one country differences in the costs of nontradable goods, such as physician 5 Original economic analysis by authors following the volume editors' time or hospitals, but this methodology is appropriate because guidelines in one region results are presented only at the level of the region.4 Source: Authors. Interventions may differ in cost-effectiveness because they are Note: Standard region-specific age structures and underlying mortality rates were used to calcu- targeted more appropriately to some age groups rather than late DALYs. Nontradable inputs were converted into U.S. dollars at the market exchange rate. The costs of tradable inputs were assumed to be internationally consistent, as were the costs associ- others, and important gender differences may also exist in cost- ated with surgical treatments. Intervention Cost-Effectiveness: Overview of Main Messages | 39 quality of evidence on which the cost-effectiveness estimates high prevalence. Figure 2.2 reflects sets of interventions dealing are based. The tables in annex 2.B and annex 2.C indicate the with high-burden diseases, and figure 2.3 deals with relatively quality of evidence associated with each intervention. low-burden diseases. Second, almost without exception, the cost-effectiveness Within each figure, intervention clusters are displayed in the numbers do not vary with the scale at which the intervention order of increasing cost-effectiveness. Additional information is undertaken, and this is probably not the case in reality on the setting, objective, and target population of each interven- (Birch and Donaldson 1987; Johns and Torres 2005). Some tion cluster for which cost-effectiveness has been calculated is interventions, such as vaccination programs, have large setup provided in annex 2.B. The tables in annex 2.B also provide costs but marginal costs of extending coverage that decline at information on the quality of the evidence on which the data least initially. Other interventions, such as educational cam- presented are based. Furthermore, the annex tables present paigns for condom use, may be easy to target to urban popula- information on potentially avertable deaths and DALYs if the tions, but the marginal costs of expanding such interventions coverage of these interventions were expanded by a further to relatively inaccessible populations increase with coverage. 20 percentage points of the relevant population (scaling up from Therefore, many of the cost-effectiveness ratios presented here 62 percent means reaching 82 percent,not 74 percent,of the per- are useful only for modest increments in coverage, and separate tinent population). Care should be taken not to confuse this analyses may have to be conducted to determine their applica- information with the current burden of the underlying disease, bility to program start-ups and larger-scale intervention on which basis interventions were divided into high-burden changes. and low-burden diseases (figures 2.2 and 2.3, respectively).5 For Third, the cost-effectiveness numbers presented apply to example, a cost-effective treatment for CVD has only limited countries whose institutional and technical capacity in relation scope for increased scale of intervention in countries with a low to health is close to the average for the region. This evaluation burden of this disease. At the same time, in many parts of Asia is restricted to what countries could do more (or less) of, and Sub-Saharan Africa, even though HIV treatment is not a and clearly a more ambitious analysis would also cover what highly cost-effective intervention, it deserves attention because countries could do better. This issue is discussed in detail in of its sizable potential for lowering the disease burden. chapters 3 and 70. The tables in annex 2.C summarize information on inter- Finally, the estimates are based on the best available data, vention clusters for which cost-effectiveness was evaluated with which in many cases are somewhat weak. Statistically derived a metric other than DALYs. For these interventions too, details confidence bounds for the cost-effectiveness estimates are not of setting, objective, target population, and quality of the evi- provided, and in most cases, uncertainty analysis has not been dence have been provided. Given the difficulty in comparing carried out. Readers are encouraged to pay attention to the these intervention clusters with those evaluated using DALYs, order of magnitude of each estimate rather than to the specific they are excluded from figures 2.2­2.5. number presented. Observations about specific interventions follow. Ranges of cost-effectiveness estimates shown reflect geographical varia- tions across regions. ASSESSING THE EVIDENCE ON THE Prevention and Control of Tuberculosis COST-EFFECTIVENESS OF INTERVENTIONS The treatment of all forms of active tuberculosis (TB) using the Figures 2.2 and 2.3 display results gathered from other chapters directly observed treatment strategy based on short-course on cost-effectiveness ratios. In some cases, interventions are chemotherapy is among the most cost-effective of all interven- grouped on the basis of their similarity and whether they were tions available to improve health in LMICs (US$5 to US$35 per personal interventions or population-based interventions. For DALY averted except in Europe and Central Asia) (box 2.1). instance, all population-based programs to prevent HIV trans- The bacillus Calmette-Guérin (BCG) vaccination for children mission via contaminated blood and needles were grouped as is also cost-effective (US$40 to US$170 per DALY averted), but a single intervention. Note that the cost-effectiveness ranges its main effect is to reduce the burden of severe TB in children should not be interpreted as statistical confidence intervals but (TB meningitis and miliary TB). Because BCG has relatively rather as a range of "best estimates" of cost-effectiveness incor- little effect on the huge burden of pulmonary TB in adults-- porating variation across interventions included in the cluster. which constitutes the major cause of ill health resulting from Ranges for the cost-effectiveness ratios are also attributable to Mycobacterium tuberculosis--development of a new vaccine variations in the epidemiological settings in which these inter- that targets adults is highly desirable. The treatment of latent ventions were evaluated. For example, a population-based pri- TB in patients uninfected with HIV is relatively cost-ineffective mary intervention in an area of low prevalence is likely to be (US$4,000 to US$25,000 per DALY averted), but it is more less cost-effective than the same intervention in a region of cost-effective for groups of patients who are coinfected with TB 40 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles HIV/AIDS: treatment of Kaposi's sarcoma Ischemic heart disease: coronary artery bypass graft Myocardial infarction: acute management with tissue plasminogen activator, with aspirin and beta-blocker Tuberculosis (endemic, latent): isoniazid treatment Diarrheal disease: improved water and sanitation at current coverage of amenities and other interventions Diarrheal disease: cholera or rotavirus immunization Diabetes, ischemic heart disease, and stroke: media campaign to reduce saturated fat Stroke and ischemic and hypertensive heart disease: polypill by absolute risk approach Ischemic heart disease: statin, with aspirin, beta-blocker, and angiotensin-converting enzyme inhibitor Stroke (ischemic): acute management with heparin and recombinant tissue plasminogen activator Diabetes, ischemic heart disease, and stroke: legislation with public education to reduce salt content Depression: drugs with optional episodic or maintenance psychosocial treatment Alcohol abuse: 25-50% increase in excise tax rate Diarrheal disease: oral rehydration therapy for package costing US$5.50 per episode Diarrheal disease: breastfeeding promotion HIV/AIDS: antiretroviral therapy Coronary artery disease: legislation substituting 2% of trans fat with polyunsaturated fat at US$6 per adult Ischemic heart disease: aspirin, beta-blocker, with optional angiotensin-converting enzyme inhibitor HIV/AIDS: home care Myocardial infarction: acute management with streptokinase, with aspirin and beta-blocker Alcohol abuse: brief advice by primary health care doctor Alcohol abuse: excise tax, advertising ban, with brief advice Myocardial infarction and stroke: secondary prevention with polypill Alcohol abuse: advertising ban and reduced access to beverage retail Lower acute respiratory infection (0-4 years): case management package at community, facility, and hospital levels Tobacco addiction: nicotine replacement therapy Tobacco addiction: nonprice interventionsa Tuberculosis (endemic): management of drug resistance Tuberculosis (endemic, infectious or noninfectious): directly observed short-course chemotherapy Haemophilus influenzae type B, hepatitis B, diphtheria, pertussis, and tetanus: pentavalent vaccine Tuberculosis (epidemic): management of drug resistance Tuberculosis (epidemic, latent): isoniazid treatment HIV/AIDS: mother-to-child transmission prevention Diarrheal disease: hand pump, standpost, or house connection where clean water supply is limited HIV/AIDS: opportunistic infection treatment Congestive heart failure: angiotensin-converting enzyme inhibitor and beta-blocker, with diuretics Stroke (ischemic): acute management with aspirin Diarrheal disease: construction and promotion of basic sanitation where facilities are limited Problems requiring surgery: surgical ward or services in district hospital or community clinic HIV/AIDS: tuberculosis co-infection prevention and treatment Emergency medical care: staffed community ambulance Tuberculosis (epidemic, infectious): directly observed short-course chemotherapy HIV/AIDS: blood and needle safety HIV/AIDS: condom promotion and distribution Tuberculosis (endemic): BCG vaccine HIV/AIDS: sexually transmitted infections diagnosis with treatment Coronary artery disease: legislation substituting 2% of trans fat with polyunsaturated fat at US$0.50 per adult HIV/AIDS: voluntary counseling and testing Diarrheal disease: water sector regulation with advocacy where clean water supply is limited Underweight child (0-4 years): child survival program with nutrition Childhood illness: integrated management of childhood illness Intervention HIV/AIDS: peer and education programs for high-risk groups Tobacco addiction: taxation causing 33% price increase Malaria: intermittent preventive treatment in pregnancy with sulfadoxine-pyrimethamine Malaria: residual household spraying Myocardial infarction: acute management with aspirin and beta-blocker Malaria: insecticide-treated bed nets Cost-Effectiveness: Tuberculosis, diphtheria-pertussis-tetanus, polio, measles: traditional EPIb Malaria: intermittent preventive treatment in pregnancy with drugs other than sulfadoxine-pyrimethaminec Emergency medical care: training volunteer paramedics with lay first responders Diarrheal disease: hygiene promotion 0 10 100 1,000 10,000 100,000 Cost-effectiveness ratio (US$ per DALY averted) Overview Source: Authors. Note: Diseases were considered high burden for LMICs if their total avertable burden was greater than 35 million DALYs. Bars represent the range in point estimates of cost-effectiveness ratios for specific interventions included in each intervention cluster and do not represent variation across regions or statistical confidence intervals. Point estimates for LMICs were obtained directly from the relevant chapters, calculated as the midpoint of range estimates reported in of the chapters, or calculated from a population-weighted average of the region-specific estimates reported in the chapters. For details of these intervention clusters, including the specific interventions covered in each, see annex tables 2.B.1 Main and 2.B.2. Only interventions with cost-effectiveness reported in terms of DALYs are included in this figure. For interventions with cost-effectiveness reported in other units, see annex tables 2.C.1 and 2.C.2. a. Nonprice interventions for tobacco addiction include advertising bans, smoking restrictions, supply reduction, and information dissemination. Messages b. EPI = Expanded Program on Immunization. c. Chloroquine as first-line drug, artemisinin combination therapy as second-line drug and sulfadoxine-pyrimethamine as first- or second-line drug. | 41 Figure 2.2 Cost-Effectiveness of Interventions Related to High-Burden Diseases in LMICs 42 Parkinson's disease: levodopa or carbidopa with deep brain stimulation | Disease Hepatitis B: vaccination Schizophrenia: antipsychotic drugs with optional psychosocial treatment, hospital-based Control Schizophrenia: antipsychotic drugs with optional psychosocial treatment, community-based Trachoma: tetracycline or azithromycin treatments Priorities Lower acute respiratory infection (0-4 years): case management of severe and very severe cases at hospital level Bipolar disorder: lithium, valproate, with optional psychosocial treatment, hospital-based in Bipolar disorder: lithium, valproate, with optional psychosocial treatment, community-based Developing Epilepsy: second-line treatment of phenobarbital with lamotrigine or surgery Dengue: vector control Countries Traffic accidents: enforcement of seatbelt laws, promotion of child restraints, and random driver breath testing Stroke: secondary prevention with carotid endarterectomy Dengue: immunization | Ramanan Tetanus: tetanus toxoid vaccination, mix of strategies Parkinson's disease: ayurvedic treatment and levodopa or carbidopa Panic disorder: drugs with optional psychosocial treatment Laxminarayan, Haemophilus influenzae type B: H. influenzae type B-containing vaccine Dengue: improved case management Undernutrition and malnutrition (0-4 years): sustained child health and nutrition package Jeffrey Cataract: extracapsular surgery Lower acute respiratory infection (0-4 years): case management of non-severe cases at community or facility level Chow Unwanted pregnancies: family-planning programs ,and Epilepsy: first-line treatment with phenobarbital Sonbol Stroke: secondary prevention with aspirin and dipyridamole Zinc deficiency (0-4 years): supplements with oral rehydration salts A. Trachoma: trichiasis surgery Shahid-Salles Adolescent health and nutrition: school health and nutrition programs Onchocerciasis: treatment with ivermectin Traffic accidents: increased speeding penalties, enforcement, media campaigns, and speed bumps Down syndrome: prenatal screening with option of pregnancy termination Leishmaniasis: case finding with treatment Measles: second opportunity vaccination in a fixed facility Soil-transmitted helminths: albendazole 0 10 100 1,000 10,000 100,000 Cost-effectiveness ratio (US$ per DALY averted) Source: Authors. Note: Diseases were considered low burden for LMICs if their total avertable burden was less than 35 million DALYs. Bars represent the range in point estimates of cost-effectiveness ratios for specific interventions included in each intervention cluster and do not represent variation across regions or statistical confidence intervals. Point estimates for LMICs were obtained directly from the relevant chapters, calculated as the midpoint of range estimates reported in the chapters, or calculated from a population-weighted average of the region-specific estimates reported in the chapters. For details of these intervention clusters, including the specific interventions covered in each, see annex tables 2.B.1 and 2.B.2. Only interventions with cost-effectiveness reported in terms of DALYs are included in this figure. For interventions with cost-effectiveness reported in other units, see annex tables 2.C.1 and 2.C.2. Figure 2.3 Cost-Effectiveness of Interventions Related to Low-Burden Diseases in LMICs and HIV. In the context of TB control, antiretroviral therapy effective (US$10 to US$500 per DALY averted), with treatment for HIV/AIDS is likely to be useful in extending the lives of becoming significantly more cost-effective for patients who patients successfully treated for TB. also have access to antiretroviral treatment. Few studies evalu- Multidrug-resistant TB is much more expensive to treat than ate the cost-effectiveness of providing antiretroviral treatment, drug-susceptible TB--2 to 10 times the cost of standard first- and even these are limited to clinical trial settings and are not line regimens for drug-susceptible TB--and this is one reason directly applicable to the resource-poor settings in which anti- why priority should be given to preventing its emergence and retroviral treatment is being expanded. Economic evaluation of spread. The management of drug resistance through the use of the cost-effectiveness of antiretroviral treatments based only on a standardized regimen that includes second-line drugs costs health outcomes for the treated patient is incomplete because roughly US$70 to US$450 per DALY averted. Individualized of the large nonhealth impacts of HIV/AIDS and the effect of treatment regimens for multidrug-resistant TB--that is, with treatment on prevention of HIV transmission. drug combinations adjusted to the resistance pattern of each The cost-effectiveness of antiretroviral treatments is highly patient--are more costly but usually yield higher cure rates. variable across settings as a function of drug prices and adher- Individualized treatment is harder to implement on a large scale ence rates. In low-cost settings with high adherence rates, anti- but may not be less cost-effective than standardized treatment retroviral treatment is moderately cost-effective (US$350 to with regimens that include second-line drugs. The set of US$500 per DALY averted); however, it can be a significantly interventions needed to manage drug-resistant TB and TB poor value for resources spent in low-adherence settings if drug associated with HIV requires higher levels of investment than resistance is allowed to emerge and proliferate. Little is known the basic directly observed treatment strategy, but its cost is still about how to achieve necessary adherence levels (80 to 90 per- typically less than US$1 for each day of healthy life gained. cent) at large scale at an affordable cost in low-income settings. Thus, a strong economic argument exists for integrating such To this end, research on effective, low-cost interventions to interventions into an enhanced strategy for TB control. achieve long-term adherence to antiretroviral treatments (using support groups and other complementary interven- Prevention and Treatment of HIV/AIDS tions) in resource-poor settings is an urgent priority. Despite the scale and relentless progression of the HIV/AIDS epidemic, important strides have been made in developing Childhood Illnesses and Mortality among cost-effective interventions for both prevention and treatment. Children under Five Neonatal mortality rates and mortality rates for children under Prevention. Although remarkably little rigorous evaluation five can be reduced by large margins, at an affordable cost, by has been conducted, population-based programs to prevent using interventions proven effective in low-income settings. HIV/AIDS appear to be highly cost-effective approaches in Improvements are likely to come from increasing the coverage countries with high HIV/AIDS prevalence where the epidemic of preventive measures, such as breastfeeding, and from is generalized. These programs include voluntary testing and expanding the scope of existing childhood vaccines beyond the counseling (US$14 to US$261 per DALY averted); peer-based traditional six antigens in areas where existing coverage is rela- programs to educate high-risk groups, including sex workers tively high and where new antigens address diseases of signifi- and injecting drug users (US$1 to US$74 per DALY averted); cant burden, particularly pneumococcal vaccines. Curative and social marketing, promotion, and distribution of condoms interventions--including case management of acute respira- (US$19 to US$205 per DALY averted). Programs to improve tory infections, malaria, and diarrhea--hold promise for lower- blood and needle safety, while highly cost-effective (US$4 to ing the 6 million preventable deaths each year in this age US$51 per DALY averted), are limited in terms of the burden of group. disease they can avert. Prevention of mother-to-child transmission using a single dose of nevirapine in generalized epidemic settings (US$6 to Neonatal Mortality. An estimated 4 million deaths occur dur- US$12 per DALY averted) stands out for its combination of well- ing the first 28 days of life, accounting for 38 percent of all documented high cost-effectiveness and significant avertable deaths of children under five. Causes include infections (36 per- infections and deaths. Treatment of sexually transmitted cent, including neonatal sepsis, pneumonia, diarrhea, and infections to lower the risk of HIV transmission, although less tetanus), preterm birth (27 percent), and asphyxia (23 percent). well proven, also appears to be highly cost-effective (US$16 to Intensive care is not required to save most of these babies. US$105 per DALY averted). Developed countries and some low-income countries--for instance, Sri Lanka--have achieved neonatal mortality rates of Treatment. For care of people living with HIV/AIDS, 15 per 1,000 without intensive care, which is less than a third of treatment of most infectious opportunistic infections is cost- current neonatal mortality rates in Sub-Saharan Africa. Intervention Cost-Effectiveness: Overview of Main Messages | 43 Adding a set of community-based interventions--including than US$1,600 in Eastern Europe. Cost-effectiveness ratios are promoting healthy behaviors, such as breastfeeding, and less than US$20 per DALY averted in all regions other than providing extra care of moderately small babies at home Europe and Central Asia. The cost-effectiveness of the tetanus through cleanliness, warmth, and exclusive breastfeeding, plus toxoid vaccine also varies widely by region from under US$400 community-based management of acute respiratory infec- per death averted and under US$14 per DALY averted in Sub- tions--to the standard maternal and child health package is Saharan Africa and South Asia to more than US$190,000 per likely to be highly effective. The cost of a year of life saved using death averted and more than US$15,000 per DALY averted in this approach could be as low as US$100 to US$257 in India Europe and Central Asia. (US$221 to US$568 per DALY averted) and US$100 to US$270 Adding additional antigens to national programs has been in Sub-Saharan Africa (US$183 to US$493 per DALY averted). successfully accomplished in many countries. Expanding the Use of these approaches is feasible now in most countries. vaccination schedule to include a second opportunity for Adding a clinical package that includes essential newborn care measles through either routine or campaign-based approaches (warmth, cleanliness, and immediate breastfeeding); neonatal costs between US$23 and US$228 per death averted and under resuscitation; facility-based care of small newborns; and emer- US$4 per DALY averted in regions other than Europe and gency care of ill newborns to the maternal and child health Central Asia. Other new vaccines are less cost-effective because package has been shown to be highly cost-effective in India of their high unit costs per dose, but they may be worthwhile, (US$11 to US$265 per year of life saved, or US$24 to US$585 especially in regions of high disease prevalence. For instance, per DALY averted) and Sub-Saharan Africa (US$25 to US$360 the pentavalent vaccine (diphtheria, pertussis, tetanus, per year of life saved, or US$46 to US$657 per DALY averted); hepatitis B, and Haemophilus influenzae type B) was estimated however, clinical care will require significant initial investment to have a cost per death averted ranging from US$1,433 to to raise coverage. greater than US$40,000 and cost-effectiveness of US$42 per Basic resuscitation of newborns using a self-inflating bag DALY averted in Sub-Saharan Africa and greater than US$245 that is available for as little as US$5 in LMICs can save lives at per DALY averted in other regions. The cost of adding a yellow low cost in areas where a midwife is available. Providing two fever vaccine ranges from US$834 per death averted and US$26 tetanus toxoid immunizations costing less than US$0.20 each per DALY averted in Sub-Saharan Africa to US$2,810 per death to all pregnant women would avert more than 250,000 deaths averted and US$39 per DALY averted in Latin America and the at low cost and is eminently achievable. Improving maternal Caribbean. and child health services delivered through a combination of Because certain regions and countries contain the largest family- and community-level care, outreach, and clinical care burden of disease, such as measles in India and Nigeria, target- will improve the survival of newborns and children and reduce ing scarce public health resources to those geographic areas stillbirths and maternal deaths. could potentially yield high returns to investment. Although immunization may have relatively low incremental cost- Vaccinations. Childhood vaccinations, long recognized as effective ratios, the total budget requirements for maintaining among the most cost-effective uses of limited health resources or increasing coverage rates, as well as for introducing new in low-income countries, prevented more than 3 million vaccines, can account for a large share of government health deaths in 2001. National immunization programs traditionally budgets. have included vaccines against TB, diphtheria, tetanus, pertus- The cost-effectiveness ratios of vaccination interventions sis, poliomyelitis, and measles at a cost per fully immunized presented here are based on estimates of their current costs child of US$13 to US$24, depending on coverage levels and and effectiveness; but they could change substantially with type of delivery strategy. The total cost in developing countries changing costs and the development of new interventions. For for national programs in 2001 ranged from US$717 million to instance, multivalent pneumococcal conjugate vaccines have US$1.4 billion, with an estimated cost per death averted rang- shown the potential to reduce the incidence of invasive pneu- ing by region from under US$275 (under US$10 per DALY mococcal disease while lowering the need for antibiotic use and averted) in Sub-Saharan Africa and South Asia to US$1,754 the likelihood of drug resistance. The current price of these (US$20 per DALY averted) in Europe and Central Asia. vaccines makes them expensive to most people in the develop- The cost-effectiveness of scaling up immunization coverage ing world. However, with future price decreases, these vaccines with the traditional Expanded Program on Immunization could be adopted widely and could markedly lower the impact (EPI) vaccines is highly dependent on the underlying preva- of the most common causes of morbidity and mortality in lence of illness, starting coverage levels and trajectories, and children under five (excluding the neonatal period). Moreover, mix of delivery strategies (whether facility-based strategies, new vaccines being developed could be included in the EPI campaigns, or mobile and outreach modalities). The cost per schedule, including vaccines that protect against rotavirus, death averted varies by region, from US$162 in Africa to more malaria, human papilloma virus associated with cervical 44 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles cancer, HIV/AIDS, and dengue. With future demonstrations of US$12,000 per death averted, or US$300 to US$400 per DALY reasonable cost-effectiveness, these vaccines could become a averted). Expensive interventions, such as bone marrow component of the set of attractive interventions. transplantation or repeated transfusions, are seldom needed. At US$10,000 or more per DALY averted, treatment for Acute Respiratory Infections. Even though vaccination transfusion-dependent thalassemias is expensive and probably strategies can be cost-effective in lowering the disease burden unaffordable to all but the rich in LMICs. A feasible strategy to related to acute respiratory infections, case management may deal with the thalassemias is to screen couples to determine also be an efficient use of financial resources, although more their risk of having an affected child, followed by prenatal test- demanding of health system capacity. Moreover, community ing--a relatively expensive proposition--of couples at high case management and case management at a health care facil- risk. Information is then available to parents to help them ity may be of comparable cost-effectiveness. In fact, treating determine whether to terminate the pregnancy. Such strategies nonsevere pneumonia at health care facilities using a combina- appear to have worked in Cyprus, Greece, and Italy, all coun- tion of oral antimicrobials and acetaminophen (US$24 to tries that formerly had a high incidence of thalassemias. US$424 per DALY averted) is more cost-effective than a similar treatment administered at home by a health care worker (US$139 to US$733 per DALY averted). Treating severe pneu- Ongoing Challenges: Malaria and Other monia in a hospital facility is more expensive (US$1,486 to Tropical Diseases US$14,719 per DALY averted). Despite health researchers' relative neglect of diseases predom- inantly found in the tropics, interventions to control--and in Diarrheal Disease. Among interventions against diarrheal some cases even eliminate--these diseases rank among the disease during the first year of life, breastfeeding promotion most cost-effective of all available options. programs (US$527 to US$2,001 per DALY averted), measles immunization (US$257 to US$4,565 per DALY averted), and Malaria. In countries where malaria is prevalent, both preven- oral rehydration therapy (US$132 to US$2,570 per DALY tion and effective treatment of this disease are highly cost- averted) are relatively cost-effective compared with rotavirus effective and can result in large health gains. Prevention tools immunizations (US$1,402 to US$8,357 per DALY averted) and include insecticide-treated bednets (US$5 to US$17 per DALY cholera immunizations (US$1,658 to US$8,274 per DALY averted) and indoor residual spraying where DDT, malathion, averted). The cost-effectiveness of oral rehydration therapy deltamethrin, or lambda-cyhalothrin is applied to surfaces is extremely sensitive to the cost of the package. The cost- inside homes as a spray or deposit for prolonged action (US$9 effectiveness of this intervention can be as low as US$132 per to US$24 per DALY averted for Sub-Saharan Africa). DALY averted for an assumed cost per child of US$0.70. An Intermittent preventive treatment of malaria during preg- important reason for the relatively unfavorable cost- nancy using sulfadoxine-pyrimethamine is a highly cost- effectiveness ratios for diarrheal disease is that significant effective intervention (US$13 to US$24 per DALY averted) to reductions in mortality from this condition have already been decrease neonatal mortality and reduce severe maternal ane- achieved and further gains are likely to be more expensive. mia. Changing first-line treatment for malaria from chloro- Further improvements in water and sanitation (US$1,118 to quine, a drug that is ineffective in many parts of the world, to US$14,901 per DALY averted from diarrheal disease) are artemisinin-based combinations offers the advantage of faster generally less cost-effective in regions where access to these cures and potential reductions in transmission, with cost- amenities is adequate and other interventions against diarrheal effectiveness ratios of less than US$150 per DALY averted. disease exist. However, in areas with little access to water and Changing to sulfadoxine-pyrimethamine may be slightly more sanitation facilities, improving access can be highly cost- cost-effective initially because of the lower cost of this drug effective (US$94 per DALY averted for installation of hand relative to artemisinin-based combinations; however, this pumps and US$270 per DALY averted for provision and pro- advantage is likely to be eroded quickly because of the rapid motion of basic sanitation facilities). expected growth of parasite resistance. Inherited Disorders of Hemoglobin. Inherited hemoglobin Lymphatic Filariasis, Onchocerciasis, and Chagas Disease. disorders, including sickle cell anemia and the thalassemias, Annual mass drug administration to treat the entire population affect roughly 500,000 babies born each year and cause early at risk for a period long enough to interrupt transmission is a death for many of them. Prenatal screening for sickle cell dis- cost-effective approach for eliminating lymphatic filariasis in ease, which is expensive, can be replaced by much cheaper areas of high prevalence (US$4 to US$8 per DALY averted). An newborn screening. Antibiotic prophylaxis is moderately cost- alternative approach is to fortify salt with diethylcarbamazine effective at preventing death in the first few years (US$8,000 to (US$1 to US$3 per DALY averted) and to use ivermectin in Intervention Cost-Effectiveness: Overview of Main Messages | 45 countries where onchocerciasis is coendemic. Onchocerciasis tion of the overall burden of disease in LMICs. For the most control programs have been highly successful in West Africa: part, interventions to provide micronutrient supplementation investigators have estimated the cost-effectiveness of can prevent malnutrition in children at a fairly low cost. They community-directed ivermectin treatment programs at include breastfeeding support programs (US$3 to US$11 per roughly US$6 per DALY averted when the drug has been pro- DALY averted and US$100 to US$300 per death averted) and vided free of charge. The cost of vector control to prevent--and growth monitoring and counseling (US$8 to US$11 per perhaps eliminate--Chagas disease has been estimated at DALY averted). Specific micronutrient supplementation pro- US$260 per DALY averted. grams can be implemented either by distributing capsules or by fortifying sugar, salt, water, or other essentials. In address- Leishmaniasis and African Trypanosomiasis. Feasible inter- ing vitamin A deficiencies, capsule distribution (US$6 to vention opportunities exist even for tropical diseases for which US$12 per DALY averted) is more cost-effective than sugar control measures are relatively less effective. Improved case fortification (US$33 to US$35 per DALY averted), especially management and immunization (currently undergoing clinical in countries where the prevalence of vitamin A deficiency is trials) for dengue (US$587 to US$1,440 per DALY averted) low. However, fortification of salt, sugar, and cereal in the case are relatively cost-effective compared with environmental of iron deficiency and fortification of water and salt in the vector control (more than US$2,000 per DALY averted). case of iodine deficiency is less expensive than distributing Leishmaniasis treatment is also extremely cost-effective supplements for mild deficiency, though pregnant women (US$315 per death averted and US$9 per DALY averted), as is and severely anemic or iodine-deficient people may still treating African trypanosomiasis patients in the second stage of require supplementation. Overall cost-effectiveness is US$66 the disease using melarsoprol or eflornithine (US$10 to US$20 to US$70 per DALY averted for iron fortification programs per DALY averted). and US$34 to US$36 per DALY averted for iodine fortifica- tion programs. Helminthic Infections. Helminthic infections, although not a major contributor to deaths in tropical regions, have a signifi- cant effect on health, growth and physical fitness, school Cancer Prevention and Treatment attendance, worker productivity, and earning potential. Mass Screening for breast cancer using clinical breast examination school-based treatment of soil-transmitted helminths (Ascaris, (CBE) is estimated to be cost-effective at US$552 per life year Trichuris, and hookworm) using albendazole costs US$2 to saved for biennial screening of women from age 40 to 60. This US$9 per DALY averted. Although the cost of treating efficacy of CBE is related to the large percentage of tumors with schistosomiasis with praziquantel is significantly greater a poor prognosis observed in developing countries. In this set- (US$336 to US$692 per DALY averted), a combination of ting, CBE is estimated to be more cost-effective than mam- albendazole and praziquantel is extremely cost-effective (US$8 mography: mammograms every two years result in 10 percent to US$19 per DALY averted). more life years saved than annual CBE, but the cost is more than 100 percent greater. As with any screening program, cost- effectiveness is greater with higher underlying prevalence of Maternal and Neonatal Health disease. Given the hugely disproportionate burden of maternal and In general, cancer prevention, when feasible, is far more neonatal deaths in LMICs, identifying affordable, easy-to- cost-effective than treatment. The cost-effectiveness of initial implement interventions to prevent these deaths is a priority. treatment is between US$1,300 and US$6,200 per year of life Evidence from South Asia and Sub-Saharan Africa suggests saved for the more treatable cancers of the cervix, breast, oral that improved primary-level coverage with a package of inter- cavity, colon, and rectum and between US$53,000 and ventions is extremely cost-effective (US$3,337 to US$6,129 per US$163,000 per year of life saved for the less treatable cancers death averted and US$92 to US$148 per DALY averted). of the liver, lung, stomach, and esophagus. Postmastectomy Improvements in the quality of prenatal and delivery care are radiation might be more cost-effective in developing coun- of similar cost-effectiveness (US$2,729 to US$5,107 per tries, where the cost of radiation treatment can be relatively death averted and US$82 to US$142 per DALY averted). An low compared to developed countries. Palliative care for ter- important finding is that improving the quality of care and minally ill cancer patients can be a challenge in resource- expanding coverage are of comparable cost-effectiveness. constrained settings, where opioid drugs, a cost-effective option, may be in short supply. Studies from developed coun- Improving Nutrition tries indicate that more advanced treatments to relieve pain The direct and indirect effects of undernutrition and and side effects of chemotherapy may be cost-effective under micronutrient deficiencies account for a significant propor- certain conditions. 46 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles Mental and Neurological Disorders more cost-effective than when combined with psychosocial Mental disorders are a heterogeneous group of conditions with treatment. Psychosocial treatment without drug treatment considerable variation in both the cost of the interventions and was of comparable cost-effectiveness (US$338 to US$927 per the burden reduction associated with such interventions. DALY averted). Interventions to treat depression, bipolar disorder, and schizo- The use of tricyclic antidepressants was more cost-effective phrenia rank among the least cost-effective of interventions than benzodiazepines, which are still commonly prescribed for considered in this volume. However, the potentially significant anxiety disorders and produce dependence in many patients. benefits to family members and to society as a whole are not Overall, the cost-effectiveness of a package of mental health captured by the DALY methodology and should be balanced interventions that addressed all four sets of disorders is against the relatively high cost of improving health of people between US$1,429 and US$2,902 per DALY averted, depending with these disorders. For many disorders, drug treatment has on the region. been shown to be effective, especially when combined with psychosocial treatment that includes cognitive-behavioral Parkinson's Disease and Epilepsy. Ayurvedic treatment, a approaches to managing symptoms and improving adherence form of traditional medicine used in India, is relatively cost- to medications, group therapy, and family interventions. effective in treating Parkinson's disease (US$750 per DALY averted). Less cost-effective interventions include a combina- tion of levodopa and carbidopa (US$1,500 per DALY averted), Schizophrenia and Bipolar Disorder. Drug treatment which are used to treat the debilitating symptoms and delay the accompanied by psychosocial treatment delivered through a progress of the disease, and deep-brain stimulation (US$31,000 community-based service was found to be the most cost- per DALY averted). effective approach for severe mental disorders such as Cost-effective options for treating epilepsy are available, schizophrenia and bipolar disorder. Newer antipsychotic and especially the use of phenobarbital to help control seizures mood-stabilizing drugs have recently become less expensive; (US$89 per DALY averted), but few eligible patients receive even so, they are less cost-effective than drugs that have been treatment. More expensive options, such as lamotrigine or sur- available for many years. For example, family psychoeducation gery, are significantly less cost-effective than phenobarbital for was much more cost-effective with haloperidol (US$1,743 to first-line treatment; however, they are cost-effective for the US$4,847 per DALY averted) compared with a newer anti- small proportion of epilepsy patients who do not respond to psychotic drug (risperidone) in treating schizophrenia phenobarbital. (US$10,232 to US$14,481 per DALY averted). For bipolar affective disorder, the combination of family psychoeducation with the older medication lithium (US$1,587 to US$4,928 per Multipronged Strategy to Prevent and Treat CVD DALY averted) is more cost-effective than the combination of CVD, including ischemic heart disease, congestive heart failure, family psychoeducation with the newer sodium valproate and stroke, is the single most important cause of death world- (US$2,765 to US$5,908 per DALY averted). wide; interventions to treat CVD are likely to account for increasingly greater proportions of health care expenditures in Depression and Panic Disorder. Treating the more common developing countries. depressive and anxiety disorders was more cost-effective than treating the more severe disorders; the interventions were less Population-Based Primary Prevention. Interventions to expensive, and the reduction in disability was greater. For modify lifestyles can effectively lower the risk of coronary depression, drug therapy with tricyclic antidepressants artery disease and stroke without expensive health infrastruc- (imipramine or amitriptyline) costs US$478 to US$1,288 per ture. They include lowering the fat composition of the diet, DALY averted. Managing depression as a chronic illness with limiting sodium intake, avoiding tobacco use, and engaging in case management to reduce relapses did not greatly decrease regular physical activity. The costs and the effectiveness of these the cost-effectiveness (US$749 to US$1,760 per DALY approaches vary widely with the socioeconomic and cultural averted). Using newer medications with fewer side effects and context in which they are contemplated. potentially greater compliance (an advantage if medications Replacing dietary trans fat from partial hydrogenation with need to be taken long term)--for example, a generic selective polyunsaturated fat is likely to be extremely effective in popu- serotonin reuptake inhibitor (SSRI) such as fluoxetine-- lations in South Asia, where the intake of trans fat is high. If increased the cost somewhat (US$1,229 to US$2,459 per such replacement is done during manufacture at a relatively DALY averted). Finally, the treatment of panic disorder using low cost rather than through changes in individual behavior, a tricyclic antidepressants (US$305 to US$619 per DALY cost-effectiveness ratio of US$25 to US$73 per DALY averted averted) and SSRIs (US$567 to US$865 per DALY averted) was can be attained. Replacing saturated fat with monounsaturated Intervention Cost-Effectiveness: Overview of Main Messages | 47 fat in manufactured foods accompanied by a public education the combination of aspirin and a beta-blocker is highly cost- campaign is relatively expensive in the base case (US$1,865 to effective (US$386 to US$545 per DALY averted). In all regions, US$4,012 per DALY averted), although the cost per DALY treating congestive heart failure using enalapril and the beta- averted is highly sensitive to both the relative risk reduction in blocker metoprolol is also highly cost-effective (approximately CVD events as well as the cost per individual. Reducing salt in US$200 per DALY averted). manufactured foods through a combination of legislation and education campaigns is also relatively expensive in the base Acute Management and Secondary Prevention of Stroke. case (US$1,325 to US$3,056 per DALY averted), but could be The cost of treating acute ischemic stroke using aspirin is much more cost-effective in high-density populations with a US$150 per DALY averted. Relatively cost-ineffective interven- high salt intake. Little evidence is available on the cost- tions involve the use of a tissue plasminogen activator effectiveness of programs to encourage exercise and other (US$1,300 per DALY averted) and anticoagulants such as behavior changes by individuals. heparin or warfarin (US$2,700 per DALY averted). Aspirin is the lowest-cost option for secondary prevention of stroke Personal Interventions. Prevention strategies targeted at indi- (US$3.80 per single percentage point decrease in the risk of a viduals at high risk for CVD--measured as a combination of second stroke within two years or US$70 per DALY averted). nonoptimal blood pressure and cholesterol, lifestyle, and The combination of the antiplatelet medication dipyridamole genetic risk factors--can be effective, especially when imple- and aspirin is equally cost-effective (US$93 per DALY averted). mented in tandem with population-based measures. A previ- In contrast, carotid endarterectomy is expensive for secondary ous cardiovascular event is a reliable predictor of a second prevention (US$1,500 per DALY averted). event. The cost-effectiveness of primary prevention of CVD may vary greatly depending on the underlying risk factors, the Strategies for Injury Prevention age of the patient, and the cost of medications. Increasing economic development and use of motor vehicles Single-pill combinations of blood pressure­lowering med- has resulted in increases in traffic-related deaths and injuries; ications, statins, and aspirin offer the potential dual benefit of these events account for roughly a third of the burden from all being highly effective at lowering the risk of CVD and facilitat- unintentional injuries in LMICs. ing patient compliance with the ongoing drug regimen. A Speed bumps appear to be the most cost-effective and cost hypothetical multidrug regimen that includes generic aspirin, less than US$5 per DALY averted in all regions if installed at the a beta-blocker, a thiazide diuretic, an angiotensin-converting most dangerous junctions that account for 10 percent of junc- enzyme (ACE) inhibitor, and a statin may be implemented at a tion deaths. Increased speeding penalties, media coverage, and cost-effectiveness ratio of US$721 to US$1,065 per DALY enforcement of traffic laws are only slightly less cost-effective. averted compared with a baseline of no treatment in a popula- Motorcycle helmet legislation (US$467 per DALY averted tion with an underlying 10-year CVD risk of 35 percent. The in Thailand), bicycle helmet legislation (US$107 per DALY use of the multidrug regimen for prevention in patients with a averted in China), and improved enforcement of traffic codes lower underlying CVD risk improves health benefits, but costs through a combination of enforcement and information cam- increase more than proportionately. paigns (US$5 to US$169 per DALY averted) are relatively more expensive but deserve greater attention, given the growing Acute Management of CVD. The cost of treating acute health burden associated with rising levels of vehicle owner- myocardial infarction using aspirin and beta-blockers is less ship. Research has demonstrated that seat belts and child than US$25 per DALY averted in all regions. Relatively more restraints are effective in the developed world, and lowering expensive interventions that offer marginally greater effective- their costs and encouraging their routine use may improve ness include the use of thrombolytics such as streptokinase their cost-effectiveness in LMICs. (US$630 to US$730 per DALY averted) and tissue plasminogen Key interventions to reduce intentional violence, both self- activator (US$16,000 per DALY averted). inflicted (suicides) and interpersonal (homicides and war- The combination of aspirin and the beta-blocker atenolol related deaths), include changing cultural norms, reducing has been shown to be highly cost-effective in preventing the access to guns, and improving criminal justice and social wel- recurrence of a vascular event. The incremental cost- fare systems, but these interventions are difficult to evaluate effectiveness ratio of sequentially adding an ACE inhibitor such using a cost-effectiveness framework, and a cost-benefit analy- as enalapril (US$660 to US$866 per DALY averted), a statin sis is more appropriate. Studies of interventions targeting inter- such as lovastatin (US$1,700 to US$2,000 per DALY averted), personal violence in developed countries show that behavioral, and coronary artery bypass graft (more than US$24,000 per legal, and regulatory interventions cost less than the money DALY averted) to the baseline therapy is greater when hospital they save, in some cases by an order of magnitude. Providing facilities are available. In regions with poor access to hospitals, shelters for victims of domestic violence in the United States 48 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles has a benefit-cost ratio of 6.8 to 18.4. Implementing a gun reg- (US$480 to US$819 per DALY averted) in all regions, but istration law in Canada involved a one-time cost of US$70 mil- combining this intervention with a tax on alcohol increases lion, compared with annual health-related costs of US$50 mil- cost-effectiveness (US$260 to US$533 per DALY averted) in all lion for firearm-related injuries in that country. Interventions regions except Sub-Saharan Africa. for troubled youths to reduce criminal activity include mentor- ing (with net benefits ranging from US$231 to US$4,651 per participant), family therapy (US$14,545 to US$60,721), and Packaging of Interventions and Services aggression replacement therapy (US$8,519 to US$34,071). This section examines the overall cost-effectiveness of a service level, including all conditions addressed as part of a package of services, rather than evaluating individual interventions Policy Interventions to Lower Alcohol and Tobacco Use separately. The growing prevalence of smoking, especially among women in LMICs, is a serious threat to health. Interventions to reduce Emergency and Hospital Care. The cost per death averted of tobacco use are noteworthy not just because they are highly training lay first responders and volunteer paramedics is cost-effective but also because the burden of deaths and dis- between US$130 and US$283 (or US$5 to US$11 per DALY ability that they can avert is large. Tobacco control through tax averted) depending on the region. Ambulances outfitted with increases often has dual benefits of increasing tax revenues as trained paramedics can avert deaths at a cost of US$1,148 to well as discouraging smoking initiation and encouraging US$3,479 (US$46 to US$137 per DALY averted) in urban set- smokers to quit. The cost-effectiveness of a policy to increase tings and US$3,457 to US$10,449 (US$140 to US$410 per cigarette prices by 33 percent ranges from US$13 to US$195 DALY averted) in rural settings. Although the evidence for the per DALY averted globally, with a better cost-effectiveness ratio cost-effectiveness of district and referral hospitals is very (US$3 to US$42 per DALY averted) in low-income countries. limited, it does indicate that basic hospital care at the district In comparison, nicotine replacement therapy (US$55 to level could be highly cost-effective (US$13 to US$104 per US$751 per DALY averted) and nonprice interventions, includ- DALY averted). ing banning advertising, providing health education informa- tion, and forbidding smoking in public places, are relatively Surgery. Some types of surgery are highly cost-effective as part less cost-effective (US$54 to US$674 per DALY averted) in low- of a country's health strategy. These include providing surgical income countries but are still important components of any care to injury victims, including those suffering from head tobacco control program. trauma and burns; handling obstetric complications, such as In regions with a relatively high prevalence of high-risk obstructed labor or hemorrhage; and undertaking elective sur- alcohol use--that is, Europe and Central Asia, Latin America gery to address conditions such as cataracts and otitis media and the Caribbean, and Sub-Saharan Africa--tax increases to that have a significant impact on the quality of life. In areas lower alcohol use are extremely cost-effective (US$105 to of high prevalence, cataract surgery can be extremely cost- US$225 per DALY averted). However, in regions with a lower effective at roughly US$100 per DALY averted. prevalence of high-risk use--namely, East Asia and the Pacific Many of these surgical interventions--including improved and South Asia--tax-based policies can be among the least resuscitation and airway management using relatively simple cost-effective interventions (more than US$2,500 per DALY procedures such as chest tubes and tracheostomy, improved averted). Advertising bans are among the most cost-effective fracture management, and improved management of burns (but least studied) of all interventions to reduce high-risk covering less than 30 percent of the body--require only the drinking in all regions (US$134 to US$280 per DALY averted). basic facilities offered by district hospitals. The quality of sur- In East Asia and the Pacific, a comprehensive ban on advertis- gery and the risk of complications vary widely, and adequate ing and reduced access to retail outlets are highly cost-effective health system capacity is an important consideration. For the interventions (US$123 to US$146 per DALY averted). Random typical surgical facility located in a district hospital in an LMIC, breath testing is one of the least cost-effective interventions the average cost per DALY averted for a representative set of to reduce the alcohol-related disease burden (US$973 to surgical procedures is between US$70 and US$230. General US$1,856 per DALY averted). In Sub-Saharan Africa, however, surgery at the district hospital is cost-effective relative to other averting the burden of disease associated with drunk driving is interventions in South Asia and Sub-Saharan Africa because of an important priority and is addressed effectively through such the relatively low input costs related to infrastructure and the policies as random breath testing and stricter enforcement of high level of the avertable disease burden. Examples of surgical drunk-driving laws (US$531 per DALY averted). Providing interventions with poor cost-effectiveness include first-line high-risk drinkers with brief advice from a physician in pri- treatment of epilepsy with surgery, which is useful only mary care settings is of intermediate cost-effectiveness to patients who do not respond to drug treatment, and Intervention Cost-Effectiveness: Overview of Main Messages | 49 percutaneous transluminal coronary angioplasty for cardiovas- cost-effective interventions addressing high-burden diseases, 8 cular events. are associated with these two sources of ill health alone. Other interventions that are both cost-effective and address high- Integrated Management of Childhood Illnesses. An inter- burden diseases include nutritional support (including breast- vention package consisting of exclusive breast feeding; vitamin feeding advice for mothers) for children under the age of four, A and zinc supplementation; screening for immunization; and and increasing coverage of the EPI. Oral rehydration therapy case management of pneumonia, malaria, and diarrhea, for diarrheal disease can be cost-effective if the cost of the including oral rehydration therapy, costs approximately package is relatively low (that is, less than US$1 per child per US$4.10 per child in Sub-Saharan Africa and is a cost-effective treatment). approach (US$38 per DALY averted) to improving the health Table 2.2 identifies interventions relevant to South Asia and of children under five when program coverage is 50 percent. Sub-Saharan Africa that have been evaluated in this volume and have the greatest potential to reduce the burden of disease in those regions at an affordable price.6 The table also high- Value of Doing Things Better lights interventions that address conditions that account for a Intervention quality is an important determinant of cost- moderate to high burden of disease but at a relatively high cost. effectiveness, and improving quality can be an efficient way to use resources. Community health status tends to be correlated Personal versus Population-based Interventions with the quality of health service facilities, which can be Figure 2.6 displays a histogram of intervention clusters catego- enhanced even in resource-constrained settings. Indeed, rized as either population based or personal (see annex 2.A for resource-poor settings have the greatest potential for improving definitions). A greater number of personal intervention clusters quality at low cost. In the case of acute respiratory infections, than population-based intervention clusters are categorized as for example, the cost-effectiveness of improving the quality of being highly cost-effective. Although this result may be partly care by implementing an educational activity for providers an artifact of the way in which we have grouped interventions ranges from US$132 to US$800 per life saved (US$4 to US$28 into clusters, it lends some support to the observation first per DALY averted) when initial intervention quality is poor made in the first edition of Disease Control Priorities in and infections are widespread. Quality improvements can cost Developing Countries (Jamison and others 1993) that personal between US$2,000 and US$5,000 per life saved (US$70 to interventions are not necessarily less cost-effective than US$176 per DALY averted) with improved baseline quality, low population-based interventions. Population-based interven- disease prevalence, or both. Educational interventions to tions are cost-effective when effectively targeted to populations improve the quality of diarrhea treatment can be extremely in which disease prevalence (or the potential prevalence and cost-effective (less than US$18 per DALY averted) depending subsequent mortality if the interventions are not imple- on these two factors. mented) is high. For example, primary prevention of acute myocardial infarction using aspirin is not nearly as cost- effective as secondary prevention in patients who have already Regional Analyses for South Asia and Sub-Saharan Africa suffered a stroke or myocardial infarction, because this latter Given the significant health burden borne by countries in category has, by virtue of the first event, identified itself as South Asia and Sub-Saharan Africa, cost-effectiveness informa- being at higher risk than the general population. Similarly, tion for interventions related to high-burden health conditions malaria prevention programs will be highly cost-effective in is presented for these two regions. In South Asia (figure 2.4), areas where malaria is a serious problem but less so in countries CVD-related interventions, including tobacco taxes, treatment where the burden of this disease is less and people are better of acute myocardial infarction with aspirin and beta-blockers, served by treatment with an effective antimalarial. and increasing coverage of the EPI program, rank among the most cost-effective interventions. Treatment of latent TB, coro- DISCUSSION nary artery bypass graft for ischemic heart disease, treatment of depression, and cholera immunization to prevent diarrheal dis- Since the publication of the previous edition of this book, the ease rank among the least cost-effective. Vitamin A deficiency, epidemiological and demographic profiles of many LMICs and leprosy, and epilepsy are important conditions that impose a the range of available health interventions have changed signif- relatively lower burden of DALYs on this region, but a number icantly. This edition has the benefit of hindsight in looking of highly cost-effective interventions to deal with each of these back at the variety and affordability of interventions that were conditions could be scaled up. evaluated in the previous edition, both to see how the optimal In Sub-Saharan Africa (figure 2.5), HIV/AIDS and malaria mix of strategies may have changed in the intervening period rank among the highest-burden conditions. Of the 16 most and to ascertain trends. 50 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles Ischemic heart disease: coronary artery bypass graft Myocardial infarction: acute management with tissue plasminogen activator, with aspirin and beta-blocker Tuberculosis (endemic, latent): isoniazid treatment Diarrheal disease: rotavirus or cholera immunization Diarrheal disease: improved water and sanitation at current coverage of amenities and other interventions Ischemic heart disease: statin, with aspirin, beta-blocker, and angiotensin-converting enzyme inhibitor Diabetes, ischemic heart disease, and stroke: media campaign to reduce saturated fat Stroke and ischemic and hypertensive heart disease: polypill with absolute risk approach Depression: drugs with optional episodic or maintenance psychosocial treatment Diabetes, ischemic heart disease, and stroke: legislation with public education to reduce salt content Diarrheal disease: breastfeeding promotion Stroke and ischemic and hypertensive heart disease: aspirin, beta-blocker, statin by absolute risk approacha Neonatal mortality: maternal and child health package, with no neonatal care after birth Neonatal mortality: combined maternal and child health and neonatal packages Diarrheal disease: oral rehydration therapy for package costing US$2.91 per episode Myocardial infarction: acute management with streptokinase, with aspirin and beta-blocker Ischemic heart disease: aspirin, beta-blocker, with optional angiotensin-converting enzyme inhibitor Coronary artery disease: legislation substituting 2% of trans fat with polyunsaturated fat at US$6 per adult Haemophilus influenzae type B, hepatitis B, diphtheria, pertussis, and tetanus: pentavalent vaccine Tobacco addiction: nicotine replacement therapy Neonatal mortality: family, community, or clinical neonatal package Myocardial infarction and stroke: secondary prevention with polypill Lower acute respiratory infection (0­4 years): case management package at community, facility, and hospital levels Tobacco addiction: nonprice interventionsb Tuberculosis (endemic): management of drug resistance Maternal conditions: improved quality of care and coverage Congestive heart failure: angiotensin-converting enzyme inhibitor and beta-blocker, with diuretics Maternal conditions: improved overall quality of care Maternal conditions: increased primary care coverage Maternal conditions: improved quality of comprehensive emergency obstetric care Problems requiring surgery: surgical ward or services in district hospital or community clinic Tuberculosis (epidemic, latent): isoniazid treatment Tuberculosis (epidemic): management of drug resistance Emergency medical care: staffed community ambulance Lower acute respiratory infection (0­4 years): case management of non-severe cases at community or facility level Intervention Tuberculosis (epidemic, infectious): directly observed short-course chemotherapy Tuberculosis (endemic, infectious or noninfectious): directly observed short-course chemotherapy Tuberculosis (endemic): BCG vaccine Coronary artery disease: legislation substituting 2% of trans fat with polyunsaturated fat at US$0.50 per adult Cost-Effectiveness: Tobacco addiction: taxation causing 33% price increase Myocardial infarction: acute management with aspirin and beta-blocker Tuberculosis, diphtheria-pertussis-tetanus, polio, measles: traditional EPIc Emergency medical care: training volunteer paramedics with lay first responders 0 10 100 1,000 10,000 100,000 Overview Cost-effectiveness ratio (US$ per DALY averted) Source: Authors. Note: Diseases were considered high burden for South Asia if their total avertable burden was greater than 10 million DALYs. Bars represent the range in point estimates of cost-effectiveness ratios for specific interventions included in of each intervention cluster and do not represent variation across regions or statistical confidence intervals. Point estimates for LMICs were obtained directly from the relevant chapters, calculated as the midpoint of range estimates Main reported in the chapters, or calculated from a population-weighted average of the region-specific estimates reported in the chapters. For details of these intervention clusters, including the specific interventions covered in each, see annex tables 2.B.1 and 2.B.2. Only interventions with cost-effectiveness reported in terms of DALYs are included in this figure. For interventions with cost-effectiveness reported in other units, see annex tables 2.C.1 and 2.C.2. Messages a. Cost-effectiveness range of aspirin, beta-blockers, and statin to prevent stroke and ischemic and hypertensive heart disease is incremental to salt reduction legislation and health education. b. Nonprice interventions for tobacco addiction include advertising bans, smoking restrictions, supply reduction, and information dissemination. c. EPI = Expanded Program on Immunization. | 51 Figure 2.4 Cost-Effectiveness of Interventions Related to High-Burden Diseases in South Asia 52 Diarrheal disease: improved water and sanitation at current coverage of amenities and other interventions | Disease Stroke and ischemic and hypertensive heart disease: polypill with absolute risk approach Diarrheal disease: rotavirus or cholera immunization Control Diarrheal disease: oral rehydration therapy for package costing US$5.51 per episode Neonatal mortality: maternal and child health package, with no neonatal care after birth HIV/AIDS: antiretroviral therapy Priorities Neonatal mortality: combined maternal and child health and neonatal packages HIV/AIDS: home care in Diarrheal disease: breastfeeding promotion Developing Neonatal mortality: family, community, or clinical neonatal package Stroke and ischemic and hypertensive heart disease: aspirin, beta-blocker, statin by absolute risk approacha Lower acute respiratory infection (0­4 years): case management package at community, facility, and hospital levels Countries HIV/AIDS: mother-to-child transmission prevention Emergency medical care: staffed community ambulance Problems requiring surgery: surgical ward or services in district hospital or community clinic | Ramanan HIV/AIDS: tuberculosis coinfection prevention and treatment Maternal conditions: increased primary care coverage Maternal conditions: improved quality of comprehensive emergency obstetric care Laxminarayan, Maternal conditions: improved quality of care and coverage HIV/AIDS: blood and needle safety Maternal conditions: improved overall quality of care Lower acute respiratory infection (0­4 years): case management of non-severe cases at community or facility level Jeffrey HIV/AIDS: condom promotion and distribution HIV/AIDS: sexually transmitted infections diagnosis with treatment Chow HIV/AIDS: voluntary counseling and testing ,and Haemophilus influenzae type B, hepatitis B, diphtheria, pertussis, and tetanus: pentavalent vaccine Childhood illness: integrated management of childhood illness Sonbol Malaria: intermittent preventive treatment in pregnancy with sulfadoxine-pyrimethamine Malaria: residual household spraying A. Shahid-Salles Malaria: insecticide-treated bed nets HIV/AIDS: peer programs for high-risk groups Underweight child (0­4 years): breastfeeding promotion and support Emergency medical care: training volunteer paramedics with lay first responders Malaria: intermittent preventive treatment in pregnancy with drugs other than sulfadoxine-pyrimethamineb Tuberculosis, diphtheria-pertussis-tetanus, polio, measles: traditional Expanded Program on Immunization Measles: second opportunity vaccination in a fixed facility 0 10 100 1,000 10,000 Cost-effectiveness ratio (US$ per DALY averted) Source: Authors. Note: Diseases were considered high burden for Sub-Saharan Africa if their total avertable burden was greater than 10 million DALYs. Bars represent the range in point estimates of cost-effectiveness ratios for specific interventions included in each intervention cluster and do not represent variation across regions or statistical confidence intervals. Point estimates for LMICs were obtained directly from the relevant chapters, calculated as the midpoint of range estimates reported in the chapters, or calculated from a population-weighted average of the region-specific estimates reported in the chapters. For details of these intervention clusters, including the specific interventions covered in each, see annex tables 2.B.1 and 2.B.2. Only interventions with cost-effectiveness reported in terms of DALYs are included in this figure. For interventions with cost-effectiveness reported in other units, see annex tables 2.C.1 and 2.C.2. a. Cost-effectiveness range of aspirin, beta-blockers, and statin to prevent stroke and ischemic and hypertensive heart disease is incremental to salt reduction legislation and health education. b. Chloroquine as first-line drug, artemisinin combination therapy as second-line drug and sulfadoxine-pyrimethamine as first- or second-line drug. Figure 2.5 Cost-Effectiveness of Interventions Related to High-Burden Diseases in Sub-Saharan Africa Number of intervention clusters dropped as key drugs have gone off patent. Acute management 35 of stroke and myocardial infarction using aspirin, beta- Personal intervention blockers, and nitroglycerin costs as little as US$15 to US$30 per 30 Population-based intervention DALY averted and ranks among the most cost-effective inter- Total 25 ventions available in LMICs. Even though many of the interventions were first developed in the industrial world, their 20 benefits are now largely available in the developing world. 15 Thus, the challenge lies in the ability of health care systems in LMICs to adopt these interventions on a large scale. 10 Technological Progress. Much progress has been made in sci- 5 entific understanding and in the availability of affordable, 0 population-based and personal interventions for preventing 0­100 100­500 500­1,000 1,000­3,000 3,000 and treating HIV/AIDS; however, adequate scaling up of these Cost-effectiveness ratio (US$ per DALY averted) interventions remains a challenge, with a few notable excep- Source: Authors. tions. The international health system has shown remarkable Note: Point estimates for the LMICs were obtained directly from the relevant chapters, calculated as the midpoint of range estimates reported in the chapters, or calculated technological agility in responding to this epidemic, demon- from a population-weighted average of the region-specific estimates reported in the strating that the world's scientific-industrial machinery is capa- chapters. ble of rising to the challenge of emerging diseases when there is Figure 2.6 Distribution of Interventions in LMICs by Cost- sufficient economic motivation for doing so. For instance, com- Effectiveness Ratio bination antiretroviral treatments are currently available for as little as US$150 for a year's supply in some countries. In con- trast, monotherapy with zidovudine, or AZT, which was the Lessons standard of care 10 years ago, was less effective, more expensive, Three lessons are broadly applicable. They relate to communi- and much more prone to drug resistance. As before, the chal- cable diseases, noncommunicable diseases, and technological lenge does not appear to be in the availability of interventions progress. either to prevent infection in adults or to effectively ensure against transmission from infected mothers to newborns. Communicable Diseases. Interventions to treat communica- Rather, the challenge lies in the willingness and ability to fund ble diseases have been highly cost-effective in the past and and deploy the interventions effectively. Clearly, much more remain so despite new challenges, such as drug-resistant remains to be done to develop affordable treatments. However, pathogens and vectors. Although much progress has been without a vaccine, the only feasible solution appears to be to made in lowering the burden of disease associated with aggressively prevent further transmission while treating vaccine-preventable illnesses, diarrhea, and to a lesser extent patients under well-implemented programs that can achieve with acute respiratory infections, progress made on other dis- the high rates of treatment adherence required to maintain the eases, such as malaria and TB, has been rolled back by such continued effectiveness of drug therapy.7 More generally, the challenges as parasite resistance in the case of malaria and the challenge of motivating technological advances for diseases that HIV epidemic in the case of TB. An important exception may do not threaten the developed world remains to be addressed. be diseases for which vaccines have been available, where sig- nificant gains in health have been achieved. In general, discern- ing a link between the availability of effective, affordable inter- Importance of Health Systems ventions in 1993 and a significant effect on the disease burden In describing efficient means of producing health, this chap- since that time is difficult because of the problem in defining ter has said little about how such efficiency may be translated the appropriate counterfactual of what would have happened into practice. The overall cost-effectiveness of a service level in the absence of interventions that were implemented. or package of interventions, rather than the cost-effectiveness of individual interventions, is the appropriate indicator to Noncommunicable Diseases. Compared with 13 years ago, determine which interventions should be used. From a plan- many more cost-effective interventions have been evaluated ning point of view, taking the infrastructure as fixed, at least and are being used for noncommunicable diseases, which con- in the immediate future, and then asking how it can best be tinue to grow in importance as populations undergo the epi- used to deliver the most cost-effective interventions might be demiological transition. Many of these interventions have been sensible. Where infrastructure is limited, expanding access will available for more than a decade; however, their costs have have to take priority. Other factors related to health system Intervention Cost-Effectiveness: Overview of Main Messages | 53 Table 2.2 Neglected Low-Cost Opportunities and High-Cost Interventions in South Asia and Sub-Saharan Africa Neglected low-cost opportunities in South Asia Cost Thousands of DALYs Burden of target per DALY averteda,b per 20% diseasesa (millions averteda (US$) increase in coverage of DALYs) CHILDHOOD IMMUNIZATION Additional coverage of traditional Expanded Program on Immunization 8 n.e. 28.4 (tuberculosis, diphtheria-pertussis-tetanus, polio, measles) HIV AND AIDS Voluntary counseling and testing Peer-based programs targeting at-risk groups (e.g., commercial sex workers) to disseminate information and teach specific skills 9­126 n.e. 7.4 School-based interventions that disseminate information to students Prevention of mother-to-child-transmission with antiretroviral therapy SURGICAL SERVICES AND EMERGENCY CARE Surgical ward in a district hospital, primarily for obstetrics, trauma and injury 6­212 at least 1.8 48.0­146.3 Staffed community ambulance Training of lay first responders and volunteer paramedics TUBERCULOSIS Childhood vaccination against endemic TB Directly observed short-course chemotherapy 8­263 n.e. 13.9 Isoniazid treatment of epidemic TB Management of drug resistance LOWER ACUTE RESPIRATORY ILLNESSES OF CHILDREN UNDER FIVE Community- or facility-based case management of non-severe cases Case management package including community- and facility-based 28­264 0.7­1.8 9.7­26.4 care for non-severe cases and hospital-based care for severe cases CARDIOVASCULAR DISEASE Management of acute myocardial infarction with aspirin and beta-blocker Primary prevention of coronary artery disease with legislation substituting 2% of trans fat with polyunsaturated fat, at $0.50 per adult Secondary prevention of congestive heart failure with 9­304 at least 0.1 25.9­39.1 angiotensin-converting enzyme inhibitors and beta-blockers incremental to diuretics Secondary prevention of myocardial infarction and stroke with polypill containing aspirin, beta-blocker, thiazide diuretic, angiotensin-converting enzyme inhibitor, and statin TOBACCO USE AND ADDICTION Tax policy to increase price of cigarettes by 33 percent Non-price interventions such as advertising bans, health information 14­374 at least 2.5 15.7 dissemination, tobacco supply reductions, and smoking restrictions Nicotine replacement therapy MATERNAL AND NEONATAL CARE Increased primary care coverage Improved quality of comprehensive emergency obstetric care 127­394 at least 1.3 37.7­47.8 Improved overall quality and coverage of care Neonatal packages targeted to families, communities, and clinics 54 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles Table 2.2 (Continued) Neglected low-cost opportunities in Sub-Saharan Africa Cost Thousands of DALYs Burden of target per DALY averteda,b per 20% diseasesa (millions averteda (US$) increase in coverage of DALYs) CHILDHOOD IMMUNIZATION Second opportunity measles vaccinationc Additional coverage of traditional Expanded Program on Immunization 1­5 n.e. 13.5­31.3 (tuberculosis, diphtheria-pertussis-tetanus, polio, measles) TRAFFIC ACCIDENTS Increased speeding penalties, media, and law enforcement 2­12 n.e. 6.4 Speed bumps at the most dangerous traffic intersections MALARIA Insecticide-treated bed netsc Residual household sprayingc 2­24 20.8­37.6 35.4 Intermittent preventive treatment during pregnancyc SURGICAL SERVICES AND EMERGENCY CARE Surgical ward in a district hospital, primarily for obstetrics, trauma and injury 7­215 1.6­21.2 25­134.2 Staffed community ambulance Training of lay first responders and volunteer paramedics CHILDHOOD ILLNESSES Integrated management of childhood illnessesc Case management of non-severe lower acute respiratory illnesses at the community or facility level Case management package including community- or facility-based 9­218 at least 1.2 9.6­45.1 care for non-severe cases and hospital-based care for severe lower acute respiratory illnesses Breastfeeding support to prevent underweight childrenc CARDIOVASCULAR DISEASE Management of acute myocardial infarction with aspirin and beta-blocker Primary prevention of coronary artery disease with legislation substituting 2% of trans fat with polyunsaturated fat, at $0.50 per adult Secondary prevention of congestive heart failure with angiotensin- 9­273 at least 0.04 4.6 converting enzyme inhibitors and beta-blockers incremental to diuretics Secondary prevention of myocardial infarction and stroke with polypill containing aspirin, beta-blocker, thiazide diuretic, angiotensin-converting enzyme inhibitor, and statin HIV AND AIDS Peer-based programs targeting at-risk groups (e.g., commercial sex workers) to disseminate information and teach specific skills Voluntary counseling and testing Diagnosis and treatment of sexually-transmitted diseasesc 6­377 n.e. 56.8 Condom promotion and distributionc Prevention and treatment of tuberculosis co-infectionc Blood and needle safety programsc Prevention of mother-to-child transmission with antiretroviral therapy MATERNAL AND NEONATAL CARE Increased primary care coverage Improved quality of comprehensive emergency obstetric care 82­409 at least 2.8 29.8­37.7 Improved overall quality and coverage of care Neonatal packages targeted to families, communities, and clinics (Continues on the following page.) Intervention Cost-Effectiveness: Overview of Main Messages | 55 Table 2.2 (Continued) High-cost interventions in South Asia Cost Thousands of DALYs Burden of target per DALY averteda,b per 20% diseasesa (millions averteda (US$) increase in coverage of DALYs) DEPRESSION Episodic treatment with newer antidepressant drug (selective serotonin reuptake inhibitors) 1,003­1,449 0.4­0.8 14.6 Episodic or maintenance psychosocial treatment plus treatment with newer antidepressant drug (selective serotonin reuptake inhibitors) HIGH BLOOD PRESSURE AND CHOLESTEROL Primary prevention of stroke and ischemic and hypertensive heart disease with aspirin, beta-blocker, and statin, incremental to policy-induced behavior change, at 15 percent risk of CVD event over 10 years 1,120­1,932 at least 6.7 48.6 Primary prevention of stroke and ischemic and hypertensive heart disease with a polypill, containing aspirin, beta-blocker, thiazide diuretic, angiotensin-converting enzyme inhibitor, and statin, at 15 percent risk of CVD event over 10 years LIFESTYLE DISEASES Primary prevention of diabetes, ischemic heart disease, and stroke through policy that replaces saturated fat with monounsaturated fat in manufactured foods, accompanied by a public education campaign 1,325­1,865 1.3­1.8 39.5 Primary prevention of diabetes, ischemic heart disease, and stroke through legislation that reduces salt content plus public education STROKE (ISCHEMIC) Acute management with recombinant tissue plasminogen activator with 48 hours of onset Acute management with heparin within 48 hours of onset 1,630­2,967 0.03­0.4 2.2­9.2 Secondary prevention with carotid endarterectomy DIARRHEAL DISEASES Oral rehydration therapy if the package cost is greater than US$2.30 per child per episode 500­6,390 0.02­2.5 22.3 Rotavirus or cholera immunization TUBERCULOSIS Isoniazid treatment for latent endemic TB in patients uninfected with HIV 5,588­9,189 n.e. 13.9 SCHIZOPHRENIA AND BIOPOLAR DISORDER Antipsychotic medication and psychosocial treatment for schizophrenia 1,743­17,702 0.02­0.12 2.2­2.9 Valproate and psychosocial treatment for bipolar disorder CARDIOVASCULAR DISEASE Management of acute myocardial infarction with streptokinase or tissue plasminogen activator, incremental to aspirin and beta-blocker Secondary prevention of ischemic heart disease with statin, 638­24,040 0.04­0.3 25.9 incremental to aspirin, beta-blocker, and angiotensin-converting enzyme inhibitor Secondary prevention of ischemic heart disease with coronary artery bypass graft 56 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles Table 2.2 (Continued) High-cost interventions in Sub-Saharan Africa Cost Thousands of DALYs Burden of target per DALY averteda,b per 20% diseasesa (millions averteda (US$) increase in coverage of DALYs) DIARRHEAL DISEASES Oral rehydration therapy if the cost per episode is greater 500­1,658 0.1­4.6 22 than US$2.80 per child Rotavirus or cholera immunization HIV AND AIDS Home care treatmentc 673­1,494 n.e. 56.8 Antiretroviral therapy in populations with low adherencec TRAFFIC ACCIDENTS Random driver breath tests Enforcement of seatbelt laws 973­2,146 at least 0.05 6.2­6.4 Child restraint promotion HIGH BLOOD PRESSURE AND CHOLESTEROL Primary prevention of stroke and ischemic and hypertensive heart disease 1,920 n.e. 10.6 with aspirin, beta-blocker, and statin, incremental to policy-induced behavior change, at 15 percent risk of CVD event over 10 years LIFESTYLE DISEASES Primary prevention of diabetes, ischemic heart disease, and stroke through policy that replaces saturated fat with monounsaturated fat in manufactured foods, accompanied by a public education campaign 1,766­2,356 1.4­1.8 9.6 Primary prevention of diabetes, ischemic heart disease, and stroke through legislation that reduces salt content plus public education STROKE (ISCHEMIC) Acute management with recombinant tissue plasminogen activator within 48 hours of onset 1,284­2,940 0.02­0.3 0.9­3.6 Acute management with heparin within 48 hours of onset Secondary prevention with carotid endarterectomy TUBERCULOSIS Isoniazid treatment for latent endemic TB in patients uninfected 4,129­5,506 n.e. 8.1 with HIV CARDIOVASCULAR DISEASE Management of acute myocardial infarction with streptokinase or tissue plasminogen activator, incremental to aspirin and beta-blocker Secondary prevention of ischemic heart disease with statin, incremental 634­26,813 0.03­0.2 4.6 to aspirin, beta-blocker, and angiotensin-converting enzyme inhibitor Secondary prevention of ischemic heart disease with coronary artery bypass graft Source: Authors. n.e. not evaluated. a. Ranges represent variation in point estimates of cost-effectiveness, DALYs averted, or burden of disease among the different interventions listed in each group. Point estimates of cost-effectiveness and DALYs averted were obtained directly from the relevant chapters or calculated as the midpoint of range estimates reported in the chapters. Burden of disease were obtained from the relevant chapters and from Mathers and others 2006. b. Avertable DALYs per 20% increase in treatment coverage in a hypothetical sample population of one million people. c. Only evaluated for Sub-Saharan Africa. Intervention Cost-Effectiveness: Overview of Main Messages | 57 capacity and infrastructure may play a key role in determin- internationally agreed-upon development goals such as the ing the adoption of interventions. The current evidence on Millennium Development Goals. These objectives are comple- the cost-effectiveness of service levels such as district or refer- mentary. ral hospitals is weak. Even though part of the problem lies The lack of reliable data on costs and effectiveness is an with the difficulty of valuing the health benefits these facilities important obstacle to efficient priority setting. Despite the rel- produce, more could be done. Chapter 3 presents a more atively good data on the efficacy of interventions in clinical trial detailed discussion of issues pertaining to health systems, but settings, reliable effectiveness data are generally lacking. the broader questions of why some cost-effective interven- Furthermore, not enough is known about the costs, extent of tions are used while others are not is a subject for future coverage, and institutional capacity requirements of interven- inquiry. tions in developing countries. The messages presented in this Even though much of the technology to significantly reduce chapter represent the best available information about the the burden of disease already exists, few cost-effective interven- relative costs of purchasing health through a wide range of tions are available for some diseases. Shaping research priorities interventions. The challenge that lies ahead is for these mes- in a manner that is responsive to the treatment needs of the sages to move beyond the academic realm: ultimately, it is the millions of HIV/AIDS patients and of people suffering from extent to which policy makers make the commitment to act on mental disorders across the range of LMICs is a challenge. them that will save lives. Setting intervention priorities efficiently can make a dollar go farther in improving health and can substantively increase available resources. Moreover, without demonstrably improved ACKNOWLEDGMENTS efficiency in health spending, aid agencies and development partners are unlikely to be persuaded to dig deeper into their We are grateful to the many authors and the nine editors of pockets to pay for further expansions of health programs. this volume, whose work, guidance, and feedback were essen- Improving efficiency should not, however, detract from the tial inputs to this chapter. Pamela Maslen provided valuable importance of increasing resources that are available for assistance in compiling annex tables. Any remaining errors are implementing these interventions and of meeting broader ours alone. 58 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles ANNEX 2.A: INTERVENTION CATEGORIES AND Palliation aims at reducing pain and suffering from a condi- PERTINENT POLICY INSTRUMENTS tion for which no cure or means of rehabilitation is currently available. It may range from the use of aspirin for headaches to The term intervention is used to denote actions taken by or for the use of opiates to control terminal cancer pain. individuals to reduce the risk, duration, or severity of an Policy instruments are activities that governments or other adverse health condition. Policy instruments encourage, dis- entities that wish to encourage or discourage interventions or courage, or undertake interventions. Stopping smoking, for to expand the potential interventions could undertake. The fol- example, is an intervention that an individual can take to lowing are five major instruments of policy: reduce his or her risk of a range of diseases, and taxing tobacco products is a potential instrument of government policy to · Information, education, and communication seek to encourage this intervention. Interventions are divided into improve the knowledge of individuals and service providers those that are population based and those that are personal as about the consequences of their choices. follows: · Taxes and subsidies on commodities, services, and pollu- tants seek to effect appropriate behavioral responses. · Population-based primary prevention is directed toward · Regulation and legislation seek to limit the availability of entire populations or population subgroups. These inter- certain commodities, to curtail certain practices, and to ventions fall into three broad categories: personal behavior define the rules governing the financing and provision of change, control of environmental hazards, and population- health services. oriented medical interventions (for example, immuniza- · Direct expenditures seek to provide or to finance the provi- tion, mass chemoprophylaxis, and screening and referral). sion of selected interventions (such as immunizations); to · Personal interventions are directed toward individuals and provide infrastructure (for instance, medical schools) that can be provided at home; at clinics (community, private, facilitates the provision of a range of interventions; or to work-based, or school-based); at district hospitals; or at alter infrastructure so as to influence behavior (for example, referral hospitals. by installing speed bumps). · Research and development, either undertaken directly or encouraged through subsidies, are central to the goal of Primary prevention aims at reducing the level of one or expanding the range of interventions available and reducing more identified risk factors to reduce the probability of the ini- their costs. tial occurrence of a disease (for instance, providing medication for established hypertension to prevent stroke or myocardial Source: This annex was prepared by Thomas Gaziano, Dean infarction). Jamison, and Sonbol Shahid-Salles. Cure of a condition aims at removing the cause and restor- ing function to what it was before. Acute management consists of time-limited interventions ANNEX 2.B: SUMMARY OF INTERVENTIONS that decrease the severity of acute events or the level of established risk factors to minimize their long-term effect (for Table 2.B.1 summarizes personal interventions. A summary of instance, providing thrombolytics for acute myocardial population-based interventions is shown in table 2.B.2. infarction or angioplasty to reduce stenosis in coronary arteries). Secondary prevention (or chronic care) consists of ongoing ANNEX 2.C: SUMMARY OF OTHER interventions aimed at decreasing the severity and frequency of INTERVENTIONS recurrent events of chronic or episodic diseases (for instance, Table 2.C.1 summarizes personal interventions for which cost- providing selective serotonin reuptake inhibitors for severe effectiveness is evaluated using a measure other than unipolar depression). US$/DALY averted. A summary of population-based interven- Rehabilitation aims at restoring or partially restoring physi- tions evaluated using measures other than DALYs is shown in cal, psychological, or social function resulting from a previous table 2.C.2. condition. Intervention Cost-Effectiveness: Overview of Main Messages | 59 60 | Disease Control Table 2.B.1 Summary of Personal Interventions Priorities Quality Cost- Number of cost- Cost- effectiveness of DALYs Number effectiveness in Developing Intervention Target effectiveness rangeb avertedb of deaths analysis Condition Intervention Intervention description setting Objective populationa (US$/DALY) (US$/DALY) (hundreds) avertedb evidencec African Case finding and Identification and treatment of Clinic Primary All ages 15 (Sub-Saharan -- -- -- 2 (Sub-Saharan Countries trypanosomiasis treatment Trypanosoma brucei gambiense prevention, Africa) Africa) using the card agglutination try- cure panosomiasis test with parasito- | Ramanan logical confirmation, allowing for rapid diagnosis and treatment African Melarsoprol Used in the second stage of the Clinic Secondary All ages 10 (Sub-Saharan -- -- -- 2 (Sub-Saharan Laxminarayan, trypanosomiasis disease prevention Africa) Africa) African Eflornithine Used in the second stage of the Clinic Secondary All ages 20 (Sub-Saharan -- -- -- 2 (Sub-Saharan trypanosomiasis disease prevention Africa) Africa) Jeffrey Alcohol abuse Brief advice to During primary health care visits, Clinic Primary Adolescents 642 -- 1.75 -- 5 heavy drinkers by provision of advice by physicians prevention and adults Chow primary health through education sessions and care providers psychosocial counseling ,and Bipolar disorder Lithium, valproate, Episodic treatment in a hospital District or Secondary Adults over 4,417 3,590­5,244 1.00 -- 5 Sonbol with optional setting with lithium or valproate referral prevention 15 psychosocial with or without maintenance or hospital A. treatment, episodic psychosocial treatment Shahid-Salles hospital-based Bipolar disorder Lithium, valproate, Episodic treatment of bipolar District or Secondary Adults over 3,113 2,498­3,728 1.35 -- 5 with optional psy- disorder in a community setting referral prevention 15 chosocial treat- using lithium or valproate with hospital ment, or without maintenance or community-based episodic psychosocial treatment Cataract Extracapsular Extracapsular cataract extraction District or Cure Adults over 183 -- -- -- 3 surgery with implantation of a posterior referral 40 chamber intraocular lens; hospital removal of the lens and the front portion of the capsule, which are then replaced with an artificial lens Congestive heart ACE inhibitor and Use of ACE inhibitor and an District Secondary Adults 150 27­274 11.59 -- 5 failure beta-blocker, with optional beta-blocker (metopro- hospital prevention diuretics lol), incremental to diuretics Dengue Improved case No specific treatment: early Clinic or Acute man- All ages 587 -- -- -- 2 management recognition of symptoms such as district agement intense continuous abdominal hospital pain, persistent vomiting, rest- lessness or lethargy; supportive treatment includes fluid replace- ment and electrolytic therapy Depression Drugs with Antidepressant drugs (tricyclic District or Secondary Adults over 1,699 657­2,741 3.96 -- 5 optional episodic antidepressant or selective sero- referral prevention 15 or maintenance tonin reuptake inhibitor) used hospital psychosocial alone or in combination with treatment psychosocial treatment for episodic depression or mainte- nance treatment Diarrheal disease Oral rehydration Case management of acute diar- Clinic Acute Children 1,062 -- 16.57 58.20 5 therapy for pack- rheal infection with oral rehydra- management age costing tion salt solutions, for package US$5.50 per costing US$5.50 per child per episode episode Epilepsy First-line First line treatment with pheno- District Secondary All ages 89 -- 2.99 3.32 5 treatment with barbital to treat epilepsy hospital prevention phenobarbital patients Epilepsy Second-line Antiepileptic drugs, phenobarbi- Referral Secondary All ages 3,027 2,994­3,060 0.29 0.32 5 (refractory) treatment with tal and lamotrigine, or a combi- hospital prevention phenobarbital nation of phenobarbital and sur- and lamotrigine gery to treat epilepsy patients or surgery unresponsive to phenobarbital Intervention HIV/AIDS Mother-to-child All pregnant women offered Clinic Primary Mothers 192 7­377 -- -- 2 transmission screening to prevent mother-to- prevention and infants prevention child transmission; administra- Cost-Effectiveness: tion of a short-course of AZT, lamivudine, or nevirapine to mothers prepartum and intra- partum and to newborns postpar- tum to reduce the risk of mother- to-child transmission; also Overview includes breastfeeding advice HIV/AIDS Sexually trans- Sexually transmitted infection Clinic Primary pre- Adolescents 57 (Sub- 9­105 (Sub- -- -- 2 (Sub-Saharan of mitted infection screening and treatment promo- vention, cure and adults Saharan Africa) Saharan Africa) Africa) Main diagnosis and tion to prevent future infection Messages treatment and to identify and treat high- risk populations (Continues on the following page.) | 61 62 | Disease Table 2.B.1 (Continued) Control Quality Priorities Cost- Number of cost- Cost- effectiveness of DALYs Number effectiveness in Intervention Target effectiveness rangeb avertedb of deaths analysis Developing Condition Intervention Intervention description setting Objective populationa (US$/DALY) (US$/DALY) (hundreds) avertedb evidencec HIV/AIDS Treatment of Treatment before or after anti- Clinic or Primary All ages 52,449 34,968­69,930 -- -- 3 Countries Kaposi's sarcoma retroviral treatment, including district prevention, failed antiretroviral treatment; hospital palliation local or systemic treatment of | lesions to provide largely cos- Ramanan metic benefit HIV/AIDS Treatment of Opportunistic infection prophy- Clinic or Primary All ages 156 3­310 -- -- 3 Laxminarayan, opportunistic laxis; necessary for patients district prevention, without access to antiretroviral hospital cure infections treatment, for immunosup- pressed patients waiting for Jeffrey antiretroviral treatment to take effect, for patients who refuse Chow or cannot take antiretroviral treatment, for patients for whom ,and antiretroviral treatment fails, Sonbol and for groups of patients who are unable to recover sufficient A. CD4 cells despite good inhibition Shahid-Salles of viral replication HIV/AIDS Tuberculosis Preventive therapy, short-course Clinic Primary All ages 121 (Sub- 6­235 (Sub- -- -- 2 (Sub-Saharan coinfection pre- chemotherapy, or co-trimoxazole prevention, Saharan Africa) Saharan Africa) Africa) vention and prophylaxis cure treatment HIV/AIDS Home care Home visits providing basic care Household Secondary All ages 673 (Sub- -- -- -- 2 (Sub-Saharan to sick AIDS patients or compre- prevention, Saharan Africa) Africa) hensive schemes that provide palliation palliative care, nutrition, psy- chosocial support and counsel- ing, and links to primary and secondary health care HIV/AIDS Antiretroviral Combination therapy with Clinic Primary All ages 922 (Sub- 350­1,494 (Sub- -- -- 3 (Sub-Saharan therapy multiple antiretroviral drugs prevention Saharan Africa) Saharan Africa) Africa) associated with prolonged survival in treated patients Integrated Integrated Integration of effective interven- Clinic Primary Children 39 (Sub- -- -- -- 3 (Sub-Saharan management of management of tions to improve child health and prevention, Saharan Africa) Africa) childhood illness childhood illness nutrition into a coordinated secondary strategy by improving health prevention, worker performance, child health cure service delivery, and family and community practices Ischemic heart Aspirin, beta- Aspirin plus beta-blocker District or Secondary Adults 688 451­926 8.40 -- 5 disease blocker, and (atenolol) with optional ACE referral prevention optional ACE inhibitor (enalapril), with or with- hospital inhibitor out hospital availability Ischemic heart Statin, with Statin (lovastatin), incremental District or Secondary Adults 2,028 1,864­2,193 3.54 -- 5 disease aspirin, beta- to aspirin, beta-blocker referral prevention blocker and ACE (atenolol), and ACE inhibitor hospital inhibitor (enalapril), with or without hos- pital availability Ischemic heart Coronary artery Placement of grafts (usually Referral Secondary Adults 36,793 -- 0.76 -- 5 disease bypass graft saphenous vein or internal mam- hospital prevention mary artery) to bypass stenosed coronary arteries, while main- taining cerebral and peripheral circulation by cardiopulmonary bypass Leishmaniasis Case finding and Combination of identification Clinic or Primary All ages 9 -- -- -- 2 treatment and treatment, vector control district prevention where feasible, and (in zoonotic hospital foci) control of animal reservoirs Intervention Lower acute Case Nonsevere infection diagnosed Clinic, Cure Children 129 50­208 5.15 17.36 5 respiratory infec- management at by breath rate and treated by a community under 5 tions (nonsevere) community or community health worker or at a Cost-Effectiveness: facility level health facility, with amoxicillin, acetaminophen, and possibly salbutamol Lower acute Case Severe or very severe infection District Cure Children 4,530 2,916­6,144 0.48 1.57 5 respiratory infec- management at diagnosed by breath rate and hospital under 5 Overview tions (severe and hospital level with x-ray tests and treated at a very severe) hospital with antibiotics and possibly salbutamol, oxygen, and of prednisolone Main (Continues on the following page.) Messages | 63 64 | Disease Table 2.B.1 (Continued) Quality Control Cost- Number of cost- Cost- effectiveness of DALYs Number effectiveness Priorities Intervention Target effectiveness rangeb avertedb of deaths analysis Condition Intervention Intervention description setting Objective populationa (US$/DALY) (US$/DALY) (hundreds) avertedb evidencec in Developing Lower acute Case- Comprehensive case-manage- Clinic or Cure Children 398 -- 11.26 37.86 5 respiratory management ment strategy covering nonse- district under 5 infections package at com- vere infection being treated by a hospital Countries munity, facility, community health worker or at a and hospital levels health facility, severe infection treated at a hospital, and very | severe infection treated at a Ramanan hospital Malaria Intermittent pre- Intermittent preventive Clinic Primary Pregnant 19 (Sub-Saharan 13­24 (Sub- 208.00 (Sub- 827.80 (Sub- 5 (Sub-Saharan Laxminarayan, ventive treatment treatment in areas with high prevention women Africa) Saharan Africa) Saharan Saharan Africa) Africa) in pregnancy with and stable transmission of Africa) sulfadoxine- Plasmodium falciparum malaria; pyrimethamine two curative doses of Jeffrey sulfadoxine-pyrimethamine given during the second and third Chow trimesters of pregnancy during prenatal care visits ,and Malaria Intermittent pre- Intermittent preventive treat- Clinic Primary Pregnant 7 (Sub-Saharan 2­11 (Sub- -- 77,500.00 (Sub- 5 (Sub-Saharan Sonbol ventive treatment ment in areas with high and sta- prevention women Africa) Saharan Africa) Saharan Africa) Africa) in pregnancy with ble transmission of Plasmodium A. Shahid-Salles drug other than falciparum malaria; two curative sulfadoxine- doses of antimalarial treatment pyrimethamine given with a possible change in first-line therapies from chloroquine to sulfadoxine- pyrimethamine, chloroquine to artemisinin combination therapy, or sulfadoxine-pyrimethamine to artemisinin combination therapy Maternal Increased primary Increased percentage of women Clinic or Primary Pregnant 132 (South -- 13.09 (South 32.00 (South 5 (South Asia), mortality care coverage accessing routine prenatal, district prevention women Asia), 88 (Sub- Asia), 27.88 Asia), 77.00 5 (Sub-Saharan intranatal, and postnatal care hospital Saharan Africa) (Sub-Saharan (Sub-Saharan Africa) Africa) Africa) Maternal Improved quality Increased percentage of women Clinic or Acute Pregnant 127 (South -- 13.28 (South 32.00 (South 5 (South Asia), mortality of comprehensive with severe complications district management women Asia), 87 (Sub- Asia), 28.28 Asia), 78.00 5 (Sub-Saharan emergency receiving comprehensive hospital Saharan Africa) (Sub-Saharan (Sub-Saharan Africa) obstetric care emergency obstetric care Africa) Africa) Maternal Improved overall Improvements to quality of Clinic Primary Pregnant 147 (South 133­160 (South 21.90 (South 56.20 (South 5 (South Asia), mortality quality of care prenatal and delivery care; prevention, women Asia), 83 (Sub- Asia), 82­85 (Sub- Asia), 53.05 Asia), 153.20 5 (Sub-Saharan enhanced package including acute Saharan Africa) Saharan Africa) (Sub-Saharan (Sub-Saharan Africa) availability of doctor and full management Africa) Africa) range of basic and comprehen- sive emergency obstetric care (all six essential obstetric func- tions: administering antibiotics intravenously or intramuscularly, administering oxytocics intra- venously or intramuscularly, manually removing the placenta, administering anticonvulsants intravenously or intramuscularly, carrying out instrumental deliv- ery, and removing retained prod- ucts of conception; optional nutritional supplementation Maternal Improved quality Improvements to quality of pre- Clinic Primary Pregnant 152 (South 138­167 (South 23.51 (South 60.29 (South 5 (South Asia), mortality of care and natal and delivery care and prevention, women Asia), 86 (Sub- Asia), 85­86 (Sub- Asia), 56.93 Asia), 164.14 5 (Sub-Saharan coverage increase in the proportion of acute Saharan Africa) Saharan Africa) (Sub-Saharan (Sub-Saharan Africa) women receiving needed care; management Africa) Africa) enhanced package including availability of doctor and full range of basic and comprehen- sive emergency obstetric care (all six essential obstetric func- tions noted above); optional Intervention nutritional supplementation Myocardial Aspirin and Aspirin with or without beta- District or Acute Adults 14 13­15 1.04 -- 5 infarction beta-blocker blocker (atenolol) referral management Cost-Effectiveness: hospital Myocardial Streptokinase, Incremental use of streptoki- District or Acute Adults 671 -- 1.04 -- 5 infarction with aspirin and nase, in addition to aspirin and referral management beta-blocker beta-blocker (atenolol) hospital Myocardial Tissue plasmino- Incremental use of tissue plas- District Acute Adults 15,869 -- 0.42 -- 5 Overview infarction gen activator, minogen activator in addition to hospital management with aspirin and aspirin and beta-blocker of beta-blocker (atenolol) Main Myocardial Polypill Combination treatment with District Secondary Adults 409 -- -- -- 5 Messages infarction and aspirin, beta-blocker, thiazide hospital prevention stroke diuretic, ACE inhibitor and statin, based on 10-year risk of cardio- vascular disease | 65 (Continues on the following page.) 66 | Disease Control Table 2.B.1 (Continued) Priorities Quality Cost- Number of cost- in Developing Cost- effectiveness of DALYs Number effectiveness Intervention Target effectiveness rangeb avertedb of deaths analysis Condition Intervention Intervention description setting Objective populationa (US$/DALY) (US$/DALY) (hundreds) avertedb evidencec Countries Neonatal Maternal and Mother and child health package Clinic or Primary Mothers 1,060 (South -- -- -- 4 (South Asia), mortality child health pack- that includes family planning, district prevention and Asia), 924 (Sub- 5 (Sub-Saharan age with no prenatal care, and comprehen- hospital infants Saharan Africa) Africa) | Ramanan neonatal care sive obstetric care after birth -- -- Laxminarayan, Neonatal Family, communi- Healthy home care practices, Clinic, com- Primary Mothers 349 (South 305­394 (South 4 (South Asia), mortality ty, or clinical including exclusive breastfeed- munity or prevention and Asia), 345 (Sub- Asia), 338­351 5 (Sub-Saharan neonatal package ing, warmth protection, clean household infants Saharan Africa) (Sub-Saharan Africa) cord care, care seeking for Africa) Jeffrey emergencies; if birth outside a facility, then clean delivery kit Chow Neonatal Combined mater- Family planning, prenatal care, Clinic, com- Primary Mothers 839 (South -- -- -- 4 (South Asia), mortality nal and child and comprehensive obstetric munity or prevention and Asia), 789 (Sub- 5 (Sub-Saharan ,and health with care packages, as well as household infants Saharan Africa) Africa) Sonbol neonatal healthy home care practices, packages including exclusive breastfeed- A. ing, warmth protection, clean Shahid-Salles cord care, care seeking for emergencies; if birth outside a facility, then clean delivery kit Panic disorder Drugs with Anxiolytic drugs (benzodi- District or Secondary Adults 734 384­1,084 0.83 -- 5 optional azepine), tricyclic antidepres- referral prevention over 15 psychosocial sants or selective serotonin hospital treatment reuptake inhibitor used with or without psychosocial treatment Parkinson's Ayurvedic Levodopa (l-dopa), carbidopa, or District Secondary Adults 1,132 752­1,512 0.13 -- 5 disease treatment and ayurvedic therapy for partial hospital or prevention over 45 levodopa or relief of symptoms referral carbidopa hospital Parkinson's Levodopa or car- Levodopa or carbidopa and deep District Secondary Adults 31,114 -- 0.15 -- 5 disease bidopa and deep brain stimulation hospital or prevention over 45 brain stimulation referral hospital Schizophrenia Antipsychotic Maintenance treatment in a hos- District Secondary Adults 11,920 4,105­19,736 0.60 -- 5 drugs with pital setting with antipsychotic hospital or prevention over 15 optional psy- drugs, neuroleptic antipsychotic referral chosocial treat- drug, or an atypical antipsychotic hospital ment, hospital- drug, with or without psychoso- based cial treatment Schizophrenia Antipsychotic Maintenance treatment in a Community Secondary Adults 9,834 2,472­17,197 0.70 -- 5 drugs with community-based setting with prevention over 15 optional psy- antipsychotic drugs, neuroleptic chosocial treat- antipsychotic drug, or an atypical ment, community- antipsychotic drug, with or with- based out psychosocial treatment Stroke (ischemic) Aspirin Aspirin dose within 48 hours of Clinic or dis- Acute man- Adults 149 -- 1.62 0.12 5 onset of acute stroke trict hospital agement over 15 Stroke (ischemic) Heparin and Heparin within 48 hours of onset District Acute man- Adults 1,977 1,278­2,675 1.22 1.70 5 recombinant of stroke or thrombolytic therapy hospital agement over 15 tissue plasmino- using recombinant tissue plas- gen activator minogen activator within 3 hours of onset Stroke (recurrent) Aspirin and Daily aspirin dose or combina- Clinic or dis- Secondary Adults 81 70­93 1.77 14.29 5 dipyridamole tion of aspirin and extended trict hospital prevention over 15 release dipyridamole Stroke (recurrent) Carotid Carotid endarterectomy surgery Referral Secondary Adults -- 4.93 39.82 5 endarterectomy to remove harmful plaque from hospital prevention over 15 1,458 the carotid arteries Stroke and Polypill by Combination treatment with District or Primary Adults 773­3,483 61.65 -- 5 ischemic and absolute risk aspirin, beta-blocker, thiazide referral prevention 2,128 Intervention hypertensive approach diuretic, ACE inhibitor, and statin hospital heart disease based on 10-year risk of cardio- vascular disease Cost-Effectiveness: Tobacco Nicotine replace- Smoking cessation treatments in Clinic Primary Adults 396 -- 37.14 452.05 5 addiction ment therapy the form of nicotine replacement prevention therapy Trachoma Trichiasis surgery Trichiasis surgery (eyelid correc- District Secondary Adults 39 -- -- -- 3 tion) to prevent blindness and hospital or prevention over 40 Overview reduce likelihood of other condi- referral tions hospital of Main (Continues on the following page.) Messages | 67 68 | Disease Table 2.B.1 (Continued) Control Quality Cost- Number of cost- Priorities Cost- effectiveness of DALYs Number effectiveness Intervention Target effectiveness rangeb avertedb of deaths analysis Condition Intervention Intervention description setting Objective populationa (US$/DALY) (US$/DALY) (hundreds) avertedb evidencec in Developing Trachoma Tetracycline or Tetracycline or azithromycin to Clinic Primary Children 6,269 3,752­8,785 -- -- 3 azithromycin treat the initial trachoma infec- prevention and adults tion through either mass treat- Countries ment of all children younger than 10 or through targeted treatment of infected children | Ramanan and household members Tuberculosis Management of Introduction of resistance test- District Secondary Adults 318 208­429 -- -- 5 Laxminarayan, (endemic) drug resistance ing, second-line drugs, longer hospital prevention, over 15 treatment regimen (12­18 cure months), and rigorous bacterio- logical and clinical monitoring; Jeffrey standardized or individualized regimen Chow Tuberculosis Directly observed Short-course chemotherapy of Clinic Primary Adults 301 84­551 -- -- 5 (endemic, infec- short-course infectious or noninfectious prevention, over 15 ,and tious or nonin- chemotherapy tuberculosis (with or without cure Sonbol fectious) transmission, non-HIV-positive), diagnosed via directly observed A. treatment strategy Shahid-Salles Tuberculosis Isoniazid Isoniazid treatment of latent District Secondary Adults 13,158 9,450­16,867 -- -- 5 (endemic, latent) treatment infection (with or without x-ray hospital prevention over 15 exclusion of active cases; non- HIV-infected population) Tuberculosis Management of Management of drug resistance District Secondary Adults 207 201­212 -- -- 5 (epidemic) drug resistance (standard regimen) for epidemic hospital prevention, over 15 TB conducted via introduction of cure resistance testing, second-line drugs, longer treatment regimen (12­18 months), and rigorous bacteriological and clinical moni- toring Tuberculosis Directly observed Short-course chemotherapy of Clinic Primary Adults 102 15­189 -- -- 5 (epidemic, short-course infectious TB (allowing for trans- prevention, over 15 infectious) chemotherapy mission, non-HIV positive) car- cure ried out for epidemic TB Tuberculosis (epi- Isoniazid Isoniazid treatment of latent District Secondary Adults 197 45­348 -- -- 5 demic, latent) treatment infection (x-ray exclusion of hospital prevention over 15 active cases; non-HIV-positive population) is conducted for epidemic tuberculosis Unwanted Family-planning Intrauterine devices, voluntary Clinic Primary Women of 117 -- -- -- 3 pregnancy programs sterilization, condoms and other prevention childbear- barrier methods, implants, and ing age oral contraceptives Zinc deficiency Supplements Provision of zinc as an adjunct to Clinic or Primary Children 73 -- -- -- 3 with oral rehy- oral rehydration salts in treating district prevention under 5 dration salts diarrhea in young children hospital Source: Authors. ACE angiotensin converting enzyme Note: -- not available. a. Refers to the age group to whom the intervention is targeted and not necessarily the one that is benefiting. b. Ranges in cost-effectiveness reflect the variation in point estimates for specific interventions included in each intervention cluster and do not represent either variation across regions or statistical confidence intervals. Point estimates were obtained directly from the relevant chapters, calculated as the midpoint of range estimates reported in the chapters, or calculated from a population-weighted average of the region-specific estimates reported in the chapters. DALYs and deaths potentially avertable are for a 20 percentage point increase in intervention coverage in a hypothetical sample population of 1 million. c. See table 2.1. Intervention Cost-Effectiveness: Overview of Main Messages | 69 70 | Disease Control Priorities Table 2.B.2 Summary of Population-Based Interventions in Quality Developing Cost- Number of cost- Target Cost- effectiveness of DALYs Number effectiveness Intervention popula- effectiveness rangeb avertedb of deaths analysis Countries Condition Intervention Intervention description setting Objective tiona (US$/DALY) (US$/DALY) (hundreds) avertedb evidencec Adolescent School health Inclusion of deworming of intes- Community, Population- School-age 37 -- -- -- 3 | health and and nutrition tinal worms and schistosomia- school oriented children Ramanan nutrition programs sis; prompt recognition and medical treatment of malaria; intervention Laxminarayan, insecticide-treated bednets; micronutrient supplements; breakfast, snacks, other meals; first-aid kits; referral to youth- Jeffrey friendly clinics; and counseling and psychosocial support Chow Alcohol abuse Excise tax 25 to 50 percent increase in the Policy level Instrument Adolescents 1,377 1,249­1,504 0.62 -- 5 current excise tax rate on of policy and adults ,and alcoholic beverages Sonbol Alcohol abuse Advertising ban Reduced access to alcoholic Policy level Instrument Adolescents 404 367­441 0.44 -- 5 and reduced beverage retail outlets by reduc- of policy and adults A. access to ing the hours of sale or advertis- Shahid-Salles beverage retail ing bans on television, radio, and billboards Alcohol abuse Excise tax, adver- 50 percent increase in the cur- Policy level Instrument Adolescents 631 601­661 2.85 -- 5 tising ban, with rent excise tax rate on alcoholic of policy and adults brief advice beverages, combined with advice, education sessions, and psychosocial counseling; possi- ble inclusion of random driver breath testing and advertising bans Chagas disease Vector control Vector control activities includ- Community Control of All ages 284 (Latin -- -- -- 1 (Latin ing spraying combined with environmen- America and the America and housing improvement, tal hazard Caribbean) the Caribbean) community involvement in sur- veillance, and strong programs of health education Coronary Legislation sub- Legislation replacing 2% of Policy level Instrument Adults 48 -- -- -- 5 artery disease stituting 2% of dietary trans fat from partial of policy trans fat with hydrogenation in manufactured polyunsaturated foods with polyunsaturated fat, fat at US$0.50 at a cost of US$0.50 per adult, per adult and assuming a 7% reduction in coronary artery disease Coronary Legislation sub- Legislation replacing 2% of Policy level Instrument Adults 838 199­1,478 -- -- 5 artery disease stituting 2% of dietary trans fat from partial of policy trans fat with hydrogenation in manufactured polyunsaturated foods with polyunsaturated fat, fat at US$6 per at a cost of US$6 per adult, and adult assuming a 7­40% reduction in coronary artery disease Dengue Vector control Chemical vector control using Community Control of All ages 2,566 1,992­3,139 -- -- 2 larvicides and insecticide space or district environmen- sprays (including emephos, per- hospital tal hazard methrin, methoprene, pyriprox- yfen, and Bacillus thuringiensis israelensis) to protect drinking water, or environmental vector control, such as removal of standing water Dengue Immunization Dengue immunization (a vaccine Community Population- Children 1,440 -- -- -- 2 is currently undergoing clinical clinic or oriented trials in Southeast Asia) district medical hospital intervention -- -- Intervention Diabetes, Legislation with Legislated reduction in salt con- Policy level Instrument All ages 1,937 18.73 5 ischemic heart public education tent of manufactured foods and of policy disease, and to reduce salt an accompanying public educa- stroke content tion campaign Cost-Effectiveness: Diabetes, Media campaign Media campaign to reduce satu- 2,617 Instrument All ages 2,617 -- 13.86 -- 5 ischemic heart to reduce rated fat content in manufac- of policy disease, and saturated fat tured foods and replace part of stroke the saturated fat with polyunsaturated fat Overview Diarrheal disease Breastfeeding Promotion of exclusive breast- Community, Personal Adult 930 -- 0.43 1.33 5 promotion feeding (recommended for six clinic, or behavior women of months) to new mothers, in district change Main which no other food or drink, hospital Messages including water, is permitted, except for supplements of vitamins and minerals and | necessary medicines 71 (Continues on the following page.) 72 | Disease Control Table 2.B.2 (Continued) Priorities Quality Cost- Number of cost- in Target Cost- effectiveness of DALYs Number effectiveness Developing Intervention popula- effectiveness rangeb avertedb of deaths analysis Condition Intervention Intervention description setting Objective tiona (US$/DALY) (US$/DALY) (hundreds) avertedb evidencec Countries Diarrheal disease Cholera or Immunization for endemic Clinic Population- Children 2,712 2,478­2,945 0.62 1.98 5 rotavirus cholera with live oral vaccine or oriented immunization rotavirus immunization with medical | rhesus-human rotavirus intervention Ramanan reassortant-tetravalent vaccine (currently under development) in Laxminarayan, populations at risk of an outbreak Diarrheal disease Improved water Improved water supply and Community Control of All ages 4,185 1,974­6,396 3.52 315.30 5 and sanitation at excreta disposal where estab- environmen- Jeffrey current coverage lished infrastructure currently tal hazards of amenities exists, in urban or rural settings Chow and other for at least five years interventions ,and Diarrheal disease Hand pump, Installation of hand water pump, Community Control of All ages 159 -- -- -- 1 Sonbol standpost, or standpost, or house connection environmen- house connection where clean water supply is tal hazards A. where clean limited and associated infra- Shahid-Salles water supply is structure currently do not exist limited Diarrheal disease Water sector Surveillance of drinking water Policy level, Instrument All ages 47 -- -- -- 1 regulation with quality and quality of service by community of policy, advocacy where the water supply utility in terms control of clean water of coverage, quantity, continuity, environmen- supply is limited control of sanitary hazards, and tal hazards cost, as well as advocacy of lower connection charges Diarrheal disease Construction and Construction of low-cost excreta Policy level, Instrument All ages 141 11­270 -- -- 1 promotion of disposal facilities such as house- community of policy, basic sanitation hold pit latrines, ventilation- control of where facilities improved latrines, or pour-flush environmen- are limited toilets, combined with public tal hazards promotion of sanitation and hygiene Down syndrome Prenatal screen- Prenatal genetic screening pro- Clinic, Population- Pregnant 15 -- -- -- 5 ing with option of gram, incorporating maternal district oriented women pregnancy serum triple screening of all hospital medical termination pregnant women, for trisomy of intervention chromosome 21, to allow par- ents to determine whether to continue with an affected pregnancy Emergency Training volun- Identification and training of Policy level Instrument All ages 6 -- 18.42 74.00 5 medical care teer paramedics community member first respon- of policy with lay first ders and paramedics to act in responders health emergencies, recognize life- or limb-threatening situa- tions, transport patients, and provide basic first aid Emergency Staffed Introduction or promotion of Policy level Instrument All ages 120 60­179 34.84 140.00 5 medical care community training programs for emergency of policy ambulance responders and ambulance driv- ers in urban or rural settings for countries that lack ambulances and training programs Haemophilus Vaccine Hib vaccination (three or four Clinic Population- Infants and 733d -- 29.25 113.83 5 influenzae type B containing Hib doses), given concurrently with oriented children (Hib) diphtheria-pertussis-tetanus medical intervention Hib, and Pentavalent Hib vaccination (three or four Clinic Population- Infants and 296d -- -- -- 5 hepatitis B, vaccine doses) and hepatitis B (three or oriented children Intervention diphtheria, four doses) given concurrently medical pertussis, and with diphtheria-pertussis- intervention tetanus tetanus vaccine Cost-Effectiveness: Hepatitis B Hepatitis B Hepatitis B (three or four doses) Clinic Population- Infants and 23,520d -- -- -- 5 vaccination given through intramuscular oriented children injection medical intervention Overview (Continues on the following page.) of Main Messages | 73 74 | Disease Control Table 2.B.2 (Continued) Priorities Quality Cost- Number of cost- Target Cost- effectiveness of DALYs Number effectiveness in Developing Intervention popula- effectiveness rangeb avertedb of deaths analysis Condition Intervention Intervention description setting Objective tiona (US$/DALY) (US$/DALY) (hundreds) avertedb evidencec HIV/AIDS Condom promo- Targeted distribution and place- Community Personal Adolescents 82 (Sub-Saharan 52­112 (Sub- -- -- 1 (Sub- Countries tion and ment of condoms in locations or behavior and adults Africa) Saharan Africa) Saharan distribution such as bars or brothels; distri- Community, change Africa) bution linked to voluntary coun- clinic | Ramanan seling and testing and sexually transmitted infection care to ensure universal access; infor- Laxminarayan, mation, education, and commu- nication, including education through literature, classroom, and clinical settings and radio, Jeffrey newspapers, and television HIV/AIDS Blood and Screening of all blood for trans- All levels, Population- All ages 84 (Sub-Saharan 7­161 (Sub- -- -- 2 (Sub- Chow needle safety fusions; sterilization for all injec- including oriented Africa) Saharan Africa) Saharan ,and tions; harm reduction for inject- community medical Africa) ing drug users, including needle clinics to intervention Sonbol exchange and drug substitution referral programs hospitals A. Shahid-Salles HIV/AIDS Voluntary Routine and voluntary confiden- Clinic Population- Adults 47 10­85 -- -- 2 counseling tial HIV counseling and testing oriented and testing medical intervention HIV/AIDS Peer and educa- Targeting community members Community Personal Adolescents 37 6­68 -- -- 2 tion programs for (for example, students or com- behavior and adults high-risk groups mercial sex workers) to dissemi- change nate information and teach specific skills Lymphatic Annual mass Two annual, single-dose, two- Clinic, Population- All ages 15 (South Asia) 4­27 (South Asia) -- -- 4 (South Asia) filariasis drug drug regimens are recommend- community oriented administration ed: ivermectin plus albendazole medical in African countries that are intervention coendemic for onchocerciasis, and diethylcarbamazine plus albendazole for all other endemic countries Lymphatic Diethyl carba- Fortification of salt with diethyl Policy level Instrument All ages 22 (South Asia) 1­43 (South Asia) -- -- 4 (South Asia) filiariasis mazine salt carbamazine of policy Lymphatic Vector control Integrated vector control to Policy level Instrument All ages 160 (South Asia) 43­277 (South -- -- 4 (South Asia) filariasis reduce overall prevalence of of policy Asia) microfilaria parasites, such as polystyrene beads in vector (mosquito) breeding habitats Malaria Insecticide- Impregnation of bednets with Household Control of All ages 11 (Sub-Saharan 5­17 (Sub- 376.00 (Sub- 1,429.60 (Sub- 5 (Sub- treated bednets deltamethrin, one treatment of environmen- Africa) Saharan Africa) Saharan Saharan Africa) Saharan permethrin, or two treatments of tal hazards Africa) Africa) permethrin, with the bednets either purchased or subsidized Malaria Residual house- One or two doses of malathion, Household Control of All ages 17 (Sub-Saharan 9­24 (Sub- 376.00 (Sub- 1,429.60 (Sub- 5 (Sub- hold spraying DDT, deltamethrin, or lambda- environmen- Africa) Saharan Africa) Saharan Saharan Africa) Saharan cyhalothrin applied to household tal hazards Africa) Africa) surfaces Measles Second opportu- Second opportunity to receive a Clinic Population- Infants 4 -- -- -- 5 nity vaccination dose of measles vaccine (either oriented and in a fixed facility through routine or supplemental medical children immunization activities) at a intervention fixed facility Meningitis Neisseria menin- Neisseria meningitidis vaccine for Clinic Population- Children 12,632 (Sub- -- -- -- 5 (Sub- gitidis vaccine serogroups A, C, Y, Wi35 only; oriented Saharan Africa) Saharan unconjugated polysaccharides medical Africa) given subcutaneously; one dose intervention with repeat three to five years later for those at high risk Intervention Problems Surgical ward or Surgical ward in a district District Instrument All ages 136 54­217 -- -- 5 requiring surgery services in dis- hospital or community clinic to hospital, of policy trict hospital or provide care for a wide range of clinic Cost-Effectiveness: community clinic conditions, such as trauma, childbirth, and abdominal conditions Onchocerciasis Ivermectin Annual dose of ivermectin Clinic, Population- Adults 37 -- -- -- 3 community oriented over 40 Overview medical intervention of Soil-transmitted Albendazole Annual albendazole anti- Community, Population- School-age 3 -- 127.76 1.98 5 Main helminthic helminthic drug treatment to school oriented children Messages infections reduce morbidity through the medical deworming of Ascaris, Trichuris, intervention and hookworm in school-age | children 75 (Continues on the following page.) 76 Table 2.B.2 (Continued) | Disease Quality Control Cost- Number of cost- Target Cost- effectiveness of DALYs Number effectiveness Priorities Intervention popula- effectiveness rangeb avertedb of deaths analysis Condition Intervention Intervention description setting Objective tiona (US$/DALY) (US$/DALY) (hundreds) avertedb evidencec in Tetanus Tetanus toxoid Tetanus toxoid vaccination via a Clinic, Population- Infants and 1,411 -- -- -- 5 Developing vaccination, mix mix of strategies depending on community oriented children of strategies local needs, including fixed facil- medical ities, immunization campaigns, intervention Countries mobile delivery, and community outreach Tobacco Taxation causing A 33 percent price increase due Policy level Instrument Adolescents 22 -- 37.27 1,905.99 5 | Ramanan addiction 33% price to tobacco taxes to discourage of policy and adults increase tobacco use, prevent initiation (and subsequent addiction) Laxminarayan, among youths, increase the likelihood of cessation among current users, reduce relapse among former users, and reduce Jeffrey consumption among continuing users Chow Tobacco Nonprice Advertising bans on television, Policy level Instrument Adolescents 353 -- -- -- 5 ,and addiction interventions radio, and billboards; health of policy and adults information and advertising in Sonbol the form of health warning labels on tobacco products; A. Shahid-Salles interventions to reduce tobacco supply, such as smuggling con- trol; restrictions on smoking Traffic accidents Increased speed- Minimizing exposure to high-risk Policy level Instrument Adults 21 3­38 0.67 197.16 5 ing penalties, scenarios by installation of of policy enforcement, speed bumps at hazardous junc- media cam- tions, increased penalties for paigns, and speeding, and other effective speed bumps road-safety regulations com- bined with media coverage and better enforcement of legislation Traffic accidents Enforcement of Mandatory seat belt and child- Policy level Instrument Adults 2,449 999­3,899 0.32 93.87 5 seatbelt laws, restraint laws, enforcement of of policy promotion of child drunk-driving laws, and random restraints and breath testing of drivers random driver breath testing Tuberculosis BCG vaccine Live attenuated vaccine, BCG; Clinic or Primary Children 68 55­82 -- -- 5 (endemic) recommended at birth or at first district prevention contact with health services in hospital Population- areas of high incidence oriented medical intervention Tuberculosis, Traditional Scaling up of EPI; a fixed incre- Community Population- Infants 7 -- -- -- 5 diphtheria, per- Expanded ment of coverage added for each oriented and tussis, tetanus, Program on year 2002­11 to reach 90 per- medical children polio, measles Immunization cent; coverage increases intervention (EPI) assumed to result from switch- ing to more effective and inten- sive implementation strategies rather than additional infrastruc- ture investments Undernutrition Sustained child Possible inclusion of prenatal Community, Population- Children 225 -- -- -- 3 and malnutrition health and nutri- care, women's health and nutri- clinic oriented under five tion program tion, breastfeeding promotion medical and counseling, complementary intervention feeding, growth monitoring and promotion, micronutrient supple- mentation, micronutrient fortifi- cation, supplementary feeding using local supplies, oral rehy- dration, and immunization and deworming; actual mix depends on local capabilities and conditions Intervention Underweight Child survival Community-based nutrition pro- Community Population- Children 42 -- -- -- 2 children program with grams to prevent growth falter- oriented under five nutrition ing, control morbidity, and medical Cost-Effectiveness: component improve survival by promoting intervention breastfeeding, providing educa- tion and counseling on optimal child feeding, preventing diar- rheal disease, and monitoring Overview growth Source: Authors. of BCG bacillus Calmette Guérin; DDT dichlorodiphenyl trichloroethane; EPI Expanded Program on Immunizations Main Note: -- not available. a. Refers to the age group to whom the intervention is targeted and not necessarily the one that is benefiting. Messages b. Ranges in cost-effectiveness reflect the variation in point estimates for specific interventions included in each intervention cluster and do not represent either variation across regions or statistical confidence intervals. Point estimates were obtained directly from the relevant chapters, calculated as the midpoint of range estimates reported in the chapters, or calculated from a population-weighted average of the region-specific estimates reported in the chapters. DALYs and deaths potentially avertable are for a 20 percentage point increase in intervention coverage in a hypothetical sample population of 1 million. c. See table 2.1. | 77 d. Cost-effectiveness ratio calculated from deaths averted only. 78 | Table 2.C.1 Cost Effectiveness of Other Personal Interventions Disease Cost- Quality Control Intervention Target effectiveness of CEA Condition Intervention Intervention description setting Objective population estimate (US$)a evidenceb Priorities Breast cancer Clinical breast exam Examination of the breast performed by Clinic Secondary prevention Women ages 7,125­9,907 per death 4 doctors or other trained health care pro- 40­60 prevented (India); in fessionals; annually, biennially, or every 522­722 per LYS (India) Developing five years; for women ages 40­60 Breast cancer Screening mammography Examination of the breasts performed by Clinic Secondary prevention Women ages 12,262­24,493 per death 2 (USA, Europe); compressing the breast firmly between a 40­70 prevented (India); 4 (India) Countries plastic plate and an X-ray cassette that 902­1846 per LYS (India); contains special X-ray film; one lifetime or 2,450­14,790a per YLS biennially (Europe); 28,600­47,900 | Ramanan (USA) Breast cancer Chemotherapy and/or Tamoxifen and/or chemotherapy for District hospital Secondary prevention Women age 45 12,820­171,700 (USA) 3 tamoxifen 45-year-old premenopausal women with Laxminarayan, early-stage breast cancer; for node- positive, node-negative, estrogen- receptor-positive, and estrogen- Jeffrey receptor-negative patients Breast cancer Radiation therapy Radiation therapy following mastectomy District hospital Secondary prevention Premenopausal 23,300­44,000 per QALY 2 Chow and chemotherapy for node-positive breast women (USA) cancer in premenopausal women ,and Cervical cancer Nationwide Pap screening Nationwide Pap screening program based District hospital Secondary prevention Adult women 769 per YLS (Vietnam) 2 Sonbol program based on five- on five-year intervals year intervals A. Shahid-Salles Cervical cancer Conventional or liquid- Conventional cytology using the District hospital Secondary prevention Adult women 126,500 (USA); 162,400 2 based cytology testing Papanicolaou (Pap) smear and HPV testing (Thailand) every 1 to 10 years; or Liquid-based cytol- ogy using the Papanicolaou (Pap) smear and HPV testing every 1 to 5 years Cervical cancer Two-visit HPV testing HPV DNA testing during the first visit fol- District hospital Secondary prevention Adult women 122 per YLS (Brazil); 1 lowed by treatment of screen-positive 167 per YLS women during the second visit (Madagascar); 41 per YLS (South Africa); 117 per YLS (Zimbabwe) Cervical cancer One-visit VIA Cervix is viewed after the application of District hospital Secondary prevention Women age 56 per YLS (Brazil); 54 per 1 an acetic acid solution; screening and 35­42 YLS (Madagascar); 43 per treatment conducted during the same visit YLS (Zimbabwe) Cervical cancer Three-visit cytology Cytology sample obtained during the first District hospital Secondary prevention Women age 589 per YLS (Brazil); 379 1 visit, colposcopy for screen-positive women 35­48 per YLS (Madagascar); conducted during the second visit, and 331 per YLS (Zimbabwe) treatment provided during the third visit Cervical cancer Chemoradiation therapy Cisplatin-based chemoradiation regimens District hospital Secondary prevention Adult women 337­31,400 per LYS 1 on the basis of published and estimated (USA) survival Chronic obstruc- Inhaled medication Inhaled ipratroprium bromide or cortico- Clinic/district Palliation Adults 7,800­13,400 per QALY 1 tive pulmonary steroid such as fluticasone propionate hospital (High-income countries) disease Chronic obstruc- A-1 antitrypsin augmenta- Intravenous treatment of chronic obstruc- Clinic/district Palliation Adults 14,400­215,000 per 1 tive pulmonary tion therapy tive pulmonary disease related to severe hospital QALY (High-income disease deficiency; ranges with age and efficacy countries) Chronic obstruc- Mechanical ventilation or Mechanical ventilation with inspiratory Clinic/district Palliation Adults 15,000­19,000 per YLS 1 tive pulmonary oxygen therapy support, invasive respiration in intensive hospital (High-income countries); disease, asthma, care unit, or long-term home oxygen 32,350­47,850 per QALY and cardiovascu- therapy (High-income countries) lar disease Colorectal cancer Flexible sigmoidoscopy Flexible sigmoidoscopy enables the physi- District hospital Secondary prevention Adults 18,700­25,954 (USA) 2 every 5 years with or cian to look at the inside of the large without fecal occult blood intestine from the rectum through the last test part of the colon, called the sigmoid or descending colon; fecal occult blood test checks stool samples for traces of blood. Colorectal cancer Double-contrast barium A series of x-rays of the colon and rectum District hospital Secondary prevention Adults 11,503­26,393 per YLS 2 enema every 5 years taken after the patient is given an enema, (USA) followed by an injection of air. The barium outlines the intestines on the x-rays, allowing many abnormal growths to be visible. This is conducted every 5 years. Colorectal cancer Colonoscopy every Colonoscopy allows the physician to look District hospital Secondary prevention Adults 9,309­22,672 per YLS 2 Intervention 10 years inside the entire large intestine, from the (USA) lowest part, the rectum, all the way up through the colon to the lower end of the Cost-Effectiveness: small intestine. The procedure is used to look for early signs of cancer in the colon and rectum. Colorectal cancer Chemotherapy Adjuvant chemotherapy for stage three District hospital Secondary prevention Adults 3,000­7,000 per YLS 1 colon cancer (High-income countries) Overview Colorectal cancer Radiation therapy Preoperative radiation therapy for rectal District hospital Secondary prevention Adults 908­15,228 per YLS 1 cancer patients; with varying rates of (Sweden) of recurrence and survival advantage with (Continues on the following page.) Main and without radiation treatment Messages | 79 80 | Table 2.C.1 (Continued) Disease Cost- Quality Control Intervention Target effectiveness of CEA Condition Intervention Intervention description setting Objective population estimate (US$) evidence Priorities Diabetes Smoking cessation Counseling and medication such as the Clinic Primary prevention Adolescents and 870 per QALY (EAP); 5 nicotine patch adults 1,170 per QALY (ECA); in 1,450 per QALY (LAC); Developing 1,230 per QALY (MNA); 730 per QALY (SAR); 660 per QALY (SSA) Countries Diabetes Annual eye examination Dilated eye examination to detect prolifer- Clinic Secondary prevention Adults 420 per QALY (EAP); 560 5 ative diabetic retinopathy and macular per QALY (ECA); 700 per | edema followed by appropriate photoco- QALY (LAC); 590 per Ramanan agulation therapy to prevent blindness QALY (MNA); 350 per QALY (SAR); 320 per QALY (SSA) Laxminarayan, Diabetes ACE inhibitor ACE inhibitors for blood pressure control Clinic Secondary prevention Adults 620 per QALY (EAP); 830 5 per QALY (ECA); 1,020 per QALY (LAC); 870 per Jeffrey QALY (MNA); 510 per QALY (SAR); 460 per Chow QALY (SSA) ,and Diabetes Metformin intervention for Metformin therapy for preventing type 2 Clinic Primary prevention Adults 2,180 per QALY (EAP); 5 preventing type 2 diabetes diabetes among people at high risk, such 2,930 per QALY (ECA); Sonbol as those with prediabetes 3,630 per QALY (LAC); 3,080 per QALY (MNA); A. Shahid-Salles 1,820 per QALY (SAR); 1,640 per QALY (SSA) Diabetes Cholesterol control Cholesterol control for people with total Clinic Secondary prevention Adults 4,420 per QALY (EAP); 5 cholesterol higher than 200 5,940 per QALY (ECA); milligrams/deciliter 7,350 per QALY (LAC); 6,240 per QALY (MNA); 3,680 per QALY (SAR); 3,330 per QALY (SSA) Diabetes Intensive glycemic control Intensive glucose control to lower the Clinic Secondary prevention Adults 2,410 per QALY (EAP); 5 level of glucose in the person with dia- 3,230 per QALY (ECA); betes to a level close to that of a person 4,000 per QALY (LAC); without diabetes, for people with HbA1c 3,400 per QALY (MNA); higher than 8 percent, in order to prevent 2,000 per QALY (SAR); or delay long-term diabetes complications 1,810 per QALY (SSA) Kidney disease Hemodialysis Most common method used to treat Clinic, home Secondary prevention Adults 42,700­70,000 per YLS 1 advanced and permanent kidney failure; (USA); 61,000­99,400 per conducted in a treatment center or home. QALY (USA) Kidney disease Kidney transplant Kidney transplant surgery District or referral Cure Adults 10,000 per LYS (USA); 2 hospital 11,000 per QALY (USA) Kidney disease ACE inhibitors ACE inhibitors for all type-1 diabetics Clinic Secondary prevention Adults 1,100­7,700 per QALY 4 with macroproteinuria and all type-2 (USA) diabetics Mild to moderate Quick-releavers in addition Rapid-acting bronchodilators that act to Clinic or district Palliation Adults 10,600-13,900 per QALY 1 asthma to inhaled corticosteriods relieve bronchoconstriction and accompa- hospital (High-income countries) nying acute symptoms of wheeze, chest tightness, and cough, e.g., salbutamol; incremental to inhaled corticosteroid treatment Opioid abuse Naltrexone induced rapid Patient is given naltrexone under general District hospital Rehabilitation Adults 2,498 per week of absti- 1 opioid detoxification under anesthetic. nence (Australia) sedation (RODS) Opioid abuse Conventional outpatient Conventional outpatient detoxification is Clinic Rehabilitation Adults 12,764 per abstinent 1 detoxification supervised withdrawal from a drug of patient (Australia) dependence that attempts to minimize withdrawal symptoms. Opioid abuse Drug-free treatments Residential or outpatient drug-free Clinic Rehabilitation Adults 7,000­13,000a (USA) 1 treatments as well as self-help group attendance Opioid abuse Methadone maintenance Substitution of short-acting heroin with Clinic Rehabilitation Adults 6,800 per LYS (High- 1 substitution long-acting orally administered opioid income countries); 9,000 Methodone; includes heroin users living in per QALY (High-income communities with high HIV prevalence countries, high-HIV prevalence) Intervention Opioid abuse Buprenorphine mainte- Buprenorphine substitution maintenance Clinic Rehabilitation Adults 49,000 per QALY (High- 1 nance substitution treatment for non-methadone patients. income countries) Cost-Effectiveness: Osteoarthritis Lifestyle change Exercise (aquatic) and calcium Clinic Primary/secondary Adults age 96,119­498,700 per 1 supplements prevention 55­75; post- QALY (High-income menopausal countries) women Overview Osteoarthritis Replacement surgery Synovial fluid replacement (hylan G-F 20) Clinic, district Secondary preven- Elderly men and 5,233­6893 per QALY 1 for patients with osteoarthritis of the hospital tion; tertiary inter- women (High income countries) knee; or complete hip joint or knee vention of replacement with implant Main Osteoporosis Hormone replacement Estrogen replacement from age 50, 60, or Clinic Primary/secondary Postmenopausal 5,088­23,734 per QALY 1 Messages therapy menopause for healthy women; 5-year to prevention women and lifetime treatment women age 50 and up | 81 (Continues on the following page.) 82 Table 2.C.1 (Continued) | Disease Cost- Quality Control Intervention Target effectiveness of CEA Condition Intervention Intervention description setting Objective population estimate (US$)a evidenceb Priorities Osteoporosis Calcium supplements with Calcium with or withouth vitamin D Clinic Primary prevention Women age 50 37,633­149,705 per 1 or without vitamin D supplements, based on evidence that it to 80 QALY in reduces appendicular fractures; assumes a Developing compliance rate of 70% Osteoporosis Nonestrogen drug Raloxifene or calcitonin to reduce bone Clinic Secondary prevention Postmenopausal 34,166­835,622 per 1 treatments fractures; with or without 5 years of women age 50 QALY (High income Countries therapy to 80 Countries) Osteoporosis Fluoride Fluoride appears to decrease the risk of Clinic Secondary prevention Postmenopausal 46,684 per QALY (UK) 1 | vertebral fracture for women with estab- women Ramanan lished osteoporosis; assumes neutral effect on hip fractures Laxminarayan, Pain Morphine Providing oral morphine and necessary Clinic Palliation All ages 210­408 per year of pain 4 associated drugs free life added (Chile, Romania, Uganda) Primary care Limited care Includes treatment of infection and minor Clinic Cure All ages 253­380 per DALY (Low 1 Jeffrey ailments trauma; for more complicated condition, income countries); includes diagnosis, advice and pain relief, 507­760 per DALY Chow and treatment as resources permit (Middle income countries) ,and Respiratory Lung transplant Surgical replacement with donor lung District hospital Cure Adults 238,200­464,000 per 1 disease (end QALY (High-income Sonbol stage) countries) A. Source: Authors. Shahid-Salles a. Currency units in US$, but not necessarily 2001. b. See table 2.1. Table 2.C.2 Cost-Effectiveness of Other Population-Based Interventions Cost- Quality Intervention Target effectiveness of CEA Condition Intervention Intervention description setting Objective population estimate (US$)a evidenceb Asthma Education Education in addition to exercise program Clinic Personal behavior Adults 71,500/QALY (High-income 1 change countries) Colorectal cancer Fecal occult blood test Fecal occult blood test to check stool District hospital Secondary prevention Adults 3,200­12,100 per YLS (High- 2 samples for traces of blood; conducted income countries) annually or biennially Diarrhea Improved quality of care Educational interventions to improve Clinic/district Instrument of policy All ages 14­6000a per death averted 4 quality of care and encourage oral rehy- hospital dration therapy in hospitals; varies with marginal improvement; low to average prevalence Diabetes Lifestyle intervention Behavioral change for weight reduction by Clinic Personal behavior Adults 80 per QALY (EAP); 100 per 5 (type 2 prevention) means of a combination of a low-calorie change QALY (ECA); 130 per QALY diet and moderate physical activity (LAC); 110 per QALY (MNA); 60 per QALY (SAR); 60 per QALY (SSA) Diabetes Influenza and pneumococ- Influenza and pneumococcal vaccinations Clinic Population-oriented Elderly 220 per QALY (EAP); 290 per 5 cal vaccinations for elderly individuals with type 2 diabetes medical intervention QALY (ECA); 360 per QALY (LAC); 310 per QALY (MNA); 180 per QALY (SAR); 160 per QALY (SSA) Diabetes Screening Screening of individuals at increased risk Clinic, district Population-oriented Adults over 25 5,140 per QALY (EAP); 6,910 5 Intervention for undiagnosed diabetes hospital medical intervention per QALY (ECA); 8,550 per QALY (LAC); 7,260 per QALY (MNA); 4,280 per QALY (SAR); 3,870 per QALY (SSA) Cost-Effectiveness: Diabetes Annual screening for Screening for microalbuminuria and treat- Clinic, district Population-oriented Adults 3,310 per QALY (EAP); 4,450 5 microalbuminuria ing those who test positive hospital medical intervention per QALY (ECA); 5,510 per QALY (LAC); 4,680 per QALY (MNA); 2,760 per QALY (SAR); 2,500 per QALY (SSA) Overview Indoor air Liquefied petroleum gas Substitution of wood, dung, and crop Policy level Instrument of policy All ages 103­1,746 per healthy year 5 pollution-related residues with liquefied petroleum gas for (EAP); 1,258­1,361 per of illness cooking and heating healthy year (ECA); Main 806­1,447 per healthy year Messages (LAC); 779­785 per healthy year (MNA); 321­558 per healthy year (SA); 534­736 | per healthy year (SSA) 83 (Continues on the following page.) 84 | Table 2.C.2 (Continued) Disease Cost- Quality Control Intervention Target effectiveness of CEA Condition Intervention Intervention description setting Objective population estimate (US$)a evidenceb Priorities Indoor air Kerosene Substitution of wood, dung, and crop Policy level Instrument of policy All ages 12­232 per healthy year 5 pollution-related residues with kerosene for cooking and (EAP); 172­188 per healthy in illness heating year (ECA); 109­650 per Developing healthy year (LAC); 98 per healthy year (MNA): 37­65 per healthy year (SAR); Countries 62­87 per healthy year (SSA) Indoor air Improved stove Replacement of traditional open stoves Policy level Instrument of policy All ages 306­605 per healthy year 5 | pollution-related with enclosed stoves that are more effi- (EAP); 975­1,134 per healthy Ramanan illness cient and/or have flues for ventilation year (LAC); 379­471 per healthy year (MNA): 13­15 Laxminarayan, per healthy year (SAR); 21­26 per healthy year (SSA) Indoor air Improved stove with Replacement of traditional open stoves Policy level Instrument of policy All ages 26­85 per healthy year (EAP); 5 pollution-related kerosene or LPG with enclosed stoves that use kerosene or 522­1,416 per healthy year Jeffrey illness liquified petroleum gas (LPG) (ECA); 305­784 per healthy year (LAC); 227­624 per Chow healthy year (MNA): 27­182 per healthy year (SAR); ,and 46­304 per healthy year Sonbol (SSA) Lung cancer Early detection screening Screening of high-risk individuals, such as District hospital Population-oriented Adults 20,000­100,000 per YLS 3 A. current and former smokers, for lung can- medical intervention (USA) Shahid-Salles cer using helical computed tomography Pneumonia Improved quality of care Improved quality of care, including educa- Clinic/district Instrument of policy Children 132­5,000a per death 4 tion for health providers and treatment of hospital averted non-severe and severe pneumonia; varies with marginal improvement; low to aver- age prevalence Pollution-related Control of toxins related Interventions include coal-fired power Policy level Instrument of policy All ages less than 0 per LYS (USA) 1 illness to energy industry plant emissions controls, gasoline lead reduction, and desulphuring of residual fuel oil Pollution-related Control of toxins related Interventions include targeted pesticide Policy level Instrument of policy All ages less than 0 per LYS (USA) 1 illness to agriculture and forestry bans and emissions standards at process- ing facilities Pollution-related Control of toxins related Interventions include radon remediation Policy level Instrument of policy All ages 5320­7730 per LYS (USA) 1 illness to residential sector and sedimentation, filtration, and chlorina- tion of drinking water Pollution-related Control of toxins related Interventions include arsenic emissions Policy level Instrument of policy All ages less than 45,600 per LYS illness to industrial sector standards at copper smelters and (USA) asbestos ban for brake linings Silicosis Engineering control Wet method e.g. spraying a surface or Policy level Control of environ- Working adults 105a per DALY (USA and 3 wetting a blade to reduce dust; local mental hazards Canada); 109a per DALY exhaust ventilation; total plant ventilation (Western Pacific) Silicosis Comfort or dust mask Comfort or dust mask with associated Policy level Instrument of policy Working adults 111­191a per DALY (USA 3 training and Canada); 117­174a (Western Pacific) Silicosis Respirator Half-face or full-face respirator and asso- Policy level Instrument of policy Working adults 300­305a per DALY (USA 3 ciated training and Canada); 266­274a per DALY (Western Pacific) Source: Authors. a. Currency units in US$, but not necessarily 2001. b. See table 2.1. Intervention Cost-Effectiveness: Overview of Main Messages | 85 NOTES Cookson, R., and P. Dolan. 1999. "Public Views on Health Care Rationing: A Group Discussion Study." Health Policy 49 (1­2): 63­74. 1. Few other cost-effectiveness studies have covered a similarly exten- ------. 2000. "Principles of Justice in Health Care Rationing." Journal of sive set of health interventions (Dixon and Welch 1991; Jamison and oth- Medical Ethics 26 (5): 323­29. ers 1993; Tengs and others 1995), and only one of those studies makes these Dixon, J., and H. G. Welch. 1991. "Priority Setting: Lessons from Oregon." comparisons on a global scale (Jamison and others 1993). The current Lancet 337 (8746): 891­94. World Health Organization project CHOICE (Choosing Interventions That Are Cost-Effective) is a parallel effort to make such global compar- Jamison, D. T., W. H. Mosley, A. R. Measham, and J. L. Bobadilla. 1993. isons (Murray and others 2000; http://www.who.int/evidence/cea). Disease Control Priorities in Developing Countries. New York: Oxford 2. Of these 319 cost-effectiveness estimates, 257 were in terms of University Press. U.S. dollars per DALY and therefore comparable. Interventions with cost- Joesoef, M. R., P. L. Remington, and P. T. Jiptoherijanto. 1989. effectiveness in terms of dollars per DALY were grouped into 121 inter- "Epidemiological Model and Cost-Effectiveness Analysis of vention clusters to facilitate analyses and presentation. Tuberculosis Treatment Programs in Indonesia." International Journal 3. Health system capacity is often used to describe both the level of care of Epidemiology 18 (1): 174­79. (primary, secondary, and tertiary) and the institutional and organizational Johns, B., and T. T. Torres. 2005. "Costs of Scaling up Health Interventions: capacities. We use the term to refer to the latter. A Systematic Review." Health Policy and Planning 20 (1): 1­13. 4. Chapter 15 presents a fuller discussion of these methods. Note that not all chapters have used these standardized costs. Furthermore, the Kamolratanakul, P., B. Chunhaswasdikul, A. Jittinandana, V. analyses have used U.S. dollars rather than purchasing-power parity dol- Tangcharoensathien, N. Udomrati, and S. Akksilp. 1993. "Cost- lars (which provide a better measure of input resource intensity and are Effectiveness Analysis of Three Short-Course Anti-tuberculosis less susceptible to exchange rate fluctuations) in order to provide a mone- Programs Compared with a Standard Regimen in Thailand." Journal of tary estimate that may be more useful to policy makers and donors. Clinical Epidemiology 46 (7): 631­36. 5. Noneconomic reasons for maintaining certain interventions can Mathers, C. D., C. J. L. Murray, and A. D. Lopez. 2006. "The Burden of include retaining key technical skills that may be required in the future Disease and Mortality by Condition: Data, Methods and Results for the and may lead to the development of new methods that may be more Year 2001." In Global Burden of Disease and Risk Factors, ed. A. D. cost-effective (see chapter 66 on referral hospitals for a more in-depth Lopez, C. D. Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. New discussion). York: Oxford University Press. 6. Some interventions with high potential to reduce the burden of dis- Murray, C. J. L., D. B. Evans, A. Acharya, and R. M. P. M. Baltussen. 2000. ease may have been excluded due to the way their cost-effectiveness ratios "Development of WHO Guidelines on Generalized Cost-Effectiveness were calculated. For example, nutrition-related interventions are excluded Analysis." Health Economics 9: 235­51. from the table because those evaluated in the volume address either vita- min A deficiency or iodine deficiency both of which are associated with Murray, C. J. L., K. Styblo, and A. Rouillon. 1993."Tuberculosis." In Disease low avertable burden. Also, only the burden of children age 0 to 4 was con- Control Priorities in Developing Countries, ed. D. T. Jamison, W. H. sidered, further lowering the avertable burden. Another example is of Mosley, A. R. Measham, and J. L. Bobadilla, 233­59. New York: Oxford the integrated management of infant and childhood illness, which is eval- University Press. uated for Sub-Saharan Africa but not for the South Asia Region. Musgrove, P. 1999. "Public Spending on Health Care: How Are Different 7. The second and third observations speak more generally to the Criteria Related?" Health Policy 47 (3): 207­23. global public goods nature of health research (see chapter 4 for an in- Tengs, T. O. 1997. "Dying Too Soon: How Cost-Effectiveness Can Save depth discussion). In relation to both HIV/AIDS and noncommunicable Lives." NCPA Policy Report 204, National Center for Policy Analysis, diseases, the responsiveness of the medical research system to threats to Dallas, TX. populations in developed countries has the potential to bring great bene- Tengs, T. O., M. E. Adams, J. S. Pliskin, D. G. Safran, J. E. Siegel, M. C. fits to people living in LMICs. Weinstein, and J. G. Graham. 1995. "Five Hundred Life Saving Interventions and Their Cost-Effectiveness." Risk Analysis 15 (3): 369­90. REFERENCES WHO (World Health Organization). 1999. TB Advocacy: A Practical Guide 1999. Geneva: WHO Global Tuberculosis Program. Birch, S., and C. Donaldson. 1987."Cost-Benefit Analysis: Dealing with the ------. 2004. Global Tuberculosis Control: Surveillance, Planning, Problems of Indivisible Projects and Fixed Budgets." Health Policy 7 Financing. Geneva: WHO. (1): 61­72. 86 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Jeffrey Chow, and Sonbol A. Shahid-Salles Chapter 3 Strengthening Health Systems Anne Mills, Fawzia Rasheed, and Stephen Tollman Interventions are not generally provided as freestanding activi- Such inefficiencies have two main causes. First, they may ties but are delivered in a variety of packages and through dif- occur because decision makers lack incentives to behave effi- ferent levels of a health system.1 For this reason, this book--in ciently; for example, their promotion chances may not depend addition to including the disease- and program-specific on how well they perform in managing a hospital. Second, chapters--addresses not only the cost-effectiveness of levels of decision makers may be constrained in their ability to make care, packages of care, and services but also the strengthening efficient choices; for instance, they may lack knowledge or of the management of health systems as a whole. experience of what to do or political factors may affect whether Cost-effectiveness data reflect largely what can be achieved they can dismiss underperforming staff members or determine in a reasonably well-functioning health system. In that which company they must buy drugs from. Evidence on the sense, they can be considered to represent potential cost- quality of care (chapter 70) demonstrates that health systems effectiveness and need to be supplemented with evidence and may not merely be inefficient in failing to minimize costs but guidance on how health systems can be strengthened to pro- may also fail to deliver effective care. vide interventions effectively, efficiently, and equitably. This The extent of inequities is also a major concern. Recent argument is given added weight by evidence on inadequacies analyses show that even when interventions are provided, the in the performance of health institutions in countries at all lev- poorest members of society usually have the least access to els of development (Hensher 2001; Preker and Harding 2003). them (Gwatkin and others 2000). In many countries, gaps in Hensher (2001) documents the extensive inefficiencies in low- child mortality between the poor and the better off widened and middle-income countries, including the following: during the 1990s (World Bank 2004). Thus, health systems need to have the capacity not only to deliver interventions effi- · failure to minimize the physical inputs used--for example, ciently but also to sustain high levels of coverage, especially of prescribing excessive quantities of drugs the poorest and most vulnerable. · failure to use the mix of inputs that costs the least--for Awareness has grown that international targets, such as the instance, allocating a high proportion of expenditure to Millennium Development Goals (MDGs) and the provision of staff salaries and only a small share to operating costs and antiretroviral treatment for HIV/AIDS patients cannot be maintenance achieved without the key elements of a functioning health · failure to operate at the appropriate scale--for example, system. The example of the reduction of maternal mortality running extremely large hospitals that suffer from scale in Sri Lanka (chapter 8) demonstrates the improvements in inefficiencies health outcomes that are possible once a basic platform of · failure to pay staff enough to encourage good performance. functioning health services is available on which targeted ini- tiatives can build (Levine 2004). Hensher estimates that hospital inefficiencies could easily Thus, the aim of this chapter is to review how health systems account for up to 10 percent of total health spending. can be strengthened in differing country contexts to deliver 87 interventions effectively, efficiently, and equitably. The chapter services in a number of postindependence Sub-Saharan is mainly concerned with strengthening health services: issues African countries and the health successes of countries such as in managing core public health functions are reviewed else- China and Cuba influenced a new international emphasis on a where (Khaleghian and Das Gupta 2004). Although the chap- broadly based definition of primary health care. Quickly, how- ter seeks to draw valuable lessons from all parts of the world, it ever, the advocates of more focused disease-specific efforts focuses on countries with the least capacity, especially the responded with the notion of selective primary health care poorer countries in Sub-Saharan Africa and Asia. (Walsh and Warren 1980), focused on a limited number of pre- sumed cost-effective interventions. Since 1980, this tension in international health policy has HISTORY AND CURRENT THEMES persisted, with four main strands, namely: Efforts to improve health in low- and middle-income countries over the past 50 years can be divided into a number of periods, 1. The health care reform movement of the 1990s, which with pendulum swings between focused, disease-specific sup- has continued into the new millennium in a somewhat port and broader health service or health system support. The attenuated form, has focused almost exclusively on financ- 1950s, 1960s, and 1970s witnessed a number of successful dis- ing and organizational changes, largely neglecting the ques- ease control efforts, often termed mass campaigns--notably tion of whether improved health outcomes have been smallpox eradication, but also, for example, malaria and yaws achieved. control (Walt 2001). These mass campaigns built on earlier 2. The definition and development of cost-effective packages efforts, including those of the Rockefeller Foundation from the of care has progressed, as reviewed in chapter 64, with some 1920s in controlling hookworm, yellow fever, and malaria. attention given to their implications for services and Despite regional differences in the degree of progress--for systems. example, malaria control was not attempted in most of Sub- 3. The emphasis on specific disease-focused international Saharan Africa--successes in regional and global control of programs, as reflected in the Global Fund for AIDS, diseases such as Chagas disease and measles in Latin America Tuberculosis, and Malaria, has been increasing, and and the Caribbean and, more recently, polio worldwide have resources for such programs have been expanding. continued since the 1960s. 4. The effort to encourage investment in integrated health From early in the history of mass campaigns, the terminol- services has continued. ogy used was of vertical and horizontal approaches (Gonzalez 1965), referring essentially to two key dimensions in program Recent events indicate that these tensions remain unre- organization (Mills 2005): the extent to which program man- solved. For example, Molyneux and Nantulya (2004) call for agement was integrated into general health systems manage- combining community-driven, global health initiatives ment, especially at lower management levels, as opposed to (including drug distribution for schistosomiasis, filariasis, and kept strictly separate, and the extent to which health workers onchocerciasis; trachoma control; bednet distribution for had one function as opposed to many functions. Vertical pro- malaria control; and immunization), with little mention of grams (also known as categorical programs) had their own how community-based efforts might link with the general financing, management structures, and staff, even down to the health infrastructure. In contrast, Unger, de Paepe, and Green service delivery level, instead of relying on existing systems. In (2003) examine how best to implement disease control pro- contrast, horizontal programs delivered a number of services grams so as to strengthen existing health systems and propose through the general health service structure. a code of best practice for such programs. In malaria control, for example, the World Health This debate is being given a new urgency by the introduc- Organization defined a process whereby an initial vertical tion of treatment for HIV/AIDS. Immunization can be deliv- approach would evolve into a horizontal approach as the inci- ered using either a vertical or a horizontal approach. dence of malaria fell. Initially, the effort required to detect and HIV/AIDS treatment, which requires continuing care, calls for treat cases demanded dedicated and mobile workers. As trans- strong health service backup. Nonetheless, such treatment mission was reduced, these workers would detect fewer and services could be organized so that they isolate themselves fewer cases, so on efficiency grounds, detection and treatment from the broader health system--say, through separate clinics activities needed to be handed over to the general health serv- with their own workers and separate laboratories--or they ice infrastructure. However, this approach faced the dilemma could contribute to a greater degree of integration by sharing that such services were often not strong enough to carry the resources. control efforts forward. The implications of these different approaches for health The Alma Ata Declaration of 1978 was a turning point. The system change are not purely academic. Table 3.1 compares increasing emphasis on building networks of peripheral health the responses to health system constraints that derive from a 88 | Disease Control Priorities in Developing Countries | Anne Mills, Fawzia Rasheed, and Stephen Tollman Table 3.1 Typical Health System Constraints and Possible Disease-Specific and Health System Responses Constraint Disease-specific response Health system response Financial inaccessibility: inability Allowing exemptions or reducing prices for Developing risk-pooling strategies to pay, informal fees focal diseases Physical inaccessibility: distance Providing outreach for focal diseases Reconsidering long-term plans for capital investment and siting to facility of facilities Inappropriately skilled staff Organizing in-service training workshops Reviewing basic medical and nursing curricula to ensure that to develop skills in focal diseases basic training includes appropriate skills Poorly motivated staff Offering financial incentives for the delivery Instituting performance review systems, creating greater clarity of particular priority services about roles and expectations, reviewing salary structures and promotion procedures Weak planning and management Providing ongoing education and training Restructuring ministries of health, recruiting developing a cadre workshops to develop planning and of dedicated managers management skills Lack of intersectoral action and Creating disease-focused, cross-sectoral Building systems of local government that incorporate represen- partnership committees and task forces at the tatives from health, education, and agriculture, promoting the national level accountability of local governance structures to the people Poor-quality care among private Offering training for private sector providers Developing accreditation and regulation systems sector providers Source: Travis and others 2004. disease-specific focus as opposed to a health systems focus. A At the community and household level, lack of demand can disease-specific focus leads to solutions for the specific pro- limit coverage. This lack may stem from cultural factors, such gram, whereas a health systems focus identifies a somewhat dif- as low acceptability of immunization or prenatal care, but it ferent set of reform priorities that relate to system-level may also result from financial and physical barriers to access. changes and affect disease management across multiple For example, estimates indicate that, in Niger, children under diseases and conditions. The disease-focused responses can five average only 0.5 visits to a health provider per year, and in generally be implemented relatively quickly, whereas the Bangladesh, only 8 percent of ill children were taken to a qual- systems-focused actions take longer. However, numerous, ified provider (see chapter 63). Many barriers can be reduced separate disease-specific responses can rapidly overwhelm by increasing accessibility, for example, by expanding the serv- frontline workers and managers. ice infrastructure closer to communities. In Cameroon, Litvack and Bodart's (1993) study finds that a combination of user fees and improved quality, including a better drug supply and HEALTH SYSTEM CONSTRAINTS ON THE improved geographic access, led to increased use despite the DELIVERY OF PUBLIC AND PERSONAL user fees. HEALTH SERVICES Low use may stem less from inaccessibility than from low quality at the level of health care delivery. Low quality can The challenge of scaling up services to meet the health-related result from human resource shortages, limited incentives for MDGs and concerns that the multiple international efforts may the staff to provide good quality care, training inappropriate to overwhelm countries' fragile infrastructures have encouraged local needs, poor drug supply systems, and lack of simple efforts to think systematically about health system constraints equipment such as that needed to measure blood pressure on achieving the MDGs, and the extent to which additional (Southern Africa Stroke Prevention Initiative Project Team funding can readily and quickly improve services (Ranson, 2004). In Tanzania, an analysis of the treatment-seeking deci- Hanson, and Mills 2003; Travis and others 2004). Weaknesses in sions of those who later died from malaria showed that the service delivery--for example, at the health center level--may great majority had preferred modern medicine, even for cere- stem from problems at that level, such as staff shortages, or may bral malaria, which according to a substantial body of evidence be affected by factors higher up the system, such as a poor drug mothers view as a condition best treated by traditional healers distribution system. Ranson and others (2003) therefore ana- (de Savigny and others 2004). Yet despite high rates of seeking lyze constraints by five different levels: community and house- modern medicine, malaria mortality remained high, whether hold, health services delivery, health sector policy and strategic because of delay in seeking treatment, poor quality care, or management, public policies cutting across sectors, and envi- poor patient adherence. Treatment quality can be improved ronmental and contextual characteristics (see table 3.2). by increasing resources, although it may also demand change Strengthening Health Systems | 89 Table 3.2 Constraints on Improving Access to Essential Health Interventions, by Level Level of constraint Types of constraints Community and household Lack of demand for effective interventions Barriers to the use of effective interventions (physical, financial, social) Health services delivery Shortages and inadequate distribution of appropriately qualified staff Weak technical guidance, program management, and supervision Inadequate drugs and medical supplies Lack of equipment and infrastructure, including poor accessibility of health services Health sector policy and strategic management Weak and overly centralized planning and management systems Weak drug policies and drug supply system Inadequate regulation of pharmaceutical and private sectors and improper industry practices Lack of intersectoral action and partnership for health between government and civil society Weak incentives to use inputs efficiently and to respond to users' needs and preferences Reliance on aid agency funding, which reduces flexibility and ownership Aid agency practices that overload country management capacity Public policies cutting across sectors Government bureaucracy (civil service rules and remuneration, centralized management system) Poor availability of communications and transportation infrastructure Environmental and contextual characteristics Governance and overall policy framework: · Corruption, weak government, weak rule of law, weak enforceability of contracts · Political instability and insecurity · Low priority attached to social sectors · Weak structures for public accountability · Lack of a free press Physical environment: · Climatic and geographic predisposition to disease · Physical environment unfavorable to service delivery Source: Hanson and others 2003. higher up the system; for example, better health worker ple, wage policies for the public sector health staff are usually performance may not be possible without reforming human set centrally and linked to overall levels of pay for the public resource management systems. sector. Even if funds are available, increasing the wages of only Performance at the third level, health sector policy and health staff members may not be possible. strategic management, can have a pervasive influence on per- At this highest level, constraints also reflect much broader formance at lower levels and is less easy to address through institutional influences, as was demonstrated by recent analyses additional funding alone. Some improvements, such as orient- of the results of efforts to build state capacity in Africa. Levy's ing management more toward good performance and reduced (2004) review points out that the results are mixed at best. For corruption, may require a change in organizational culture or a example, of all World Bank civil service reform projects com- change in structures--for instance, decentralizing authority or pleted by 1997, only 29 percent were rated as satisfactory by the creating autonomous agencies. Such changes can be difficult operations evaluation department. Levy argues that a key rea- and may take time to implement (Preker and Harding 2003). son for the limited success was an implicit presumption that Other improvements require action outside the country--for the weakness of public administration was managerial and example, a change in aid agency practices so that weak country could be remedied through organizational change and finan- management structures are not overloaded by multiple cial support for technical advice, hardware, and training. demands and reporting structures. However, public administrations are part of political institu- Finally, at the highest levels, broad multisectoral public poli- tions and of social, economic, and political interests more cies and environmental and contextual characteristics set limits broadly, and they do not change readily or quickly. Never- on what the health sector can change without help. For exam- theless, windows of opportunity may open that drastically 90 | Disease Control Priorities in Developing Countries | Anne Mills, Fawzia Rasheed, and Stephen Tollman affect the chances of change within a few years. Consider the which its activities are viewed by the general public." The cases of Mozambique, Rwanda, and Uganda, all countries that importance of the stewardship role is indicated by analyses that experienced many years of conflict and economic collapse but suggest that, in countries with good governance, a relationship that have since made significant progress in reforming govern- is apparent between increased health spending and reduced ment institutions and performance. Apart from those excep- child mortality (chapter 9), but that such a relationship is not tional cases, Levy argues that the way forward for administra- apparent in countries that scored less well on indicators of tive reform is likely to be an incremental one. good governance. In general, straightforward shortages of buildings, equip- Strengthening structures of accountability to communities ment, and drugs and a lack of specific skills on the part of health and introducing mechanisms to ensure that users have a voice workers and managers can be addressed fairly rapidly with addi- in the local health system and can influence priorities are likely tional funding. Remedying staff shortages takes somewhat to be important in encouraging good performance. Methods to longer, especially if the education system is producing insuffi- increase the transparency of resource allocation to peripheral cient numbers of people with the qualifications needed to enter services are also needed. In Burkina Faso, participation by com- health training programs. The constraints most impervious to munity representatives in public primary health care clinics has additional funding are likely to relate to broader systems and increased the coverage of immunization, the availability of institutional deficiencies, such as a bureaucratic culture that essential drugs, and the percentage of women who get two or does not reward good performance and political systems that more prenatal visits. In Ceara, Brazil, strengthened community ignore the voices of the poor. Long-term and carefully phased accountability mechanisms helped improve service delivery capacity building in the broadest sense, including political (chapter 9). Factors identified as important to the success of development and strengthened governance structures, is likely community-based health and nutrition programs in chapter 56 to be required to relax these constraints (Mills and others 2001). include the existence of an effective, respected, and socially inclusive organization at the community level that builds on ASSESSMENT OF APPROACHES TO established community procedures. STRENGTHEN HEALTH SYSTEM CAPACITY Because of the substantial role that private sources of care play in almost all low- and middle-income countries, regulat- Strengthening health system capacity to improve performance ing and developing creative ways to work with the private sec- is a wide-ranging subject, likely to require action--often simul- tor are important. This effort needs to be seen as part of the taneously or appropriately sequenced--on many fronts. In stewardship role. Even though most countries have a network particular, it requires attention to the various functions of the of regulations controlling private providers and products such health system, especially to the various dimensions of manage- as drugs, the regulations are often outdated and poorly ment, as well as to the relationships between the health system, enforced and can even be counterproductive (box 3.1). its patients (clients), and their communities. Evidence on Evidence is growing that using a mix of measures to influ- which approaches work best is limited. The coverage of this ence both consumers and providers can improve the quality of section is therefore selective, drawing on chapters in part III care obtained through private providers. Chapter 70 provides and focusing on stewardship and regulation, organizational several examples, including introducing total quality manage- structures and their financing, and general management ment practices and training with peer review feedback. functions--namely, human resources and quality assurance. Providers in the informal sector are some of the hardest to When possible, we identify general lessons and note reach because of their wide distribution, small scale, and mini- instances of relevant country experiences. In interpreting them, mal education; however, some evidence indicates that their dis- readers will need to keep in mind the strengths and weaknesses pensing practices can be improved (box 3.2). of their own country's health system. For example, in South Regulation can be used as an intervention in its own right, Africa, where basic hospital supplies are good, improved train- as well as a way to improve health service delivery. The list ing of health staff members reduced case-fatality rates for of interventions identified as success stories (chapter 8) severe malnutrition, whereas in settings that experience short- includes these in which a change in regulation was at the root ages of antibiotics, potassium, and milk powder and that lack a of success: doctor, training alone is highly unlikely to reduce high case- fatality rates (Ashworth and others 2004). · regulations requiring all sex workers in brothels to use con- doms in Thailand Stewardship and Regulation · tobacco control legislation in Poland and South Africa Saltman and Ferroussier-Davis (2000, 735) explain stewardship · provision of a legal and regulatory framework for adding as a "function of governments responsible for the welfare of fluoride to salt in Jamaica populations and concerned about the trust and legitimacy with · legislation banning the sale of noniodized salt in China. Strengthening Health Systems | 91 Box 3.1 Bringing de Jure and de Facto Regulations in Line In Tanzania, local drug shops are important sources of were more reliable than those in government facilities. drugs. They are required to obtain a permit each year and Revising the regulations to permit drug shops to stock a to meet certain conditions related to premises, qualifica- small set of oral antibiotics, for example, would allow more tions of the seller, and products (nonprescription medi- constructive engagement between sales staff members and cines only). A study in three districts found that, despite regulators, including the provision of information on regular inspections of drug shops, infringement of the reg- essential drugs, registered brands, appropriate dosing, and ulations was widespread--including the sale of prohibited consumer advice. The Strategies for Enhancing Access to or inappropriately packaged drugs, which inspectors must Medicines Project is experimenting with allowing a wider have known about. Illegal drug sales may have contributed range of drugs to be provided in one region using accred- to poor-quality treatment and encouraged the develop- ited outlets for dispensing drugs (drug shops that meet ment of drug resistance, but they had important benefits in specified quality criteria and whose staff members have terms of accessibility, because drug supplies in drug shops been trained by the project). Source: Goodman 2004. Box 3.2 Improving the Quality of Drug Dispensing by Private Sector Shops In Kilifi district, Kenya, an education program piloted by Between 1998 and 1999, the proportion of antimalarial the Kenya Medical Research Institute­Wellcome Trust drug users obtaining an adequate dose rose from 8 to 33 Collaborative Research Programme worked with district percent, and by 2001, with a national change to sulfadox- health managers to train and inform rural drug retailers ine pyrimethamine, to 64 percent. The proportion of those and communities. Its effect was evaluated by means of with malarial fevers who received an adequate dose of a annual household surveys of drug use and shop surveys in recommended antimalarial drug within 24 hours rose an early and a late implementation area. The program from 1 to 28 percent by 2001. showed major improvements in drug-selling practices. Source: Marsh and others 2004. Given that enforcement is the Achilles heel of regulation, a In recent years, the approach known as new public manage- noteworthy point is that these countries are all middle-income ment, explained further in chapter 73, has encouraged a countries with a reasonable level of enforcement capacity. In rejection of traditional, hierarchical forms of public sector other countries, approaches such as that outlined in box 3.2, management, whereby a single organization both finances and where the authorities work with the private sector rather provides health services. For example, the U.K. health service than seeking to control it, may have a better chance of has introduced a clear separation between the entities purchas- succeeding. ing services (deciding what services are required for a given population and allocating funds for them) and those providing services. One aim of such arrangements is to ensure that Organizational Structures and Financing providers' interests--as opposed to users' interests--do not The appropriate configuration of health system structures can dominate decisions on what services are funded. In addition, ensure a clear delineation of responsibilities and accountabili- separating purchasers and providers allows competition to be ties inside organizations, linking performance with rewards. introduced in service provision. Although introducing compe- Governance and organizational structures can also help ensure tition is widely considered desirable to encourage efficiency, organizations' accountability to the public. debate continues on the magnitude of potential adverse effects. 92 | Disease Control Priorities in Developing Countries | Anne Mills, Fawzia Rasheed, and Stephen Tollman Examples of new organizational structures include remov- run into major opposition from public sector workers, who ing national health services from civil service control, intro- oppose changes in their terms and conditions of service. ducing executive agencies to manage health services, and using Some of the more successful elements of new public contracts to govern relationships, both within the public sector management reforms are those that involve contracting out (between public purchasers and public providers) and between services, especially to nongovernmental organizations the public and private sectors (Preker and Harding 2003). (NGOs). Early evaluation of contracting experiences indicated Colombian reforms introducing competition in both insur- that, even though contracting had been perceived as a way to ance and provider markets are among the most comprehensive. avoid the inefficiencies inherent in public sector provision, it Another reform example is Ghana's creation of the Ghana nonetheless required public sector capacity to manage the Health Service, which is separate from the Ministry of Health. contracting process (Mills 1998). This situation was particu- The high transaction costs involved in creating and manag- larly a problem if the contractor was a commercial firm or ing these types of arrangements and the lack of evidence that individual provider with incentives to maximize profits competition improves the quality of care have moderated (box 3.3). Contracting with individuals and firms that are initial enthusiasm for new forms of public management. In strongly influenced by a profit motive requires a certain level addition, critics argue that such arrangements are more of state capacity to ensure that the arrangements work in the demanding on management capacity than is direct service pro- interests of the state and the general public. In some countries, vision (Mills and others 2001). Moreover, implementation has therefore, NGOs may be more appropriate service providers proved challenging. For example, in Trinidad and Tobago and (Palmer and Mills 2003). A number of quite positive results in Zambia, reforms to create new health service agencies have from contracting with NGOs are now available (World Bank Box 3.3 The Importance of Government Capacity: Contracting Out Health Services in South Africa Successive studies have evaluated experiences in contract- A similar study evaluated the performance of contracts ing out hospital care and primary care services in South with general practitioners for primary care in two Africa. The hospital study compared three district hospi- provinces and compared their performance with that of tals whose management had been contracted out to the public clinics. General practitioners' costs were similar to same private company with three nearby, comparable, those of small public clinics, but the service was generally publicly managed district hospitals. Overall, the contrac- of poorer quality. Exploration of the relationship between tor hospitals were able to provide care of more or purchasers and providers found that the contract was less equivalent quality at significantly lower cost to incomplete and open to interpretation and that monitor- themselves--in major part because their productivity was ing was constrained both by a lack of capacity and more than double that of the public hospitals as a result of resources and by the difficulty of monitoring a complex their effective human resource policies. However, the con- service delivered in remote locations. Sanctions were tractor captured all the efficiency gains as profit, leading to vaguely specified and rarely used because of a sense of a situation where contracting out was actually more costly mutual dependence between parties to the contract that for the government than direct provision. The contractor's lessened their willingness to enter into disputes. In addi- capacity to profit from the arrangement was due mainly to tion, the two provinces varied in terms of their capacity to its ability to secure highly favorable contract terms and monitor performance. The province with lesser capacity prices and to ensure a high total number of days of care. had little information about general practitioners' per- Interview data confirmed a substantial imbalance between formance and little contact with them, which seemed to the government and the contractor in relation to the increase suspicions of what general practitioners were skills, capacities, and information required to negotiate doing. In contrast, the province with greater capacity had contracts. In addition, government officials underesti- a better information system and a decentralized manage- mated the extent of potential competition for contracts ment system that led to greater contacts between man- and therefore overestimated their dependence on the one agers and general practitioners and an apparently greater contractor. degree of understanding between the parties. Sources: Broomberg, Masobe, and Mills 1997; Mills and others 2004; Palmer and Mills 2003. Strengthening Health Systems | 93 2004), and the example of Cambodia is one of the most fre- Human Resources quently quoted (chapter 13). Nevertheless, most evidence Achieving health policy goals depends on being able to train, comes from programs with substantial external financial and recruit, and retain a staff with the necessary bundles of skills. In technical resources, and long-term experience of sustainability planning for human resource needs, countries must relate is lacking. the numbers and levels of each category of staff members to Management decentralization has been another continuing health policy goals and the priorities that are set, given the theme in recent years. One variant is its application to hospi- overall availability of resources and local labor market tal management, which involves giving hospitals autonomous constraints. or corporate status along with much greater responsibilities In recent years, concerns about the international brain drain for raising income and managing their own affairs. A second have increased greatly, with evidence indicating that migration variant is the creation of autonomous government agencies, by doctors and nurses is severely affecting health services in some and a third is decentralization to general management Sub-Saharan African countries (Physicians for Human Rights structures at lower levels, such as a health authority or local 2004). Actions by developing countries to improve recruitment government. and retention should either raise the rewards, both financial and Some pushing down of the locus of control over decision nonfinancial, of local employment or reduce the attractiveness making is a prerequisite for effective management at the local of alternative employment--for example, by making qualifica- and facility levels. However, without the necessary resources tions less portable across countries (chapter 71). Raising the and management expertise at these levels and the right incen- remuneration of health workers may be difficult because it is tives, adverse consequences may arise for both efficiency and likely to lead to demands for increased pay from other public sec- equity. For example, experience with hospital autonomy in tor employees. There is a long history of making use of local low-capacity settings suggests that making the hospital partially cadres, which can also allow training that is more specific to the dependent on fees for its income will restrict access by the poor needs of the local health system and its priorities. to the hospital and also worsen the care they receive when Examples include nurses with extended training and roles admitted (Castaño, Bitrán, and Giedion 2004). However, for and people working at subnurse levels with training of a few nonpatient care services, whose functions are easier to specify weeks to three years. For example, Bangladesh employs family and monitor, autonomous agencies may have some advantages. welfare visitors, health assistants, and medical assistants; For example, the Tamil Nadu Medical Supply Corporation has Uganda provides three years of training to clinical officers, who greatly improved the efficiency and effectiveness of drug pur- function as subdoctors, and three months of training to nurs- chasing and distribution (Mills and others 2001). ing aides; and Malawi trains clinical officers, who carry out For decentralization of general health service management surgical procedures and administer anesthetics in addition to to succeed, attention must be paid to the entire management providing medical care. Despite widespread use of such work- system, including management skills, information, analytical ers, evidence on how they perform relative to more qualified tools, and accountability mechanisms both to the community staff members is limited, though a study of clinical officers in and to higher levels of management. Because decentralization Malawi suggests that well-trained clinical officers can safely is a complex process, takes a variety of forms, and is affected by substitute for doctors in performing cesarean deliveries the local context, research on its merits and demerits has been (Fenton, Whitty, and Reynolds 2003). inconclusive (Alliance for Health Policy and Systems Research The salaries necessary to recruit and retain staff members 2004). Some evidence indicates that decentralization to local will depend on the opportunities they have for other employ- governments can lead to neglect of broader public health func- ment both within the country and in other countries. Salary tions and disease control, because these types of care are less levels will also depend on health workers' preferences between visible to the public than curative care, as Khaleghian and Das financial and nonfinancial incentives. Evidence suggests that Gupta (2004) indicate occurred in the Philippines. influences on motivation, though reflecting universal princi- Reviews of the merits of integrating services and of the ples, will vary considerably from place to place (Brown 2002). effect of vertical programs on health systems have also been Therefore, compensation and incentive structures need to be inconclusive. Some positive examples are available, such as the adapted to countries' circumstances; however, evidence is strengthening of health infrastructure and surveillance systems scanty on how countries have attempted to adapt such struc- by the polio elimination campaign in Latin America and the tures and whether they have been successful in improving Caribbean (Levine 2004). Nonetheless Briggs, Capdegelle, and recruitment and retention. Garner's (2001) review of the effects of strategies for integrat- One approach to improving health workers' performance is ing primary health care services on performance, costs, and to link performance and remuneration. The Chinese national patient outcomes finds too few studies of good enough quality tuberculosis (TB) program, identified as a success story (chap- to draw firm conclusions. ter 8), provided village doctors with incentives to treat TB 94 | Disease Control Priorities in Developing Countries | Anne Mills, Fawzia Rasheed, and Stephen Tollman patients. However, performance-related pay requires a good the availability of facilities and materials; of opportunities for regulatory framework, skilled managerial resources, and care- learning and career progression; of subsidized housing and ful monitoring to counter adverse effects--all features that are education for dependents; and of a culture that values the con- unlikely to be available in countries with limited capacity. Even tribution of health workers to the achievement of organiza- in China, other experiences are much less positive because tional and system goals. In addition, the methods and levels of managers were not required to take likely adverse health conse- funding, the extent of organizational autonomy, the nature of quences into account (box 3.4). Similar comments apply to the support and supervisory systems, the role of the organization widespread practice of allowing doctors to work in both the and of providers in the health system, and the regulation and public and the private sectors to increase their incomes. accountability structures all influence how organizations and Doctors may exploit their private practice rights by encourag- individuals function. Thailand provides an example in which ing patients to attend privately if they want better quality the provision of both monetary and nonmonetary rewards has care--or even by diverting government resources, such as improved the recruitment, retention, and status of rural doc- drugs, to private patients. Thus, the effects of private practice tors (box 3.5). on incentives in public practice tend to be negative unless care- The introduction of well-funded disease control programs fully monitored and controlled. runs the risk of attracting the most able staff members away Nonmonetary rewards to encourage staff retention can be from other positions. Past programs have successfully used useful in such settings, as well as easier to manage. They include combinations of financial and other incentives to encourage Box 3.4 Incentive Payments in China China has made wide use of incentive payments in quality of care, the bonus system was clearly designed to hospitals--and even in public health programs. Research achieve financial goals rather than quality goals. suggests that such payments have deleterious effects when Furthermore, during the 1980s and 1990s, the government their ability to skew behavior is not controlled. In provided a decreasing share of the income of public health Shandong province, studies found that a change in the institutions, and the share of service charges greatly bonus system for hospital doctors from one that was tied increased. As a result, public health institutions became to the quantity of services provided to one that was tied to heavily dependent on generating their own income. revenue generated was associated with a significant Negative effects included duplicate inspections of factory increase in hospital revenue. About 20 percent of hospital premises by different public health units, excessively fre- revenue was generated by the provision of unnecessary quent inspections, and neglect of less profitable factories care. Although data did not permit linking bonus type to that were less able to pay inspection charges. Source: Liu and Mills 2002, 2003. Box 3.5 The Role of Financial and Nonfinancial Incentives in Thailand Thailand has experienced periods of severe medical brain self-esteem has been increased by providing career oppor- drain from the public to the private sector and has had tunities up to the post of deputy director general, an great difficulties in staffing hospitals in rural areas. Since annual award for rural doctors, and membership in the the late 1970s, policies have been directed at making serv- rural doctor society. Substantial experience as a rural doc- ice in rural areas more attractive. Measures include sub- tor is explicitly valued by leading public health specialists, stantial salary increases, good working conditions in who themselves have spent substantial periods working as district hospitals, and provision of housing. Professional rural doctors. Source: Wibulpolprasert and Pengpaiboon 2003. Strengthening Health Systems | 95 Box 3.6 Improving Staff Performance in Cambodia An NGO contracted to manage a district in Cambodia more. Subcontracts were made competitive: if a health introduced contracts between the NGO's managers and center's management or output was poor, other health facility staff members involving a monthly incentive pay- workers or managers were asked to apply to take over the ment, a punctuality incentive, and a performance bonus. contract. During 2001, four contracts were replaced. The contracts were initially introduced in three facilities, Monitoring activities, especially spot checks at the which subsequently experienced significantly higher use household level to verify that recorded visits had taken levels than those that did not have the incentive payments. place, were considered vital to ensuring quality and trans- Because individual contracts were too demanding to man- parency. Soeters and Griffiths (2003) argue that out- age and excluded health center and hospital directors from siders--in this case the NGO--are better able to introduce staff management, the system was changed to one of sub- new management procedures than a ministry of health, contracting with facility managers. Output improved even which tends to be risk averse. Source: Soeters and Griffiths 2003. good worker productivity and program performance (chap- in staff performance by establishing clear agreements with staff ter 71). Incentives have included better salaries; field and trans- members concerning issues such as the working hours that portation allowances; and nonfinancial incentives such as would be expected and the informal charges that staff members streamlined management, specialized training, availability of were not to demand from patients. In return, staff members facilities and material resources, and results-oriented manage- received substantial incentive payments (box 3.6). ment that provides effective administrative and technical sup- port. Governments need to find ways of benefiting and learning from these experiences. For example, governments might allow Quality Assessment and Assurance periods of secondment to externally funded programs, after The quality of health services has a number of important impli- which staff members return to the government with enhanced cations. It affects the outcomes that a health system can skills. The success of such an approach will depend on remuner- achieve--both directly, through patient treatment, and indi- ation not differing too greatly and on government bureaucracies rectly, by encouraging or discouraging use of the services. It also providing the scope for staff members to use their new skills. affects staff morale, because working in an environment where However, the history of civil service reform is not encourag- employees know the treatment quality is poor is not motivating. ing (Nunberg 1999). Reforms have sought to reduce the size of Substantial evidence, reviewed in chapter 70, indicates that the civil service and to improve productivity using incentive the quality of care is often suboptimal and varies widely with- schemes such as performance-based pay and promotion struc- in countries. In part this suboptimal quality is attributable to tures. Such reforms have been largely unsuccessful because of resource constraints, but providing good-quality care is possi- the political difficulties of reducing the size of the civil service. ble even in resource-poor settings. Structural and organizational changes are typically unpopular Evidence on how providers' practices can be improved can with labor unions, especially if they perceive such changes as be grouped into two categories: policies that indirectly affect threatening workers' well-being. Experience demonstrates the providers' practices by changing structural conditions, includ- difficulties of aligning system and organizational objectives ing the practice environment, and policies that directly affect with individual workers' objectives (Martineau and Buchan individual and group practices. 2000) and suggests that solutions need to be sought that do not In the first category, legal mandates and administrative reg- involve radical reform of employment patterns unless the ulations can be used to bar unqualified workers from practic- country setting is particularly propitious. ing; professional oversight and clinical guidelines can encour- Where contracting with NGOs or other private providers is age good practices; contracts can specify and monitor quality an option, doing so may permit changed employment patterns standards, such as immunization coverage targets; and accred- and improved performance without the widespread disruption itation can stimulate quality improvements. Among policies that can result from attempting to change government workers' that directly affect providers' behavior, training with peer terms and conditions of service. In Cambodia, a project con- review feedback has been shown to improve quality, as have tracted to an NGO (HealthNet) obtained some improvements total quality management approaches; remuneration can be 96 | Disease Control Priorities in Developing Countries | Anne Mills, Fawzia Rasheed, and Stephen Tollman made dependent on performance subject to the caveats raised oritization process. A more acceptable strategy in most settings earlier. Measures that improve quality can increase use, is to constrain the overall public sector resource envelope in strengthen the public sector's capabilities, and be highly cost- terms of staff, buildings, equipment, and drugs and to leave effective--even cost saving. rationing decisions within the envelope to clinical discretion (Segall 2003). However, clinicians may implicitly ration services in inequitable ways--for example, on the basis of age TARGETING RESOURCES or social status--and supplementary measures are likely to be needed to ensure that health workers do not discriminate An important dimension of health system capacity that has not against poorer and marginalized members of society. been considered explicitly so far, is the ability to ensure that Resource allocation formulas have an important role to play resources are used in ways that meet health system objectives. in the public sector in directing resources to underserved geo- As noted earlier, many health systems fail to perform as well as graphic areas and population groups and to underfunded pro- they might on effectiveness, efficiency, and equity criteria. This grams (Musgrove 2004). Given the typical shortages of health section addresses what policy instruments might be available to workers in more remote areas, such formulas should include ensure that additional resources are used to the greatest effect, remote area allowances or allow for the higher costs of deliver- first at the systems level and then at the level of service delivery. ing services in such areas. A formula in Zambia, for example, used distance from the railway line as a proxy for remoteness. Systems-Level Mechanisms A similar approach to ensuring that resources go where they At the systems level, tools available to decision makers include are most needed is the "marginal budgeting for bottlenecks" regulation and legislation, resource allocation formulas, and approach of the World Bank (see chapter 9). This country- financial incentives. based planning and budgeting approach assesses health sector Decision makers can use regulation and legislation to set impediments to faster progress toward the MDGs, identifies minimum standards of care that insurance packages must ways to remove them, and estimates both the costs and the cover, for instance. They can influence the availability of drugs likely effects of their removal on MDG outcomes. by, for example, liberalizing prescribing and introducing In targeting resources to specific programs, expansion of accompanying measures to educate providers and users so as to one area of health provision should not occur at the expense of increase the use of certain drugs that are safe to distribute on a another priority area. For example, where staff capacity and large scale. One approach that has worked in Uganda is a facilities are limited, targeting additional funding to TB case social-marketing program making subsidized and clearly pack- detection and treatment may simply take staff time away from aged drugs for sexually transmitted diseases available through child health. This problem of the systemwide effects of disease- the retail sector (Mills and others 2002). specific programs was discussed earlier. Addressing this prob- In some settings, explicit rationing of the provision of care lem requires empowering a central body, such as a ministry of in the public and private sectors can be used to prioritize the health or a regional or district health authority, to take an over- most cost-effective interventions and limit the provision of less all view of priorities so that resource conflicts can be resolved. cost-effective ones. However, regulatory controls are unlikely to Even though financial incentives need to be used cautiously, be effective in low-capacity settings and will simply encourage they can be powerful tools for influencing providers' behavior, as illicit activities. Moreover, explicit rationing requires a high indicated earlier. They can also be an important influence on degree of public acceptance and public involvement in the pri- users'behavior. Experience in South Africa and Uganda (box 3.7) Box 3.7 Removal of Fees at the Primary Care Level in Uganda In February 2001, the government of Uganda abolished of drugs worsened during the first year of implementation cost sharing in public facilities at the community level. but gradually improved during the second year. A study This move was followed by a marked increase in the use of concluded that before the policy change user fees were health services by all population groups. For villages near probably a major deterrent to the use of public health public health centers, the increase was greatest among the services and that their removal was especially beneficial to poorest groups. The frequency with which centers ran out the poor. Source: Nabyonga and others 2005. Strengthening Health Systems | 97 suggests that, in some settings, removing or reducing user fees at shows key indicators for the most constrained and other the primary care level may be an important element in encour- countries. aging greater take-up of primary care. Further studies of the The most constrained group has significantly worse health effects of fee removal are needed. indicators and much worse access to health resources. For example, countries in this group have almost twice the infant Service-Level Mechanisms mortality rate and more than twice the maternal mortality rate of other countries but only one-sixth as many nurses. In At the service level, evidence suggests the value of providing a absolute terms, the most constrained group represents a rela- framework of resources and guidance within which managers tively small share of the total population of countries analyzed and health workers can prioritize their efforts. The experience and consists, for the most part, of small countries (more than of the Tanzania Essential Health Interventions Program (chap- half have populations of less than 10 million) in Sub-Saharan ter 54) highlights the health gains that a decentralized manage- Africa. ment structure can achieve when district managers are pro- The key question in relation to improving health outcomes is vided with the information, tools, and training to enable them what financing and delivery strategy might work best in these to match services and additional resources with the local bur- settings. Should it take the form of a limited number of pro- den of disease. Berwick (2004) draws similar lessons from the grams, each addressing one or a few diseases? Or should efforts experience of several highly successful projects in resource- be devoted to building up the basic health service infrastructure poor settings: set clear aims and targets, use a team approach, on which targeted efforts to address specific health problems can build an infrastructure of human resources and data systems, then be built? Given the lack of evidence, providing guidance is engage with the policy environment, and develop simple difficult, and the chapters in this book present different views. approaches to rapid scaling up. Chapter 63 firmly dismisses the option of bypassing organized Patient education on major causes of ill health is also impor- health services altogether in the poorest countries and promot- tant to ensure that people know when to seek care (for example, ing the delivery of child health interventions directly to house- in the case of childhood illness); understand their rights to var- holds through, for example, community-based projects dis- ious services and the official level of charges; and can make pensing antimalarials or antibiotics. It argues that, though this appropriate decisions about drug purchases. Patient charters approach may be a short-term solution, successes using it largely may play a role in making explicit what patients have the right occur in small-scale pilots with strong managerial backup. to expect from their health services and what level of service Chapter 56 suggests that in the poorest societies basic preventive providers should achieve. Local policies on service provision services should be introduced first--especially immunization, need to relate to community preferences: if they do not, clients' access to basic drugs,and management of the most severe threats confidence in the public health system will be undermined. One to health such as emergency care for traffic injuries. At slightly simple example is the pervasive view in some South African higher levels of development, the introduction of community- communities that public clinics water down medicines, thereby based activities may be cost-effective if coverage by the formal rendering them ineffective (Schneider and Palmer 2002). health service is poor. Both chapters imply that the issue is not Indeed, generic medicines used by the public sector are often which approach to use but how to phase approaches and use a perceived as less effective than name brand drugs. Accurate mix that depends on the intervention and the local context. public information is needed to counter that perception. Accomplishing this requires not only service delivery capacity but also management capacity to plan and evaluate the mix of SOLUTIONS IN LOW-CAPACITY ENVIRONMENTS approaches and make adjustments over time. Molyneux (2004) suggests that disease control programs Developing countries possess a range of capacities to improve can be used to build capacity for the long term and that, with the functioning of their health systems, but one group of coun- time, such programs can become more advisory and less man- tries faces the greatest constraints to doing so. Analyses under- agerial. For example, in Pakistan, primary health care was built taken for the Commission on Macroeconomics and Health on the experience of TB and leprosy clinics. In China, the ver- used the framework presented in table 3.2 to understand the tical programs for disease control purchased time from health dimensions of the constraints problem in 79 low-income service operational staff members, thereby ensuring that funds countries. Using proxies for the various types and levels of flowed into the health service infrastructure (Dean Jamison, constraints--gross domestic product per capita, female literacy personal communication, 2004). In seven countries in rate, number of nurses per population ratio, diphtheria- southern Africa, a successful combined strategy for measles pertussis-tetanus immunization coverage, access to health serv- immunization started with a single, nationwide catch-up cam- ices, control of corruption, and government effectiveness-- paign in which mobile teams vaccinated all children in a par- countries can be classified as more or less constrained. Table 3.3 ticular age group (Levine 2004), an action that can sharply 98 | Disease Control Priorities in Developing Countries | Anne Mills, Fawzia Rasheed, and Stephen Tollman Table 3.3 Health Indicators by Country Level of Constraint Most constrained Indicator Unit countriesa Other countries Total population (in 2000) Millions 401 3,525 Population living on less than US$1/day Millions 123 (9 countries) 886 (29 countries) Population living on less than US$1/day Percentb 30 25 Population living on less than US$2/day Millions 192 (9 countries) 2,128 (30 countries) Population living on less than US$2/day Percentb 48 60 Physicians Per 100,000 population 8.9 101.7 Nurses Per 100,000 population 39.6 208.7 Hospital beds Per 1,000 population 0.78 3.00 Maternal mortality Per 100,000 births 1,134 565 Births with skilled attendant Percent 30.6 59.8 Low birthweight infants Percent 16.4 13.9 Infant mortality (in 1998) Per 1,000 live births 105.3 61.2 Mortality among children under five Per 1,000 live births 171.2 91.9 Measles immunization coverage Percent 48.4 75.3 Diphtheria-pertussis-tetanus Percent 40.3 76.3 immunization coverage TB Directly observed therapy short Percent 31.15 42.10 course (DOTS) detection TB DOTS treatment success Percent 68.4 77.1 Number of countries included n.a. 20 59 Source: Ranson, Hanson, and Mills 2003. n.a. not applicable. Note: Calculations were performed for a constraints index with up to three missing variables. Values for missing variables were imputed using a method described in the source. a. These are the bottom quartile of countries, according to the constraints indicators, compiled into an index. The constraints index was calculated by normalizing each of the variables (subtracting the mean and dividing by the standard deviation) and then summing the normalized values. This calculation gives each variable equal weight in the index. b. These averages are population weighted, whereas all other means in the table are unweighted. reduce the spread of the virus. Routine services were used to system setting is, the more important the provision of good continue measles immunization but with a follow-up cam- technical and management backup will be to service delivery. paign three to four years later to prevent the number of sus- The authors ascribe some of the difficulties that integrated ceptible cases from rising to the level required for transmission. management of infant and childhood illness (IMCI) faced in Over five years, measles virtually disappeared from southern several countries to the absence of full-time IMCI coordina- Africa. However, maintaining this achievement requires that tors, operational plans, and specific budget lines. They suggest routine services be able to reach more than 80 percent cover- that, when health systems are extremely weak, vertical pro- age, a level many countries find hard to sustain. Moreover, in grams may be required; however, as health systems strengthen, low-capacity environments, campaigns can divert attention financing and delivery strategies can become less vertical and and resources from routine primary health care services. more horizontal and less selective and more integrated. Schreuder and Kostermans (2001) indicate that this problem occurred in southern Africa, particularly with respect to divert- ing scarce management capacity, implying that reducing deaths RESEARCH PRIORITIES from one cause may risk worsening services for other diseases and conditions. A notable paucity of evidence is apparent in relation to most Victora and others (2004) suggest that the most appropriate key areas discussed in this chapter. This lack of evidence is illus- mix of vertical and horizontal approaches depends on the trated by a recent review of the evidence on the equity of uti- human and financial resources available, the urgency with lization and financing strategies (box 3.8). This and other which results need to be achieved, the existing organization of reviews of the available evidence have led the Lancet to call for health services, and the natural development of programs over a new health systems research specialty ("Mexico 2004: Global time. Within a horizontal approach, the weaker the health Health Needs a New Research Agenda" 2004). Strengthening Health Systems | 99 Box 3.8 Gaps in the Evidence on Equity of Health Financing and Utilization in Low-Income Settings A recent review evaluated evidence on the effect of various range of settings, including nonrandomized designs when financing strategies on use of health care. It found that randomization is impossible or inappropriate and multi- most research was small scale and had findings of limited center case studies that examine why arrangements do or applicability. Well-designed, large-scale evaluations of the do not work in different settings. The study proposed effect of alternative financing interventions were lacking, developing and applying quality criteria for quantitative and a multitude of case studies described specific experi- and case study research along the lines of guidelines ences but with little methodological rigor. The review recently developed for randomized group trials. recommended larger-scale, more systematic studies in a Source: Palmer and others 2004. Areas where evidence is especially limited that are identified doctoral level, and researchers with doctoral degrees made up in this chapter--where research is a high priority--include the only a quarter of the research workforce. An analysis of studies following: cited in Medline showed that only 5 percent of the health sys- tems research literature concerned developing countries. · Evidence on most health system reforms--for example, Given the importance of influencing policy and practice, the hospital autonomy reforms and decentralization--is inade- approach to research needs to encompass solving operational quate to draw conclusions about the circumstances under problems in real-life settings. Ethical issues arise in using limited which reforms are likely to improve the efficiency and supplies of talent to study problems unrelated to the local con- equity of service delivery. text when the human resources and systems required to · Few studies relate a reform to health outcomes, and even improve operational programs are lacking. Moreover, the qual- evidence on intermediate outcome measures, such as costs ity and effect of investigations are much improved when they and quality of service provision, is often lacking. are based on dialogue with the primary users and set in real-life · Virtually no information is available about the costs of contexts. The concept of the cycle from research to policy and strengthening capacity or the effectiveness of different practice needs to be emphasized more strongly. It encompasses approaches to capacity strengthening, even though the lack not only generating knowledge but also managing the research of system capacity is widely noted. agenda, including setting priorities, and promoting the use of · Evidence is largely lacking on the characteristics of delivery evidence through means such as advocacy channels and spe- strategies capable of achieving and maintaining high cover- cific mechanisms designed to link producers and users of age for specific interventions in various epidemiological, research (Alliance for Health Policy and Systems Research health system, and cultural contexts. 2004). Given the importance of context in translating research · Evidence is lacking on what types of governance and evidence into service and system practice, operational research institutional arrangements will support the achievement of and program evaluation capacity must be built among coun- widespread health improvements, especially for the poorest try-based scientists and practitioners. members of society. CONCLUSIONS Addressing the deficiencies in the evidence base requires developing better study designs and analytical methods and This chapter has sought to address the question of how health building expertise in and understanding of health systems systems can be strengthened to deliver cost-effective and equi- research. Capacity for research and analysis in health policy and table interventions and services. Recent cross-country analysis health systems is currently limited. A recent survey (Alliance for on the association between health expenditure by government Health Policy and Systems Research 2004) estimated that proj- and health outcomes has suggested that the effectiveness of ect funding for health systems research accounted for less than increased health expenditure depends heavily on governments 0.02 percent of the total annual health expenditure of develop- adopting the right policies (World Bank 2004). What are the ing countries. More than half of research projects had budgets right policies, and what are effective implementation processes? of less than US$25,000. Of institutions identified as engaged in The review in this chapter suggests that in many areas not health systems research, a third had no staff qualified at the enough is known to recommend particular approaches and 100 | Disease Control Priorities in Developing Countries | Anne Mills, Fawzia Rasheed, and Stephen Tollman also that recommendations need to be adapted to local con- ACKNOWLEDGMENTS texts. Nonetheless, six key points can be identified in relation to improving health systems: Discussions at a workshop sponsored by the Disease Control Priorities Project in South Africa during June 30­July 3, 2004, · Health systems face numerous constraints in low-income contributed considerably to the development of ideas for this countries, but they are the basis for the long-term future of chapter. sustained health improvements. The health of the system must, therefore, be carefully considered whenever major NOTE new programs are put in place. 1. The health system is understood to encompass all activities whose · If capacity constraints are such that a focused disease- or prime intent is to improve health. program-specific effort is desirable to address an urgent problem, the effort should be designed to contribute to the REFERENCES long-term system strengthening, rather than detracting from it. Countries must avoid having multiple vertical pro- Alliance for Health Policy and Systems Research. 2004. Strengthening grams competing for limited human resources and manage- Health Systems: The Role and Potential of Policy and Systems Research. Geneva: Alliance for Health Policy and Systems Research. rial capacity. Over time, as horizontally organized services Ashworth, A., M. Chopra, D. McCoy, D. Sanders, D. Jackson, N. Karaolis, strengthen, the need for more vertical financing and deliv- and others. 2004. "WHO Guidelines for Management of Severe ery strategies will lessen. Malnutrition in Rural South African Hospitals: Effect on Case Facility · Reforms affecting organizational structures and human and the Influence of Operational Factors." Lancet 363: 1110­15. resource management are likely to play an important role in Berwick, D. M. 2004. "Lessons from Developing Nations on Improving improved performance. However, emerging evidence sug- Health Care." British Medical Journal 328: 1124­29. gests in most settings that changes are most likely to be suc- Briggs, C. J., P. Capdegelle, and P. Garner. 2001. "Strategies for Integrating Primary Health Services in Middle- and Low-Income Countries: cessfully implemented if they are incremental and gradual Effects on Performance, Costs, and Patient Outcomes. Cochrane rather than "big bang" reforms. Stability of policies and con- Database of Systematic Reviews (4) CD003318. sistent implementation are also required. Broomberg, J., P. Masobe, and A. Mills. 1997. "To Purchase or to Provide? · Linking financial incentives to performance, whether The Relative Efficiency of Contracting Out versus Direct Public Provision of Hospital Services in South Africa." In Private Health through contracts with health care providers or through Providers in Developing Countries: Serving the Public Interest?, ed. S. performance-related pay, may bring rewards if careful mon- Bennett, B. McPake, and A. Mills, 214­36. London: Zed Press. itoring is possible; however, evidence on the sustainability of Brown, K. 2002. "Improving Organisational and Individual Performance such arrangements is lacking, and effective monitoring may for Service Delivery: How Can Officials Become More Responsive to the Needs of the Poor?" Paper presented at the Department for require long-term external involvement. Evidence is needed International Development workshop on Improving Service Delivery on alternative approaches to improving performance. in Developing Countries, Eynsham Hall, Oxfordshire, U.K., November · Organizational reforms must keep the goal of improved 24­30. health outcomes, equity, and responsiveness in sight. Doing Castaño, R., R. Bitrán, and U. Giedion. 2004. Monitoring and so requires paying special attention to users' demands, to Evaluating Hospital Autonomization and Its Effects on Priority Services. Bethesda, MD: Partners for Health Reform Plus Project, Abt primary care and first-level hospitals, to quality of care, and Associates. to technical backup for disease control programs. de Savigny, D., E. Mwageni, C. Mayombana, H. Masanja, A. Minhaj, D. · Capacity-strengthening efforts in most settings must Momburi, and others. 2004. "Care-Seeking Patterns in Fatal Malaria." encompass action at all levels, from increasing leadership of Background paper prepared for the Institute of Medicine study on the Economics of Antimalarial Drugs, Washington, DC. the ministry of health at the national level through strength- Fenton, P. M., C. J. Whitty, and F. Reynolds. 2003. "Caesarean Section in ening support for peripheral levels. Malawi: Prospective Study of Early Maternal and Perinatal Mortality." British Medical Journal 327 (7415): 587­91. The current body of knowledge represents a largely ad hoc Goodman, C. 2004. "An Economic Analysis of the Retail Market for Fever and disjointed collection of facts, figures, and points of view. and Malaria Treatment in Rural Tanzania." Ph.D. thesis, University of London. Making confident recommendations relevant to strengthening Gonzalez, C. L. 1965. "Mass Campaigns and General Health Services." health system capacity is thus difficult. Although international Public Health Paper 29. Geneva: World Health Organization. financing is vital, countries need flexibility to develop solutions Gwatkin, D., S. Rutstein, K. Johnson, R. Pande, and A. Wagstaff. 2000. based on local assessments and experience and to progress at a Socioeconomic Differences in Health, Nutrition, and Population: 45 pace commensurate with their situations. Sustained investment Countries. Washington, DC: World Bank. in analytical and operational research capacity is needed as part Hanson, K., K. Ranson, V. Oliveira-Cruz, and A. Mills. 2003. "Expanding Access to Health Interventions: A Framework for Understanding the of program and systems support, to serve national priority set- Constraints to Scaling Up." Journal of International Development 15 ting and policy formulation. (1): 1­14. Strengthening Health Systems | 101 Hensher, M. 2001. "Financing the Health System through Efficiency Nunberg, B. 1999. "Rethinking Civil Service Reform." Poverty Reduction Gains." Background paper prepared for Working Group 2 of the and Economic Management Notes 31, World Bank, Washington, DC. Commission on Macroeconomics and Health, World Health http://www1.worldbank.org/ prem/PREMNotes/premnote31.pdf. Organization, Geneva. http://www.cmhealth. org/docs/wg3_paper2.pdf. Palmer, N., and A. Mills. 2003. "Classical versus Relational Approaches to Khaleghian, P., and M. Das Gupta. 2004. "Public Management and the Understanding Controls on a Contract with GPs in South Africa." Essential Public Health Functions." Working Paper 25, Disease Control Health Economics 12: 1005­20. Priorities Project, Bethesda, MD. Palmer, N., D. Mueller, L. Gilson, A. Mills, and A. Haines. 2004. "Health Levine, R. 2004. What's Worked? Accounting for Success in Global Health. Financing and Equity of Utilisation in Low-Income Settings: Is There Washington, DC: Center for Global Development. an Evidence Base?" Lancet 364 (9442): 1365­70. Levy, B. 2004. "Governance and Economic Development in Africa: Physicians for Human Rights. 2004. An Action to Prevent Brain Drain: Meeting the Challenge of Capacity Building." In Building State Building Equitable Health Systems in Africa. Boston: Physicians for Capacity in Africa: New Approaches, Emerging Lessons, ed. B. Levy and Human Rights. S. Kpundeh, 1­42. Washington, DC: World Bank Institute. Preker, A. S., and A. Harding. 2003. Innovations in Health Service Delivery. Litvack, J., and C. Bodart. 1993. "User Fees plus Quality Equals Improved Washington, DC: World Bank. Access to Health Care: Results of a Field Experiment in Cameroon." Ranson, K., K. Hanson, and A. Mills. 2003. "Constraints to Expanding Social Science and Medicine 37 (3): 369­83. Access to Health Interventions: An Empirical Analysis and Country Liu, X., and A. Mills. 2002. "Financing Reforms of Public Health Services Typology." Journal of International Development 15 (1): 15­40. in China: Lessons for Other Nations." Social Science and Medicine 54 Saltman, R. B., and O. Ferroussier-Davis. 2000. "On the Concept of (11): 1691­98. Stewardship in Health Policy." Bulletin of the World Health ------. 2003. "The Influence of Bonus Payments to Doctors on Hospital Organization 78 (6): 732­39. Revenue: Results of a Quasi-Experimental Study." Applied Health Schneider, H., and N. Palmer. 2002. "Getting to the Truth? Researching Economics and Health Policy 2 (2): 91­98. User Views of Primary Health Care." Health Policy and Planning 17: Marsh, V. M., W. M. Mutemi, A. Willetts, K. Bayah, S. Were, A. Ross, and 32­41. K. Marsh. 2004. "Improving Malaria Home Treatment by Training Schreuder, B., and C. Kostermans. 2001. "Global Health Strategies versus Drug Retailers in Rural Kenya." Tropical Medicine and International Local Primary Health Care Priorities: A Case Study of National Health 9 (4): 451­60. Immunization Days in Southern Africa." South African Medical Journal Martineau, T., and J. Buchan. 2000. "Human Resources and the Success of 91 (3): 249­54. Health Sector Reform." Human Resources Development Journal 4 (3): Segall, M. 2003. "District Health Systems in a Neoliberal World: A Review 174­83. of Five Key Policy Areas." International Journal of Health Planning and "Mexico 2004: Global Health Needs a New Research Agenda." 2004. Lancet Management 18 (S1): S5­26. 364: 1555­56. Soeters, R., and F. Griffiths. 2003."Improving Government Health Services Mills, A. 1998. "To Contract or Not to Contract? Issues for Low- and through Contract Management: A Case from Cambodia." Health Policy Middle-Income Countries." Health Policy and Planning 13 (1): 32­40. and Planning 18 (1): 74­83. ------. 2005. "Mass Campaigns versus General Health Services: What Southern Africa Stroke Prevention Initiative Project Team. 2004. Have We Learnt in 40 Years about Vertical versus Horizontal "Secondary Prevention of Stroke: Results from the Southern Africa Approaches." Bulletin of the World Health Organization 83 (4): 315­16. Stroke Prevention Initiative (SASPI) Study." Bulletin of the World Mills, A., S. Bennett, S. Russell, with N. Attanayake, C. Hongoro, V. E. Health Organization 82 (7): 503­8. Muraleedharan, and P. Smithson. 2001. The Challenge of Health Sector Travis, P., S. Bennett, A. Haines, T. Pang, Z. Bhutta, A. Hyder, and others. Reform: What Must Governments Do? Oxford, U.K.: Macmillan Press. 2004. "Overcoming Health Systems Constraints to Achieve the Mills, A., R. Brugha, K. Hanson, and B. McPake. 2002. "What Can Be Done Millennium Development Goals." Lancet 364: 900­6. about the Private Health Sector in Low-Income Countries?" Bulletin of Unger, J.-P., P. de Paepe, and A. Green. 2003. "A Code of Best Practice to the World Health Organization 80 (4): 325­30. Avoid Damaging Health Care Services in Developing Countries." Mills, A., N. Palmer, L. Gilson, D. McIntyre, H. Schneider, E. Sinanovic, and International Journal of Health Planning and Management 18 (S1): H. Wadee. 2004. "The Performance of Different Models of Primary S27­40. Care Provision in Southern Africa." Social Science and Medicine 59 (5): Victora, C., K. Hanson, J. Bryce, and J. P. Vaughan. 2004. "Achieving 931­43. Universal Coverage with Health Interventions." Lancet 364: 1541­48. Molyneux, D. 2004. "`Neglected' Disease but Unrecognised Successes: Walsh, J., and K. Warren. 1980."Selective Primary Health Care: An Interim Challenges and Opportunities for Infectious Disease Control." Lancet Strategy for Disease Control in Developing Countries." New England 364: 380­83. Journal of Medicine 301: 967­74. Molyneux, D., and V. Nantulya. 2004. "Linking Disease Control Walt, G. 2001. "Global Cooperation in Global Public Health." In Programmes in Rural Africa: A Pro-Poor Strategy to Reach Abuja International Public Health, ed. M. Merson, R. Black, and A. Mills, Targets and Millennium Development Goals." British Medical Journal 667­99. Gaithersburg, MD: Aspen Publishers. 328 (7448): 1129­32. Wibulpolprasert, S., and P. Pengpaiboon. 2003. "Integrated Strategies to Musgrove, P. 2004. "Compensatory Finance in Health: Geographic Equity Tackle the Inequitable Distribution of Doctors in Thailand: Four in a Federal System." In Health Economics in Development, ed. P. Decades of Experience." Human Resources for Health 1: 12. Musgrove, 133­42. Health, Nutrition, and Population Series. World Bank. 2004. The Millennium Development Goals for Health: Rising to Washington, DC: World Bank. the Challenges. Washington, DC: World Bank. Nabyonga, J., M. Desmet, H. Karamagi, P. Y. Kadama, F. G. Omaswa, and O. Walker. 2005. "Abolition of Cost-Sharing Is Pro-Poor: Evidence from Uganda." Health Policy and Planning 20 (2): 100­8. 102 | Disease Control Priorities in Developing Countries | Anne Mills, Fawzia Rasheed, and Stephen Tollman Chapter 4 Priorities for Global Research and Development of Interventions Barry R. Bloom, Catherine M. Michaud, John R. La Montagne, and Lone Simonsen It is a profound and necessary truth that the deep things in science are not found because they are useful; they are found because it was possible to find them. --J. Robert Oppenheimer In R. Rhodes, The Making of the Atomic Bomb NEEDS, CONTEXT, OPPORTUNITIES, the United States has defined the realm of global health AND MAJOR CHALLENGES research as "problems, issues, and concerns that transcend national boundaries and may best be addressed by sharing Half of the entire increase in human life expectancy--a crude knowledge and cooperative action." but easily defined measure of the health of populations-- An important corollary of this definition is that global realized over recorded history occurred in the 20th century. health research is derived from individuals and institutions From 1900 to 2000, life expectancy at birth increased from rather than from nation states. Thus, global health knowledge 48.0 to 77.1 years in the United States and from 48.0 to 77.7 should be available to everyone, not just to the country in years in the United Kingdom, gains of almost 30 years. which it is done or that sponsored it. As a result, knowledge However, improvements in life expectancy were not limited to derived from health research is a true public good, which by the industrial nations. For example, between 1900 and 1990, definition possesses the following two special properties life expectancy in India increased from 27 to 59 years, a gain of (Commission on Macroeconomics and Health 2001): 32 years (Fogel 2004). These increases are largely attributable to a better understanding of biology, medicine, and public · Nonexclusivity. Thus, when supplied it does not require pay- health--that is, to the benefits of research. ment to benefit individuals or groups (for example, the ben- Research is traditionally defined as the generation of new efit to the world community of eliminating smallpox). knowledge and the development of new and enabling tech- · Nonrivalry. Hence, the use of the benefits by an individual, nologies to identify or respond to major gaps in current group, or country will not diminish others' ability to bene- knowledge. Research includes the development of new fit from the same good or service.1 tools, methodologies, and strategies. The World Health Organization (WHO) and its Advisory Committee on Health Research have suggested two other defining aspects Investments in research have produced remarkable improve- of health research--namely, the verification of knowledge in ments in global health, especially over the past 20 years. different contexts and the creation and dissemination of prod- Immunization programs have led to unprecedented progress in ucts of knowledge. The Institute of Medicine (1997, page 1) in the fight against common childhood diseases (such as measles, 103 pertussis, poliomyelitis, and tetanus) and in the eradication of accurately, reliably, and cheaply. With these technologies, global smallpox. At the same time, vaccination programs have cat- research relationships that once would have been impossible alyzed the construction of a global infrastructure for epidemio- are now commonplace. For example, in a matter of weeks, logical monitoring and research, especially in the Americas and researchers in China could sequence the gene for the surface in Asia. Moreover, researchers are rapidly developing many pre- protein of the coronavirus associated with severe acute respira- ventive, diagnostic, and therapeutic tools, and the growing tory syndrome (SARS) and then produce the surface protein power of genomics and proteomics will accelerate the pace. as diagnostic antigen. Such speedy reaction would have been The 21st century will see a continuation of the inexorable inconceivable just a decade ago. trend toward the globalization of travel, trade, and communi- Daunting challenges remain, however, that health research cations. At one level, economic globalization has increased dis- alone is unlikely to solve. The context for health is very complex parities between countries in terms of gross domestic product and varies in different countries of the world (box 4.1). The per capita. On a population basis, however, economic gains by predictable outcome of current trends is an increase in the China, India, and other developing countries have made vast health and technology gaps between the rich and poor coun- numbers of people substantially wealthier than ever before tries. Tip O'Neill, the colorful speaker of the U.S. House of (Fischer 2003). Cell phones and radios are ubiquitous, even in Representatives for 10 years, often said that "all politics is the most remote parts of Africa and Asia, and the Internet has local." A provocative thesis we present here is that (a) all health permitted the transmission of data across long distances rapidly, care is national, and (b) all health research is global. Box 4.1 Context of Global Health Advances include the following: · Disparities have increased. The richest 20 percent of · the globalization of knowledge and the increased the world's population now accounts for 150 times the mobility of the world's population income of the poorest 20 percent. The ratio of the · the expansion of knowledge about disease problems in income of the top 20 percent to that of the poorest most of the developing world 20 percent rose from 30 to 1 in 1960 to 61 to 1 in 1991 · the remarkable progress achieved in the control of and to 78 to 1 in 1994. Evidence of global environmen- infectious diseases in most parts of the world tal degradation is apparent, especially in the developing · the worldwide penetration of new forms of world. For example, 45 percent of tropical rain forest communication has already been lost, at least 20 percent of current · the promise of new technologies in biomedical research species will be extinct by 2030 and 50 percent by the (for example, in the fields of genomics, transgenic end of the century, and half of China's and many other organisms, informatics, robotics, and nanotechnology) countries' cities already face water shortages. · the increasing flow of private resources devoted to · With global warming, temperatures will likely rise understanding health problems related to development. 1.0°C to 4.5°C this century, threatening coastal areas and changing patterns of vectorborne and epidemic The following concerns are pertinent: disease. · The world's population continues to grow, numbering · The pace of migration from rural to urban environ- 6.3 billion people in 2004, with 200,000 added each ments is speeding up, giving rise to more megacities. day. In at least 68 countries, more than 40 percent of · The period 1955­98 witnessed 31 civil and foreign the population is younger than 15, whereas in wealthy wars, 35 million displaced people and refugees, and countries, the proportion of elderly people in the 127 instances of state failures--ethnic wars, revolution- population is expanding quickly. ary wars, and disruptive regime changes--in 96 states. · Despite falling global poverty rates, progress is uneven: · Terrorism has become a global threat. 1.2 billion people still live on less than US$1 a day, and · Gender discrimination persists. 2.8 billion live on less than US$2 a day. One in six children is chronically hungry. Sources: King and Zeng 2001, 2002; Wilson 2003; World Bank 2004; World Revolution (http://www.worldrevolution.org). 104 | Disease Control Priorities in Developing Countries | Barry R. Bloom, Catherine M. Michaud, John R. La Montagne, and others Global Burden of Disease disease and major unipolar depressive disorders, for exam- Tables 4.1 and 4.2 summarize some of the key findings that are ple, but also these two chronic conditions already account most relevant to a discussion of global health research priori- for an increasing burden of disease and death in developing ties. The magnitude and distribution of the burden of disease countries. Demographic changes in many low- and middle- across different regions in 2001 reveals a great deal about income countries are driving the observed transition toward unmet research needs (Mathers and others 2003), in particular: patterns of disease previously seen only in the industrial countries. The incidence of both ischemic heart disease and · Communicable diseases and maternal, perinatal, and nutri- cerebrovascular disease increases rapidly with age; thus, tional conditions remain the major contributors to the bur- countries in which the proportion of elderly people in the den of disease in Sub-Saharan Africa and in parts of East population increases will also experience increases in the Asia and the Pacific. These regions differ significantly from relative importance of noncommunicable illness. Unlike all the other low- and middle-income regions and call for a communicable illnesses among younger people, chronic ill- unique set of priorities in relation to global health research. ness associated with aging cannot be entirely prevented, · Noncommunicable diseases are already the leading only delayed. These factors must be considered when setting contributors to the disease burden in all other low- and priorities for global health research. middle-income regions, which are undergoing rapid demo- graphic, economic, and epidemiological transitions. Not Children under five still account for an unnecessarily large only does the world face an epidemic of cardiovascular share of the disease burden in many low- and middle-income Table 4.1 Broad Patterns of the Disease Burden, by World Bank Region, 2001 East Asia Europe and Latin America Middle Sub- High- and the Central and the East and South Saharan income Category Pacific Asia Caribbean North Africa Asia Africa countries World Population (millions) 1,851 477 526 310 1,388 668 929 6,150 Communicable, maternal, 22.2 9.4 21.8 27.1 44.3 70.4 5.7 36.7 perinatal, and nutritional conditions (prevalence, percent) Noncommunicable diseases 65.8 76.4 65.0 59.3 44.4 21.2 86.7 52.6 (prevalence, percent) Injuries (prevalence, percent) 12.0 14.3 13.2 13.7 11.4 8.4 7.5 10.7 Source: Mathers and others 2003. Table 4.2 Leading Causes of the Disease Burden, by World Bank Region, 2001 East Asia Europe and Latin America Middle East Sub-Saharan High-income Rank and the Pacific Central Asia and the Caribbean and North Africa South Asia Africa countries 1 Cerebrovascular Ischemic heart Perinatal conditionsa Ischemic heart Perinatal HIV/AIDS Ischemic heart diseases disease disease conditionsa disease 2 Perinatal Cerebrovascular Unipolar depressive Perinatal conditionsa Lower respiratory Malaria Cerebrovascular conditionsa diseases disorders infections diseases 3 Chronic obstruc- Unipolar depressive Homicide and Traffic accidents Ischemic heart Lower Unipolar tive pulmonary disorders violence disease respiratory depressive disease infections disorders 4 Ischemic heart Self-inflicted injuries Ischemic heart Lower respiratory Diarrheal diseases Diarrheal Alzheimer's disease disease infections diseases disease and other dementias 5 Unipolar Chronic obstructive Cerebrovascular Diarrheal diseases Unipolar depressive Perinatal Tracheal and depressive pulmonary disease diseases disorders conditionsa lung cancer disorders Source: Mathers and others 2003. a. Perinatal conditions include low birthweight, birth asphyxia, and birth trauma. Priorities for Global Research and Development of Interventions | 105 regions where lower respiratory infections, diarrheal diseases, enormous new disease burdens in those countries, while at the and perinatal conditions persist. Because these diseases can same time providing early warning to industrial countries to largely be prevented through relatively low-cost interventions, stimulate new research on vaccines and drugs. Even though research into how best to implement these interventions and industry in developing countries may currently be devoting reduce the infectious disease burden in lower-income countries more effort to creating look-alike drugs, which are unlikely to remains a priority. add a great deal to the duration or quality of life, than to creat- ing drugs for major global killers, and even though market incentives for interventions in resource-poor countries are The 10:90 Issue lacking for most diseases, the solution is not to balkanize Efforts over the past two decades by the Commission on Health research and science, but to stimulate scientific capacity in all Research in Development, the WHO human reproduction countries. Local researchers and industries in developing coun- and tropical disease research programs, the WHO Ad Hoc tries might be able to create interventions that can find a niche Committee on Research Relating to Future Intervention in markets in developing countries or that public sector or Options, and--more recently--the Global Forum for Health public-private partnerships are prepared to support. Research have been largely responsible for the increasing focus on the role of health research in economic and social develop- ment. At a time when few health research resources were being Global Health Agendas devoted to the specific health problems of developing coun- Some major global health problems cannot be addressed with tries, these entities played a critical role in making the case the available knowledge and existing tools. Major challenges for that more should be done. The Global Forum for Health R&D remain to reduce the unfinished burden of infectious dis- Research took the most effective advocacy position, arguing eases; address the rapidly increasing burden of chronic diseases that 90 percent of the US$70 billion per year devoted to health in aging populations; and reduce the unnecessary burden caused research and development (R&D) was spent on diseases of the by injuries, casualties of war, and humanitarian emergencies. rich countries and only 10 percent was spent on the diseases uniquely afflicting poor countries. This advocacy has been The Unfinished Agenda of Infectious Diseases. In 1969, the effective and has galvanized global recognition that more U.S. surgeon general issued a now famous, if less than pre- research funding should be devoted to improving the health of scient, pronouncement: "The time has come to close the book the 85 percent of the world's population who live in developing on infectious diseases" (WHO 2000). In 2001, infectious dis- or transition countries. eases still accounted for 32 percent of the global burden of An absolute divergence in gross domestic product persists mortality and 37 percent of the global burden of disease. In between industrial and developing countries and, thus, what Sub-Saharan Africa, they are responsible for 68 percent of they can reasonably devote to research. Infectious diseases con- deaths. The HIV/AIDS epidemic continues to spread, affecting tinue to exact their highest tolls in the poorest countries, and large proportions of populations in Sub-Saharan Africa, but it new tools to prevent and treat HIV/AIDS, tuberculosis (TB), is at only an early stage in Asia, when effective prevention malaria, respiratory and diarrheal disease, SARS, influenza, and efforts could make a difference, as they did in Thailand. AIDS more exotic infections such as Ebola are urgently needed. is responsible for the decline in life expectancy to less than A longer-term view of global health problems recognizes the 40 years of age in five Sub-Saharan African countries. Yet recent increasing convergence of health problems, particularly chronic promising results of public health efforts in Brazil, Senegal, diseases and injuries. It is no longer true that research on car- Thailand, and Uganda demonstrate that HIV/AIDS can be pre- diovascular disease, diabetes, or depression, for example, is not vented and controlled on a nationwide scale. relevant to developing countries. Vast knowledge is available on Even though the public in industrial countries often seems how to prevent a major portion of heart disease, lung cancer, surprised by each new outbreak of infectious disease, the pat- type 2 diabetes, sexually transmitted infections, and injuries in tern of emerging infectious diseases worldwide is continuing the elderly, yet most countries do not implement that knowl- and, at the same time, is constantly changing. Since 1970, peo- edge effectively. Thus, more research is needed to successfully ple have been afflicted by 32 new diseases that had never transfer that knowledge from industrial to developing previously been reported in humans, such as hepatitis C, countries. For example, if monitored carefully, cost-effective, Legionnaire's disease, Ebola, Vibrio cholerae 0139 epidemic, community-based antihypertensive, antiretroviral, and antide- Nipah virus encephalitis, SARS, and the highly pathogenic avian pressive treatments could have an enormous effect in most influenza. The 1918 influenza epidemic killed 20 million to developing countries. 40 million people worldwide. Precisely when human-to-human Increased surveillance and diagnostic capacity for emerging transmission of the avian influenza viruses will occur is impos- infectious diseases in developing countries will prevent sible to predict, but it is likely to happen eventually. This 106 | Disease Control Priorities in Developing Countries | Barry R. Bloom, Catherine M. Michaud, John R. La Montagne, and others eventuality underscores the importance of encouraging system- the third largest contributor to the global burden of disease. atic collaboration on emerging and reemerging infections and Clearly public health sectors have a great deal to contribute in strengthening global surveillance and laboratory capability. A terms of reducing injuries from motor vehicle crashes, falls, syndromic approach to diagnosis and surveillance, as for the and workplace injuries. Less amenable to intervention by pub- identification of flaccid paralysis, which accelerated the elimi- lic health systems will be wars and humanitarian emergencies. nation of poliomyelitis in the Western hemisphere, may be cru- Obtaining accurate figures is difficult, but as Murray and others cial when laboratory diagnosis is not readily available. (2002) note, available statistics have greatly underestimated the Finally, the deliberate dissemination of anthrax spores in burden of war and civil strife on health systems. September and October 2001 in the United States has raised the specter of biological terrorism, either with pathogens natural to the environment that took years to eradicate, like smallpox, or The Crisis in Health Systems genetically engineered pathogens of unknown capability. Unprecedented advances in the development of health care Thus, the global infectious disease agenda remains unfin- technologies, drugs, vaccines, and new diagnostics, which hold ished. Given the continuing emergence of new infectious dis- the promise of healthier and longer lives for many, have pro- eases and the increasing resistance of microbial pathogens to found influences on health systems worldwide; in rich and existing drugs and of insect vectors to pesticides, as well as low poor countries alike, they raise expectations and demand for compliance with treatments, it is likely to remain so. health services along with difficult issues relating to access to information, costs, quality of care, equity, organization, and The Coming Epidemic: Chronic Diseases and Aging accountability. All systems are challenged by the need for Populations. In 1998, for the first time, chronic diseases con- quality improvement and self-learning. tributed more to the global burden of disease than infectious The overall cost of health care has increased so much that diseases, indicating the emergence of a convergence between fewer individuals can afford to pay for the best available care; the principal diseases of the developing countries and the thus, the financing of health systems has become central to industrial countries. Worldwide, cardiovascular disease is the national policy debates worldwide. Access and equity consider- major cause of mortality and morbidity (13.6 percent of total ations pose particularly daunting challenges in poor countries, disability-adjusted life years) followed by cancer (6.6 percent of where access to treatment may be a matter of life and death total disability-adjusted life years). Diabetes type 2 is increasing for entire populations. This situation now prevails in the Sub- in most countries of the world at an alarming rate. An unan- Saharan African countries, where the continuing spread of the ticipated finding from the global burden of disease analysis was HIV/AIDS epidemic has resulted in a sharp drop in life that psychiatric illness, particularly depression, is a major cause expectancy. of disability everywhere (Murray and Lopez 1996a, 1996b). Comparative analysis of health systems worldwide seeks to Depression is now the most important disability among understand the determinants of their performance--for women in the United States, and globally it is projected to be instance, financing, human resources, health information, the second largest contributor to the burden of disease by 2020. and quality of care--and to find ways to correct failures. The success of public health and childhood immunization Strengthening such research is one obvious way to tackle the in reducing the number of childhood deaths from infectious current crisis in health systems.Another less obvious but impor- disease is partially responsible for the increasing burden of tant implication of the current situation is the need to pursue chronic diseases. However, part of the increase is caused by the best possible science to develop new and better tools and the poor eating habits, lack of exercise, smoking, and other concomitant need to ensure the availability and affordability of unhealthy lifestyle choices that tend to increase with a nation's drugs and technologies where they are needed to address major income. Even though noncommunicable diseases associated health problems. with aging are increasingly contributing to the global burden of disease, the emergence of a highly virulent infectious disease pandemic could allow communicable illnesses to reassert their New Frontiers for R&D primacy. The extraordinary advances in science provide unprecedented opportunities for both industrial and developing regions. The The Unnecessary Epidemic: Injuries, Casualties of War, and following sections highlight promises as well as potential pit- Humanitarian Emergencies. Before the analysis of the global falls of frontier research. burden of disease, the contribution of injuries to the burden of disease and disability was unclear. The most rapidly rising cat- Genomics, Molecular Epidemiology, and Preventive egory of injuries is that resulting from motor vehicle crashes. If Medicine. Probably the most exciting area of biomedical present trends continue, by 2020 motor vehicle crashes will be research for at least the next decade derives from the Human Priorities for Global Research and Development of Interventions | 107 Genome Project and other efforts to sequence entire genomes to provide such knowledge on a scale that was inconceivable of mammals, birds, insects, and microbial pathogens. even five years ago. Examination of these genome sequences will allow investiga- The hope is that genomics and related biomedical research tors to define and understand intrinsic risks for disease as well on stem cells will give rise to new therapies for repairing and as interactions between genes and environmental threats. The remodeling tissue damaged by chronic disease, from heart dis- sequencing of the major microbial pathogens has given rise to ease to diabetes and chronic neurological diseases. The possi- molecular targets for new drugs against specific pathogens that bility of preventive treatment has now also arisen--that is, the are distinct from their host counterparts and unique antigenic identification of risks for chronic disease early in life and the fragments that may become effective components of new vac- implementation of preventive strategies--behavioral, nutri- cines. Researchers have sequenced the genomes of virtually all tional, or medical--to avert or overcome intrinsic risks and major viral, bacterial, and parasitic infectious disease agents thereby prevent disease. and placed the results in databases available to everyone, a Despite the optimism and enthusiasm, a darker side of the true public good (see The Institute of Genomic Research at Human Genome Project is emerging. Because individuals http://www.tigr.org). face different risks, the focus on "boutique medicine" will The availability of these genome sequences has catalyzed increase--that is, the focus on risks for individuals and the ambitious research efforts. For example, a project is under way development of niche interventions targeting those risks, to genetically engineer the Anopheles mosquito to render it rather than the focus on populations. Identification of those unable to transmit malaria. Even if this effort fails, knowledge intrinsic risks at birth, for example, will for some time be a lux- of the mosquito's genome has given new life to medical ento- ury available to better-off children in rich countries but not to mology and will likely help reduce vectorborne diseases in babies in poor countries or to poor populations of rich coun- other ways. In a second example, the growing number of avail- tries. Ultimately the Human Genome Project and the rapid able influenza virus sequences will greatly aid understanding of advances in biomedical research in the industrial world have the epidemiology and evolution of pandemic and interpan- the unintended potential to increase the gap between rich and demic influenza viruses and will be a powerful tool for guiding poor. If, however, most complex diseases have multigenic sus- vaccine strain selection. ceptibilities, the magic bullet approach of boutique medicine The genome project has already changed the understanding may not fulfill current expectations of rich or poor countries. of health and disease (box 4.2). Until now, epidemiology has New therapies, whether they arise from genomics or from dealt largely with external and environmental risks for disease. more traditional pharmacology, must be tested carefully to What the Human Genome Project offers is knowledge of the ensure that people in developing countries are not unfairly other side of the health equation--that is, the intrinsic risks for treated in clinical trials. Contract research organizations now disease. Previously undreamed of molecular and cellular tools carry out 60 percent of clinical trials. Many of these organiza- to explore gene expression and function are becoming available tions already test products in developing countries that will Box 4.2 Uncovering Individual Risks for Specific Diseases Genomic information makes possible predicting individ- a poor prognosis and are creating the first generation ual risks for certain diseases and to certain components of of drugs effective against mutated genes causing specific the environment. One level relates to polymorphisms in cancers. individual genes that represent intrinsic risks for certain The promise of the genome is first and foremost a conditions (for example, breast cancer). A second level greater knowledge about disease, risks for disease, and relates to differences in the expression patterns of multiple mechanisms of pathogenesis. The exploitation of knowl- genes on DNA chips that make it possible, for example, edge from the genome is just beginning, and practical to distinguish melanomas from lymphomas from colon ramifications and many effective products have yet to be cancers or stages within these cancers that no pathologist realized. Despite the hyperbole about its promise, the could duplicate for accuracy. Within patterns for breast genome does represent a new frontier, beyond random cancer or certain types of leukemia, experts can now dis- testing of compounds, for rational and evidence-based tinguish those likely to survive five years from those with design of effective interventions. Source: Authors. 108 | Disease Control Priorities in Developing Countries | Barry R. Bloom, Catherine M. Michaud, John R. La Montagne, and others have anticipated markets in rich countries but are unlikely, Many of the lessons of social epidemiology--and the flourishing should they be licensed, to be available or affordable to popu- world of advertising--indicate that most behaviors, including lations in developing countries. This practice is both an ethical risky or unhealthy behaviors, are socially patterned. Science has and a practical health problem. unfortunately not done a good job of learning how to change Finally, in countries where testing for genetic risks becomes social patterns. For example, merely targeting individuals at high available, the likelihood of risk adjustment--that is, the exclu- risk for HIV/AIDS without changing the social context that sion of people with some risks from insurance and discrimina- might reinforce stigmatization is not the best way to prevent dis- tion in relation to jobs, marriage, and housing--can be antici- ease.Indeed,in many developing countries that now provide free pated. In this information age, personal genetic information counseling, testing, and antiretroviral drugs for people with will certainly present an unprecedented challenge to privacy HIV/AIDS, the biggest barrier remains the social stigma of being and confidentiality. HIV positive.Health systems must widen their view beyond indi- vidual patients to target entire communities and the media to A Faint Hope: Population-Based Research. The focus of change unhealthy socially patterned behavior. In the United future research in the rich countries will likely be on individual States, epidemiological estimates indicate that 50 percent of the risks and on interventions tailored to those risks. Yet from the 2.3 million annual deaths are preventable or postponable. point of view of the world as a whole, the most effective inter- McGinnis and Foege (2004) find that in 2000, 19 percent of ventions are population-based interventions, such as vaccines, deaths were caused by tobacco, about 14 percent were attributa- insecticide-impregnated bednets, environmental modifica- ble to poor diet and lack of exercise, and about 12 percent to tions, antismoking campaigns, clean water, and safe sex. With injuries. One of the great challenges is to learn how to communi- knowledge derived from biomedical science and the Human cate what is known about the prevention of such conditions as Genome Project, it is hoped that some interventions will heart disease, obesity, and diabetes more effectively. emerge that do not require knowing any individuals' intrinsic Reliable and comparative analysis of health risks is key for genetic risks and that may apply to entire populations at risk. preventing disease and injury. A recently published study The hope is that they could be comparable to existing (Ezzati and others 2003) reports estimates of the disease bur- population-based interventions--for example, vaccines rec- den caused by the joint effect of 20 selected leading risk factors ommended for all children to prevent major infectious dis- in 14 subepidemiological regions of the world. In regions eases, treatment of schoolchildren once a year with ivermectin where high mortality persists, four risk factors--underweight to prevent onchocerciasis, and antismoking campaigns. in childhood, micronutrient deficiency, indoor smoke from In the rich countries, research has shown that aspirin and solid fuels, and tobacco--caused 35 to 42 percent of lower res- a combination of inexpensive antihypertensive drugs reduce piratory infections in 2000. In the same regions, the combined deaths from heart attacks by 30 percent and from strokes by risks of high blood pressure, high cholesterol, high body mass 50 percent. Even though they are off patent, these interventions index, low fruit and vegetable intake, and physical inactivity are currently not widely used in developing countries. These caused 82 to 89 percent of the burden of ischemic heart disease. findings are the products of basic research, but their effective Important gaps in scientific evidence about the effects of mul- use will depend on operational research. tiple risk factors and risk factor interactions persist and require further exploration (Ezzati and others 2003). In this context, The Next Frontier: Human Behavior and Social investigators should not underestimate social and behavioral Determinants of Disease. Another revolution in research is determinants of disease, including poverty, environment, emerging: understanding the functioning of the human brain culture, and so on. and, ultimately, human behavior. Biomarkers for neuropsychi- atric disease and environmental stresses are being sought, and with MRI and positron emission tomography technology, "Appropriate Science" for the Developing World researchers can see areas of the brain that are thinking, remem- Although much discussion about "appropriate technology" for bering, or enjoying music. Within the next 50 years, science will developing countries has taken place in recent decades, curi- have the technical ability to begin to untangle the processes of ously little discussion has occurred about appropriate science. thinking in molecular terms, with exciting or frightening pos- Much of the past debate assessed the imbalance of research sibilities to alter or affect them. Anticipating quantifiable bio- relevant to developing countries' health problems largely as a markers for stresses and psychopathology as well as objective function of the projected affordability of the products of the tools for measuring the effectiveness of new psychotropic inter- research--drugs, diagnostics, and new technologies developed ventions in changing behavior is not unreasonable. in the industrial countries--rather than considering the poten- Thefactors that lead people to engage in unhealthy or destruc- tial contributions that scientists from developing countries tive behaviors are more complex than simple individual choices. could make both to advancing science and to addressing their Priorities for Global Research and Development of Interventions | 109 countries' health problems. Although some technologies are of communication between individuals that makes trans- more or less appropriate to contexts in developing countries for mission of knowledge possible. One thinks of a few great reasons of cost, maintenance, or skill requirements, no limita- clinical teachers who simply "know" the diagnosis without tion exists on what science or knowledge is appropriate in laboratory tests, or health care professionals who can put developing countries. their colleagues in developing countries at ease and bring Some might argue that people in developing countries out the best in everyone being taught rather than being should restrict their research focus to diseases that principally condescending or patronizing. Tacit knowledge is intuitive, affect their countries. If that were generalized to all countries, breaks down barriers of culture or training, is highly moti- rich countries would not carry out research on tropical diseases, vating, and is often transformational in people's lives. and the developing countries would do little research on chronic diseases. This strategy would violate two fundamental princi- A few examples illustrate the importance of contextual ples of science. First, connectivity in science is unpredictable: knowledge. Many ideas and interventions are available, but research on one disease or problem often brings conceptual or knowledge on their effectiveness in different populations and technological advances that are vital to progress in others; on how to increase their usefulness is limited. The need to therefore, to the extent possible, every country should support define best practices in different circumstances is urgent in a relatively broad spectrum of research. Second, creative science relation to health. For example, data from the industrial coun- requires the freedom to pursue ideas. Progress in science is not tries indicate that providing a three-drug package containing fostered by restricting freedom of inquiry. There is every reason aspirin to people with hypertension as preventive treatment to believe that scientists in developing countries will create might be possible on a population-based model as well as by knowledge of value to diseases that primarily afflict people in individual physicians or medical personnel. However, Asians industrial countries, both because of the convergence of health are more predisposed to hemorrhagic strokes than Europeans; problems and because scientific knowledge is a public good. therefore, treatment with such a regimen in Asia might have a Epistemology is the formal study of knowledge, and theories significantly increased risk of adverse effects. of how knowledge is generated abound. One such theory par- In another example, antiretroviral drugs are responsible ticularly relevant in the context of health research holds that for the 50 percent decline in mortality from HIV/AIDS in three basic kinds of knowledge exist:2 the United States, and the Global Fund to Fight AIDS, Tuberculosis, and Malaria; bilateral agencies; and the pharma- · Public knowledge. This knowledge is generally published in ceutical industry are engaging in major efforts to make them the scientific literature, available in principle to all (with the available to resource-poor endemic countries. Despite encour- glaring exception of those who cannot afford the major sci- aging examples in Brazil and Haiti, it is unclear whether--as in entific journals). Because this knowledge is available to the DOTS (directly observed treatment short course), a supervised entire global scientific community, it is a true public good. method of administering drugs used for treating TB in Indeed, publication in such journals is the basis for most resource-poor countries--these drugs can be given safely and judgments of academic and scientific achievement and is a effectively by community-based treatment programs, be precondition for scientific support and advancement. appropriately monitored, and prevent the emergence of drug · Contextual knowledge. This kind of knowledge is absolutely resistance or toxicity and thus provide cures for a high percent- essential to bringing the fruits of public knowledge to a par- age of the patients in poor countries. However, if this method ticular country or people. It requires learning and experi- can be used, it will strengthen the fight against HIV/AIDS. ence and involves cultural, social, and economic knowledge Research on community-based programs for treating chil- of a place, without which effectively implementing public dren with epilepsy or adults with depression provides another knowledge or scientific discoveries in that context or evalu- example. The provision of ivermectin (Mectizan) to prevent ating the success of programs within national contexts is and treat onchocerciasis revealed that even making a drug to be often impossible. In this case, research may have to be car- taken only once a year available and providing it free of charge ried out in relation to how to implement interventions, as had an almost negligible effect initially, because in some areas WHO's Special Programme for Research and Training in of Sub-Saharan Africa an effective health delivery system was Tropical Diseases programs in leprosy and malaria have simply not available. It is to Merck's credit that the Mectizan done in the absence of full scientific evidence. As essential as program invested considerable resources to create a delivery it may be, the global scientific and academic communities and monitoring system that has moved onchocerciasis to the do not widely recognize or value contextual knowledge. category of diseases targeted for elimination as public health · Tacit knowledge. In contrast to public knowledge and con- problems by WHO. textual knowledge, tacit or intrinsic knowledge is impossible The flow of knowledge is not unidirectional. Reciprocity to write down or teach because it depends on a special kind between research in different fields and different countries is 110 | Disease Control Priorities in Developing Countries | Barry R. Bloom, Catherine M. Michaud, John R. La Montagne, and others vital for the expansion of knowledge, and the unique contribu- three research needs that had not previously been articulated as tions of developing countries to global health research are often essential to development. overlooked and not always appreciated. For example, DOTS · The first is a need for new knowledge through research to was initially developed in Tanzania, where researchers found develop new tools for addressing continually emerging that the best drug combination given with supervision, even global health problems. Some of this knowledge will be gen- though more costly, was both more effective in preventing eralizable, but much will be context specific and perhaps relapse and emerging drug resistance and more cost-effective country specific. than the cheapest combinations. Similarly, artemisinin, the · The second is the recognition that in many developing most rapidly acting drug for treating cerebral malaria, derives countries research capacity--that is, people with the train- from an ancient Chinese medicine, qinghaosu, and is now a ing to carry out surveillance and laboratory and operational major tool in the armamentarium of malaria treatments. research--is limited, indicating an enormous need for Research on isolated populations in developing countries can training. Career structures and incentives to retain trained further the understanding of some of the genetic determinants professionals in public health, medical sciences, and health of a variety of diseases, and transnational research on almost systems in developing countries are also needed. An enor- any disease has the potential to provide important insights into mous brain drain is under way for nursing and other health differences in risk factors in different contexts. professionals. The inducements to leave developing coun- Such reciprocity depends critically on the development of tries for higher salaries and better working conditions in the scientific capacity. In terms of resource allocation, research industrial world are compelling, even though many in the funders often appear to have overlooked the necessary connec- health field would prefer to help alleviate their own coun- tion between research and training the next generations of tries' health problems if it were feasible for them to do so. researchers. Scientific and health capacity building and training · The third is that all the key priorities depend on the are inseparable from research, yet funders seldom recognize the strengthening of institutions: universities, schools of public training aspect, and it is difficult to ensure that funding for health and medicine, centers for disease control, and training will be recognized as integral to research. research institutions for health policy and economics. As the report indicates, remarkably few high-level institutions for "Appropriate Technology" for the Developing World research and training in public health have been created in developing countries during the past 25 years. Little The development community has long debated the nature progress has been made since the mid 1990s. Thus, their of appropriate technology for resource-poor countries. large needs for human capacity as well as laboratory and Innumerable instances exist of high-tech biomedical equip- research infrastructure for public health are not surprising. ment standing unused in laboratories and hospitals through- For this situation to improve in a timely way, a new basis for out the developing world, serving as status symbols but not as cooperation in support of people and institutions must be tools to further knowledge or alleviate illness. However, the best forged between the developing and industrial countries. tools appropriate for learning from the research should be made available when the primary purpose of research is to Clearly governments should make greater commitments to acquire knowledge, particularly if human subjects are engaged health training and institutions. Whereas the international as volunteers in clinical studies to help develop that new knowl- community has focused primarily on access to drugs in edge. For example, researchers studying the effectiveness of resource-poor countries, new partnerships in research are clear- antiretroviral drugs in resource-poor countries should have ly needed. It is gratifying that programs that encourage and sup- access to technology that can measure CD4 cells, viral loads, port cross-national, North-South, and South-South scientific and antiviral drug resistance, which are critical for analyzing collaborations and institutional links have been increasing, and the drugs' effectiveness. Sophisticated technology may be vital the tremendous effect of the Bill & Melinda Gates Foundation to establish the scientific principle of effectiveness, thereby in supporting research has emphasized the value of public and enabling implementation of the most cost-effective treatment private commitments and partnerships in research. Regrettably, program in settings where the high-tech methodology may no concomitant commitment has been lacking on the part of longer be necessary on a large scale but may remain useful for governments and many foundations to support training in validating the effectiveness of lower-tech surrogate markers. research; career path opportunities; and institutions such as university science departments, medical schools, and schools of public health, all of which are critical for reducing the research Strengthening Capacity and Institutions and capacity gap between rich and poor countries. A 1996 WHO report (Ad Hoc Committee on Health Research No simple answer is available regarding the best ways to Relating to Future Intervention Options 1996) emphasizes ensure effective collaboration in relation to global heath. Global Priorities for Global Research and Development of Interventions | 111 Box 4.3 SARS and Influenza: A Paradigm Shift for Global Research Collaboration Outbreaks of emerging infectious diseases are by their unprecedented rapid response to the potentially devastat- nature unpredictable. They can be contained when they ing bird H5N1 influenza A. are detected early and the number of cases is small. When These examples represent an important paradigm shift they are not contained, they can have enormous human in global research collaboration in that they required and economic consequences. Economic losses attributed national surveillance at the epidemiological and laboratory to SARS, which infected 8,000 people and killed 774, have levels; unprecedented sharing of information at all levels of been estimated at US$30 million per day in Canada and a the health system; and close cooperation among clinicians, total of US$16 billion to US$30 billion in Asia. epidemiologists, and bench scientists, as well as those The global response to the SARS epidemic demon- involved in veterinary surveillance, for the rapid develop- strated the power of international collaboration under ment of effective intervention strategies. Integrated global leadership of WHO among public health professionals, responses raise difficult issues pertaining to information researchers, and institutions in several countries to halt sharing and ownership of specimens and reagents, which the progression of a new disease (La Montagne and others have profound implications for future global health R&D. 2004). Another example is influenza: an existing interna- They also underscore that, despite the political temptations tional network of influenza research sites, which is criti- of denial and the economic threats of epidemic disease, cal for defining the strains to be used each year for honest and accurate information is essential for early warn- immunization, was instrumental in developing an ing and for making effective health policy. Source: Authors. collaborations can be difficult,they are not inexpensive,and their Inherent Difficulties in Setting Priorities. The first part of successes are limited in number, but they can potentially have a this chapter underscored the immense scope of health prob- major effect (box 4.3). Questions arise about how to develop lems and the potential of global health research to make a economies of scale in R&D and institutional capabilities; difference. Given the complexity of the task and the multiple how many research centers are optimal in a developing country; participants involved in the process, defining priorities for the how much should be done by country partners and how much global health research agenda is daunting. in the developing countries; and what the roles of basic scientific Scientists tend to argue that more research is urgently need- versus epidemiological, clinical, and operational research should ed on the diseases they are studying. Their research may cer- be. The experience of working together in true partnerships tainly include worthwhile issues, but they may not be priorities appears to be generally rewarding for scientists in both industrial in the wider context of global health R&D. and developing countries and seems to be an effective way of Some hold the view that the choice of priorities should increasing research capacity. One set of lessons still to be learned begin with a statement defining topics that should not be is what the best forms of collaboration are: individual scientist, priorities--for example, the development of vaccines (such as institutional, transnational, or multinational. a leprosy or hookworm vaccine) when cost-effective treatments are available. Others strongly disagree, given the interconnect- PRIORITY SETTING edness and unpredictability of science. The failure of the U.S. "war on cancer" offers a useful cau- Setting priorities for R&D of interventions is both complex and tion on the limitations of rational planning of science. In the critical in the context of severely constrained resources. A sys- 1960s, a group of distinguished scientists developed a set of tematic approach that takes into account the disease burden as future research priorities for the National Cancer Program. well as scientific opportunities has been proposed to guide Despite the importance of the problem, the requisite scientific decisions. knowledge was not then available to develop the modern tools that have recently been successful in treating and preventing Approaches cancer. Planning for where the new innovations and discover- The challenge is to ensure that available resources are targeted ies will come from is hard, and planners have to be open to at major health problems. changing their priorities and incorporating new approaches. 112 | Disease Control Priorities in Developing Countries | Barry R. Bloom, Catherine M. Michaud, John R. La Montagne, and others A key challenge is the problematic nature of anticipating sci- ments of the burden of disease; developed countries' threat entific connections in advance. For example, the sequencing of assessments, for example, in relation to bioterrorism and epi- a mouse leukemia virus genome as part of the National Cancer demic potential; scientific or technical opportunities; advocacy; Program is what enabled scientists years later to classify HIV as political commitment; ethical considerations; and funding a related member of the retrovirus family. Indeed, who would availability. have predicted that research on the once arcane coronavirus Using a systematic and evidence-based approach to priority would become essential to control the spread of SARS? Or that setting, the WHO Ad Hoc Committee for Health Research the esoteric question of whether tumor cells extinguished dif- Relating to Future Intervention Options (1996) undertook the ferentiated functions of normal body cells would lead to the first broadly based, systematic effort to formulate "best buys" discovery of monoclonal antibodies? Or that the study of sex in for health R&D (table 4.3). The steps included assessments of bacteria would give rise to the entire genetic revolution of the the following: past half century? The need to recognize the unpredictability of science and the limitations of scientists at any time is best · size of the disease burden illustrated by Oppenheimer's statement at the beginning of this · reasons the disease burden persisted chapter. · adequacy of the current scientific knowledge base · cost-effectiveness of potential interventions and the prob- Systematic and Evidence-Based Approach to Priority Setting. ability of successful development of new tools The process of setting priorities for the global health research · adequacy of the current level of ongoing research and agenda is complex and includes accurate or perceived assess- funding. Table 4.3 R&D Best Buys Category Key R&D investments Maternal and child health Strategic research Understand the relative importance, in different environments, of increased nutrient intake and of control of infectious disease as a way to reduce malnutrition Package development Evaluate and refine the package for the integrated management of the sick child and evaluation Develop, evaluate, and refine the mother-baby package for pregnancy, delivery, and neonatal care Evaluate the implementation of a range of family planning packages offering a wide choice of methods New tools to improve Evaluate the efficacy and optimal dosage of candidate rotavirus vaccines in low-income countries package content Evaluate the efficacy of optimal dosage of candidate conjugate pneumococcal vaccines and the effectiveness of existing vaccine against influenza B in low-income countries Develop and evaluate ways to increase efficiency in the Expanded Program on Immunizations by simplifying delivery and maximizing use of opportunities for immunization Evaluate the promotion of insecticide-impregnated bednets for inclusion in a future healthy household package Develop new contraceptive methods, particularly to widen the choice of long-term but reversible methods, postcoital methods for regular and emergency use, and methods for men Microbial threats Strategic research Screen drugs on molecular targets predicted by the genome sequence of major pathogens Investigate influences on the spread of antimicrobial resistance and approaches to monitoring resistant strains with the aim of identifying ways of slowing their emergence Intervention development Develop an effective prophylaxis for TB--for example, depot (or long-acting), or a vaccine chemoprophylaxis Develop a malaria vaccine Develop an HIV vaccine Develop improved methods for the diagnosis, prevention, and treatment of sexually transmitted diseases, including vaginal microbicides Noncommunicable Establish a special program for research and training on noncommunicable diseases and healthy aging diseases and injuries Establish a special program or initiative for research, training, and capacity building on injuries Health policy Establish a special program for research and training on health systems and policy Source: Ad Hoc Committee on Health Research Relating to Future Intervention Options 1996. Priorities for Global Research and Development of Interventions | 113 This five-step approach has been influential. The Global Infectious Diseases of the U.S. National Institutes of Health and Forum for Health Research and the Special Programme for was conducted in May 2004 were leading scientists from low- Research and Training in Tropical Diseases have endorsed it and middle-income developing countries and worked in the and further developed it. The Global Forum's combined same region in which they held their citizenship. The survey approach matrix links the five steps with four actors or factors highlighted their views about key factors influencing research determining the health status (Global Forum for Health priority setting as well as major barriers that hampered Research 2002): stronger participation by scientists from developing countries in global health research. According to the survey, the most · individual, family, and community important factors determining research priorities were the · health ministry, health research institutions, and health magnitude of disease burdens and the needs of the industrial systems and services countries. Major barriers to the success of research collabora- · sectors other than health tion in global health were the lack of sustained funding; the dif- · central government macroeconomic policies. ficulty of linking research, programs, and policy; the weak research leadership; and the absence of a science culture The five steps also provide the basis for the strategic (Harley, Simonsen, and Breman 2004). emphases matrix for tropical diseases research (Remme and A more balanced participation of scientists from industrial others 2002). and developing countries, a better gender mix, and the inclu- The individual disease chapters in this volume used a sion of major stakeholders are essential to the successful devel- slightly modified version of the framework developed by the opment of a truly global health research agenda. The challenge WHO Ad Hoc Committee to identify gaps and guide the for- is to develop creative mechanisms for addressing current mulation of research priorities on the basis of the following shortcomings. premise: even though the current mix of available cost- The process for selecting the best research projects and pro- effective interventions averts a proportion of the burden of grams within each priority area is well established and is any particular disease and the remaining burden could be fur- grounded in scientific merit, based primarily on trust in peer ther reduced with improved application of existing technolo- review and expert judgment. Keeping this process independent gies to affected populations, a fraction of disease remains that from political pressures is extremely important. However, the cannot be averted. Two reasons account for this fact. First, the peer review process has limitations, including a natural conser- cost for extending the existing technology to the remainder of vatism and risk aversion by scientists, given the responsibility the population would be prohibitive. Second, the existing for the allocation of public funding, their often narrow base of interventions may simply not be sufficiently effective. These expertise in one discipline, and their specific cultural perspec- two categories define the magnitude of the need for new or tive. Alternative models of project selection from industry and better tools and, in essence, serve as a rationale and indicate other scientific, mission-oriented entities might offer interest- priorities for research. ing alternatives--for example, managerial systems or strategic A clear example is the case of HIV/AIDS. Neither behavioral planning processes, particularly for translating knowledge into interventions, such as exhortations for abstinence and fidelity successful interventions, an area that research is currently and the provision of condoms, nor antiretroviral therapy has emphasizing. stopped the global spread of HIV, which challenges the scien- Ethical considerations and pressures exerted by advocacy tific community to undertake more research on preventive vac- groups--such as public-private partnerships for targeted drug cines. The availability of highly active antiretroviral therapy or vaccine development, fresh looks at "orphan drug" legisla- challenges the research community to find ways of providing tion, patent rules ensuring financial returns to industry as well effective and life-saving treatment for HIV/AIDS patients in a as the affordability of new products in developing countries, manner that ensures proper use and compliance, averts the and commitment before their development by the public and development of drug resistance, and thereby becomes a finan- private sectors to subsidize their development or ensure mar- cially sustainable policy. kets for the products--are likely to counterbalance to some extent the lack of incentives for the pharmaceutical industry to Participants and Decision Makers. Two main concerns lie at develop drugs, diagnostics, and vaccines for which markets do the core of most discussions of the priority-setting processes: not exist or are not profitable. In addition, one might hope that the predominance of the industrial countries and the predom- the growing pharmaceutical and vaccine industry in develop- inance of the scientific community in formulating research ing countries might place a higher priority on addressing agendas. Two-thirds of respondents in a survey of researchers nationally and regionally important health problems than do that was funded by the National Institute of Allergy and multinational companies. 114 | Disease Control Priorities in Developing Countries | Barry R. Bloom, Catherine M. Michaud, John R. La Montagne, and others The share of total R&D funds allocated to major causes Since the mid 1970s, the WHO Special Programme for of the disease burden in developing countries remains Research and Training in Tropical Diseases and a few other insufficient. As a result, the availability of funding to support institutions have been key players in strengthening research global health R&D is ultimately the defining factor regarding and research capacity for tropical diseases that are endemic in the implementation of selected R&D priorities. Thus, the Bill & specific developing regions--African trypanosomiasis, Chagas Melinda Gates Foundation has become a major driving force in disease, dengue, leishmaniasis, leprosy, lymphatic filariasis, defining priorities for global health R&D through its support malaria, onchocerciasis, and schistosomiasis. As a result, effec- of promising public-private partnerships. The new US$200 mil- tive control measures are now available for Chagas disease, lep- lion it provided to finance the Grand Challenges in Global rosy, lymphatic filariasis, and onchocerciasis--but questions Health represents the newest large influx of funds in support of remain regarding effective implementation strategies. The global health research (see Foundation of National Institutes of other diseases still lack effective control measures and, thus, Health at http://www.grandchallengesgh.org). require further research to develop better tools and effective control strategies (http://www.who.int/tdr/grants/strategic- emphases/default.htm). Findings The process that led to the formulation of the Grand Research agendas proposed in the various chapters fall into Challenges in Global Health represents two important depar- three broad categories: tures from earlier approaches to priority setting. First, the announcement of the call for ideas in May 2003 had an · priorities that are already on the global health agenda unprecedented dissemination worldwide and resulted in over · important topics that are not yet on the global agenda, but 1,000 submissions from scientists and institutions in 75 coun- should be pursued tries. Second, the formulation of a grand challenge, described · promising research topics that are not yet priorities, but as "a call for specific scientific or technological innovation that should be pursued. would remove a critical barrier to solving an important health problem in the developing world with a high likelihood of Michaud and others (2005) provide a more exhaustive global impact and feasibility" (Varmus and others 2003) was account of the research priorities summarized here and recom- broad and had a clear goal. mended in the volume. The research agendas proposed in chapters 16, 18, and 21 are extensive and encompass research on basic epidemiology Priorities Already Part of the Global Health Agenda. and risk factors and the development of new or better drugs, Priorities that are already the most prominent part of the vaccines, diagnostics, and intervention methods. The fact that global health agenda relate almost exclusively to the unfinished these priorities do not represent a marked departure from agenda of infectious diseases and to the continuous threats previous research priorities for these conditions attests to the of emerging infectious diseases, including bioterrorism. The complexity of these diseases and their importance in the largest investments pertain to the development of new drugs poorest countries. They will require a broadly based and sus- and vaccines that are needed to reduce the burden of tained global research effort to overcome the rapid spread of HIV/AIDS, malaria, and TB; to the early detection and control antibiotic and insecticide resistance, limited human resources, of new highly pathogenic viral agents (for example, SARS); and poorly developed health systems that severely constrain and to the prevention and treatment of infectious diseases the health community's ability to reduce the burden of resulting from microbial terrorism (for instance, anthrax and disease. smallpox). In 2001, the National Institute of Allergy and Infectious Important Topics That Are Not Yet on the Global Research Diseases developed a global research plan for HIV/AIDS, Agenda but Should Be Pursued. Cardiovascular diseases, malaria, and TB. The plan outlines a comprehensive approach neuropsychiatric disorders, obesity, diabetes, and cancers are for fighting infectious diseases that involves building a sustain- causing a rapidly increasing share of the disease burden in all able research capability domestically and internationally and developing regions, with the exception of Sub-Saharan Africa; enhancing global partnerships. It comprises short-, medium-, however, they do not yet figure prominently on the global and long-term goals for research that "will lead to prevention health research agenda. The research priorities recommended and treatment strategies that are effective, feasible, and realistic independently by the authors of various chapters pertaining to for individual countries struggling with the burden of numer- major causes of noncommunicable diseases converge. Indeed, ous infectious diseases" (National Institute of Allergy and diet, lifestyle, obesity, tobacco, and alcohol are common Infectious Diseases 2001). risk factors for cardiovascular diseases, certain cancers, and Priorities for Global Research and Development of Interventions | 115 diabetes. These diseases and risk factors represent a cluster of Millennium Development Goals, particularly in resource-poor conditions that pose similar research challenges. countries with weak health systems. In this context, further The first important cross-cutting theme emerging from this research is critical to elucidate neglected areas of health system cluster of chapters is the issue of portability, or how to bring reforms, including the following (Mills 2004): knowledge and programs from one location and define how they can become best practices elsewhere. Cost-effective pre- · improving public service provision ventive strategies and therapeutic approaches to reduce the · enhancing human resources burden of cardiovascular diseases, cancer, diabetes, and mental · ensuring accountability for health outcomes, funds, and disorders have been developed and tested in industrial coun- medicines tries. Much of the extensive knowledge base accumulated in · ensuring a functioning central government industrial countries to prevent the development of cardiovas- · providing evidence for policy. cular diseases, diabetes, and cancers is likely to be relevant to developing countries, yet few epidemiological studies have Promising Research Topics Not Yet Global Priorities. Other quantified the impact of major risk factors for chronic diseases important research topics emerge from the various chapters in developing regions, and few trials have been conducted to that are not yet global priorities but that are nevertheless assess the effectiveness of different intervention strategies. worthwhile pursuing. Major themes pertain to the following: Research to explore the transferability of cost-effective inter- ventions from industrial to developing countries therefore · epidemiology of injuries and cost-effective interventions to figures prominently in several chapters. reduce the burden resulting from both intentional and The primary prevention for noncommunicable diseases unintentional injuries, particularly motor vehicle crashes in industrial countries rests on the reduction of major risk and road and vehicle safety factors--namely, diet, lifestyles, and tobacco and alcohol con- · major risk factors for disease in different contexts (for sumption. Research priorities include the development of epi- example, tobacco, obesity, physical activity) demiological databases and of intervention studies to identify · medical and surgical errors cost-effective strategies to reduce the prevalence of major risk · occupational and environmental health factors in different contexts in developing countries. The trans- · risk analysis and risk communication fer of personal and population-based interventions to reduce · delivery of care at different levels of the health system the risk of cardiovascular disease, which are based on decades of · performance of health systems research in the industrial countries, is particularly promising. · management of health research Research priorities include evaluating a range of intervention · reproductive and sexual health strategies,from simple dietary interventions to reduce the risk of · health effects of global warming. cardiovascular disease (for example, food supplementation with folic acid and linoleic acid and reduction in the salt, saturated The importance of strengthening the research agenda in fat, and trans fat content of processed foods), to the hypotheti- those and other areas and the resultant opportunities to make a cal "polypill," which would combine drugs to lower cholesterol, real difference have not been sufficiently recognized in the past. clotting, and blood pressure. Reducing the risk of cardiovascu- lar disease is particularly important for diabetes, which is itself an important risk factor for cardiovascular disease. KEY RECOMMENDATIONS The second theme pertains to lifelong medical management of chronic conditions that cannot be cured but could be im- The priority-setting process should focus initially on defining a proved through the development and testing of public health small number of key priorities that have a reasonable chance of prevention and treatment algorithms. This issue has been little succeeding and yielding cost-effective outcomes in resource- considered in past discussions of priorities for global health constrained environments and that are,thus,least likely to divert research but now appears to be reasonably cost-effective. limited resources from being more effectively directed else- Examples include unipolar depression, bipolar disorders, schiz- where. Five broad recommendations emerge from this chapter. ophrenia, epilepsy, diabetes, and secondary prevention of ischemic heart disease and stroke. The third theme pertains to crucial implementation Invest More Wisely in Health R&D research that combines operations research and health services The focus should be on how best to invest limited resources for and systems research. Such research is becoming central to health R&D. This approach raises hard questions about select- ensuring the success of the rapid scaling-up of cost-effective ing priorities and the extent to which the burden of disease and interventions that is required to meet the health targets of the scientific opportunity to play a role. A telling example of the 116 | Disease Control Priorities in Developing Countries | Barry R. Bloom, Catherine M. Michaud, John R. La Montagne, and others dilemma may be a vaccine against bird flu. Because bird flu, The shift in thinking in relation to the convergence of health A (H5N1) or other strains, is not yet a major human disease, burdens and research opportunities in both industrial and the setting of priorities by the disease burden criterion would developing countries has far-reaching implications for the for- not accord a bird flu vaccine high priority. But knowing that mulation of research priorities. In addition to emphasizing the between 20 to 40 million people died in the 1918 influenza epi- commonality of health problems, it also emphasizes the impor- demic, and that with a transmission time of 1.5 days, there tance of stronger global research collaboration in tackling would be few public health measures other than a vaccine that major health problems and underscores the need for much would make a difference in preventing a pandemic, developing stronger public-private partnerships to ensure that affordable and testing multiple candidate vaccines should be an urgent drugs and vaccines will be developed and made available in research priority. resource-constrained environments. Despite the paucity of past analysis of the relationship between the cost of research in an area and its success in im- proving the level of health, the amount invested is unlikely to Maximize the Potential of Information Technology have been the most important determinant of success. The No advances in science have more potential for improving What Works Working Group has developed 17 case studies of health globally than the information and communication sci- success stories, all of which were supported by public finance ences. At the scientific level, the ability to handle increasingly mostly in resource-constrained settings (Levine and What massive amounts of data, whether from genetics, epidemiolo- Works Working Group 2004). A review of lessons learned from gy, or clinical trials, offers the opportunity to mine the world of the 20 biggest research successes in improving health in low- and knowledge in ways that could not be contemplated a decade middle-income countries would be a worthwhile undertaking. ago. Knowledge can be transferred instantaneously through the Internet; through access to open databases; and through the new public libraries of science and medicine, such as the U.S. Shift the Paradigm for Priority Setting National Library of Medicine PubMed Central. With informa- The paradigm shift from dividing the world's health problems tion technology, procedures can be put in place to minimize into those of the industrial countries and those of the devel- medical and pharmaceutical errors and to provide greater oping countries toward creating a better understanding of the accounting for medical costs and outcomes. Finally, research commonality of health problems between the industrial and with partners in many parts of the world can now be carried most developing countries lies at the core of priority setting out in real or in lag time, as in the case of clinical research on for global health research. Implicit in this shift is the recogni- malaria (Royall and others 2004). The tools, hardware, and tion that the health problems of Sub-Saharan Africa are software for this informatics revolution must be made available urgent and require special emphasis on the devastating bur- as widely as possible to universities and health systems in devel- den of infectious diseases, particularly HIV/AIDS, and on the oping countries. need to develop effective infrastructure for health. In time, the expectation is that the health needs of most Sub-Saharan African countries will similarly converge with those of other Increase Global Research Capacity regions of the world and that knowledge developed in these Research capacity continues to limit the successful implemen- regions will be transferable and helpful to accelerating devel- tation of those interventions most needed to improve health in opment there. resource-constrained environments. The number of people As sociologists have long recognized, scientific and medical trained to carry out the surveillance and the laboratory and technology diffuses from the industrial to the developing operational research that are so essential to the succes- countries, and it will, in the short term at least, increase the sful implementation of cost-effective interventions remains disparities between rich and poor countries--and perhaps to woefully inadequate. Redressing this limitation is a daunting a comparable extent between the rich and the poor within task that will require substantial financial investment and countries--because more affluent and better-educated popu- creative approaches to create conditions that will reverse the lations tend to have greater access to new technologies. brain drain and strengthen academic and research institutions However, even though disparities in health between affluent in developing countries. urban dwellers and poor rural populations in China and India have increased over the past decade, the overall quality of health of the entire population has increased during the same Create a Global Health Architecture period. The hope is that information derived from research Health is not the sole provenance of the health sector, and yet will be a great leveler over time and will contribute to reduc- there is no forum or architecture for coordinating the increas- ing global inequities. ingly important multisectoral interactions to improve health. Priorities for Global Research and Development of Interventions | 117 Cardiovascular and pulmonary disease in Europe and the and Infectious Diseases, U.S. National Institutes of Health, U.S. United States are increasingly determined by China's energy Department of Health and Human Services, Bethesda, MD. sector, and global warming is impacted by the policy of the U.S. Institute of Medicine. 1997. America's Vital Interest in Global Health: Protecting Our People, Enhancing Our Economy, and Advancing Our President. Health is critically affected by education, energy, International Interests. Washington DC: National Academy Press. transport, finance, trade, immigration, communication, and the King, G., and L. Zeng. 2001. "Improving Forecasts of State Failure." World environment. Major health problems will be most successfully Politics 53 (July): 623­58. addressed if partnerships can be developed between sectors, ------. 2002. Consolidated State Failure Events, 1955­2001. College Park, governments, NGOs, business and industry, and academe. MD: University of Maryland. La Montagne, J. R., L. Simonsen, R. J. Taylor, and J. Turnbull. 2004. "SARS Research Working Group Co-Chairs--Severe Acute Respiratory Support Freedom of Scientific Inquiry Syndrome: Developing a Research Response." Journal of Infectious Diseases 189 (4): 634­41. No country has a monopoly on ideas, and every country has Levine, R., and What Works Working Group with Molly Kinder. 2004. something important to contribute to knowledge about health. Millions Saved: Proven Successes in Global Health. Washington, DC: The universality of science requires that scientists everywhere Center for Global Development. strive for the highest level of rigor and quality and that every Mathers, C., A. Lopez, C. Stein, D. Ma Fat, R. Chalapati, M. Inoue, and country have some sustainable level of scientific research and others. 2005. "Death and Disease Burden by Cause: Global Burden of Disease Estimates for 2001 by World Bank Country Groups." Disease problem-solving capacity. Encouraging and supporting scien- Control Priorities Project Working Paper No. 18. Fogarty International tists with the ability and passion to contribute to knowledge Center and National Institutes of Health, Washington, DC. about health, globally or locally, must become one of the key McGinnis, J. M., and W. H. Foege. 2004. "The Immediate versus the aims of the global health and development agendas. Important." Journal of the American Medical Association 291 (10): 1263­64. Michaud, C. M., P. Maslen, S. Sahid-Salles, and J. Breman. 2005. "Synthesis ACKNOWLEDGMENT of Priorities for Global Research and Development." Disease Control Priorities Project Working Paper, Fogarty International Center and This chapter is dedicated to the memory of John La Montagne, National Institutes of Health, Washington, DC. who was a tireless supporter of health research into problems of Mills, A. 2004. "Missing Areas in Health Sector Reform." Unpublished paper, London School of Hygiene and Tropical Medicine. developing countries, a good friend, and an inspiration to us all. Murray, C. J., G. King, A. D. Lopez, N. Tomijima, and E. G. Krug. 2002. "Armed Conflict as a Public Health Problem." British Medical Journal 324 (7333): 346­49. NOTES Murray, C. J., and A. D. Lopez, eds. 1996a. The Global Burden of Disease: A 1. Obviously, nonrivalry does not pertain to knowledge that is propri- Comprehensive Assessment of Mortality and Disability from Diseases, etary, as in the pharmaceutical industry, although the system of patents Injuries, and Risk Factors in 1990 and Projected to 2020. Vol. 1 of Global was created to make such enabling knowledge available to all by providing Burden of Disease and Injury. Cambridge, MA: Harvard School of a limited monopoly for its exploitation by discoverers or inventors. Public Health for the World Health Organization and the World Bank. 2. For this formulation, we are indebted to Suwit Wibulpolprasert, ------. 1996b. Global Health Statistics: A Compendium of Incidence, deputy permanent secretary of the Ministry of Health, Thailand. Prevalence, and Mortality for over 200 Conditions. Vol. 2 of Global Burden of Disease and Injury. Cambridge, MA: Harvard School of Public Health for the World Health Organization and the World Bank. REFERENCES National Institute of Allergy and Infectious Diseases. 2001. "NIAID Global Health Research Plan for HIV/AIDS, Malaria, and Tuberculosis." Ad Hoc Committee on Health Research Relating to Future Intervention National Institute of Allergy and Infectious Diseases, National Options. 1996. Investing in Health Research and Development. Geneva: Institutes of Health, U.S. Department of Health and Human Services, World Health Organization. Bethesda, MD. Commission on Macroeconomics and Health. 2001. Macroeconomics and Remme, J. H., E. Blas, L. Chitsulo, P. M. Desjeux, H. D. Engers, T. P. Kanyok, Health. Geneva: World Health Organization. and others. 2002. "Strategic Emphases for Tropical Diseases Research: Ezzati, M., S. V. Hoorn, A. Rodgers, A. D. Lopez, C. D. Mathers, and A TDR Perspective." Trends in Parasitology 18 (10): 421­26. C. J. Murray. 2003. "Estimates of Global and Regional Potential Health Royall, J., M. Bennett, I. van Schayk, and M. Alilio. 2004. "Tying up Lions: Gains from Reducing Multiple Major Risk Factors." Lancet 362 (9380): Multilateral Initiative on Malaria Communications: The First Chapter 271­80. of a Malaria Research Network in Africa." American Journal of Tropical Fischer, S. 2003. "Socialist Economy Reform: Lessons of the First Three Medicine and Hygiene 71 (2 Suppl.): 259­67. Years." American Economic Review 93 (May): 390­95. Varmus, H., R. Klausner, E. Zerhouni, T.Acharya,A. S. Daar, and P.A. Singer. Fogel, R. W. 2004. The Escape from Hunger and Premature Death, 2003."Grand Challenges in Global Health." Science 302 (5644): 398­99. 1700­2100. Cambridge, U.K.: Cambridge University Press. Wilson, E. O. 2003. The Future of Life. New York: Vintage Books. Global Forum for Health Research. 2002. The 10/90 Report on Health WHO (World Health Organization). 2000. Report on Global Surveillance Research 2001­2002. Geneva: Global Forum for Health Research. of Epidemic-Prone Infectious Diseases. WHO/CDS/CSR/ISR/2000.1. Harley, L., L. Simonsen, and J. Breman. 2004. "Perceptions of Health Geneva: WHO. Research Priority Setting and Barriers: A Survey of International World Bank. 2004. World Development Indicators. Washington, DC: World Health Researchers." Unpublished paper, National Institute of Allergy Bank. 118 | Disease Control Priorities in Developing Countries | Barry R. Bloom, Catherine M. Michaud, John R. La Montagne, and others Chapter 5 Science and Technology for Disease Control: Past, Present, and Future David Weatherall, Brian Greenwood, Heng Leng Chee, and Prawase Wasi As we move into the new millennium it is becoming increas- medical sciences. The elegant anatomical dissections of Andreas ingly clear that the biomedical sciences are entering the most Vesalius swept away centuries of misconceptions about the rela- exciting phase of their development. Paradoxically, medical tionship between structure and function of the human body; practice is also passing through a phase of increasing uncer- the work of Isaac Newton, Robert Boyle, and Robert Hooke tainty, in both industrial and developing countries. Industrial disposed of the basic Aristotelian elements of earth, air, fire, and countries have not been able to solve the problem of the spiral- water; and Hooke, through his development of the microscope, ing costs of health care resulting from technological develop- showed a hitherto invisible world to explore. In 1628, William ment, public expectations, and--in particular--the rapidly Harvey described the circulation of the blood, a discovery that, increasing size of their elderly populations. The people of many because it was based on careful experiments and measurement, developing countries are still living in dire poverty with dys- signaled the beginnings of modern scientific medicine. functional health care systems and extremely limited access to After steady progress during the 18th century, the biological basic medical care. and medical sciences began to advance at a remarkable rate Against this complex background, this chapter examines the during the 19th century, which saw the genuine beginnings role of science and technology for disease control in the past of modern scientific medicine. Charles Darwin changed the and present and assesses the potential of the remarkable devel- whole course of biological thinking, and Gregor Mendel laid opments in the basic biomedical sciences for global health care. the ground for the new science of genetics, which was used later to describe how Darwinian evolution came about. Louis Pasteur and Robert Koch founded modern microbiology, and MEDICINE BEFORE THE 20TH CENTURY Claude Bernard and his followers enunciated the seminal prin- ciple of the constancy of the internal environment of the body, From the earliest documentary evidence surviving from the a notion that profoundly influenced the development of phys- ancient civilizations of Babylonia, China, Egypt, and India, it is iology and biochemistry. With the birth of cell theory, modern clear that longevity, disease, and death are among humanity's pathology was established. These advances in the biological oldest preoccupations. From ancient times to the Renaissance, sciences were accompanied by practical developments at the knowledge of the living world changed little, the distinction bedside, including the invention of the stethoscope and an between animate and inanimate objects was blurred, and spec- instrument for measuring blood pressure, the first use of x-rays, ulations about living things were based on prevailing ideas the development of anesthesia, and early attempts at the about the nature of matter. classification of psychiatric disease as well as a more humane Advances in science and philosophy throughout the 16th approach to its management. The early development of the use and 17th centuries led to equally momentous changes in of statistics for analyzing data obtained in medical practice also 119 occurred in the 19th century, and the slow evolution of public Epidemiology and Public Health health and preventive medicine began. Modern epidemiology came into its own after World War II, Significant advances in public health occurred on both sides when increasingly sophisticated statistical methods were first of the Atlantic. After the cholera epidemics of the mid 19th applied to the study of noninfectious disease to analyze the pat- century, public health boards were established in many terns and associations of diseases in large populations. The European and American cities. The Public Health Act, passed emergence of clinical epidemiology marked one of the most in the United Kingdom in 1848, provided for the improvement important successes of the medical sciences in the 20th century. of streets, construction of drains and sewers, collection of Up to the 1950s, conditions such as heart attacks, stroke, refuse, and procurement of clean domestic water supplies. cancer, and diabetes were bundled together as degenerative dis- Equally important, the first attempts were made to record basic orders, implying that they might be the natural result of wear health statistics. For example, the first recorded figures for the and tear and the inevitable consequence of aging. However, United States showed that life expectancy at birth for those who information about their frequency and distribution, plus, in lived in Massachusetts in 1870 was 43 years; the number of particular, the speed with which their frequency increased in deaths per 1,000 live births in the same population was 188. At association with environmental change, provided excellent evi- the same time, because it was becoming increasingly clear that dence that many of them have a major environmental compo- communicable diseases were greatly depleting the workforce nent. For example, death certificate rates for cancers of the required to generate the potential rewards of colonization, con- stomach and lung rose so sharply between 1950 and 1973 that siderable efforts were channeled into controlling infectious dis- major environmental factors must have been at work generat- eases, particularly hookworm and malaria, in many countries ing these diseases in different populations. under colonial domination. The first major success of clinical epidemiology was the However, until the 19th century, curative medical tech- demonstration of the relationship between cigarette smoking nology had little effect on the health of society, and many of the and lung cancer by Austin Bradford Hill and Richard Doll in improvements over the centuries resulted from higher stan- the United Kingdom. This work was later replicated in many dards of living, improved nutrition, better hygiene, and other studies, currently, tobacco is estimated to cause about 8.8 per- environmental modifications. The groundwork was laid for a cent of deaths (4.9 million) and 4.1 percent of disability- dramatic change during the second half of the 20th century, adjusted life years (59.1 million) (WHO 2002c). Despite this although considerable controversy remains over how much we information,the tobacco epidemic continues,with at least 1 mil- owe to the effect of scientific medicine and how much to con- lion more deaths attributable to tobacco in 2000 than in 1990, tinued improvements in our environment (Porter 1997). mainly in developing countries. This balance between the potential of the basic biological The application of epidemiological approaches to the study sciences and simpler public health measures for affecting the of large populations over a long period has provided further health of our societies in both industrial and developing coun- invaluable information about environmental factors and dis- tries remains controversial and is one of the major issues to be ease. One of the most thorough--involving the follow-up of faced by those who plan the development of health care ser- more than 50,000 males in Framingham, Massachusetts-- vices for the future. showed unequivocally that a number of factors seem to be linked with the likelihood of developing heart disease (Castelli and Anderson 1986). Such work led to the concept of risk SCIENCE, TECHNOLOGY, AND MEDICINE factors, among them smoking, diet (especially the intake of ani- IN THE 20TH CENTURY mal fats), blood cholesterol levels, obesity, lack of exercise, and elevated blood pressure. The appreciation by epidemiologists Although rapid gains in life expectancy followed social change that focusing attention on interventions against low risk factors and public health measures, progress in the other medical sci- that involve large numbers of people, as opposed to focusing ences was slow during the first half of the 20th century, possi- on the small number of people at high risk, was an important bly because of the debilitating effect of two major world wars. advance. Later, it led to the definition of how important envi- The position changed dramatically after World War II, a time ronmental agents may interact with one another--the that many still believe was the period of major achievement in increased risk of death from tuberculosis in smokers in India, the biomedical sciences for improving the health of society. for example. This section outlines some of these developments and the effect A substantial amount of work has gone into identifying risk they have had on medical practice in both industrial and devel- factors for other diseases, such as hypertension, obesity and oping countries. More extensive treatments of this topic are its accompaniments, and other forms of cancer. Risk factors available in several monographs (Cooter and Pickstone 2000; defined in this way, and from similar analyses of the pathological Porter 1997; Weatherall 1995). 120 | Disease Control Priorities in Developing Countries | David Weatherall, Brian Greenwood, Heng Leng Chee, and others role of environmental agents such as unsafe water, poor sanita- updating results have made it possible to provide broad-scale tion and hygiene, pollution, and others, form the basis of The analyses combining the results of many different trials. World Health Report 2002 (WHO 2002c), which sets out a pro- Although meta-analysis has its problems--notably the lack of gram for controlling disease globally by reducing 10 conditions: publication of negative trial data--and although many poten- underweight status; unsafe sex; high blood pressure; tobacco tial sources of bias exist in the reporting of clinical trials, these consumption; alcohol consumption; unsafe water, sanitation, difficulties are gradually being addressed (Egger, Davey-Smith, and hygiene; iron deficiency; indoor smoke from solid fuels; and Altman 2001). high cholesterol; and obesity. These conditions are calculated to More recent developments in this field come under the gen- account for more than one-third of all deaths worldwide. eral heading of evidence-based medicine (EBM) (Sackett and The epidemiological approach has its limitations, however. others 1996). Although it is self-evident that the medical pro- Where risk factors seem likely to be heterogeneous or of only fession should base its work on the best available evidence, the limited importance, even studies involving large populations rise of EBM as a way of thinking has been a valuable addition continue to give equivocal or contradictory results. to the development of good clinical practice over the years. Furthermore, a major lack of understanding, on the part not It covers certain skills that are not always self-evident, just of the general public but also of those who administer including finding and appraising evidence and, particularly, health services, still exists about the precise meaning and inter- implementation--that is, actually getting research into pretation of risk. The confusing messages have led to a certain practice. Its principles are equally germane to industrial and amount of public cynicism about risk factors, thus diminishing developing countries, and the skills required, particularly the effect of information about those risk factors that have been numerical, will have to become part of the education of physi- established on a solid basis. Why so many people in both indus- cians of the future. To this end, the EBM Toolbox was estab- trial and developing countries ignore risk factors that are based lished (Web site: http://www.ish.ox.ac.uk/ebh.html). However, on solid data is still not clear; much remains to be learned evidence for best practice obtained from large clinical trials about social, cultural, psychological, and ethnic differences may not always apply to particular patients; obtaining a balance with respect to education about important risk factors for dis- between better EBM and the kind of individualized patient care ease. Finally, little work has been done regarding the perception that forms the basis for good clinical practice will be a major of risk factors in the developing countries (WHO 2002c). challenge for medical education. A more recent development in the field of clinical epidemiology--one that may have major implications for Partial Control of Infectious Disease. The control of commu- developing countries--stems from the work of Barker (2001) nicable disease has been the major advance of the 20th century and his colleagues, who obtained evidence suggesting that in scientific medicine. It reflects the combination of improved death rates from cardiovascular disease fell progressively with environmental conditions and public health together with the increasing birthweight, head circumference, and other meas- development of immunization, antimicrobial chemotherapy, ures of increased development at birth. Further work has and the increasing ability to identify new pathogenic organ- suggested that the development of obesity and type 2 diabetes, isms. Currently, live or killed viral or bacterial vaccines--or which constitute part of the metabolic syndrome, is also those based on bacterial polysaccharides or bacterial toxoids-- associated with low birthweight. The notion that early fetal are licensed for the control of 29 common communicable dis- development may have important consequences for disease in eases worldwide. The highlight of the field was the eradication later life is still under evaluation, but its implications, particu- of smallpox by 1977. The next target of the World Health larly for developing countries, may be far reaching. Organization (WHO) is the global eradication of poliomyelitis. The other major development that arose from the applica- In 1998, the disease was endemic in more than 125 countries. tion of statistics to medical research was the development of After a resurgence in 2002, when the number of cases rose to the randomized controlled trial. The principles of numerically 1,918, the numbers dropped again in 2003 to 748; by March based experimental design were set out in the 1920s by the 2004, only 32 cases had been confirmed (Roberts 2004). geneticist Ronald Fisher and applied with increasing success The Expanded Program on Immunization (EPI), launched after World War II, starting with the work of Hill, Doll, and in 1974, which has been taken up by many countries with slight Cochrane (see Chalmers 1993; Doll 1985). Variations on this modification, includes Bacillus Calmette-Guérin (BCG) and theme have become central to every aspect of clinical research oral polio vaccine at birth; diphtheria, tetanus, and pertussis at involving the assessment of different forms of treatment. More 6, 10, and 14 weeks; measles; and, where relevant, yellow fever recently, this approach has been extended to provide broad- at 9 months. Hepatitis B is added at different times in different scale research syntheses to help inform health care and communities. By 1998, hepatitis B vaccine had been incorpo- research. Increasing the numbers of patients involved in trials rated into the national programs of 90 countries, but an esti- and applying meta-analysis and electronic technology for mated 70 percent of the world's hepatitis B carriers still live in Science and Technology for Disease Control: Past, Present, and Future | 121 countries without programs (Nossal 1999). Indeed, among developing countries by the use of single antimicrobial agents 12 million childhood deaths analyzed in 1998, almost 4 million when combinations would have been less likely to produce were the result of diseases for which adequate vaccines are resistant strains. Finally, public health measures have been available (WHO 2002a). hampered by the rapid movement of populations and by war, The development of sulfonamides and penicillin in the famine, and similar social disruptions in developing countries. period preceding World War II was followed by a remarkable In short, the war against communicable disease is far from period of progress in the discovery of antimicrobial agents over. effective against bacteria, fungi, viruses, protozoa, and helminths. Overall, knowledge of the pharmacological mode of Pathogenesis, Control, and Management of Non- action of these agents is best established for antibacterial and communicable Disease. The second half of the 20th century antiviral drugs. Antibacterial agents may affect cell wall or pro- also yielded major advances in understanding pathophysiology tein synthesis, nucleic acid formation, or critical metabolic and in managing many common noncommunicable diseases. pathways. Because viruses live and replicate in host cells, antivi- This phase of development of the medical sciences has been ral chemotherapy has presented a much greater challenge. characterized by a remarkable increase in the acquisition of However, particularly with the challenge posed by HIV/AIDS, knowledge about the biochemical and physiological basis of a wide range of antiviral agents has been developed, most of disease, information that, combined with some remarkable which are nucleoside analogues, nucleoside or nonnucleoside developments in the pharmaceutical industry, has led to a situ- reverse-transcriptase inhibitors, or protease inhibitors. ation in which few noncommunicable diseases exist for which Essentially, those agents interfere with critical self-copying or there is no treatment and many, although not curable, can be assembly functions of viruses or retroviruses. Knowledge of controlled over long periods of time. the modes of action of antifungal and antiparasitic agents is Many of these advances have stemmed from medical increasing as well. research rather than improved environmental conditions. In Resistance to antimicrobial agents has been recognized since 1980, Beeson published an analysis of the changes that the introduction of effective antibiotics; within a few years, occurred in the management of important diseases between penicillin-resistant strains of Staphylococcus aureus became the years 1927 and 1975, based on a comparison of methods for widespread and penicillin-susceptible strains are now very treating these conditions in the 1st and 14th editions of a lead- uncommon (Finch and Williams 1999). At least in part caused ing American medical textbook. He found that of 181 condi- by the indiscriminate use of antibiotics in medical practice, tions for which little effective prevention or treatment had animal husbandry, and agriculture, multiple-antibiotic- existed in 1927, at least 50 had been managed satisfactorily by resistant bacteria are now widespread. Resistance to antiviral 1975. Furthermore, most of these advances seem to have agents is also occurring with increasing frequency (Perrin and stemmed from the fruits of basic and clinical research directed Telenti 1998), and drug resistance to malaria has gradually at the understanding of disease mechanisms (Beeson 1980; increased in frequency and distribution across continents Comroe and Dripps 1976). (Noedl, Wongsrichanalai, and Wernsdorfer 2003). The critical Modern cardiology is a good example of the evolution of issue of drug resistance to infectious agents is covered in detail scientific medicine. The major technical advances leading to a in chapter 55. better appreciation of the physiology and pathology of the In summary, although the 20th century witnessed remark- heart and circulation included studies of its electrical activity able advances in the control of communicable disease, the cur- by electrocardiography; the ability to catheterize both sides of rent position is uncertain. The emergence of new infectious the heart; the development of echocardiography; and, more agents, as evidenced by the severe acute respiratory syndrome recently, the development of sophisticated ways of visualizing (SARS) epidemic in 2002, is a reminder of the constant danger the heart by computerized axial tomography, nuclear magnetic posed by the appearance of novel organisms; more than 30 new resonance, and isotope scanning. These valuable tools and the infective agents have been identified since 1970. Effective vac- development of specialized units to use them have led to a cines have not yet been developed for some of the most com- much better understanding of the physiology of the failing mon infections--notably tuberculosis, malaria, and HIV--and heart and of the effects of coronary artery disease and have rev- rapidly increasing populations of organisms are resistant to olutionized the management of congenital heart disease. Those antibacterial and antiviral agents. Furthermore, development advances have been backed by the development of effective of new antibiotics and effective antiviral agents with which to drugs for the management of heart disease, including diuretics, control such agents has declined. The indiscriminate use of beta-blockers, a wide variety of antihypertensive agents, antibiotics, both in the community and in the hospital popula- calcium-channel blockers, and anticoagulants. tions of the industrial countries, has encouraged the emergence By the late 1960s, surgical techniques were developed to of resistance, a phenomenon exacerbated in some of the relieve obstruction of the coronary arteries. Coronary bypass 122 | Disease Control Priorities in Developing Countries | David Weatherall, Brian Greenwood, Heng Leng Chee, and others surgery and, later, balloon angioplasty became major tools. such as the H2-receptor antagonists and a wide range of drugs Progress also occurred in treatment of abnormalities of cardiac directed at bronchospasm. There have been some surprises-- rhythm, both pharmacologically and by the implantation of the discovery that peptic ulceration is almost certainly caused artificial pacemakers. More recently, the development of by a bacterial agent has transformed the management of this microelectronic circuits has made it possible to construct disease, dramatically reducing the frequency of surgical inter- implantable pacemakers. Following the success of renal trans- vention. Neurology has benefited greatly from modern diag- plantation, cardiac transplantation and, later, heart and lung nostic tools, while psychiatry, though little has been learned transplantation also became feasible. about the cause of the major psychoses, has also benefited Much of this work has been backed up by large-scale con- enormously from the development of drugs for the control of trolled clinical trials. These studies, for example, showed that both schizophrenia and the depressive disorders and from the early use of clot-dissolving drugs together with aspirin had the emergence of cognitive-behavior therapy and dynamic a major effect on reducing the likelihood of recurrences after psychotherapy. an episode of myocardial infarction (figure 5.1). The large The second half of the 20th century has witnessed major number of trials and observational studies of the effects of progress in the diagnosis and management of cancer (reviewed coronary bypass surgery and dilatation of the coronary arteries by Souhami and others 2001). Again, this progress has followed with balloons have given somewhat mixed results, although from more sophisticated diagnostic technology combined with overall little doubt exists that, at least in some forms of coro- improvements in radiotherapy and the development of power- nary artery disease, surgery is able to reduce pain from angina ful anticancer drugs. This approach has led to remarkable and probably prolong life. Similar positive results have been improvements in the outlook for particular cancers, including obtained in trials that set out to evaluate the effect of the con- childhood leukemia, some forms of lymphoma, testicular trol of hypertension (Warrell and others 2003). tumors, and--more recently--tumors of the breast. Progress The management of other chronic diseases, notably those of in managing other cancers has been slower and reflects the the gastrointestinal tract, lung, and blood has followed along results of more accurate staging and assessment of the extent similar lines. Advances in the understanding of their patho- and spread of the tumor; the management of many common physiology, combined with advances in analysis at the struc- cancers still remains unsatisfactory, however. Similarly, tural and biochemical levels, have enabled many of these although much progress has been made toward the prevention diseases to be managed much more effectively. The pharma- of common cancers--cervix and breast, for example--by pop- ceutical industry has helped enormously by developing agents ulation screening programs, the cost-effectiveness of screening for other common cancers--prostate, for example--remains controversial. Percentage dead Many aspects of maternal and child health have improved Routine hospital care alone significantly. A better understanding of the physiology and dis- 13% dead (568/4300) orders of pregnancy together with improved prenatal care and Aspirin only obstetric skills has led to a steady reduction in maternal mor- 10 tality. In an industrial country, few children now die of child- Streptokinase only hood infection; the major pediatric problems are genetic and congenital disorders, which account for about 40 percent of admissions in pediatric wards, and behavioral problems Routine care combination of 5 both Streptokinase and Aspirin (Scriver and others 1973). Until the advent of the molecular 8% dead (343/4292) era, little progress was made toward an understanding of the cause of these conditions. It is now known that a considerable proportion of cases of mental retardation result from definable chromosomal abnormalities or monogenic diseases, although 0 0 1 2 3 4 5 at least 30 percent of cases remain unexplained. Major Weeks from starting treatment improvements have occurred in the surgical management of Source: ISIS-2 Collaborative Group 1988 (with permission). congenital malformation, but only limited progress has been made toward the treatment of genetic disease. Although a few Figure 5.1 Effects of a One-hour Streptokinase Infusion Together factors, such as parental age and folate deficiency, have been with Aspirin for One Month on the 35-Day Mortality in the Second incriminated, little is known about the reasons for the occur- International Study of Infarct Survival Trial among 17,187 Patients rence of congenital abnormalities. with Acute Myocardial Infarction Who Would Not Normally Have Received Streptokinase or Aspirin, Divided at Random into Four In summary, the development of scientific medical prac- Groups to Receive Aspirin Only, Streptokinase Only, Both, or Neither tice in the 20th century led to a much greater understanding Science and Technology for Disease Control: Past, Present, and Future | 123 of deranged physiology and has enabled many of the com- such diseases may occur much more frequently and be more mon killers in Western society to be controlled, though few difficult to control. to be cured. However, although epidemiological studies of Partly because of advances in scientific medicine, industrial these conditions have defined a number of risk factors and countries have to face another large drain on health resources although a great deal is understood about the pathophysiol- in the new millennium (Olshansky, Carnes, and Cassel 1990). ogy of established disease, a major gap remains in our In the United Kingdom, for example, between 1981 and 1989, knowledge about how environmental factors actually cause the number of people ages 75 to 84 rose by 16 percent, and that these diseases at the cellular and molecular levels (Weatherall of people age 85 and over by 39 percent; the current population 1995). of males age 85 or over is expected to reach nearly 0.5 million by 2026, at which time close to 1 million females will be in this Consequences of the Demographic and Epidemiological age group. Those figures reflect the situation for many indus- Transitions of the 20th Century. The period of development trial countries, and a similar trend will occur in every country of modern scientific medicine has been accompanied by major that passes through the epidemiological transition. Although demographic change (Chen 1996; Feachem and others 1992). data about the quality of life of the aged are limited, studies The results of increasing urbanization, war and political unrest, such as the 1986 General Household Survey in the United famine, massive population movements, and similar issues States indicate that restricted activity per year among people must have had a major effect on the health of communities over the age of 65 was 43 days in men and 53 days in women; during the 20th century, but there has been a steady fall in those data say little about the loneliness and isolation of old childhood mortality throughout the New World, Europe, the age. It is estimated that 20 percent of all people over age 80 will Middle East, the Indian subcontinent, and many parts of Asia suffer from some degree of dementia, a loss of intellectual during this period, although unfortunately there has been function sufficient to render it impossible for them to care for much less progress in many parts of Sub-Saharan Africa. themselves. Scientific medicine in the 20th century has pro- Although much of the improvement can be ascribed to vided highly effective technology for partially correcting the improving public health and social conditions, the advent of diseases of aging while, at the same time, making little progress scientific medicine--particularly the control of many infec- toward understanding the biological basis of the aging process. tious diseases of childhood--seems likely to be playing an Furthermore, the problems of aging and its effect on health increasingly important part in this epidemiological transition. care have received little attention from the international public Although surveys of the health of adults in the developing health community; these problems are not restricted to indus- world carried out in the 1980s suggested that many people trial countries but are becoming increasingly important in between the ages of 20 and 50 were still suffering mainly from middle-income and, to a lesser extent, some low-income diseases of poverty, many countries have now gone through an countries. epidemiological transition such that the global pattern of dis- Although dire poverty is self-evident as one of the major ease will change dramatically by 2020, with cardiorespiratory causes of ill health in developing countries, this phenomenon is disease, depression, and the results of accidents replacing com- emphatically not confined to those populations. For example, municable disease as their major health problems. in the United Kingdom, where health care is available to all Countries undergoing the epidemiological transition are through a government health service, a major discrepancy in increasingly caught between the two worlds of malnutrition morbidity and mortality exists between different social classes and infectious disease on the one hand and the diseases of (Black 1980). Clearly this phenomenon is not related to the industrial countries, particularly cardiac disease, obesity, and accessibility of care, and more detailed analyses indicate that it diabetes, on the other. The increasing epidemic of tobacco- cannot be ascribed wholly to different exposure to risk factors. related diseases in developing countries exacerbates this prob- Undoubtedly social strain, isolation, mild depression, and lack lem. The global epidemic of obesity and type 2 diabetes is a of social support play a role. However, the reasons for these prime example of this problem (Alberti 2001). An estimated important discrepancies, which occur in every industrial coun- 150 million people are affected with diabetes worldwide, and try, remain unclear. that number is expected to double by 2025. Furthermore, diabetes is associated with greatly increased risk of cardiovas- Economic Consequences of High-Technology Medicine cular disease and hypertension; in some developing countries The current high-technology medical practice based on mod- the rate of stroke is already four to five times that in industrial ern scientific medicine must steadily increase health expendi- countries. These frightening figures raise the questions tures. Regardless of the mechanisms for the provision of health whether, when developing countries have gone through the care, its spiraling costs caused by ever more sophisticated epidemiological transition, they may face the same pattern of technology and the ability to control most chronic illnesses, diseases that are affecting industrial countries and whether combined with greater public awareness and demand for 124 | Disease Control Priorities in Developing Countries | David Weatherall, Brian Greenwood, Heng Leng Chee, and others medical care, are resulting in a situation in which most indus- tal medical practice. The potential of this approach has been trial countries are finding it impossible to control the costs of discussed in detail recently (WHO 2002c). Although the claims providing health care services. for the benefits of reducing either single or multiple risk factors The U.K. National Health Service (NHS) offers an interest- are impressive, no way exists of knowing to what extent they ing example of the steady switch to high-technology hospital are attainable. Furthermore, if, as seems likely, they will reduce practice since its inception 50 years ago (Webster 1998). Over morbidity and mortality in middle life, what of later? The that period, the NHS's overall expenditure on health has WHO report admits that it has ignored the problem of com- increased fivefold, even though health expenditure in the peting risks--that is, somebody saved from a stroke in 2001 is United Kingdom absorbs a smaller proportion of gross domes- then "available" to die from other diseases in ensuing years. tic product than in many neighboring European countries. At Solid information about the role of risk factors exists only for a the start of the NHS, 48,000 doctors were practicing in the limited number of noncommunicable diseases; little is known United Kingdom; by 1995 there were 106,845, of whom 61,050 about musculoskeletal disease, the major psychoses, dementia, were in hospital practice and 34,594 in general (primary care) and many other major causes of morbidity and mortality. practice. Although the number of hospital beds halved over the The problems of health care systems and improving per- first 50 years of the NHS, the throughput of the hospital service formance in health care delivery have been reviewed in World increased from 3 million to 10 million inpatients per year, over Health Report 2000--Health Systems: Improving Performance a time when the general population growth was only 19 per- (WHO 2000). Relating different systems of health care to out- cent. Similarly, outpatient activity doubled, and total outpatient comes is extremely complex, but this report emphasizes the visits grew from 26 million to 40 million. Because many indus- critical nature of research directed at health care delivery. As a trial countries do not have the kind of primary care referral pro- response to the spiraling costs of health care, many govern- gram that is traditional in the United Kingdom, this large skew ments are introducing repeated reforms of their health care toward hospital medicine seems likely to be even greater. programs without pilot studies or any other scientific indica- The same trends are clearly shown in countries such as tion for their likely success. This vital area of medical research Malaysia, which have been rapidly passing through the epi- has tended to be neglected in many countries over the later demiological transition and in which health care is provided on years of the 20th century. a mixed public-private basis. In Malaysia, hospitalization rates have steadily increased since the 1970s, reflecting that use is slowly outstripping population growth. The number of private Summary of Scientific Medicine in the 20th Century hospitals and institutions rose phenomenally--more than The two major achievements of scientific medicine in the 20th 300 percent--in the same period. In 1996, the second National century--the development of clinical epidemiology and the Health and Morbidity Survey in Malaysia showed that the partial control of infectious disease--have made only a limited median charge per day in private hospitals was 100 times contribution to the health of developing countries. Although in higher than that in Ministry of Health hospitals. Those figures part this limited effect is simply a reflection of poverty and dys- reflect, at least in part, the acquisition of expensive medical functional health care systems, it is not the whole story. As technology that in some cases has led to inefficient use of soci- exemplified by the fact that of 1,233 new drugs that were mar- etal resources. As in many countries, the Malaysian government keted between 1975 and 1999, only 13 were approved specifi- has now established a Health Technology Assessment Unit to cally for tropical diseases, the problem goes much deeper, provide a mechanism for evaluating the cost-effectiveness of reflecting neglect by industrial countries of the specific medical new technology. problems of developing countries. Those brief examples of the effect of high-technology prac- For those countries that have gone through the epidemio- tice against completely different backgrounds of the provision logical transition and for industrial countries, the central prob- of health care reflect the emerging pattern of medical practice lem is quite different. Although the application of public health in the 20th century. In particular, they emphasize how the rapid measures for the control of risk factors appears to have made a developments in high-technology medical practice and the major effect on the frequency of some major killers, those gains huge costs that have accrued may have dwarfed expenditure on have been balanced by an increase in the frequency of other preventive medicine, certainly in some industrial countries and common chronic diseases and the problems of an increasingly others that have gone through the epidemiological transition. elderly population. At the same time, remarkable developments A central question for medical research and health care in scientific medicine have allowed industrial countries to planning is whether the reduction in exposure to risk factors develop an increasingly effective high-technology, patch-up that is the current top priority for the control of common dis- form of medical practice. None of these countries has worked eases in both industrial and developing countries will have a out a way to control the spiraling costs of health care, and major effect on this continuing rise of high-technology hospi- because of their increasing aged populations, little sign exists Science and Technology for Disease Control: Past, Present, and Future | 125 that things will improve. Although some of the diseases that between different socioeconomic groups in many industrial produce this enormous burden may be at least partially pre- countries and to define the most effective approaches to edu- ventable by the more effective control of risk factors, to what cating the public about the whole concept of risk and what is extent such control will be achievable is unclear, and for many meant by risk factors. In addition, a great deal more work is diseases these factors have not been identified. In short, scien- required on mechanisms for assessing overall performance of tific medicine in the 20th century, for all its successes, has left a health care systems. major gap in the understanding of the pathogenesis of disease The third priority must be to focus research on the impor- between the action of environmental risk factors and the basic tant diseases that the biomedical sciences have yet to control, disease processes that follow from exposure to them and that including common communicable diseases such as malaria, produce the now well-defined deranged physiology that char- AIDS, and tuberculosis; cardiovascular disease; many forms of acterizes them. cancer; all varieties of diabetes; musculoskeletal disease; the These problems are reflected, at least in some countries, by major psychoses; and the dementias. Of equal importance is increasing public disillusion with conventional medical prac- gaining a better understanding of both the biology and patho- tice that is rooted in the belief that if modern medicine could physiology of aging, together with trying to define its social and control infectious diseases, then it would be equally effective cultural aspects. in managing the more chronic diseases that took their place. In the fields of child and maternal health, the requirements When this improvement did not happen--and when a mood for research differ widely in industrial and developing coun- of increasing frustration about what medicine could achieve tries. Industrial countries need more research into the mecha- had developed--a natural move occurred toward trying to find nisms of congenital malformation and the better control and an alternative answer to these problems. Hence, many countries treatment of monogenic disease and behavioral disorders of have seen a major migration toward complementary medicine. childhood. In developing countries, both child and maternal It is against this rather uncertain background that the role of health pose different problems, mainly relating to health edu- science and technology for medical care in the future has to be cation and the control of communicable disease and nutrition. examined. In many developing countries, some of the common mono- genic diseases, notably the hemoglobin disorders, also require urgent attention. SCIENCE, TECHNOLOGY, AND MEDICINE In short, our priorities for health care research come under IN THE FUTURE two main heads: first, apply knowledge that we already have more effectively; second, apply a multidisciplinary attack on Before considering the remarkable potential of recent develop- diseases about which we have little or no understanding. These ments in basic biological research for improvements in health issues are developed further in chapter 4. care, we must define priorities for their application. New Technologies Priorities for Biomedical Research in the Future The sections that follow briefly outline some examples of the In the setting of priorities for biomedical research in the future, new technologies that should help achieve these aims. the central objective is to restore the balance of research between industrial and developing countries so that a far Genomics, Proteomics, and Cell Biology. Without question greater proportion is directed at the needs of the latter. In the the fields of molecular and cell biology were the major devel- 1990s, it was estimated that even though 85 percent of the opments in the biological sciences in the second half of the global burden of disability and premature mortality occurs in 20th century. The announcement of the partial completion of the developing world, less than 4 percent of global research the human genome project in 2001 was accompanied by claims funding was devoted to communicable, maternal, perinatal, that knowledge gained from this field would revolutionize and nutritional disorders that constitute the major burden of medical practice over the next 20 years. After further reflection, disease in developing countries (WHO 2002b). some doubts have been raised about this claim, not in the least The second priority is to analyze in much more detail meth- the time involved; nevertheless, considerable reason for opti- ods of delivery of those aspects of health care that have already mism still exists. Although the majority of common diseases been shown to be both clinically effective and cost-effective. clearly do not result from the dysfunction of a single gene, most It is vital that the delivery of health care be based on well- diseases can ultimately be defined at the biochemical level; designed, evidence-based pilot studies rather than on current because genes regulate an organism's biochemical pathways, fashion or political guesswork. It is essential to understand why their study must ultimately tell us a great deal about pathologi- there are such wide discrepancies in morbidity and mortality cal mechanisms. 126 | Disease Control Priorities in Developing Countries | David Weatherall, Brian Greenwood, Heng Leng Chee, and others The genome project is not restricted to the human genome Expected % of births with thalassemia major but encompasses many infectious agents, animals that are 100 extremely valuable models of human disease, disease vectors, 90 and a wide variety of plants. However, obtaining a complete 80 nucleotide sequence is one thing; working out the regulation and function of all the genes that it contains and how they 70 United Kingdom interact with each other at the level of cells and complete 60 Greece organisms presents a much greater challenge. The human 50 genome, for example, will require the identification and deter- Sardinia 40 mination of the function of the protein products of 25,000 30 Cyprus genes (proteomics) and the mechanisms whereby genes are maintained in active or inactive states during development 20 Ferrara (methylomics). It will also involve the exploration of the roles of 10 the family of regulatory ribonucleic acid (RNA) molecules that 0 have been discovered recently (Mattick 2003). All this informa- 1970 1972 1974 1976 1978 1980 1982 1984 tion will have to be integrated by developments in information Year technology and systems biology. These tasks may take the rest Source: Modified from Modell and Bulyzhenkov (1998, 244). of this century to carry out. In the process, however, valuable Figure 5.2 Decline in Serious Forms of Thalassemia in Different fallout from this field is likely to occur for a wide variety of Populations after the Initiation of Prenatal Diagnosis in 1972 medical applications. Many of these are outlined in a recent Following the Development of North-South Partnerships. WHO report, Genomics and World Health 2002 (WHO 2002a). The first applications of DNA technology in clinical practice were for isolating the genes for monogenic diseases. Either by using the candidate gene approach or by using DNA markers find the most effective and economic approach. Recombinant for linkage studies, researchers have defined the genes for many DNA technology was used years ago to produce pure antigens monogenic diseases. This information is being used in clinical of hepatitis B in other organisms for the development of safe practice for carrier detection, for prenatal diagnosis, and for vaccines. More recently, and with knowledge obtained from the defining of the mechanisms of phenotypic variability. It has various genome projects, interest has centered on the utility of been particularly successful in the case of the commonest DNA itself as a vaccine antigen. This interest is based on the monogenic diseases, the inherited disorders of hemoglobin, chance observation that the direct injection of DNA into mam- which affect hundreds of thousands of children in develop- malian cells could induce them to manufacture--that is, to ing countries (Weatherall and Clegg 2001a, 2001b). Through express--the protein encoded by a particular gene that had North-South collaborations, it has been possible to set up been injected. Early experiences have been disappointing, but screening and prenatal diagnosis programs for these conditions a variety of techniques are being developed to improve the in many countries, resulting in a marked decline in their fre- antigens of potential DNA-based vaccines. quency, particularly in Mediterranean populations (figure 5.2). The clinical applications of genomics for the control of Gene therapy, that is, the specific correction of monogenic dis- communicable disease are not restricted to infective agents. eases, has been fraught with difficulties, but these are slowly Recently, the mosquito genome was sequenced, leading to the being overcome and this approach seems likely to be successful notion that it may be possible to genetically engineer disease for at least some genetic diseases in the future. vectors to make them unable to transmit particular organisms From the global perspective, one of the most exciting (Land 2003). A great deal is also being learned about genetic prospects for the medical applications of DNA technology is in resistance to particular infections in human beings (Weatherall the field of communicable disease. Remarkable progress has and Clegg 2002), information that will become increasingly been made in sequencing the genomes of bacteria, viruses, and important when potential vaccines go to trial in populations other infective agents, and it will not be long before the genome with a high frequency of genetically resistant individuals. sequence of most of the major infectious agents is available. The other extremely important application of DNA tech- Information obtained in this way should provide opportunities nology for the control of communicable disease--one of par- for the development of new forms of chemotherapy (Joët and ticular importance to developing countries--is its increasing others 2003) and will be a major aid to vaccine development place in diagnostics. Rapid diagnostic methods are being devel- (Letvin, Bloom, and Hoffman 2001). In the latter case, DNA oped that are based on the polymerase chain reaction (PCR) technology will be combined with studies of the basic immune technique to identify pathogen sequences in blood or tissues. mechanisms involved in individual infections in an attempt to These approaches are being further refined for identifying Science and Technology for Disease Control: Past, Present, and Future | 127 Box 5.1 Chronic Myeloid Leukemia: The Path from Basic Science to the Clinic 1960 An abnormal chromosome, named the Philadelphia results in a high remission rate in patients with CML and chromosome, was found in the white cells of most patients other tumors. with chronic myeloid leukemia (CML). A major advance in the understanding of the mecha- 1973 By the use of specific dyes to label chromosomes, nisms of malignant transformation followed the discov- notably quinocrine fluorescence and Giemsa staining, ery that many forms of cancer result from the acquisition the Philadelphia chromosome was found to be the result of mutations in cellular oncogenes--that is, normal of a translocation between chromosomes 9 and 22: housekeeping genes of cells that are involved in a variety t(9:22)(q34:q11). of regulatory functions. In some cases, we may be born 1983 It was found that the translocation that causes the with a mutation of this kind, but the vast majority of Philadelphia chromosome juxtaposes the c-abl oncogene cancers seem to be attributable to acquired mutations from chromosome 9 with a breakpoint cluster (bcr) region involving one or more oncogenes. The discovery of a on chromosome 22, resulting in an abnormal bcr/abl drug that was able to interfere specifically with the activ- gene. The product of this gene,an abnormal tyrosine kinase, ity of the product of an abnormal oncogene was a major has increased tyrosine kinase activity compared with the advance in oncology, offering the hope that future product of the normal c-abl gene, a major fact in causing the research will make it possible to tailor-make agents uncontrolled white cell proliferation characteristic of CML. directed at abnormal oncogene activity for the treatment 1996 A selective inhibitor of the abnormal tyrosine kinase of cancer. produced by the bcr/abl gene product was developed that Source: Bartram and others 1983; Druker and others 1996; Klein and others 1982; Nowell and Hungerford 1960; Rowley 1973. organisms that exhibit drug resistance and also for subtyping another, respond to environmental signals, regulate how many classes of bacteria and viruses. Although much remains and when they will divide, and control the other intricate to be learned about the cost-effectiveness of these approaches processes of cell biology (box 5.1). compared with more conventional diagnostic procedures, · Second are tumor suppressor genes; loss of function by muta- some promising results have already been obtained, particu- tion may lead to a neoplastic phenotype. larly for identification of organisms that are difficult to grow or in cases that require a very early diagnosis (Harris and Tanner In the rare familial cancers, individuals are born with one 2000). This type of technology is being widely applied for defective gene of this type, but in the vast majority of cases, the identification of new organisms and is gaining a place cancer seems to result from the acquisition during a person's in monitoring vaccine trials (Felger and others 2003). The lifetime of one or more mutations of oncogenes. For example, remarkable speed with which a new corona virus and its differ- in the case of the common colon cancers, perhaps up to six ent subtypes were identified as the causative agent of SARS and different mutations are required to produce a metastasizing the way this information could be applied to tracing the puta- tumor. The likelihood of the occurrence of these mutations tive origins of the infection are an example of the power of this is increased by the action of environmental or endogenous technology (Ruan and others 2003). carcinogens. Genomics is likely to play an increasingly important role Array technology, which examines the pattern of expression in the control and management of cancer (Livingston and of many different genes at the same time, is already providing Shivdasani 2001). It is now well established that malignant valuable prognostic data for cancers of the breast, blood, and transformation of cell populations usually results from lymphatic system. This technology will become an integral part acquired mutations in two main classes of genes: of diagnostic pathology in the future, and genomic approaches to the early diagnosis of cancer and to the identification of · First are oncogenes--genes that are involved in the major high-risk individuals will become part of clinical practice. It is regulatory processes whereby cells interact with one also becoming possible to interfere with the function or 128 | Disease Control Priorities in Developing Countries | David Weatherall, Brian Greenwood, Heng Leng Chee, and others products of oncogenes as a more direct approach to the treat- Table 5.1 Pharmacogenomics ment of cancer (box 5.1), although early experience indicates Gene Drug Clinical consequence that drug resistance may be caused by mutation, as it is in more conventional forms of cancer therapy. Drug metabolism The genomic approach to the study of common diseases of NAT-2 Isoniazid, hydralazine, Neuropathy, lupus middle life--coronary artery disease, hypertension, diabetes, procainamide, erythematosus and the major psychoses, for example--has been widely publi- sulfonamides cized (Collins and McKusick 2001). Except in rare cases, none CYP2D6 Beta-blockers, Arrhythmias, dyskinesia antidepressants, with antipsychotics, of them is caused by a defective single gene; rather, they appear codeine, debrisoquine, narcotic effects, changes to be the result of multiple environmental factors combined antipsychotics, many in efficacy, many others with variation in individual susceptibility attributable to the others action of several different genes. The hope is that if these sus- CYP2C9 Tolbutamide, phenytoin, Anticoagulant effects of ceptibility genes can be identified, an analysis of their products nonsteroidal anti- warfarin modified will lead to a better understanding of the pathology of these inflammatories diseases and will offer the possibility of producing more defin- RYR-1 Halothane and other Malignant hyperthermia itive therapeutic agents. Better still, this research could provide anesthetics the opportunity to focus public health measures for prevention Protection against oxidants on genetically defined subsets of populations. G6PD Primaquine, sulfonamides, Hemolytic anemia Pharmacogenomics is another potential development acetanilide, others from the genomics revolution (Bumol and Watanabe 2001) Drug targets (table 5.1). Considerable individual variability exists in the ACE Captopril, enalapril Modified response to treat- metabolism of drugs; hence, clinical medicine could reach a ment of cardiac failure, stage at which every person's genetic profile for the metabolism hypertension, renal disease of common drugs will be worked out and become part of their HERG Quinidine Cardiac arrhythmia (long physicians' toolkit. This information will also be of consider- Q­T syndrome) able value to the pharmaceutical industry for designing more HKCNE2 Clarithromycin Drug-induced arrhythmia effective and safer therapeutic agents. Source: Modified from Evans and Relling (Science 286: 487 [2001], as quoted in WHO 2002a). A word of caution is necessary: Although well-defined Note: Table shows examples of genetic polymorphisms that cause unwanted effects of drugs or genetic variation is responsible for unwanted side effects of modification of response. Currently, arrays are being developed for the rapid identification of drugs, this information is still rarely used in clinical practice; families of polymorphisms related to infection-defense genotypes, drug-metabolism genotypes, and many others. Although many polymorphisms associated with variations to drug response or a possible exception is screening for glucose-6-phosphate toxicity have been defined, the bulk of variation of response to drugs follows a multifactorial dehydrogenase (G6PD) deficiency for primaquine sensitivity, pattern of inheritance. The examples shown are as follows: NAT-2 N-acetyltransferase; CYP cytochrome P450; though the costs preclude its application in many developing RYR-1 ryanidine receptor; G6PD Glucose-6-phosphate dehydrogenase; and countries. Furthermore, plasma levels after the administration ACE angiotensin-converting enzyme. HERG and HKCNE2 are potassium channels. of most common drugs follow a normal distribution, indicat- ing that if genetic variation exists, a number of different genes must be involved. Hence, although the idea of all people having gene--has gone through long periods of slow progress and their genetic profile for handling drugs as part of their standard many setbacks, the signs are that it will be successful for at least medical care will take a long time to achieve, if it ever happens, a limited number of monogenic diseases in the long term (Kaji no doubt exists that this field will gradually impinge on med- and Leiden 2001). It is also likely to play a role for shorter-term ical research and clinical practice. objectives--in the management of coronary artery disease Many other potential applications of genomic research for and some forms of cancer, for example. DNA technology has medical practice wait to be developed. The role of DNA array already revolutionized forensic medicine and will play an technology for the analysis of gene expression in tumors has increasingly important role in this field. Although it is too already been mentioned. Advances in bioengineering, with the early to assess to what extent the application of DNA technol- development of biomicroelectromechanical systems, micro- ogy to the studies of the biology of aging will produce infor- level pumping, and reaction circuit systems, will revolutionize mation of clinical value, considering the massive problem of chip technology and enable routine analysis of thousands of our aging populations and the contribution of the aging molecules simultaneously from a single sample (Griffith and process to their illnesses, expanding work in this field is vital. Grodzinsky 2001), with application in many other fields of Current work in the field of evolution using DNA technology research. Although somatic cell gene therapy--that is, the seems a long way from clinical practice; however, it has con- correction of genetic diseases by direct attack on the defective siderable possibilities for helping us understand the lack of Science and Technology for Disease Control: Past, Present, and Future | 129 adaptation of present day communities to the new environ- intractable human diseases, particularly those involving the ments that they have created. nervous system (Institute of Medicine 2002). Stem Cell and Organ Therapy. Stem cell therapy, or, to use its Information Technology. The explosion in information tech- more popular if entirely inappropriate title, therapeutic nology has important implications for all forms of biomedical cloning, is an area of research in cellular biology that is raising research, clinical practice, and teaching. The admirable desire great expectations and bitter controversies. Transplant surgery on the part of publicly funded groups in the genomics field to has its limitations, and the possibility of a ready supply of cells make their data available to the scientific community at large to replace diseased tissues, even parts of the brain, is particu- is of enormous value for the medical application of genomic larly exciting. Stem cells can be obtained from early embryos, research. This goal has been achieved by the trio of public from some adult and fetal tissues, and (at least theoretically) databases established in Europe, the United States, and Japan from other adult cells. (European Bioinformatics Institute, GenBank, and DNA Data Embryonic stem cells, which retain the greatest plasticity, Bank of Japan, respectively). The entire data set is securely held are present at an early stage of the developing embryo, from in triplicate on three continents. The continued development about the fourth to seventh day after fertilization. Although and expansion of accessible databases will be of inestimable some progress has been made in persuading them to produce value to scientists, in both industrial and developing countries. specific cell types, much of the potential for this field so far has Electronic publishing of high-quality journals and related come from similar studies of mouse embryonic stem cells. For projects and the further development of telepathology will example, mouse stem cells have been transplanted into mice help link scientists in industrial and developing countries. The with a similar condition to human Parkinson's disease with increasing availability of telemedicine education packages some therapeutic success, and they have also been used to try will help disseminate good practices. Realizing even these few to restore neural function after spinal cord injuries. examples of the huge potential of this field will require a major Many adult tissues retain stem cell populations. Bone mar- drive to train and recruit young information technology scien- row transplantation has been applied to the treatment of a wide tists, particularly in developing countries, and the financial range of blood diseases, and human marrow clearly contains support to obtain the basic equipment required. stem cells capable of differentiating into the full complement of cell types found in the blood. Preliminary evidence indicates Minimally Invasive Diagnostics and Surgery: Changes in that they can also differentiate into other cell types if given the Hospital Practice. Given the spiraling costs of hospital care in appropriate environment; they may, for example, be a source of industrial countries and the likelihood of similar problems for heart muscle or blood vessel cell populations. Although stem developing countries in the future, reviewing aspects of diag- cells have also been found in brain, muscle, skin, and other nostics and treatment that may help reduce these costs in the organs in the mouse, research into characterizing similar cell future is important. Changes in clinical practice in the latter populations from humans is still at a very early stage. half of the 20th century have already made some headway on One of the major obstacles to stem cell therapy with cells this problem. In the U.K. NHS, the number of hospital beds derived from embryos or adult sources is that, unless they come occupied daily halved between 1950 and 1990 even though the from a compatible donor, they may be treated as "foreign" and throughput of the service, after allowance for change of defini- rejected by a patient's immune system. Thus, much research is tion, increased from 3 million to 10 million inpatients per directed at trying to transfer cell nuclei from adult sources into year. Remarkably, by 1996, of 11.3 million finished consultant an egg from which the nucleus has been removed, after which episodes, 22 percent were single-day cases. How can this the newly created "embryo" would be used as a source of efficient trend be continued? A major development with this embryonic stem cells for regenerative therapy for the particular potential is the application of minimally invasive and robotic donor of the adult cells. Because this technique, called somatic surgery (Mack 2001). Advances in imaging, endoscopic tech- cell nuclear transfer, follows similar lines to those that would be nology, and instrumentation have made it possible to convert required for human reproductive cloning, this field has raised a many surgical procedures from an open to an endoscopic number of controversies. Major ethical issues have also been route. These procedures are now used routinely for gall bladder raised because, to learn more about the regulation of differen- surgery, treatment of adhesions, removal of fibroids, nephrec- tiation of cells of this type, a great deal of work needs to be tomy, and many minor pediatric urological procedures. The carried out on human embryonic stem cells. recent announcement of successful hip replacement surgery If some of the formidable technical problems of this field using an endoscopic approach offers an outstanding example can be overcome and, even more important, if society is able to of its future potential. Although progress has been slower, a come to terms with the ethical issues involved, this field holds number of promising approaches exist for the use of these considerable promise for correction of a number of different techniques in cardiac surgery and for their augmentation by 130 | Disease Control Priorities in Developing Countries | David Weatherall, Brian Greenwood, Heng Leng Chee, and others the introduction of robotics into surgical practice. Transplant The increasing application of functional imaging, together surgery will also become more efficient by advances in the with a better understanding of biochemical function in the development of selective immune tolerance (Niklason and brain, is likely to lead to major advances in our understanding Langer 2001). of many neuropsychiatric disorders and, hence, provide op- These trends, and those in many other branches of medi- portunities for their better management. Early experience cine, will be greatly augmented by advances in biomedical with fetally derived dopaminergic neurons to treat parkinson- imaging (Tempany and McNeil 2001). Major progress has ism has already proved to be successful in some patients and already been made in the development of noninvasive diagnos- has raised the possibility that genetically manipulated stem tic methods by the use of MRI, computer tomography, positron cell treatment for this and other chronic neurological disor- imaging tomography, and improved ultrasonography. Image- ders may become a reality. Promising methods are being guided therapy and related noninvasive treatment methods are developed for limiting brain damage after stroke, and there is also showing considerable promise. increasing optimism in the field of neuronal repair based on the identification of brain-derived neuronotrophic growth Human Development and Child and Maternal Health. factors. Similarly, a combination of molecular genetic and Among the future developments in molecular and cell biology, immunological approaches is aiding progress toward an a better understanding of the mechanisms of human develop- understanding of common demyelinating diseases--notably ment and the evolution of functions of the nervous system multiple sclerosis. offer some of the most exciting, if distant, prospects Strong evidence exists for a major genetic component to the (Goldenberg and Jobe 2001). In the long term, this field may common psychotic illnesses--notably bipolar depression and well have important implications for reproductive health and schizophrenia. Total genome searches should identify some of birth outcomes. The role of a better understanding of the the genes involved. Although progress has been slow, there are monogenic causes of congenital malformation and mental reasonable expectations for success. If some of these genes can retardation was mentioned earlier in this chapter. Already be identified, they should provide targets for completely new thoughts are turning to the possibility of the isolation and clin- approaches to the management of these diseases by the phar- ical use of factors that promote plasticity of brain development, maceutical industry. Recent successes in discovering the genes and specific modulators of lung and gut development are pre- involved in such critical functions as speech indicate the dicted to start to play an increasing role in obstetric practice. A extraordinary potential of this field. Similarly, lessons learned better understanding of the mechanisms leading to vasocon- from the identification of the several genes involved in familial striction and vascular damage as a cause of preeclampsia has forms of early-onset Alzheimer's disease have provided invalu- the potential for reducing its frequency and thus for allowing able information about some of the pathophysiological mech- better management of this common condition. Similarly, an anisms involved, work that is having a major effect on studies increasing appreciation of the different genetic and metabolic directed at the pathophysiology and management of the much pathways that are involved in spontaneous preterm births commoner forms of the disease that occur with increasing should lead to effective prevention and treatment, targeting frequency in aged populations. specific components of these pathways and leading to reduc- tion in the frequency of premature births. An increasing Nutrition and Genetically Modified Crops. By 2030, the knowledge of the mode of action of different growth factors world's population is likely to increase by approximately 2.5 bil- and promoters of gut function will enhance growth and devel- lion people, with much of this projected growth occurring in opment of preterm infants. developing countries. As a consequence, food requirements are expected to double by 2025. However, the annual rate of Neuropsychiatry. Particularly because depression and related increase in cereal production has declined; the present yield is psychiatric conditions are predicted to be a major cause of ill well below the rate of population increase. About 40 percent of health by 2020 and because of the increasing problem of potential productivity in parts of Africa and Asia and about dementia in the elderly, neuropsychiatry will be of increas- 20 percent in the industrial world are estimated to be lost to ing importance in the future (Cowan and Kandel 2001). pathogens. Developments in the basic biomedical sciences will play a major Given these considerations, the genetic modification (GM) role in the better diagnosis and management of these disorders. of plants has considerable potential for improving the world's Furthermore, the application of new technologies promises to food supplies and, hence, the health of its communities. The lead to increasing cooperation between neurology and psychia- main aims of GM plant technologies are to enhance the try, especially for the treatment of illnesses such as mental retar- nutritional value of crop species and to confer resistance to dation and cognitive disorders associated with Alzheimer's and pathogens. GM technology has already recorded several suc- Parkinson's diseases that overlap the two disciplines. cesses in both these objectives. Science and Technology for Disease Control: Past, Present, and Future | 131 Controversy surrounds the relative effectiveness of GM development. These problems are particularly relevant to the crops as compared with those produced by conventional health problems of the developing countries. means, particularly with respect to economic issues of farming One of the main barriers to progress in these fields is the rel- in the developing world. Concerns are also expressed about the ative isolation of the social sciences and health care economics safety of GM crops, and a great deal more research is required from the mainstreams of medical research and practice. Better in this field. The results of biosafety trials in Europe raise some integration of these fields will be a major challenge for univer- issues about the effects of GM on biodiversity (Giles 2003). sities and national and international health care agencies. Plant genetics also has more direct potential for the control of disease in humans. By genetically modifying plants, researchers hope it will be possible to produce molecules toxic Integration of the Medical Sciences: Organizational to disease-carrying insects and to produce edible vaccines that Priorities for the Future are cheaper than conventional vaccines and that can be grown From these brief examples of the likely direction of biomedical or freeze dried and shipped anywhere in the world. A promis- research in the future, some tentative conclusions can be drawn ing example is the production of hepatitis B surface antigen about its effects on the pattern of global health care. in transgenic plants for oral immunization. Work is also well The control of communicable disease will remain the top advanced for the production of other vaccines by this approach priority. Although this goal can be achieved in part by improv- (WHO 2002a). ing nutrition and sanitation and applying related public health measures in developing countries, the search for vaccines Social and Behavioral Sciences, Health Systems, and Health or better chemotherapeutic agents must also remain a high Economics. As well as the mainstream biomedical sciences, priority. However, although optimism that new vaccines will research into providing health care for the future will require become available is well founded, many uncertainties still exist, a major input from the social and behavioral sciences and particularly in the case of biologically complex diseases like health economics. These issues are discussed in more detail in malaria. It is vital that a balance be struck between the basic chapter 4. biomedical science approach and the continued application of The World Health Report 2002 (WHO 2002c) emphasizes methods to control these diseases by more conventional and the major gaps in public perception of what is meant by health well-tried methods. and, in particular, risk factors, in both industrial and develop- For the bulk of common noncommunicable diseases, the ing countries. Epidemiological studies have indicated that situation is even less clear. Although much more humane, cost- morbidity and mortality may be delayed among populations effective, and clinically effective approaches to their manage- that are socially integrated. Increasing evidence of this kind ment seem certain to be developed, mainly by high-technology underlines the importance of psychosocial factors in the devel- and expensive procedures, the position regarding prevention opment of a more positive approach to human health, clearly a and a definitive cure is much less certain. Hence, the program valuable new direction for research on the part of the social for reducing risk factors, as outlined in the World Health Report sciences. 2002 (WHO 2002c), clearly should be followed. However, a Neither developing nor industrial countries have come to strong case exists for a partnership of the public health, epi- grips with the problems of the organization and delivery of demiological, and genomic sciences to develop pilot studies to health care. Learning more about how to build effective health define whether focusing these programs on high-risk subsets of delivery strategies for developing countries is vital. Similarly, populations will be both cost-effective and more efficient. For the continuous reorganization of the U.K. NHS, based on those many chronic diseases for which no risk factors have been short-term political motivation and rarely on carefully defined, strategies of the same type should be established to designed pilot studies, is a good example of the requirement for define potential environmental factors that may be involved. research into the optimal approaches to the provision of health Although surprises may arise along the way, such as the discov- care in industrial countries. Indeed, across the entire field of ery of the infective basis for peptic ulceration, the multilayered health provision and the education of health care professionals, environmental and genetic complexity of these diseases, com- an urgent requirement exists for research into both methodol- bined with the ill-understood effects of aging, suggests that no ogy and, in particular, development of more robust endpoints quick or easy answers to these problems will present them- for its assessment. selves; future planning for global health services must take this Similar problems exist with respect to research in health factor into consideration. economics. Many of the parameters for assessing the burden of Given these uncertainties, an important place exists for the disease and the cost-effectiveness of different parameters for involvement and integration of the social sciences and health the provision of health care are still extremely crude and con- economics into future planning for biomedical research. troversial, and they require a great deal more research and Major gaps in knowledge about public perceptions and 132 | Disease Control Priorities in Developing Countries | David Weatherall, Brian Greenwood, Heng Leng Chee, and others understanding of risk factors, a lack of information about the be exploited by current scientific research in the industrial social and medical problems of aging populations, and wide- countries. spread uncertainty about the most cost-effective and efficient This need is particularly pressing in the case of the major ways of administering health care--both in developing coun- communicable killers: malaria, tuberculosis, and AIDS. tries and in those that have gone through the epidemiological Similarly--and equally important--if developing countries are transition and already have advanced health care systems--still to make the best use of this new technology for their own exist. particular disease problems, partnerships will have to be In short, the emerging picture shows reasonable grounds for established between both academia and the pharmaceutical optimism that better and more definitive ways of preventing or industries of the North and South. curing communicable diseases will gradually become available; Although this approach should be followed as a matter of only the time frame is uncertain. Although there will be major urgency, that developing countries build up their own research improvements in management based on extensive and increas- capacity is equally important. Genomics and World Health 2002 ingly high-technology practice, the outlook for the prevention (WHO 2002a) includes some encouraging accounts of how and definitive cure of the bulk of noncommunicable diseases this capacity is being achieved in Brazil, China, and India. The is much less certain. Hence, it is vital that research in the basic establishment of the Asian-Pacific International Molecular biomedical sciences be directed at both the cause and the Biology Network is a good example. prevention of noncommunicable diseases, and that work in It is important that work start now to apply the advances the fields of public health and epidemiology continues to be stemming from the basic biological sciences for the health of directed toward better use of what is known already about their the developing world. This beginning will form a platform for prevention and management in a more cost-effective and the integration of future advances into health care programs efficient manner. for these countries. However, because of uncertainties of the time involved, more conventional public health approaches to medical care must not be neglected, and a balance should be New Technologies and Developing Countries struck between research in this area and research in the emerg- The role of genomics and related high-technology research and ing biomedical sciences. practice in developing countries is discussed in detail in Genomics and World Health 2002 (WHO 2002a). The central question addressed by the report was, given the current eco- Economic Issues for Future Medical Research nomic, social, and health care problems of developing coun- The central economic issues regarding medical research in the tries, is it too early to be applying the rather limited clinical future are how it is to be financed and how its benefits are to be applications of genomic and related technology to their health used in the most cost-effective way in both industrial and devel- care programs? The report concluded that it is not too early, oping countries. Currently, research is carried out in both and subsequent discussion has suggested that this decision private and public sectors (table 5.2). Work in the private sector was right. Where DNA technology has already proven cost- is based mainly in the pharmaceutical industry and, increas- effective, it should be introduced as soon as possible ingly, in the many large biotechnology companies that evolved (Weatherall 2003). Important examples include the common rapidly following the genomic revolution. In the public sector, inherited disorders of hemoglobin (see chapter 34) and, in par- the major sites of research are universities, government research ticular, the use of DNA diagnostics for communicable disease. The advantage of this approach is that it offers a technical base on which further applications can be built as they become Table 5.2 Estimated Global Health Research and available. It also provides the impetus to develop the training Development Funding for 1998 required, to initiate discussions on the many ethical issues that Type of funding Total (US$ billions) Percentage work of this type may involve, and to establish the appropriate regulatory bodies. The way this type of program should be Public funding: high-income and 34.5 47 transition countries organized--through North-South collaboration, local net- working, and related structures, monitored by WHO--was Private funding: pharmaceutical 30.5 42 industry clearly defined in the report. Private not-for-profit funding 6.0 8 For the full benefits of genomics to be made available to developing countries--and for these advances not to widen Public funding: low- and middle- 2.5 3 income countries the gap in health care provision between North and South-- Total 73.5 100 the most pressing and potentially exciting developments from the new technologies of science and medicine will have to Source: WHO 2002b. Science and Technology for Disease Control: Past, Present, and Future | 133 institutes, and centers--either within the universities or expenditure on health research is directed at diseases that, freestanding--that are funded through a variety of philan- numerically, affect a relatively small proportion of the world's thropic sources. The input of philanthropic sources varies population. If the enormous potential of modern biomedical greatly between countries. In the United Kingdom, the research is not to result in a widening of the gap in health care Wellcome Trust provides a portion of funding for clinical and between North and South, this situation must be corrected. basic biomedical research that approaches that of the govern- The governments of industrial countries may be able to ment, and in the United States, the Howard Hughes organiza- encourage a more global view of research activity on the part of tion also plays a major, though proportionally less important, their pharmaceutical and biotechnology industries by various role in supporting medical research. Similarly, the Bill & tax advantages and other mutually beneficial approaches. Melinda Gates Foundation and other large international philan- Progress in this direction seems likely to be slow, however. For thropic foundations are contributing a significant amount of this reason, moving quickly toward a virtual global network for funding for medical research. In developing countries, such research that would bring together the research agencies of the research funding as is available comes from government sources. North and South holds many attractions. Although those of the For example, Thailand and Malaysia spend US$15.7 million and North that rely on government and charitable funding may US$6.9 million each year, representing 0.9 percent and 0.6 per- find it equally difficult to convince their governments that cent of their health budgets, respectively (WHO 2002b). more of their budget should be spent on work in the develop- As examined in the report of the WHO Commission on ing world, they vitally need to move in this direction, possibly Macroeconomics and Health (WHO 2001), considerable dis- by turning at least some proportion of their overseas aid to cussion is taking place about how to mobilize skills and this highly effective approach to developing North-South resources of the industrial countries for the benefit of the partnerships. health of the developing world. However, how this interna- In short, to produce the funding required for medical tional effort should be organized or, even more important, research in the future and to ensure that it takes on a much funded is still far from clear. A number of models have been more global view of its objectives, a complete change in atti- proposed, including the creation of a new global institute for tude is called for on the part of the industrial countries. This health research and a global fund for health research with an transformation, in turn, will require a similar change of out- independent, streamlined secretariat analogous to the Global look on the part of those who educate doctors and medical Fund to Fight AIDS, Tuberculosis, and Malaria. Recently, a scientists. The introduction of considerable sums of research number of large donations have been given--either by govern- monies into the international scene by governments or philan- ments or by philanthropic bodies--to tackle some of the major thropic bodies as single, large donations, while welcome, will health problems of the developing world. Although many of not form the basis for the kind of sustainable research program these approaches are admirable, those that involve single dona- that is required. Rather, the attitudes of both government tions raise the critical problem of sustainability. People with funding agencies and charitable bodies in industrial countries experience in developing interactions between the North and will have to change, with a greater proportion of their funding South will have no doubts about the long period of sustained being directed at diseases of the developing world in the future. work that is often required for a successful outcome. Achieving this end will require a major program of education Because of the uncertainties about sustainability and the on the global problems of disease at every level, including efficiency of large international bodies, it has been suggested governments, industry, universities, charitable organizations, that a virtual global network for health research be established and every other body that is involved in the medical research in which the leading research agencies of the North and South endeavor. take part, together with a coordinating council (Keusch and Issues requiring the assessment of the economic value of Medlin 2003). In this scheme or in a modified form (Pang medical research are discussed in chapter 4. 2003), both government funding agencies and philanthropic bodies would retain their autonomy and mechanisms of fund- Education ing while at the same time their individual programs would be The central theme of the previous sections is that the potential better integrated and directed toward the problems of global fruits of the exciting developments in the biomedical sciences health. will be achieved only if a complete change in attitude occurs on A central problem of both private and public patterns of the part of industrial countries, with the evolution of a much funding for medical research is that industrial countries have more global attitude to the problems of medical research and tended to focus their research on their own diseases and have, health care. Change will have to start in the universities of the with a few exceptions, tended to ignore the broader problems industrial countries, which will need to incorporate a more of developing countries, a trend that has resulted in the well- global perspective in medical education so that the next gener- known 10/90 gap in which more than 90 percent of the world's ation of young people is more motivated to develop research 134 | Disease Control Priorities in Developing Countries | David Weatherall, Brian Greenwood, Heng Leng Chee, and others careers that take a more international view of the problems of Ethical Issues medical research. A major change of emphasis in education will Few advances in scientific medicine have not raised new ethical be required and will be difficult to achieve unless those who issues for society. The genomics era has encountered many control the university education and research programs can be problems in this respect, and although many of the initial fears convinced that funding is available for further development in and concerns have been put to rest by sensible debate and the these new directions (Weatherall 2003). Excellent examples of development of effective control bodies, new problems con- the value of the development of North-South partnerships tinue to appear (WHO 2002a). The ill-named field of thera- between universities and other academic institutions do peutic cloning is still full of unresolved issues regarding human already exist. embryo research, the creation of embryos for research pur- An effective approach to increasing global funding for inter- poses, and other uncertainties, but these questions should not nationally based research is through virtual global networks be overemphasized at a time when most societies face even involving the leading research agencies in the North and South. more onerous ethical issues. For example, as the size of our Hence, a similar effort will be required to educate these agen- aging population increases, many societies may have to face the cies and their governments that this approach to improving extremely difficult problem of rationing medical care. The the level of health globally is cost-effective. In particular, it will theme recurring throughout both industrial and developing be vital to persuade them that this approach may constitute an countries is how to provide an adequate level of health care effective use of their programs of aid for developing countries. equally to every income group. Carrying out a number of pilot studies showing the economic Many developing countries still lack the basic structure for value of North-South partnerships in specific areas of medical the application of ethical practices in research and clinical care, research may be necessary. Indeed, a number of these partner- including the development of institutional ethics committees, ships have already been formed in several countries and infor- governmental regulatory bodies, and independent bioethical mation of this type almost certainly exists (WHO 2002a). research bodies. Every country requires a completely inde- Of course, much broader issues involving education need to pendent bioethics council that can debate the issues uninhib- be resolved for the better exploitation of medical research. The ited by pressures from government, commerce, or pressure problems of educating the public so that developing countries groups of any kind. Our approaches to developing a more ade- can partake in the advancements of the genome revolution quate ethical framework for much of medical decision making, were set out in detail in Genomics and World Health 2002 whether it involves preventive medicine, clinical practice, or (WHO 2002a), but a great deal of work along these lines is also research, constitute another neglected area that requires required for industrial countries. People are increasingly suspi- research input from many different disciplines. cious of modern biological science and of modern high- The important question of the ethical conduct of research technology medicine, a factor that, together with concerns over in the developing countries by outside agencies has been the pastoral skills of today's doctors, is probably playing a role reviewed in detail recently (Nuffield Council on Bioethics in driving many communities in industrial countries toward 2002). complementary medicine (see Horton 2003). These trends undoubtedly are attributable to inadequacies of medical edu- cation and the way that science is taught in schools--reflected Why Do We Need Research? by the lack of scientific literacy both in the general public and It is important to appreciate that considerable public suspicion in governments. If trust is to be restored between the biomed- exists about both the activities and the value of biomedical ical sciences and the public, significant efforts will have to be research. Suspicion has been generated in part by the field's made to improve the level of scientific literacy, and a much exaggerated claims over recent years, an uneasy feeling that more open dialogue will need to be developed between scien- research is venturing into areas that would best be avoided, and tists and the community. This requirement will be increasingly a lack of understanding about the complexity of many of the important as work on basic biomedical sciences impinges on problems that it is attempting to solve. At the same time, many areas such as gene therapy, stem cell research, and the collection government departments that run national health care pro- of large DNA databases to be used for both research and ther- grams, the private sector (with the exception of the pharma- apeutic purposes in the future. ceutical industry), and many nongovernmental organizations The difficulties in achieving a more global view of medical set aside extremely small fractions of their overall expenditure research and health care on the part of industrial countries for for research. For many of those organizations, research seems the future should not be underestimated. Without a major irrelevant as they deal with the stresses of daily provision of attempt to solve these difficulties, the potential of modern programs of health care and with crisis-management scenarios biomedical sciences seems certain to simply widen the gap in that have to follow rapid change or major failures in providing health care between North and South. health care. Science and Technology for Disease Control: Past, Present, and Future | 135 One of the major challenges for the biomedical research to provide medical scientists of the future with a more global community will be to better educate the public about its activ- perspective of health and disease. If this transformation can be ities and to restore their faith in and support for the medical achieved--if it can form the basis for the establishment of net- research endeavor. Educating many governments and non- works for sustainable research programs between universities governmental organizations about the critical importance of and related bodies in the North and South--much progress decision making based on scientifically derived evidence will be will be made toward distributing the benefits of biomedical vital. Medical care will only get more complex and expensive in research and good practice among the populations of the the future; its problems will not be solved by short-term, polit- world. However, the great potential of advances in the biomed- ically driven activity. The need for good science, ranging from ical sciences for global health will not come to full fruition studies of molecules to communities, has never been greater. without much closer interaction between the fields of basic and clinical research and the fields of public health, health econom- SUMMARY ics, and the social sciences. Clearly, the most important priorities for medical research are REFERENCES development of more effective health delivery strategies for developing countries and control of the common and Alberti, G. 2001. "Noncommunicable Diseases: Tomorrow's Pandemics." intractable communicable diseases. In this context, the argu- Bulletin of the World Health Organization 79 (10): 907. ment has been that much of the medical research that has been Barker, D., ed. 2001. Fetal Origins of Cardiovascular and Lung Disease. New York: Marcel Dekker. carried out in industrial countries, with its focus on noncom- Bartram, C. R., A. deKlein, A. Hagemeijer, T. van Agthoven, A. Geurts municable disease and its outcomes in high-technology prac- van Kessel, D. Bootsma, and others. 1983. "Translocation of c-Abl tice, is completely irrelevant to the needs of developing Oncogene Correlates with the Presence of a Philadelphia Chromosome countries. This view of the medical scene, however, is short in Chronic Myelocytic Leukemia." Nature 306 (5940): 277­80. term. Although some redistribution of effort is required, every Beeson, P. B. 1980. "Changes in Medical Therapy during the Past Half Century." Medicine (Baltimore) 59 (2): 79­99. country that passes through the epidemiological transition is Black, D. 1980. Inequalities in Health: Report of a Working Party, now encountering the major killers of industrial countries. Department of Health and Society Security. London: Her Majesty's Learning more about those killers' basic causes, prevention, Stationery Office. and management is crucial. Although the initial costs of pro- Bumol, T. F., and A. M. Watanabe. 2001. "Genetic Information, Genomic viding the benefits of this research are often extremely high, Technologies, and the Future of Drug Discovery." Journal of the they tend to fall as particular forms of treatment become more American Medical Association 285 (5): 551­55. widely applied. Hence, because we cannot completely rely on Castelli, W. P., and K. Anderson. 1986. "Population at Risk: Prevalence of High Cholesterol Levels in Hypertensive Patients in the Framingham our current preventive measures to control these diseases, med- Study." American Journal of Medicine 80 (Suppl. 2A): 23. ical research must continue. Chalmers, I. 1993. "The Cochrane Collaboration: Preparing, Maintaining, Research in basic human biology and the biomedical sci- and Disseminating Systematic Reviews of the Effects of Health Care." ences is entering the most exciting phase of its development. Annals of the New York Academy of Science 703: 156­63; discussion 163­165. However, it is difficult to anticipate when the gains of this Chen, L. C. 1996."World Population and Health." In 2020 Vision: Health in explosion in scientific knowledge will become available for the 21st Century. Washington, DC: National Academy Press. the prevention and treatment of the major killers of mankind. Collins, F. S., and V. A. McKusick. 2001. "Implications of the Human Thus, medical research must strike a balance between the well- Genome Project for Medical Science." Journal of the American Medical tried approaches of epidemiology, public health, and clinical Association 285 (5): 540­44. investigation at the bedside with the application of discoveries Comroe, J. H. Jr., and R. D. Dripps. 1976. "Scientific Basis for the Support in the completely new fields of science that have arisen from the of Biomedical Science." Science 192 (4235): 105­11. genome revolution. Cooter, R., and J. Pickstone. 2000. Medicine in the Twentieth Century. Amersterdam: Harwood. If this balanced approach toward the future provision of Cowan, W. M., and E. R. Kandel. 2001. "Prospects for Neurology and health care is not to continue to worsen the gap between North Psychiatry." Journal of the American Medical Association 285 (5): and South, however, a complete change of attitude is necessary 594­600. toward health care research and practice on the part of the Doll, R. 1985. "Preventive Medicine: The Objectives." Ciba Foundation industrial countries. A major effort will be required to educate Symposium 110: 3­21. all parties--international nongovernmental organizations, Druker, B. J., S. Tamura, E. Buchdunger, S. Ohno, G. M. Segal, S. Fanning, governments, universities, and the private sector--in global and others. 1996. "Effects of a Selective Inhibitor of the Abl tyrosine kinase on the Growth of Bcr-Abl Positive Cells." Nature Medicine 2 (5): health problems (Weatherall 2003). Equally important will be a 561­66. major change of emphasis in the universities of industrial Egger, M., G. Davey-Smith, and D. G. Altman. 2001. Systematic Reviews in countries toward education programs in science and medicine Health Care: Meta-Analysis in Context. London: BMJ Publications. 136 | Disease Control Priorities in Developing Countries | David Weatherall, Brian Greenwood, Heng Leng Chee, and others Feachem, R. G. A., T. Kjellstrom, C. J. L. Murray, M. Over, and M. A. Nowell, P. C., and D. A. Hungerford. 1960. "A Minute Chromosome in Phillips. 1992. The Health of Adults in the Developing World. Oxford, Human Chronic Granulocytic Leukemia." Science 132: 1497­501. U.K.: Oxford University Press. Nuffield Council on Bioethics. 2002. The Ethics of Research Related to Felger, I., B. Genton, T. Smith, M. Tanner, and H. P. Beck. 2003. "Molecular Healthcare in the Developing Countries. London: Nuffield Council on Monitoring in Malaria Vaccine Trials." Trends in Parasitology 19 (2): Bioethics. 60­63. Olshansky, S. J., B. A. Carnes, and C. Cassel. 1990. "In Search of Finch, R. G., and R. J. Williams. 1999. Antibiotic Resistance. London: Methuselah: Estimating the Upper Limits to Human Longevity." Baillière Tindall. Science 250 (4981): 634­40. Giles, J. 2003. "Biosafety Trials Darken Outlook for Transgenic Crops in Pang, T. 2003. "Complementary Strategies for Efficient Use of Knowledge Europe." Nature 425 (6960): 751. for Better Health." Lancet 361 (9359): 716. Goldenberg, R. L., and A. H. Jobe. 2001. "Prospects for Research in Perrin, L., and A. Telenti. 1998. "HIV Treatment Failure: Testing for HIV Reproductive Health and Birth Outcomes." Journal of the American Resistance in Clinical Practice." Science 280 (5371): 1871­73. Medical Association 285 (5): 633­39. Porter, R. 1997. The Greatest Benefit to Mankind: A Medical History of Griffith, L. G., and A. J. Grodzinsky. 2001. "Advances in Biomedical Humanity from Antiquity to the Present. London: Harper Collins. Engineering." Journal of the American Medical Association 285 (5): Roberts, L. 2004. "Polio: The Final Assault?" Science 303 (5666): 1960­68. 556­61. Rowley, J. D. 1973. "A New Consistent Chromosomal Abnormality in Harris, E., and M. Tanner. 2000. "Health Technology Transfer." British Chronic Myelogenous Leukemia Identified by Quinacrine Medical Journal 321 (7264): 817­20. Fluorescence and Giemsa Staining." Nature 243 (5405): 290­93. Horton, R. 2003. Second Opinion: Doctors, Diseases and Decisions in Ruan, Y. J., C. L. Wei, A. L. Ee, V. B. Vega, H. Thoreau, S. T. Su, and Modern Medicine. London: Grant Books. others. 2003. "Comparative Full-Length Genome Sequence Analysis of Institute of Medicine. 2002. Stem Cells and the Future of Regenerative 14 SARS Coronavirus Isolates and Common Mutations Associated Medicine. Washington, DC: National Academy Press. with Putative Origins of Infection." Lancet 361 (9371): 1779­85. ISIS-2 (Second International Study of Infarct Survival) Collaborative Sackett, D. L., W. M. Rosenberg, J. A. Gray, R. B. Haynes, and W. S. Group. 1988. "Randomised Trial of Intravenous Streptokinase, Oral Richardson. 1996. "Evidence Based Medicine: What It Is and What It Aspirin, Both, or Neither among 17,187 Cases of Suspected Acute Isn't." British Medical Journal 312 (7023): 71­72. Myocardial Infarction: ISIS-2." Lancet 2 (8607): 349­60. Scriver, C. R., J. L. Neal, R. Saginur, and A. Clow. 1973. "The Frequency of Joët, T., U. Eckstein-Ludwig, C. Morin, and S. Krishna. 2003. "Validation Genetic Disease and Congenital Malformation among Patients in a of the Hexose Transporter of Plasmodium falciparum as a Novel Drug Pediatric Hospital." Canadian Medical Association Journal 108 (9): Target." Proceedings of the National Academy of Sciences of the U.S.A. 1111­15. 100 (13): 7476­79. Souhami, R. L., I. Tannock, P. Hohenberger, and J. C. Horiot, eds. 2001. Kaji, E. H., and J. M. Leiden. 2001."Gene and Stem Cell Therapies." Journal The Oxford Textbook of Oncology. 2nd ed. Oxford, U.K.: Oxford of the American Medical Association 285 (5): 545­50. University Press. Keusch, G. T., and C. A. Medlin. 2003. "Tapping the Power of Small Tempany, C. M., and B. J. McNeil. 2001. "Advances in Biomedical Institutions." Nature 422 (6932): 561­62. Imaging." Journal of the American Medical Association 285 (5): 562­67. Klein, A., A. G. van Kessel, G. Grosveld, C. R. Bartram, A. Hagemeijer, Warrell, D. A., T. M. Cox, J. D. Firth, and E. J. Benz, eds. 2003. Oxford D. Bootsma, and others. 1982. "A Cellular Oncogene Is Translocated Textbook of Medicine. 4th ed. Oxford, U.K.: Oxford University Press. to the Philadelphia Chromosome in Chronic Myelocytic Leukaemia." Weatherall, D. J. 1995. Science and the Quiet Art: The Role of Research in Nature 300 (5894): 765­67. Medicine. New York: Rockefeller University, W. W. Norton, and Oxford Land, K. M. 2003. "The Mosquito Genome: Perspectives and Possibilities." University Press. Trends in Parasitology 19 (3): 103­5. ------. 2003. "Genomics and Global Health: Time for a Reappraisal." Letvin, N. L., B. R. Bloom, and S. L. Hoffman. 2001."Prospects for Vaccines Science 302 (5645): 597­99. to Protect against AIDS, Tuberculosis, and Malaria." Journal of the Weatherall, D. J., and J. B. Clegg. 2001a. "Inherited Haemoglobin American Medical Association 285 (5): 606­11. Disorders: An Increasing Global Health Problem." Bulletin of the World Livingston, D. M., and R. Shivdasani. 2001. "Toward Mechanism-Based Health Organization 79 (8): 704­12. Cancer Care." Journal of the American Medical Association 285 (5): ------. 2001b. The Thalassaemia Syndromes. 4th ed. Oxford, U.K.: 588­93. Blackwell Scientific Publications. Mack, M. J. 2001. "Minimally Invasive and Robotic Surgery." Journal of the ------. 2002. "Genetic Variability in Response to Infection: Malaria and American Medical Assocation 285 (5): 568­72. After." Genes and Immunity 3 (6): 331­37. Mattick, J. S. 2003. "Challenging the Dogma: The Hidden Layer of Webster, C. 1998. The National Health Service: A Political History. Oxford, Non-Protein-Coding RNAs in Complex Organisms." Bioessays 25 (10): U.K.: Oxford University Press. 930­39. Modell, B., and V. Bulyzhenkov. 1998. "Distribution and Control of Some WHO (World Health Organization). 2000. World Health Report 2000-- Genetic Disorders." World Health Statistics Quarterly 41: 209­18. Health Systems: Improving Performance. Geneva: WHO. Niklason, L. E., and R. Langer. 2001. "Prospects for Organ and Tissue ------. 2001. Macroeconomics and Health: Investing in Health for Replacement." Journal of the American Medical Association 285 (5): Economic Development: Report of the Commission on Macroeconomics 573­76. and Health. Geneva: WHO. Noedl, H., C. Wongsrichanalai, and W. H. Wernsdorfer. 2003. "Malaria ------. 2002a. Genomics and World Health 2002. Geneva: WHO. Drug-Sensitivity Testing: New Assays, New Perspectives." Trends in ------. 2002b. Global Forum for Health Research: The 10/90 Report on Parasitology 19 (4): 175­81. Health Research 2001­2002. Geneva: WHO. Nossal, G. J. V. 1999. "Vaccines." In Fundamental Immunology, ed. W. E. ------. 2002c. The World Health Report 2002: Reducing Risks, Promoting Paul. Philadelphia: Lippincott-Raven. Healthy Life. Geneva: WHO. Science and Technology for Disease Control: Past, Present, and Future | 137 Chapter 6 Product Development Priorities Adel Mahmoud, Patricia M. Danzon, John H. Barton, and Roy D. Mugerwa The overall goal of this chapter is to introduce the multistep Even though investigators have explored the conceptual process that leads to new product development and use and to framework for understanding how knowledge may be trans- outline the economic and institutional context for products lated into products over the years, consensus is lacking on the developed specifically for major global diseases. In addition, it specific drivers of the process or on the effects of alternative attempts to define the major financial efforts under way to help institutional arrangements. stimulate the process. Because product development is inte- Several features of the innovation process and its environ- grally related to intellectual property issues and to regulatory ment are essential for product development (Hilleman 2000; and liability concerns, these topics are also included. Data on Nederbragt 2000; Schmid and Smith 2002). Innovation product development for the developing countries are not advances through a sequence of steps from discovery, through systematically available. We have, therefore, used information process development, to animal and human testing--a se- based on analyses for developed countries and, when possible, quence with many overlapping features. Discovery may come in made comparisons. two ways: in a nonlinear, quantum-leap fashion that results in findings of an unexpected or unpredictable nature or in a linear fashion that builds on existing knowledge. Nonlinear processes INTRODUCTION are characteristically random despite many efforts to inject varying degrees of predictability or goal definitions (Webber In recent decades, scientific advances in many disciplines, and Kremer 2001). By contrast, the goal of linear innovation is particularly molecular biology, genomics, and medicinal defined improvement of a known process or mechanism. chemistry, opened the way for developing new therapeutic agents, several vaccines, and enhanced diagnostic capabilities. The central questions for the purpose of this chapter are what Discovery drives research, discovery, and development and what institu- Product development is fundamentally anchored to the discov- tional and financing arrangements are necessary to promote ery process. In modern societies, discovery represents a societal research and development (R&D) for global diseases? Medical capability that involves multiple institutions and constituents. needs and public health imperatives constitute the logical The concept of networks of innovation has been introduced answer to the first question; however, our armamentarium for to describe one of the processes of discovery that leads to combating major global diseases suffers from certain funda- the development of pharmaceutical products or vaccines mental gaps. Innovation or discovery in the health fields is the (Galambos and Sewell 1995, 272). Original scientific observa- process whereby the findings of many sciences are translated tions are made in organizations widely distributed across soci- from basic findings into approaches to protect health (vac- ety, such as academic environments, government laboratories, cines) or reverse disease (therapeutic and diagnostic products). biotechnology companies, or the large organizations dedicated 139 to R&D. Because of the multiplicity of these settings and the market needs to determine the level of investment required for traditions of open scientific communications, combined with plant construction and operation are the two fundamental the high costs of research and the importance of incentives, components of this phase. intellectual property issues must be taken into account. One important feature of discovery and development is the The outcome is appreciably complex. Therefore, prescribing length of time it takes. Estimates indicate that the average time in a systematic way how to develop products along a planned for a new chemical entity (NCE) or vaccine to proceed from dis- pathway--particularly those intended for use in developing covery through preclinical testing,human clinical trials,and reg- countries--is challenging. Recent decades have witnessed many ulatory approval is longer than a decade (Garber, Silvestri, and attempts to develop specific drugs or vaccines to meet develop- Feinberg 2004; Hilleman 1996; Rappuoli, Miller, and Falkow ing countries'needs,and the process has been difficult.Examples 2002), including the time spent on unsuccessful attempts. This include pharmaceuticals to treat major global killers such as timeline imposes certain pressures on how decisions are made, malaria and African trypanosomiasis and vaccines for most of on the investment needed, and on competing priorities. the diarrheal diseases and respiratory infections (Nossal 2000). Development Institutions Development Cycles As indicated previously, innovation and discovery occur in a Discovery may set in motion a series of steps that eventually multiplicity of settings. Although these settings have been con- leads to the deployment of a product suitable for human use. centrated in developed counties and have served the process of The next step following discovery is process definition to map product development well, the challenges of developing new the steps of manufacturing and scalability to optimize the size of products for the developing world are considerable. Many coun- manufacturing. This process involves translating an idea discov- tries, such as Brazil, India, and Singapore, are initiating a new ered anywhere in the multiplicity of settings defined earlier, wave of fundamental research institutions (Ahmad 2001; including mobilizing the energies of many sciences, to come up Jayaramann 2003). Their involvement in the discovery of prod- with a product. For instance, for a discovery in the therapeutic ucts necessary for the health needs of developing countries is a field to be translated into a drug, the sciences of medicinal fundamental paradigm shift. Along with the developing world's chemistry, structural biology, and structure-function relation- emerging biotechnology industry, a movement toward product ships are fundamental to the process. More recently, the product discovery and development is under way. In addition, multiple development process has begun using genomics and proteomics PPPs--for example, the MMV (2002) and the GATB (2001)-- to bring about a more focused approach to defining clinical can- are adding to the total global effort (Lyles 2003; Widdus 2001). didate products. Only then are pharmacology, toxicology, and The major feature of these new settings is their ability to focus bioavailability used in the next phase of therapeutic evaluation. on the immediate needs of developing countries. The challenge, The capabilities for process definition and scalability have however, lies in sustaining their funding and ensuring their abil- traditionally been concentrated in the research-based pharma- ity to proceed from discovery to development and manufactur- ceutical industry, but several recent successful efforts in public- ing, possibly with appropriate partners. private partnerships (PPPs) have expanded these capabilities, Finally, the evaluation of a product's pharmacological, bio- such as the Medicines for Malaria Venture (MMV) and Global logical, and toxicological properties may be carried out in Alliance for TB Drug Development (GATB). Developing coun- developed or developing countries. Indeed, the evaluation of tries such as Brazil, India, and the Republic of Korea are now the safety and efficacy of products intended for developing undertaking major efforts to achieve similar capabilities countries should occur in those settings. Although quality (Biehl 2002; Lohray 2003). control standards should be applied globally (Milstien and Therapeutic evaluation may begin at an in vitro or molecu- Belgharbi 2004), specific efforts must be directed at protecting lar level before proceeding to animal testing and the usual three the rights of human subjects (Agre and Rapkin 2003; Barrett phases of human assessment (Hilts 2003). The scientific disci- and Parker 2003; Emanuel and others 2004; McMillan and plines of clinical research, epidemiology, and biostatistics have Conlon 2004). In general, clinical development is heavily regu- progressed at a significant pace in recent decades. In parallel, lated in developed countries, and additional mechanisms exist ethical and societal concerns about research involving human for monitoring other aspects of product development, such as subjects and its standards, particularly across countries, cul- animal experimentation, use of controlled substances, and so tures, and capabilities, are being extensively debated (Agre and on, but the global situation varies considerably. The time is ripe Rapkin 2003; Barrett and Parker 2003; Emanuel and others to consider the development of a global coordinated effort that 2004; McMillan and Conlon 2004). involves uniform standards and reciprocity. The engineering aspects of product development are the The analysis in the following sections focuses on the costs of next major step. Optimizing manufacturability and assessing developing drugs, vaccines, and diagnostics. The emphasis on 140 | Disease Control Priorities in Developing Countries | Adel Mahmoud, Patricia M. Danzon, John H. Barton, and others drugs and vaccines reflects both the available evidence and the multiple countries. Clinical trial out-of-pocket costs reflect fact that regulatory requirements and costs are much greater expenditures on patients' medical treatment and monitoring, for drugs and vaccines than for devices and diagnostics. data collection, and analysis. In the study by DiMasi, Hansen, and Grobowski (2003), the average expected clinical cost, adjusted for the probability of entering each clinical phase, was PHARMACEUTICAL PRODUCTS US$60.6 million per compound entering human trials. In addi- tion, the authors estimated that the out-of-pocket costs of drug The costs of developing new medicines and diagnostics reflect discovery and preclinical development account for 30 percent both the technical complexities of product development and of overall R&D costs, raising the total expected out-of-pocket costs related to regulatory approval, which requires clinical tri- cost to US$86.8 million per compound entering clinical trials. als to establish product safety and efficacy. Although the rela- The average number of clinical trial patients per compound was tive contributions of these two components are difficult to 5,303, and the average cost per patient was US$23,500 before distinguish empirically--and even conceptually--there is gen- adjusting for the probability of entering each clinical phase. eral consensus that increasing regulatory requirements have Second, in the United States, the Food and Drug contributed to the rising costs of new product development Administration (FDA) approves only roughly one in five com- in the United States. In considering the costs of new product pounds that enter human clinical trials.1 The costs incurred for development for diseases prevalent in low-income countries the four out of five compounds that failed must be included as (LICs), we attempt to identify those costs that might be influ- costs of bringing one new compound to market. Failures occur enced by regulatory policy as opposed to the unavoidable costs because of safety concerns, lack of significant efficacy, and poor resulting from the hard science of new product development. economic prospects. Even though the new technologies of drug discovery should eventually improve predictive accuracy for both safety and efficacy, success rates were no better in the 1990s R&D Costs for Drugs for Industrial Countries than in the 1980s (DiMasi,Hansen,and Grabowski 2003; DiMasi The most detailed evidence on the cost of developing new drugs and others 1991). Adjusting for failure rates raises the total out- is from DiMasi, Hansen, and Grabowski (2003), who estimate of-pocket cost from US$86.8 million to US$403 million per the cost of bringing a compound to market at US$802 million approved compound. in 2000 dollars. Their estimate is based on U.S. data from Third, the US$802 million total cost estimate includes the 10 major companies for a randomly selected sample of 68 com- opportunity cost of capital over the roughly 12-year investment pounds that entered human testing between 1983 and 1994 and period. Using an 11 percent real (net of inflation) cost of capi- reached approval between 1990 and 2001. The 68 compounds tal, DiMasi, Hansen, and Grabowski (2003) estimate the total include 61 small molecule chemical entities, 4 recombinant cost of capital at US$399 million. This figure represents the proteins, 2 monoclonal antibodies, and 1 vaccine. Together, the return that shareholders would have received had they invested 10 companies accounted for 42 percent of R&D by U.S. compa- in activities that yielded immediate returns rather than in the nies. The cost estimates are based on project-level data obtained lengthy drug discovery process. If pharmaceutical R&D is from the companies for the period 1980­99. The sample was financed by--and hence must compete for--private equity restricted to compounds that originated within these compa- capital, shareholders must be compensated for this opportu- nies to avoid omitted costs of in-licensed products. nity cost. Thus, the cost of capital is appropriately included as Earlier studies using similar data and methods found signif- a cost of R&D if the R&D is undertaken in commercial firms icantly lower R&D costs for drugs launched in the 1970s and and financed by equity capital. As discussed later, if not-for- 1980s (DiMasi and others 1991; Hansen, 1979). For the 1980s profit organizations finance R&D, the opportunity cost of drug cohort, the estimate was US$359 million per NCE (U.S. capital may be lower. If we assume financing by private equity, Congress, Office of Technology Assessment 1993). Thus, the adding the US$399 million cost of capital to the US$403 mil- estimate for the 1990s drug cohort of US$802 million repre- lion out-of-pocket cost yields US$802 million as the capitalized sents a significant increase over and above inflation. cost at launch, before taxes, per approved compound. The Three main factors contribute to this high and rising cost of after-tax estimate is considerably lower because, like any busi- R&D. Understanding the contribution of each of these factors ness expense, R&D expenses are tax deductible, plus R&D tax is important to understanding whether drug R&D costs might credits may be available in certain circumstances. However, for be lower in developing countries. purposes of comparing the costs of R&D to the revenues a First, the inputs into pharmaceutical R&D are costly, includ- commercial firm would require to cover these costs, if costs are ing highly trained scientists, highly specialized capital equip- measured net of tax, then revenues must also be measured net ment, expensive animal studies, and clinical trials involving of tax, in which case adjusting for tax makes little difference. thousands of human subjects that are often coordinated across Hereafter we use the before-tax R&D cost estimates to facilitate Product Development Priorities | 141 comparison with other estimates of R&D costs. The before-tax personnel than in the United States. The trial duration may also estimates are also most relevant to not-for-profit firms and be shorter because the target diseases are acute rather than PPPs that are not subject to taxes. chronic. To the extent that the lower out-of-pocket clinical costs If commercial firms facing a commercial cost of capital and in the GATB and MMV studies reflect fewer patients in trials and with no in-kind contributions (see the next section) undertake lower cost per patient, such savings could, in principle, apply to R&D, they might be able to save roughly 10 to 20 percent of LIC drugs regardless of whether these drugs are developed by their costs by conducting trials in developing countries and not-for-profit or commercial enterprises.3 possibly more if they adhere to the countries' regulatory Another factor contributing to the lower out-of-pocket requirements, which may permit fewer and shorter trials than costs reported by the MMV and the GATB is that these PPPs are normal for the FDA. Whether firms could realize those benefit from in-kind contributions of personnel, technologies, potential savings may be a matter of judgment depending on and other resources supplied by their industry and academic perceived liability risks. If commercial firms conduct R&D for partners. The MMV estimates these in-kind contributions as LIC diseases in not-for-profit spinoffs, they may realize tax equivalent to its own incurred costs. Thus, if these in-kind con- advantages and a lower cost of capital, which would further tributions are included, the full social cost for developing LIC reduce their cost below the US$802 million estimate. drugs increases to US$250 million to US$300 million per com- pound, or only 25 to 35 percent less than the DiMasi, Hansen, and Grabowski (2003) estimate of US$403 million. However, R&D Costs for Drugs for Developing Countries as long as such in-kind contributions are available without Recent studies by two PPPs that focus on new product devel- charge to PPPs, the actual budget cost to PPP funders is only opment for diseases in developing countries yield much lower US$150 million to US$178 million, or less than half DiMasi, cost estimates for drugs in their portfolios than those in the Hansen, and Grabowski's (2003) estimate.4 previous section. The GATB and the MMV estimate the costs The second major determinant of R&D costs is failure rates. of R&D at US$150 million (MMV 2002) and US$178 million The GATB and MMV estimates show overall drug failure rates (midpoint of the range of US$115 million to US$240 million) similar to those in DiMasi, Hansen, and Grabowski's (2003) (GATB 2001, 101) or less than a quarter of DiMasi, Hansen, study. Indeed, there is no obvious reason to expect significant and Grabowski's (2003) estimate of US$802 million. The rea- differences in failure rates if LIC drugs face similar scientific sons for these large differences are instructive. challenges and are reviewed by the FDA or the European First, the GATB and MMV estimates reflect only out-of- Medicines Evaluation Agency applying the same safety, efficacy, pocket costs,with no allowance made for the opportunity cost of and risk-benefit tradeoff standards as are applied to drugs for capital. Nevertheless, the estimates of out-of-pocket cost are less the industrial countries. However, if the regulatory review of than half of the US$403 million out-of-pocket cost estimated by LIC drugs uses risk-benefit tradeoffs that reflect conditions in DiMasi, Hansen, and Grabowski (2003). This difference in out- developing countries, then success rates might be higher, imply- of-pocket costs primarily reflects two factors: (a) fewer clinical ing a lower budget cost per approved compound for LIC drugs. trials and, hence, fewer patients in trials--namely, 1,368 patients Finally, the third major contributor to R&D costs is the per drug for the GATB compared with 5,303 in the DiMasi, opportunity cost of capital, which accounts for US$399 million, Hansen, and Grabowski (2003) study--and (b) lower costs per or almost half of DiMasi, Hansen, and Grabowski's (2003) patient of US$1,000 to US$3,000 for the GATB for trials run in US$802 million cost per compound. The GATB and MMV esti- developing countries compared with US$23,500 per patient in mates do not include the cost of capital. Whether the cost of the DiMasi, Hansen, and Grabowski (2003) study. capital should be included in estimating the cost of R&D for Some drugs for LICs may require fewer trials, fewer patients, LIC drugs depends on the circumstances and the perspective. If or both per trial because of differences in drug types and trial LIC drugs are to be developed by commercial firms that must objectives and different regulatory requirements. For example, generate a competitive return for their shareholders, then the some of the drugs in the two PPPs' portfolios are modifications cost estimates appropriately include a cost of capital at roughly of existing drugs for which some data have been established. 11 percent, as in the DiMasi, Hansen, and Grabowski (2003) R&D costs for LIC drugs may also be lower to the extent that study. However, if LIC drugs are developed by PPPs or other these drugs are tested for fewer indications, with less within- not-for-profit institutions with financing from philanthropic or sample stratification by patient subgroup and less need to test governmental agencies, the opportunity cost of capital may be for drug interactions. Clinical effects for infectious and parasitic lower if these funders typically do not require a rate of return on diseases may also be greater than for chronic diseases, which their investment to compensate them for the forgone alternative permit smaller trial sizes.2 The lower trial cost per patient for uses of the funds during the investment period. For example, LIC drugs partly reflects the lower costs of conducting trials in government investments sometimes assume a social opportu- developing countries,with much lower costs of medical care and nity cost of capital of about 5 percent. Using a 5 percent cost of 142 | Disease Control Priorities in Developing Countries | Adel Mahmoud, Patricia M. Danzon, John H. Barton, and others capital for financing from philanthropic or governmental agen- most currently available vaccines have been developed over cies implies a roughly 50 percent markup over out-of-pocket relatively long periods and multiple organizations have been R&D costs to reflect the cost of capital rather than the roughly involved in their discovery, our cost estimates are based on his- 100 percent estimated by DiMasi, Hansen, and Grabowski torical data and on many assumptions that are probably chang- (2003), assuming the same time flow of investments. ing rapidly (Agre and Rapkin 2003; Barrett and Parker 2003; Applying this markup to the US$150 million to US$178 mil- Emanuel and others 2004; McMillan and Conlon 2004). The lion estimated out-of-pocket R&D cost for the MMV and cost elements are similar to those for pharmaceutical R&D the GATB yields a total capitalized R&D cost of roughly except for the specific regulatory procedures for vaccines, such US$250 million for LIC drugs if they are developed by PPPs as the completion of plant construction before phase 3 trials. with foundation or government funding, assuming that in- As noted earlier, estimates indicate that an NCE costs kind contributions are at current levels and that trials are con- US$403 million to US$802 million in 2000 dollars (DiMasi, ducted in developing countries. Alternatively, these funders Hansen, and Grabowski 2003). Clarke (2002) estimates that a might choose to use a zero cost of capital, reflecting the impor- vaccine costs approximately US$700 million by the time the tance that they attach to developing new medicines to treat cur- product is marketed, including not only the actual costs of rently untreatable diseases and to replace existing drugs that are products, but also such items as the cost of failures and the cost increasingly ineffective because of resistance. In that case, the of funds (Grabowski 1997). In addition, the size of phase three appropriate capitalization cost is zero, and the out-of-pocket clinical trials has recently escalated along with costs. costs of US$150 million to US$178 million are the full R&D The Institute of Medicine (2004) estimates that total expen- costs per new compound for LIC diseases. diture on vaccine R&D in 1995 was US$1.4 billion. The large pharmaceutical companies accounted for approximately 50 Economics of Vaccine Discovery and Development percent of the total (Mercer Management Consulting 1995). In discussions of the economics of vaccine development, com- However, the current situation is more complex for vaccine paring the findings with those obtained for pharmaceuticals research than for drug R&D. In 2004, only five major multina- is useful. Note, however, that the two product categories are tional companies were investing in vaccine R&D and produc- different in many fundamental and practical aspects. tion (Institute of Medicine 2004). In addition, a multitude of Pharmaceuticals are used to treat an existing clinical condition smaller, new biotechnology organizations in both developed with the ultimate aim of reversing the course of disease. By and developing countries are pursuing multiple vaccine targets contrast, vaccines are used to prevent a future threat. In addi- that are of considerable value (Nossal 2004). Since September tion, pharmaceuticals may be administered over a prolonged 11, 2001, U.S. government funding for microbial threats that time frame and, in many chronic conditions, may be taken can be used as agents of terror has increased: Project Bioshield from the time of diagnosis for the rest of the patient's life, is devoting more than US$5 billion during the next 10 years to whereas most vaccines are administered once or a few times. discovering and producing vaccines and other therapeutics The costs of vaccine production consist of the traditional (Herrera 2004). These initiatives may have spillover benefits for components of discovery, process development, scale-up, and vaccines and therapeutics for developing countries. manufacturing, as well as the costs pertaining to regulatory Another barrier, in addition to complexity and costs that requirements, liability, and postlicensing studies (Andre 2002; may directly or indirectly affect investment in vaccine R&D, is Grabowski 1997). Furthermore, the economic framework for the condition of the vaccine market. Even though experts disease prevention (Kou 2002) raises many questions that are anticipate healthy growth in the total global vaccine market less clear than calculating the cost of treatment of a specific from approximately US$6 billion in 2004 to US$20 billion in pathological condition in an individual or setting priorities for 2009, the number of large private pharmaceutical compa- government budgets. Finally, the financing of vaccine purchas- nies involved in vaccine research is down to five (Mercer ing and immunization programs has traditionally been sepa- Management Consulting 2002). As a recent Institute of rated from the totality of health care financing. Although this Medicine report (2004) demonstrates, other significant barri- practice may have appeared to be advantageous at some point ers also stand in the way of a well-functioning vaccine research globally or in individual countries, the current outcome is less and production system. These barriers include the difficulties than satisfying in that the financing of vaccines is fragmented of entering the field and of financing research, plus in the (Institute of Medicine 2004) and competes at a less favorable United States they include the government's role in determin- level with other budgetary priorities. ing pricing in relation to the government's purchase of a signif- icant proportion of vaccines. Similar situations arise in other Costs of Vaccine R&D countries. All lead to an underappreciation of the value of vac- The decision to develop a new vaccine is usually based on med- cines and reduce the incentives for investment in future vaccine ical need, scientific feasibility, and market conditions. Because products. Product Development Priorities | 143 As noted earlier, whether the cost of R&D for drugs or vac- diagnostic methods. Progress has been made in securing ade- cines intended for use in developing countries is less than for quate drug supplies to treat or prevent diseases such as tubercu- products targeted to high-income markets is questionable. losis (TB), and in many instances, the most pressing need is for Certainly, developing vaccines for LICs requires investment improved diagnostics to ensure wider and wiser use of effective from both industrial and developing countries and participa- therapies. Thus, an urgent need exists to develop diagnostic tests tion by scientists from both industrial and developing coun- that are simple, cost-effective, and robust enough to be used in tries. In the case of vaccines, discovery similar to pharmaceuti- resource-constrained settings with endemic diseases. cals is a costly process. Therefore, a research infrastructure has to be supported in academic institutions and private sector and government laboratories for new ideas to emerge and to be Diagnostics Development Priorities tested. The capabilities needed to discover a new HIV or malaria for Developing Countries vaccine are different and far more complex than those used to The top priorities for developing new diagnostic methods per- manufacture traditional vaccines such as whole-cell pertussis. tain to HIV, TB, and malaria. In the field of HIV/AIDS, where Indeed, the technological know-how needed to discover new the goal is to simplify the diagnosis of HIV, the need is for a vaccines is embedded in the advancing edge of science. noninvasive, inexpensive, and simple but highly sensitive and Alternative mechanisms of financing and managing the specific HIV test for saliva, sputum, urine, or other body secre- development of new vaccines for the developing world must tions, as well as tests for monitoring highly active antiretroviral be identified and may require governmental, international, therapy. and philanthropic funding. Appropriate new institutions or In diagnosis of mycobacterium TB, the limited sensitivity alliances could evolve from the multiple PPPs now being pur- of microscopy and the diagnostic challenges posed by smear- sued. The case has repeatedly been made for a massive infusion negative, extrapulmonary, and pediatric TB emphasize the of funds and global coordination if vaccines against great killer need to find an alternative approach. In this context, the Gen- diseases such as HIV/AIDS are to be developed (Klausner and Probe Amplified Mycobacterium Tuberculosis Direct Test (Coll others 2003). and others 2003; O'Sullivan and others 2002) and nucleic acid amplifications assays, as well as serological tests (Perkins 2000), Effect and Cost of Vaccination Programs have great potential. Diagnosing latent mycobacterium TB The major societal and health effect of vaccines are realized infection using tuberculin skin testing has major limitations, mainly when immunization programs reach a significant including the inability to differentiate latent TB from active proportion of individuals in a society (Mahmoud 2004). The TB. The QuantiFERON-TB test (Mazurek and Villarino 2003), effect of vaccines in interrupting or preventing the transmission which was approved by the FDA for detecting latent mycobac- of infectious agents depends on two concepts: inducing resist- terium TB infection, and the MPB64 patch test (Perkins ance in healthy individuals before exposure and extending the 2000), a mycobacterial antigen test (Nakamura and others umbrella of prevention to the majority of the target population 1998) specific to the mycobacterium TB complex, are promis- to achieve herd immunity (Anderson and May 1990). When ing and should undergo further evaluation. deciding to mount a vaccination program, health professionals For specific diagnosis of malaria, the most useful approach face scientific, public health, and financial considerations. The would be a rapid test to determine whether patients who pres- ultimate outcome is a cost structure that has to compete against ent with fever have malaria. If this rapid test has the capability well-established budgetary constraints and comparisons. The of estimating parasite density, it may help predict those at subject of the cost-effectiveness of vaccination programs has higher risk of progression to severe disease or treatment failure. been examined at multiple levels and in many settings (Miller For the major noncommunicable diseases--for instance, and Hinman 1999). The overall conclusion derived from most cardiovascular diseases--portable imaging devices,such as radi- quantitative techniques--for example, cost-benefit analysis, ographic or ultrasound machines, are becoming the new stan- cost-effectiveness analysis, and cost utility and decision dard for diagnosis. Adaptation of these technologies to settings analysis--indicates that vaccination was one of the most effec- in developing countries is urgently needed. tive health measures of the 20th century (CDC 1999). Economics of Diagnostics R&D DIAGNOSTICS R&D for new drugs and vaccines poses major challenges in developing countries because of financial constraints and lack Evidence-based disease control strategies are now in place for of infrastructure. By contrast, both the timelines and the costs of most of the major infectious diseases affecting developing developing diagnostics are significantly lower even though the countries. Implementing these strategies depends on accurate process of developing diagnostics is in many respects similar to 144 | Disease Control Priorities in Developing Countries | Adel Mahmoud, Patricia M. Danzon, John H. Barton, and others the development of drugs or vaccines. Whereas the costs related closely with interested parties to stimulate interest; identify to clinical trials and the opportunity cost of funds are lower, the obstacles; and facilitate the development, evaluation, approval, process does have additional engineering requirements. For and appropriate use of new diagnostics for TB in LICs diseases with relatively large at-risk populations, large and small (http://www.who.int./tdr/about/resources/contributions.htm). biotechnology companies have been sufficiently attracted to Current activities include research on new diagnostic targets invest in diagnostic R&D and stand to generate adequate com- and methodologies, product development programs to facili- mercial returns even for inexpensive products. For less com- tate commercial and noncommercial R&D, and formal mon diseases or diagnostic indications, industry investment laboratory and field product evaluation trials. The Sexually has been minimal, and direct R&D investment by the Special Transmitted Diseases Diagnostic Initiative is a 10-year-old col- Programme for Research and Training in Tropical Diseases laborative project established in recognition of the critical need (TDR) (http://www.who.int/tdr) and other public sector agen- for improved diagnostic tools for common sexually transmit- cies or PPPs will be needed if products are to be developed. ted diseases. Its mission is to promote the development, evalu- Diagnostics activity in the TDR's Product Research and ation, and application of diagnostic tests appropriate for use in Development Unit currently focuses on two disease areas primary health care settings in developing countries, with a through work carried out by the TB Diagnostics Initiative focus on syphilis, chlamydia, and gonorrhea. (http://www.who.int/tdr/diseases/tb/tbdi.htm) and the Sexually Transmitted Diseases Diagnostic Initiative. This work is done in partnership with academic researchers, disease control Cost Estimates of Diagnostics R&D experts, public health officials from disease-endemic countries, No systematic estimates of the costs of developing diagnostics and industry. The TDR has recently invested substantially in its are available that are comparable to the studies for pharmaceu- capacity to support the clinical development and registration ticals. The costs of developing new diagnostics depend on the of new diagnostics and will work closely with industry, regula- type of tool; the duration from discovery to approval; and the tory agencies, and ministries of health in industrial countries technicalities involved in technology acquisition, patent fees, and disease-endemic countries to improve the quality and market research, laboratory and field trials, marketing and standardization of diagnostic trials and to facilitate the imple- product launch, and support costs. Table 6.1 summarizes costs mentation and appropriate use of proven technologies. As an related to the development of selected diagnostics for TB. Note example, the mission of the TB Diagnostics Initiative is to work that these are out-of-pocket costs and do not include the Table 6.1 Costs of Developing Selected TB Diagnostics (US$) Type of test Nucleic Drug susceptibility Item acid amplification Screening test testing Location of company Rest of world United States and United States and United States and United States and European Union European Union European Union European Union Market research costs 10,000 100,000 50,000 500,000 500,000 Technology acquisition and patent fees 275,000 250,000 50,000 200,000 Development of prototype 3,775,000 4,000,000 4,662,000 2,825,000 Consumables used during development 1,575,000 75,000 150,000 Scale-up and validation 600,000 200,000 200,000 Total product development costs 575,000 5,625,000 4,850,000 4,987,000 3,375,000 Total costs of clinical trials 180,000 1,450,000 2,000,000 294,000 (location of study sites) (disease-endemic (United States and (United States and (disease-endemic countries) European Union) European Union) countries) Regulatory approval costs (agencies) 100,000 800,000 454,000 (FDA, European (United States) Union) Marketing and launch support costs 80,000 1,500,000 200,000 Product support costs for one year 50,000 1,125,000 20,000 Total 995,000 10,600,000 7,574,000 5,781,000 3,875,000 Source: http://www.who.int.tdr/about/resources/default.htm. Product Development Priorities | 145 opportunity cost of capital (http://www.who.int./tdr/about/ Even with pricing at marginal cost, medicines may still be resources/contributions.htm). unaffordable for the poorest populations, particularly for drugs with high manufacturing costs, in which case additional subsi- dies may be necessary. However, the important conclusion is FINANCING AND INSTITUTIONAL that for drugs for global diseases, the existence of a market in ARRANGEMENTS FOR NEW PRODUCT industrial countries attracts private sector investment in R&D; DEVELOPMENT thus, differential (tiered) pricing provides a finance mechanism for developing new drugs that can achieve both dynamic effi- Research that contributes to the discovery and development of ciency (appropriate incentives for R&D) and static efficiency drugs, vaccines, and diagnostics occurs in public, private, and (appropriate incentives for use of existing products) (Danzon mixed settings, each with different funding mechanisms. and Towse 2005). By contrast, for drugs and vaccines that target diseases that occur predominantly in LICs, no HIC market exists in which to Public Sector recoup the costs of R&D, and patents and differential pricing In most high-income countries (HICs), government funding will not suffice to attract R&D for products that cannot expect from tax revenues is generally targeted to basic research--that is, to generate sufficient revenue to cover their development costs. research that advances understanding of underlying disease In 2002, annual per capita spending on drugs alone in member processes but is unlikely to yield commercially viable products states of the Organisation for Economic Co-operation and in the near term. The research may be done in government insti- Development was US$279, while developing countries typically tutions or in academic and other not-for-profit research institu- spent less than US$20 per capita for all health services (Sachs tions. Governments also stimulate private sector R&D through 2001; Troullier and others 2002). Per capita health spending on tax credits. drugs by the poorest individuals, who may be the majority of patients for communicable diseases, is even lower. Thus, for products that target LIC-only diseases, even if millions of Private For-Profit Sector patients are in need, expected revenues are insufficient to attract Applied research that targets specific products is generally private sector investment for developing new products without undertaken by the private sector using equity financing. Firms additional public subsidies. that rely on equity financing must provide a return to their For HIV/AIDS, even though the majority of the disease investors comparable to returns on other potential invest- burden is in LICs, the markets in HICs have been sufficient to ments, hereafter referred to as a competitive return. This attract private sector companies to develop several drugs and requirement applies for multinational pharmaceutical compa- to undertake considerable investment in an AIDS vaccine, nies, for biotechnology firms, and for firms in developing albeit with little success to date. In 2001, the GATB increased countries, unless they receive public subsidies. Start-up firms in the estimated size of the TB market from US$150 million HICs generally rely on equity capital from venture capitalists to US$450 million per year, with the potential to grow to and other private investors, whereas established firms issue US$700 million per year (GATB 2001). This amount is within shares in the broadly based public equity markets but finance the range normally considered necessary to attract private most of their R&D from retained earnings on existing prod- investment. However, estimated potential revenues for anti- ucts. The need to provide a competitive return to shareholders malarials and treatments for other LIC diseases are still well means that commercial firms can invest only in products that below this threshold. In addition to the limited ability to pay, they expect will generate sufficient revenues to cover all costs, some developing countries still lack the health care infrastruc- including the costs of R&D. In practice, commercial firms have ture necessary for conducting clinical trials and for delivering focused on products with a potential market in industrial medicines and vaccines effectively, which further reduces countries because of their residents' ability to pay prices suffi- incentives for R&D investment. cient to cover costs. Given the low potential revenues and lack of necessary infrastructure, R&D for tropical diseases and TB for the past Differential, or "Tiered," Pricing. For global products--that 25 years has been far less, relative to need, than for global dis- is, products targeting diseases that occur in all countries, such as eases. The number of NCEs per million disability-adjusted life cardiovascular diseases--revenues generated in HICs and in the years lost (a proxy for research relative to need) was 0.55 for more affluent sectors of middle-income countries are sufficient infectious and parasitic diseases but roughly 1.25 to 1.44 for to recoup the investment in R&D to the extent that, ideally, cardiovascular system diseases (Troullier and others 2002). prices in LICs need to cover only the incremental or marginal Between 1975 and 1999, just 16 of the 1,393 NCEs registered costs of production for these countries. were for tropical diseases or TB (Troullier and others 2002).5 146 | Disease Control Priorities in Developing Countries | Adel Mahmoud, Patricia M. Danzon, John H. Barton, and others Several of these products were fortuitous by-products of com- Fight AIDS, Tuberculosis, and Malaria, is allocated primarily to mercial research efforts initially intended for the oncology or paying for vaccinations and treatment. By paying for vaccines veterinary market (Ridley 2003). and drugs, such financing could provide additional revenues Multinational companies appear to be showing some signs to suppliers of these products and, hence, stimulate R&D. of increasing their investment in tropical disease R&D. For However, for the financing of vaccines and therapeutics to example, GlaxoSmithKline, AstraZeneca, and Novartis have serve as an effective pull mechanism for future R&D, such recently announced or established research centers devoted to financing must be sustained and must pay originators enough tropical disease. AstraZeneca's facility in Bangalore, India, will that they can recoup the costs of R&D. Thus, purchasers such focus on TB treatments and receive a commitment of person- as the United Nations Children's Fund or the Global Fund face nel and US$40 million in investment during 2003­8. The non- a tradeoff between paying the lowest possible prices so as to profit Novartis Institute for Tropical Diseases in Singapore is a maximize their ability to supply existing medicines to current US$122 million joint venture between Singapore and Novartis patients and paying somewhat higher prices so as to create that will focus on dengue fever and TB. GlaxoSmithKline has incentives for future R&D. established a research institute for TB and malaria in Spain Creating effective pull financing incentives for R&D is prob- ("Drugs for the Poor" 2003). ably best done by means of explicit purchasing commitments for specific products. Some progress has been made in identify- Orphan Drug Acts. Orphan drug acts provide additional ing the contractual and legal requirements of such commit- stimulus for private sector R&D for diseases that afflict only ments to enter into future contracts. The most promising small populations in HICs. The U.S. Orphan Drug Act grants candidates for initial implementation would be products or vac- orphan status to drugs to treat diseases that affect 200,000 or cines that are already in late stages of development or have been fewer patients per year in the United States. Orphan drug status approved for industrial countries but for which additional pur- provides additional R&D tax credits and seven years of market chasing commitments are needed to induce the investment nec- exclusivity, during which the FDA cannot approve another essary to undertake clinical trials and build the manufacturing drug to treat the same condition unless it uses a novel mecha- capacity required to extend these products to LICs. Possible can- nism of action. Such market exclusivity enhances the orphan didates are the pneumococcal vaccine and the rotavirus vaccine. drug's market power, enabling the developer to charge high For both these products, accelerated development and intro- prices that to some extent offset the low sales volumes, thereby duction plans have been created in the Global Alliance for covering the costs of R&D. The U.S. act has stimulated a sharp Vaccines and Immunization to address the many practical issues increase in the number of drugs developed to treat orphan con- surrounding the implementation of an advance purchase ditions since its passage. The European Union recently adopted contract. When advance purchasing commitments have been similar legislation. successfully demonstrated on products in the late stages of The potential for orphan status in the United States and the development, extending this promising approach to products at European Union may provide some additional stimulus for earlier stages of development may be possible. commercial firms to develop drugs and vaccines for LIC dis- eases, but the effects are likely to be minor for several reasons. First, after one product has acquired market exclusivity, firms Public-Private Partnerships have few incentives to develop other products to treat the In recent years, a growing number of initiatives involving part- same disease. Second, the value of orphan drug status in terms nerships between industry and government, nonprofit, and of annual revenue per patient is greatest for drugs to treat philanthropic organizations have been set up to stimulate trop- chronic diseases that require daily or weekly treatment. ical disease R&D. One of the oldest is the World Health Potential revenues for treatments for acute diseases, for which Organization, World Bank, and United Nations Development each patient needs only a short course of treatment, are likely Programme TDR, which has worked with industry, academia, to be smaller. Thus, though orphan drug acts may create some and research institutions to spur R&D and has contributed to additional stimulus for R&D for LIC diseases, other institu- half the new drugs developed for neglected tropical diseases tional and financing mechanisms are essential. Of these, PPPs during the past 25 years (Ridley 2003; Troullier and others are the most promising. 2002) (see table 6.2 for examples of the program's initiatives). The TDR is a relatively small program, with contributions of "Pull" Financing Mechanisms. Since the late 1990s, organiza- US$30 million in 2002. tions such as the Bill & Melinda Gates Foundation and the Since the late 1990s, increased government and foundation Rockefeller Foundation have increased their funding commit- funding, particularly from the Rockefeller Foundation and the ments to fight diseases in developing countries. This new fund- Bill & Melinda Gates Foundation, has stimulated the growth of ing, including funding coordinated through the Global Fund to product development PPPs, giving a "push" stimulus to R&D Product Development Priorities | 147 Table 6.2 Selected Initiatives of the TDR Disease Product Partner Status Uncomplicated malaria Lapdap GlaxoSmithKline Dossier submitted (2002) Lapdap, artesunate GlaxoSmithKline Phase 1 Pyronaridine, artesunate Shin Poong Preclinical trials Severe malaria Intramuscular artemether Artecef Registered (2000) Rectal artesunate Under discussion FDA approval letter received Visceral leishmaniasis Miltefosine Zentaris Registered (2002) Paromomycin Institute for OneWorld Health Phase 3 studies Sleeping sickness Intravenous eflornithine Aventis Registered (2001) River blindness Ivermectin Merck Registered (1989) Source: Ridley 2003. (Varmus and others 2004). According to the Initiative on grants of US$50 million or less, with significantly larger Public-Private Partnerships for Health, about 20 PPPs were amounts for the International AIDS Vaccine Initiative and involved in product development as of 2004. Although a few the Malaria Vaccine Initiative. Several of the organizations rely focus on a specific project, most adopt a portfolio approach heavily on the Bill & Melinda Gates Foundation and the with multiple candidates. The latter include five targeting Rockefeller Foundation for both their initial and continued HIV/AIDS vaccines or microbicides; three working with funding (Widdus 2004). Note that the dollar funding amounts malaria therapeutics or vaccines; three investigating TB thera- shown exclude in-kind contributions from industry and other peutics, vaccines, or diagnostics; and at least six targeting drugs sources, whose worth is difficult to calculate because the value for other neglected diseases (Widdus 2004). to the PPPs is presumably greater than the cost to the donor. The PPPs are heterogeneous in terms of their objectives, Table 6.4 shows the product development PPPs' portfolios structure, and financing. In general, their goal is to develop of products as of early 2004. The percentage of products still in products for use in developing countries with a public health preclinical development is higher for vaccines than for drugs, rather than a commercial goal. Their sources for promising which may reflect the scientific challenges of developing vac- compounds include modifications of existing compounds; cines for LIC diseases. In general, comparing funding amounts continued development of compounds previously abandoned with the number of products in development across PPPs is because of a lack of commercial potential; and totally new ini- inappropriate as an indicator of performance because the dif- tiatives coming out of academia, industry, or government lab- ferent PPPs target different problems and have received varying oratories. If a PPP acquires a product from another firm, the in-kind contributions. Also, some products are modest exten- other firm typically retains patent rights in HICs and middle- sions of existing therapies, whereas others are more innovative income countries, and the PPP commits to noncommercial and, hence, more risky approaches. pricing in developing countries. As table 6.3 shows, aggregate committed funding for the PPPs draw on financing from foundations and, to a lesser product development PPPs as of early 2004 was US$1.2 billion, extent, from governments. They work closely with private excluding in-kind contributions. A comparison of these fund- industry, including large pharmaceutical and biotechnology ing amounts to the costs per NCE suggests that current rates firms, obtaining a range of in-kind contributions, including of investment will produce some progress, but not rapid promising compounds; useful technologies; patent rights; and advances. Assuming optimistically that future funding for PPPs expertise and advice on discovery, clinical trials, manufactur- will be US$300 million per year, that private industry will ing, market estimation, regulatory requirements, and so on. invests similar amount, and that other sources will provide They operate largely as "virtual" firms, usually contracting out US$100 million a year (all of which are probably generous esti- actual operations to other firms or to contract research or serv- mates) would imply total investment of US$700 million per ice organizations. As compounds move into human trials, PPPs year. If this level of investment were sustained over time, it must also liaise closely with disease-endemic countries regard- might result in two or three NCEs per year, using the conserva- ing clinical trials, regulatory requirements, and product deliv- tive cost estimates of US$200 million to US$300 million per ery. Thus, they face significant scientific, managerial, financing, NCE. This development level would be significant progress, and operational challenges. although it still leaves a large shortfall, given the number of dis- Table 6.3 lists the leading product development PPPs and eases for which no good treatment or vaccine is available and their committed funding as of early 2004. Several have received the threat of resistance developing to existing treatments. It is 148 | Disease Control Priorities in Developing Countries | Adel Mahmoud, Patricia M. Danzon, John H. Barton, and others Table 6.3 Selected Endeavors by Product Development PPPs, 2004 Number of people killed Number of Committed funds annually by new cases raised to date Disease the disease per year PPP Focus (US$ millions) HIV/AIDS 2,800,000 5,500,000 International AIDS Vaccine Initiative Vaccines 350 South African AIDS Vaccine Initiative Vaccines 45 International Partnership for Microbicides Microbicides 95 Microbicides Development Programme Microbicides 27 Global Microbicide Project Microbicides 64 TB 1,600,000 8,000,000 Global Alliance for TB Drugs 42 Aeras Vaccines 108 Foundation for Innovative New Diagnostics Diagnostics 30 Malaria 1,200,000 300,000,000­ Malaria Vaccine Initiative Vaccines 150 500,000,000 European Malaria Vaccine Initiative Vaccines 18 MMV Drugs 107 Dengue fever 19,000 20,000,000 Pediatric Dengue Vaccine Initiative Vaccines 56 Hookworm 3,000 -- Human Hookworm Vaccine Initiative Vaccines 20 Leishmaniasis 51,000 1,000,000­ Drugs for Neglected Diseases Initiative Drugs 11 (Institute for 1,500,000 and Institute for OneWorld Health OneWorld Health) Chagas disease 14,000 16,000,000­ Drugs for Neglected Diseases Initiative Drugs 30 (Drugs for 18,000,000 and Institute for OneWorld Health Neglected Diseases Initiative) Total 5,700,000 351,000,000­ n.a. n.a. 1,200 353,000,000 Source: Sander 2004. n.a. not applicable; -- not available. Table 6.4 Product Development PPPs' Portfolios, 2004 Number of products in PPP Preclinical trials Phase 1 Phase 2 Phase 3 Aeras 2 Drugs for Neglected Diseases Initiativea 3 European Malaria Vaccine Initiative Foundation for Innovative New Diagnostics 1 1 1 GATBb 10 0 1 0 Human Hookworm Vaccine Initiative 2 International AIDS Vaccine Initiative 2 2 1 Institute for OneWorld Health 3 1 1 International Partnership for Microbicides Microbicides Development Programme 1 1 MMV 14 in discovery 4 2 2 Malaria Vaccine Initiative 8 6 1 Pediatric Dengue Vaccine Initiative South African AIDS Vaccine Initiativec 6 0 Total 47 16 7 8 Of which drugs 26 8 5 8 Of which vaccines 19 8 2 0 Sources: Initiative on Public-Private Partnerships for Health survey, PPPs' Web sites, interviews. a. The Drugs for Neglected Diseases Initiative has two malaria drugs in phase 3 that are partly financed by the European Union. b. The GATB anticipates a portfolio of three phase 1 trials and expects its current phase 2 trial will enter phase 3 before 2007. Its portfolio also includes platform-related investments. c. The South African AIDS Vaccine Initiative does not yet have any of its own products at phase 1 but is collaborating on two projects that are at this stage. Product Development Priorities | 149 also far short of the Commission on Macroeconomics and unclear. Second, the presumption that patents result in subop- Health's target for 2006 of US$3 billion in R&D spending for timal drug consumption because of monopoly pricing ignores diseases in developing countries (Sachs 2001). the widespread prevalence of insurance in HICs and middle- income countries, so that, in practice, consumers face out-of- pocket prices that are already close to marginal cost. Third, and Industry in Developing Countries most important, is the difficulty of estimating the value of a Pharmaceutical firms in developing countries have tradition- product before its use in the market, because both positive fea- ally focused on the generic sector, making use of their expertise tures (additional uses) and negative features (side effects) may in engineering and other skills needed for efficient drug manu- be discovered. In addition, distortions in the amounts paid for facturing. More recently, the adoption of product patents has patent rights would distort incentives for R&D. Moreover, the created incentives for LIC firms to invest in R&D. For example, proposal would reduce originator firms' incentives to invest in India adopted product patents as of 2005, and several leading postlaunch improvements. generic firms are already developing new products. However, assuming that these firms will focus their efforts on developing drugs for tropical diseases would be a mistake. As for-profit INTELLECTUAL PROPERTY firms, they face similar incentives to those of commercial firms in any country, which means focusing on the global diseases The issue of intellectual property is involved in the debate that offer the greatest expected net revenues rather than dis- about the perceived conflict between patents and access. eases specific to LICs. Nonetheless, several policies might help target R&D efforts in these countries toward tropical diseases. These policies include collaboration with the product develop- The Role of Patents in Drug Development ment PPPs, provision of special government funding or tax Under a patent system, an inventor is entitled to a limited credits for products that target LIC diseases, and provision of monopoly for a period of time, typically 20 years. This exclusiv- subsidies targeting the development and scientific testing of ity may permit high prices and, consequently, an increased products derived from local products and other traditional economic return that serves as an incentive to develop new medicines. products. The system has worked quite effectively in the phar- maceutical area, where the incentives deriving from exclusivity have resulted in important new drugs. The first generation of Other Proposed Mechanisms for Increasing Affordability patients pays a higher price than subsequent generations, which In evaluating other proposals for making drugs or vaccines provides compensation for the large research costs involved in available in developing countries, distinguishing proposals to developing a new drug. When the patent expires, the price nor- stimulate new product development from proposals to increase mally falls as generic competitors enter the market. the affordability of existing drugs is critical. One proposal Even though this approach has been extremely successful in pertaining to affordability is that multinational companies the developed world, it does not generally work for products should voluntarily license production rights to LIC producers. for which the main market is limited to the developing world. Experience with generic markets across countries indicates that The total magnitude of the market in the developing world for necessary conditions for such out-licensing to reduce prices to products for HIV, malaria, TB, or less widespread diseases is consumers are (a) the existence of competition between multi- likely to be too small to provide an adequate incentive for the ple licensees, (b) the licensees having lower production costs private sector. This fact, together with the fact that patents are than the originator firms, and (c) a mechanism that prevents likely to result in higher prices, has raised important concerns middlemen and retailers from capturing any potential savings. in the developing world. In practice, these conditions may not be met. The more proba- ble scenario of licensing to only one local generic manufacturer is unlikely to reduce prices to consumers. The Drug Access Debate Another proposal is that governments should purchase The Agreement on Trade-Related Aspects of Intellectual patent rights, paying the originator firm the estimated value of Property Rights (TRIPS) entered into force on January 1, the drug (net of production costs) and then selling the product 1995.6 This agreement requires the members of the World to consumers at the marginal cost of production. This proposal Trade Organization (WTO), which include nearly all major has several disadvantages. First, because the government would trading nations, to live up to defined standards of intellectual presumably have to raise taxes to pay for the patents, the tax- property protection. TRIPS was part of a much broader inter- induced efficiency loss could offset any efficiency gain in the national trade package negotiated during the Uruguay Round, pharmaceutical market, so the net effect on efficiency is one of a series of international trade negotiations that have 150 | Disease Control Priorities in Developing Countries | Adel Mahmoud, Patricia M. Danzon, John H. Barton, and others taken place since World War II. The United States and involved and that the problem is not patents but the inade- European nations, which were the strong proponents of TRIPS, quacy of the countries' medical infrastructure. were responding to pressure from their pharmaceutical, copy- An area of convergence has begun to emerge in relation to right content, and trademark-based industries. differential pricing: prices should be lower in developing The pharmaceutical industry's concern was that a number nations than in developed nations, permitting pharmaceutical of developing nations had made deliberate decisions to deny firms to recover their research expenditures in the developed patent protection to pharmaceutical products and to grant pro- world while making products available at near marginal pro- tection only to processes for producing pharmaceuticals. These duction cost to the poor in the developing world. This differ- nations believed that inexpensive access to pharmaceutical ential pricing is justified because potential sales in poor nations products was so important that these products should not be are so small that the market provides only a minimal incentive: patented. In its 1970 patent law, for example, India excluded total sales in the poorest nations account for only about 1 per- pharmaceuticals from product patent protection, effectively cent of global pharmaceutical sales. The research-based phar- choosing to provide low-cost pharmaceuticals for its people at maceutical industry would prefer to achieve this differential the expense of eliminating incentives to create new products. pricing by means of a donation program or simply by charging This law was one of the reasons the Indian generic pharmaceu- different prices. Critics would prefer that the patent monopoly tical industry was able to evolve to make and market copies of not be available to raise prices in the developing world, thereby drugs that were still on patent in wealthier nations. Another opening up markets to local generic producers. concern for the pharmaceutical industry arose from the com- Movement toward agreement on differential pricing was pulsory license process, a legal process available in some reflected in the Doha Declaration on TRIPS agreement and nations to authorize the use of a patented technology under public health, reached at a November 2001 WTO meeting of some circumstances even over the patent holder's objection. In trade ministers. This declaration affirmed that TRIPS "should practice, compulsory licenses are rarely granted but are instead be interpreted and implemented in a manner supportive of used as a threat to negotiate lower prices for the technology or WTO members' right to protect public health and, in particu- pharmaceutical involved. lar, to promote access to medicines for all" (TRIPS, paragraph The United States was determined to change these laws 4, 2001). It affirmed the right of nations to use the exceptions and in TRIPS achieved important requirements for expanding to TRIPS to address public health concerns, specifically stating patent protection. The most important TRIPS provision that "public health crises, including those related to HIV/AIDS, relevant to pharmaceuticals is article 27, which includes a tuberculosis, malaria and other epidemics, can represent a requirement that "patents shall be available for any inventions, national emergency" and, thus, facilitate the right to use com- whether products or processes, in all fields of technology." pulsory licensing (WTO 2003). (U.K. Commission on Intellectual Property Rights 2002). The The Doha Declaration left an issue unresolved: the manu- clear intent of this language was to prohibit exclusions of phar- facture of drugs under compulsory license for nations that do maceutical products as in the Indian law. Article 31 established not have the capability to manufacture the drugs themselves. careful procedural limitations on when a nation could grant a The problem arises from the compulsory licensing article of compulsory license. As part of the political compromise, tran- TRIPS, which contains a provision, article 31(f), requiring that sitional provisions gave developing nations extra time to com- the manufacture of products under compulsory license be pre- ply with the treaty's requirements and also set up arrangements dominantly for the domestic market. Thus, a small Sub-Sahara for the remaining parts of patent terms to be made available for African nation clearly has the right to grant a compulsory products developed during the transition period. Because of license but may have no local industry able to manufacture the these transitional provisions, developing nations were not gen- product. If it asks a foreign firm to manufacture the product, erally required to provide product patents on pharmaceuticals that firm would be manufacturing the product primarily for until January 1, 2005 (a date that has since been extended to export, a violation of TRIPS. 2016 for the least developed countries). The Doha negotiators did not find a way to resolve this During the years following the entry into force of TRIPS, problem, and article 6 of the Doha Declaration called for mem- a substantial and bitter debate over access to pharmaceutical bers of the TRIPS Council (a group of national representatives) products in developing countries focused largely on access to to find a solution by the end of 2002. By that time, all member antiretroviral agents for HIV patients in Sub-Saharan Africa. A countries except the United States had agreed to a procedure group of nongovernmental organizations argued that patents for waiving article 31(f). The new agreement covered products on these drugs in the developing world raise the prices of the needed to address public health problems recognized in the products necessary to help such patients survive. The research- Doha Declaration, but the United States feared that it would be based pharmaceutical industry countered that many of the expanded to a variety of other products and was unwilling to relevant products are not covered by patents in the nations accept it. Finally, a compromise was reached in August 2003. Product Development Priorities | 151 The United States accepted the 2002 document, provided that many cases of efforts to avoid the problem by, for example, the General Council chairperson of the WTO made an appro- modifying the research; conducting the research offshore in priate parallel statement. The chair made the statement, which locations where the relevant patents are not in force; or, in included language that the agreement would be used "in good some cases, simply ignoring the patent. faith to protect public health" and not be "an instrument to pursue industrial or commercial policy objectives," and recog- nized the need to respond to the industry's concern that REGULATORY AND LIABILITY ISSUES products produced under this agreement would not be exported to major developed world markets (WTO 2003; see Developing and registering new products are generally lengthy also UNAIDS 2003). and complicated processes (Abraham and Reed 2002; Baylor This agreement represents a step forward for access and will and McVittie 2003; FDA 2004) that are regulated both at the certainly place pressure on the research-based pharmaceutical national level and, in some circumstances, at the international industry to provide products in the developing world at low level. The role of the regulatory system extends beyond the prices. It leaves several important problems only partly launch of a new product to manufacturing and compliance resolved, however. One is the need to prevent importation of standards and to postmarketing surveillance for clinical effects the low-priced products into the developed world. Such and potential untoward outcomes. For products that are imports would cut into the patent-protected market and affect intended to be deployed in global markets, manufacturers have incentives to develop new products. A second is political back- to comply with regulatory requirements in the country of ori- lash. When the general public becomes aware that a product is gin as well as the requirements of each country where the prod- available to the poor in a developing nation at a price far below uct may be marketed. One exception is the mutual recognition that which patients in developed nations must pay, the political systems used currently by European Union countries (Pignatti, backlash for the pharmaceutical industry in the developed Boone, and Moulon 2004). The situation may be different for world may be severe. products intended for use only in developing countries; how- Most important, resolving the legal problem of article 31(f) ever, for legal and liability reasons, manufacturers in developed does not resolve the economic problem. It confirms that there countries have refrained from working with two different sets will be no patent incentive for the development of drugs for of regulatory requirements. diseases endemic to the developing nations and that public The best example for illustrating this process is the FDA funds will be needed for this purpose. Such funds are currently (2004). Over the years, FDA regulations have developed into a inadequate. clear pathway. The process is initiated through an application by the manufacturer and a step-by-step approach toward licensing. The agency gets involved in every phase of the devel- The Research Tool Issue opment process and approves in advance the experimental Another important problem arises from the changing nature of design, assays, and endpoints for clinical trials. After it has col- medical research and of patenting practice. This is the research lected all the information, the agency examines the materials tool problem: many of the basic tools used in medical research submitted and reaches a decision. The FDA process extends are now themselves patented. For example, the research use of through regulating and approving marketing materials and certain genetically modified mice is patented in the United postlicensing collection of efficacy data and information about States, as are the uses of many gene sequences and protein crys- possible side effects. tal coordinates. In the case of the malaria antigen merozoite The FDA approval process differs somewhat for pharma- surface protein 1, some 39 patent families cover various aspects ceutical products and vaccines. One of the main differences is of the protein (U.K. Commission on Intellectual Property the obligation of vaccine manufacturers to prepare materials Rights 2002). for use in phase 3 trials in the final and approved production Such patents can significantly complicate research and facility. This requirement means that the firm must invest in make it more expensive. Each one that might affect a particular completing the manufacturing plant well ahead of launching a research program requires legal analysis to determine whether specific product, a process that can take three to six years. The it is valid and actually applies to the planned research program. regulatory process for vaccines also dictates batch release for If relevant, a license must be sought or the research program every batch ready for deployment in the marketplace. This part must be redesigned. The more patents are involved, the greater of the regulatory process, although it ensures quality control, the likelihood that a patent holder will refuse to grant a license adds to costs and to the timeline. or will demand an exorbitant sum. Even though Walsh, Arora, In 1996, the European Union adopted a centralized and Cohen's (2003) study finds no cases of research programs procedure for applications and approvals through the being canceled midstream because of this problem, it finds European Medicines Evaluation Agency and through a mutual 152 | Disease Control Priorities in Developing Countries | Adel Mahmoud, Patricia M. Danzon, John H. Barton, and others recognition process (Pignatti, Boone, and Moulon 2004). In Whether or how this trend in the United States will affect many ways, the procedure parallels the FDA process, with sev- the developing world is unclear. Europe has moved toward a eral differences reflecting the fact that the European Union liability system somewhat similar to that of the United States, consists of many countries, each with a country-based process but many developing nations may not have such a tort liability that remains as an alternative or an addition to the com- system. Even if they do not have such a system, groups partici- munitywide process. The International Conference on pating in pharmaceutical development might be sued in the Harmonization of Technical Requirements for Regulation of United States for harm occurring in the developing world. Pharmaceuticals for Human Use was established to achieve Doctrines exist that restrict such suits, but firms may fear that coordination of the process of drug development between these doctrines are insufficiently effective. Hence, recognizing industry, Japan, the United States, and the European Union the potential costs of protecting against liability and, at the (Abraham and Reed 2002; Ohno 2002). The conference's activ- same time, ensuring that products are designed and manufac- ities have improved understanding of the regulatory process tured to the highest standards will be important. and reduced duplication. In contrast, the absence of a unified or harmonized approach to product registration and approval at the global NOTES level adds multiple layers of complexity. National systems 1. The phase transition probabilities in DiMasi, Hansen, and consist of complex processes with differing thresholds and Grabowski (2003) and the overall success probability of 0.215, conditional interpretations and with changing requirements in addition on entering human trials, are estimated from a larger sample of 538 inves- tigational compounds first tested in humans between 1983 and 1994. to differing Global Manufacturing Program standards and 2. The size of trial required to estimate statistical significance depends enforcement. A number of recent attempts have been made on the magnitude of the drug effect; the extent of stratification within the to resolve the issue. First among these is the World Health total sample by patient age, condition, and so on; the required statistical confidence; and other factors. Organization's effort to expand its prequalification system, to 3. DiMasi, Hansen, and Grabowski's (2003) data for average cost and develop technical standards earlier in the approval process, and average number of patients are based on actual, retrospective cost data, to expand the availability of reference reagents for international whereas the GATB estimates are prospective estimates (best guesses) based on prior clinical trials for tuberculosis drugs in the United States and a sur- calibration (Milstien and Belgharbi 2004). These efforts aim at vey of clinical trial experts to determine administrative and data manage- injecting a higher level of quality control and transparency into ment costs. the global regulatory system. The effort may have the potential 4. The MMV estimate of in-kind contributions does not include the to provide a global process that transcends national borders. value of basic research conducted by universities and foundations from which it obtains its lead compounds. Similarly, commercial firms also ben- Such a process should provide a simplified, systematic, and dis- efit from such basic research and it is omitted from the DiMasi, Hansen, ciplined system that would reduce costs and speed up market and Grabowski (2003) estimates, so comparisons are not necessarily access for new products. biased by this exclusion. 5. Tropical diseases include parasitic diseases (malaria, African try- The issue of liability in relation to harm to individuals panosomiasis, Chagas disease, schistosomiasis, leishmaniasis, lymphatic receiving pharmaceutical products has been extremely signifi- filariasis, onchocerciasis, and intestinal nematode infections); leprosy; cant in U.S. product development. It is entirely appropriate for dengue fever; Japanese encephalitis; trachoma; and infectious diarrheal diseases. those developing new products to be sued if they are negligent 6. This section is based in part on Barton (2004). in their research or product development, but in some cases pharmaceutical firms have been sued for side effects of drugs that may have been unforeseeable or may not even have been REFERENCES the result of the product. This type of liability can be a barrier Abraham, J., and T. Reed. 2002. "Progress Innovation and Regulatory to product development. Although perhaps a less serious con- Science in Drug Development: The Politics of International Standard- cern since the 1993 Daubert v. Merrell Dow Pharmaceuticals Setting." Social Studies of Science 32 (3): 337­69. lawsuit in the United States, a case that has been interpreted to Agre, P., and B. Rapkin. 2003. "Improving Informed Consent: A Comparison of Four Consent Tools." IRB: Ethics and Human Research restrict the presentation to juries of evidence determined not to 25 (6): 1­7. be "scientific," the issue is still significant. It may also be part of Ahmad, K. 2001. "Brazil and USA at Loggerheads over Production of the reason the U.S. vaccine industry has shrunk significantly, Generic Antiretrovirals." Lancet 357: 453. and it has certainly affected the direction of investment, push- Anderson, R. M., and R. M. May. 1990. "Immunization and Herd ing it away, for example, from products such as vaccines that Immunity." Lancet 335: 641­45. are used in one or a few doses in healthy people toward prod- Andre, F. E. 2002. "How the Research-Based Industry Approaches Vaccine ucts used repetitively by those who already have a chronic dis- Development and Establishes Priorities." Developmental Biology 110: 25­29. ease (Institute of Medicine 2004). It, thus, provides pressure Barrett, R. J., and D. B. Parker. 2003. "Rites of Consent: Negotiating directly contrary to public health priorities, which emphasize Research Participation in Diverse Cultures." Monash Bioethics Review prevention and, therefore, the use of vaccines. 22 (2): 9­26. Product Development Priorities | 153 Baylor, N. W., and L. D. McVittie. 2003. "Changes in the Regulations for Jayaramann, S. 2003. "Indian Biogenerics Industry Emerges." Nature Vaccine Research and Development." In The Jordan Report 20th Biotechnology 21: 1115­16. Anniversary, 45­49. Washington, DC: U.S. Department of Health and Klausner, R. D., A. S. Fauci, L. Corey, G. J. Nabel, H. Gayle, S. Berkley, and Human Services. others. 2003. "The Need for a Global HIV Vaccine Enterprise." Science Biehl, J. 2002. "Biotechnology and the New Politics of Life and Death in 300: 2036­39. Brazil: The AIDS Model." Princeton Journal of Bioethics 5: 59­74. Kou, U. 2002. Guidelines for Estimating Costs of Introducing New Vaccines CDC (U.S. Centers for Disease Control and Prevention). 1999. into the National Immunization System. Geneva: World Health "Achievements in Public Health 1900­1999: Impact of Vaccines Organization, Department of Vaccines and Biologicals. Universally Recommended for Children in the United States, Lohray, B. B. 2003. "Medical Biotechnology in India." Advances in 1990­1998." Morbidity and Mortality Weekly Report 48 (12): 243­48. Biochemical Engineering/Biotechnology 85: 215­81. Clarke, B. 2002. Presentation for the National Vaccine Advisory Lyles, A. 2003. "Public-Private Partnerships at the National Institutes of Committee, Washington DC. Health from Discovery to Commercialization." Clinical Therapeutics Coll, P., M. Garrigo, C. Moreno, and N. Marti. 2003. "Routine Use of Gen- 25 (11): 2900­2. Probe Amplified Mycobacterium Tuberculosis Direct (MTD) Test for Mahmoud, A. 2004. "The Global Vaccination Gap" (editorial). Science Detection of Mycobacterium Tuberculosis with Smear-Positive and 305: 147. Smear-Negative Specimens." International Journal of Tuberculosis and Mazurek, G. H., and M. E. Villarino. 2003. "Guidelines for Using the Lung Disease 7 (9): 886­91. QuantiFERON®-TB Test for Diagnosing Latent Mycobacterium Danzon, P., and A. Towse. 2005. "Theory and Implementation of Tuberculosis Infection." Morbidity and Mortality Weekly Report Differential Pricing for Pharmaceuticals." In International Public Goods Recommendations and Reports 52 (RR-2): 15­18. and Transfer of Technology under a Globalized Intellectual Property McMillan, J. R., and C. Conlon. 2004. "The Ethics of Research Related to Regime, ed. K. Maskus and J. Reichman. Cambridge, U.K.: Cambridge Health Care in Developing Countries." Journal of Medical Ethics University Press. 30: 204­6. DiMasi, J. A., R. W. Hansen, and H. G. Grabowski. 2003. "The Price of Mercer Management Consulting. 1995. "Report on the United States Innovation: New Estimates of Drug Development Costs." Journal of Vaccine Industry." Report for the Department of Health and Human Health Economics 22: 151­85. Services, Washington DC. DiMasi, J. A., R. W. Hansen, H. G. Grabowski, and L. Lasagna. 1991. "Cost ------. 2002. "Lessons Learned: New Procurement Strategies for of Innovation in the Pharmaceutical Industry." Journal of Health Vaccines." Report for the Global Alliance for Vaccines and Economics 10 (2): 107­42. Immunization Board, Geneva. "Drugs for the Poor: Exotic Pursuits." 2003. Economist, January 30, p. 52. Miller, M.A., and A. R. Hinman. 1999."Cost-Benefit and Cost-Effectiveness Emanuel, E. T., D. Wendler, J. Killen, and C. Grady. 2004. "What Makes Analysis of Vaccine Policy." In Vaccines, 3rd ed., ed. Stanley Plotkin and Clinical Research in Developing Countries Ethical? The Benchmarks of Walter Orenstein, 1074­87. Philadelphia: W. B. Saunders. Ethical Research." Journal of Infectious Diseases 89: 930­37. Milstien, J., and L. Belgharbi. 2004. "Regulatory Pathways for Vaccines for FDA (U.S. Food and Drug Administration). 2004. "Biological Products, Developing Countries." Bulletin of the World Health Organization Bacterial Vaccines, and Toxoids: Implementation of Efficacy Review." 82: 128­33. Federal Register 69: 1­23. MMV (Medicines for Malaria Venture). 2002. Annual Report 2002. Galambos, L., and J. C. Sewell. 1995. Networks of Innovation. New York: Geneva: MMV. Cambridge University Press. Nakamura, R. M., M. A. Velmonte, K. Kawajiri, C. F. Anf, R. A. Frias, M. Garber, D. A., G. Silvestri, and M. B. Feinberg. 2004. "Prospects for an T. Mendoza, and others. 1998. "MPB64 Mycobacterial Antigen: A New AIDS Vaccines: Three Big Questions, No Easy Answers." Lancet Skin Test Reagent through Patch Method for Rapid Diagnosis of Active Infectious Diseases 4: 397­414. Tuberculosis." International Journal of Tuberculosis and Lung Disease 2 (7): 541­46. GATB (Global Alliance for TB Drug Development). 2001. The Economics of TB Drug Development. New York: GATB. Nederbragt, H. 2000. "The Biomedical Disciplines and the Structure of Biomedical and Clinical Knowledge." Theoretical Medicine and Grabowski, H. 1997. "The Effect of Pharmacoeconomics on Company Bioethics 21 (6): 553­66. Research and Development Decisions." Pharmacoeconomics 11: 389­97. Nossal, G. J. V. 2000. "The Global Alliance for Vaccines and Immunization: A Millennial Challenge." Nature Immunology 1: 5­8. Hansen, R. W. 1979."The Pharmaceutical Development Process: Estimates of Current Development Costs and Times and the Effects of ------. 2004. "A Healthier Climate for Funding of Vaccine Research." Regulatory Changes." In Issues in Pharmaceutical Economics, ed. Nature Immunology 5: 457­59. R. I. Chien, 151­87. Lexington, MA: Lexington Books. Ohno, Y. 2002. "ICH Guidelines: Implementation of the 3Rs (Refinement, Herrera, S. 2004. "U.S. Rejiggers Bioshield Bill." Nature Biotechnology Reduction, and Replacement)--Incorporating Best Scientific Practices 22: 792. into the Regulatory Process." Institute for Laboratory Animal Research Journal 43 (Suppl.): 595­98. Hilleman, M. R. 1996."Three Decades of Hepatitis Vaccinology in Historic Perspective: A Paradigm of Successful Pursuits." In Vaccinia, O'Sullivan, C. E., D. R. Miller, P. S. Schneider, and G. D. Roberts. 2002. Vaccination, Vaccinology: Jenner, Pasteur, and Their Successors, ed. "Evaluation of Gen-Probe Amplified Mycobacterium Tuberculosis S. A. Plotkin and B. Fantini, 199­209. New York: Elsevier. Direct Test by Using Respiratory and Nonrespiratory Specimens in a Tertiary Care Center Laboratory." Journal of Clinical Microbiology ------. 2000."Vaccines in Historic Evolution and Perspective: A Narrative 40 (5): 1723­27. of Vaccines Discoveries." Vaccine 18: 1436­47. Perkins, M. D. 2000. "New Diagnostic Tools for Tuberculosis." Hilts, P. J. 2003. Protecting America's Health: The FDA, Business, and One International Journal of Tuberculosis and Lung Disease 4 (12): S182­88. Hundred Years of Regulation. New York: Alfred A. Knopf. Pignatti, E., H. Boone, and I. Moulon. 2004. "Overview of the European Institute of Medicine. 2004. Financing Vaccines in the 21st Century. Regulatory Approval System." Journal of Ambulatory Care Management Washington, DC: National Academy Press. 27: 89­97. 154 | Disease Control Priorities in Developing Countries | Adel Mahmoud, Patricia M. Danzon, John H. Barton, and others Rappuoli, R., H. Miller, and S. Falkow. 2002. "The Intangible Value of U.S. Congress, Office of Technology Assessment. 1993. Pharmaceutical Vaccination." Science 297: 937­39. R&D Costs, Risks, and Rewards. OTA-H-522. Washington, DC: U.S. Ridley, R. 2003. "Product R&D for Neglected Diseases: 27 Years of Government Printing Office. WHO/TDR Experiences with Public-Private Partnerships." European Varmus, H., R. Klausner, E. Zerhouni, T. Acharya, A. S. Daar, and Molecular Biology Organization Reports 4: S43­46. P. A. Singer. 2004. "Public Health: Grand Challenges in Global Health." Sachs, J. 2001. Macroeconomics and Health: Investing in Health for Science 303: 168­69. Economic Development. Geneva: World Health Organization. Walsh, J., A. Arora, and W. Cohen. 2003. "Working through the Patent Sander, A. 2004. "The Emerging Landscape of Public-Private Partnerships Problem." Science 299: 1021. for Product Development." In Combating Diseases Associated with Webber, D., and M. Kremer. 2001. "Perspectives on Stimulating Industrial Poverty: Financing Strategies for Product Development and the Potential Research and Development for Neglected Infectious Diseases." Bulletin Role of Public-Private Partnerships, ed. R. Widdus and K. White, 79­80. of the World Health Organization 79: 735­41. Geneva: Initiative for Public-Private Partnerships for Health, Global Widdus, R. 2001. "Public-Private Partnerships for Health: Their Main Forum for Health Research. Targets, Their Diversity, and Their Future Direction." Bulletin of the Schmid, E. F., and D. A. Smith. 2002. "Should Scientific Innovation Be World Health Organization 79: 713­20. Managed?" Drug Discovery Today 7: 941­45. ------. 2004. "Historical Context: Why Public-Private Partnerships for Trouiller, P., P. Olliaro, E. Torreele, J. Orbinski, R. Laing, and N. Ford. 2002. Product Development Emerged and How." In Combating Diseases "Drug Development for Neglected Diseases: A Deficient Market and a Associated with Poverty: Financing Strategies for Product Development Public-Health Policy Failure." Lancet 359: 2188­94. and the Potential Role of Public-Private Partnerships, ed. R. Widdus and U.K. Commission on Intellectual Property Rights. 2002. Integrating K. White. Geneva: Initiative for Public-Private Partnerships for Health, Intellectual Property Rights and Development Policy. London: U.K. Global Forum for Health Research. Commission on Intellectual Property Rights. WTO (World Trade Organization). 2003. "The General Council UNAIDS (Joint United Nations Programme on HIV/AIDS). 2003. Chairperson's Statement," August 30. http://www.wto.org/english/ Accelerating Action against AIDS in Africa. Geneva: UNAIDS. tratop_e/trips_e/t_news_e.htm. Product Development Priorities | 155 Chapter 7 Economic Approaches to Valuing Global Health Research David Meltzer Health research has contributed tremendously to advancing recent advances in understanding about the value of improve- health and welfare throughout the world. Gains in health have ments in health and in approaches to the value and priorities of already been attained, and health research continues to have health research provide powerful theoretical tools to address great potential to contribute to the well-being of persons in these issues. both low-income and high-income countries. The "10/90" Although the application of these principles is still in its reports of the Global Forum on Health Research (GFHR) have infancy, these tools provide a valuable framework for determin- helped highlight that, although the burden of disease attributa- ing the value of health research relevant to low-income coun- ble to morbidity and mortality is greatest in developing coun- tries and for maximizing the value of research that aims to pro- tries, most health research has focused on the needs of devel- mote health in those countries. In the following section, recent oped countries and relatively little on the needs of lower-income innovations in determining the value of health research at an countries (GFHR 2003). Despite the relative dearth of health aggregate level are reviewed. The conclusion is that the value of research focused specifically on their most important health health research has been immense and is likely to increase dra- problems, lower-income countries have also benefited from matically in coming years, especially in lower-income countries. health research, often as much from research originally moti- In deriving these conclusions, not only health-oriented meas- vated to serve high-income countries as from research specifi- ures of improved health, such as life years, disability-adjusted cally designed to address the needs of low-income countries. life years (DALYs), or quality-adjusted life years (QALYs), were In the future, health research aimed primarily at addressing considered, but also how increases in income may increase the the needs of affluent countries is likely to continue to produce value of such improvements. In a later section, recent methods benefits for lower-income countries, even in the absence of that can be used to estimate the value of specific research proj- efforts to manage scientific research in the interest of lower- ects are reviewed. The application of such methods promises to income nations. Indeed, as lower-income countries increase in better identify the value of health research in specific instances, income, their health problems will likely come to resemble which can enhance the case for research spending to improve more closely the health problems of higher-income countries. health in developing countries and can increase the efficiency However, "trickle-down" approaches are not an efficient way of with which available research funds are spent. producing knowledge to advance the health and well-being of the populations of lower-income nations. Instead, a rational approach is needed that appreciates both the value of health in THE VALUE OF HEALTH RESEARCH lower-income nations and the potential for rational scientific management to efficiently allocate resources for scientific Health research can be valuable for a variety of reasons. The research. The work of the GFHR and other organizations in most obvious reason is simply the value that people place on this regard is an important step in this direction. Moreover, improvements in health. Those health improvements may be 157 reflected in increased length of life or quality of life. Other rea- Although both the connection of health research to health and sons that health research can be valuable include the improve- the valuation of health differ in important ways between lower- ments in productivity, decreases in medical care costs, and income and higher-income countries, these studies provide a greater ability to plan for and invest in the future when health useful framework for understanding the past and likely future improves. Health research can also sometimes decrease costs value of health research in lower-income countries. independent of any effects on health--for example, when a lower-cost treatment can replace a higher-cost treatment of similar efficacy. The Value of Improved Health A large body of work has attempted to assess and quantify Several authors have contributed to the literature on the value the economic value of health from these perspectives. The goal of improvements in health: Becker, Philipson, and Soares of some of this literature is a comprehensive assessment of the (2003); Cutler and others (1997); Murphy and Topel (2003); cost of illness (for example, Rice 1994). Such comprehensive and Nordhaus (2003). Although these studies differ somewhat assessments are often attempted in order to assess the cost- in the analytic framework they use, the basic form of their effectiveness of medical or public health interventions, which analyses is quite similar. In essence, they all use estimates of the are frequently measured in terms of quality-adjusted life value of longevity on the basis of revealed preference tech- expectancy (Gold and others 1996) or DALYs (WHO 1994; niques and multiply these values by the increases in longevity World Bank 1993). Other aspects of this literature focus on that have been observed to determine the value of improve- single dimensions of the economic value of health. For exam- ments in health. Revealed preference estimates of the value of ple, Ram and Schultz (1979) performed an early analysis show- life are most commonly derived by assessing the wage premium ing the effects of malaria eradication on productivity. Similarly, required by workers to accept riskier jobs. This idea dates back Meltzer (1992) attempted to identify the effects of mortality at least to the work of Adam Smith, but the modern treatment on investment in education, the demographic transition, and of this issue stems primarily from the work of Thaler and the onset of sustained economic growth. In that work, it was Rosen (1975). For example, if the lifetime earnings of a risky suggested that health can increase income because increasing mining job that has a 1 1,000 increased risk of death compared life expectancy raises returns to investment in education, in with another mining job is US$1,000, then the statistical value turn encouraging decreases in family size and further increases of a life would be said to be US$1,000,000, since paying 1,000 in investment in the education of children and a transition to workers would result in one extra death on average and require sustained human capital­based economic growth. These and that the employer pay wages equal to 1,000 workers other mechanisms, including shifts in the ratio of productive US$1,000/worker US$1,000,000. adults to dependents and changes in savings rates with increas- A simple example considering the value of increases in life ing life expectancy, have been suggested to produce positive expectancy in the United States from 1970 to 2000 illustrates effects of health on per capita income in both the short and the well how these sorts of estimates have been used to estimate the long run (Bloom, Canning, and Jamison 2004). value of increased longevity. From 1970 to 2000, life expectancy Although some evidence suggests the importance of all at birth for Americans grew by a little more than 5 years, from these effects, recent work on the value of health in developed about 75 years to about 80 years. With about 300 million countries has focused primarily on the value that people place Americans living during this period, this increase constitutes on improvements in their health. Such studies for the United about 1.5 billion life years. Using a conservative revealed pref- States have suggested that increases in longevity over the past erence estimate of about US$4 million as the statistical value of several decades are valued in the tens of trillions of dollars, a life (Viscusi and Aldy 2003), assuming a life expectancy of 75 indeed, contributing about as much to increasing welfare over to 80 years, and assuming discounting is somewhere between 0 the period as increases in per capita income. Subsequently, sim- and 3 percent annually, one gets a value of about US$50,000 to ilar findings have been reported for other countries, and cross- US$100,000 per life year saved. Multiplying this gain in life national studies have found that the value of growth in life years and the estimate of the value of a life year, one gets a value expectancy compared with the value of income growth has of increased longevity over this period of about US$75 trillion been even greater for low-income countries (Becker, Philipson, to US$150 trillion, or about US$2 trillion to US$4 trillion per and Soares 2003). These studies have been influential in devel- year. The average of these numbers translates into almost oping a broad consensus that the returns to health research are US$10,000 per person per year, which is about as large as the large and in producing dramatic increases in the budget for the increase in per capita income in the United States during this National Institutes of Health (NIH) in the United States. The period. In essence, this analysis is a simplified form of that per- conclusions of these studies in the United States have been formed by Murphy and Topel (2003). based on assessments of both the high value of health and the An alternative approach to obtaining a similar number may role that health research has played in improvements in health. help some readers better understand the intuition behind these 158 | Disease Control Priorities in Developing Countries | David Meltzer effects. Over this 30-year period, the 5-year increase in life gains in life expectancy of about 25 years since 1950 and a expectancy that occurred raised life expectancy by about two population in 2000 of about 1 billion persons, the gains in life months (or one-sixth year) per year. With each life year valued expectancy in India alone over the period are worth about at about US$75,000, these gains in life years are readily seen to US$400 trillion, or US$8 trillion per year, equivalent to about be worth about US$10,000 per person per year. US$800 per person per year. Understanding the value of improvements in health in Thus, as large as the US$75 trillion to US$150 trillion esti- lower-income countries requires data both on increase in mate for the health gains for the United States is, such gains are longevity and on how such increases in longevity might be val- only about one-fourth as large as the gains for India. The value ued. Data on increases in longevity are relatively easy to obtain of the gains for India is larger for several reasons. One is just the for most countries and vary substantially across countries over longer period considered--30 years for the United States com- time. For example, many countries that had low income in pared with 50 for India--although, in fact, gains in the United 1950, including the Arab Republic of Egypt, Mexico, and States were small enough between 1950 and 1970 that the Thailand, have had an increase in life expectancy since 1950 of longer period explains relatively little of the difference. Much more than 20 years, or about three months per year, which is larger factors are the fivefold greater increase in life expectancy about 50 percent greater than the rate of growth in life in India than in the United States over the period (25 years expectancy in the United States. Although some countries that compared with 5 years) and a population about five times as had low incomes in 1950 and did not experience much eco- large. Thus, if health gains are valued in U.S. dollars, the value nomic growth during this period did not increase in life of such gains for India over this period far exceeds that for the expectancy over this period, the vast majority of countries with United States. Given the 5:1 size of India and the gains in life low incomes in 1950 experienced increases in life expectancy expectancy, the ratio of those gains would be more like 25:1 far exceeding the rate of increase in the United States and other than 4:1 (US$400 trillion as opposed to US$75 trillion to developed countries. This fact suggests that the potential for US$150 trillion) were it not that the statistical value of life in increases in life expectancy is greater for lower-income coun- the United States is greater than that in India. Even if one does tries than for higher-income ones, so the increase in life not feel comfortable assigning differential valuation to lives expectancy for lower-income countries might be expected to across countries, the fact that the valuation of gains in India is slow in coming years. If the rate of change in the value of health so much larger than the valuation of the gains in the United gains were determined primarily by the rate of increase in life States using this method provides a powerful reminder of cer- expectancy, growth in the value of health in lower-income tain basics of the value of improving health in lower-income countries would be expected to slow in the future. versus higher-income countries: many more people live in low- Assessing the importance of changes in life expectancy income than in high-income countries, and the potential for relative to changes in the value of increased life expectancy in large increases in life expectancy is thus far greater. influencing the value of health requires data on the value of Such statistics provide some insights into the value of his- increased longevity. Such data on the value of mortality reduc- torical improvements in health, but in terms of considering tions are limited but do exist. Viscusi and Aldy (2003) reviewed future investments, considering what can be said about the the literature on the statistical value of life across countries that growth of the value of health in lower- versus higher-income vary in income. One key finding is that the statistical value of countries in the future is useful. To do so, one finds it useful to life increases with income, with an elasticity of approximately decompose growth in the value of health into growth in life 0.5 to 0.6, so that a 100 percent increase in income leads to a expectancy and growth in the value of longevity (growth in 50 to 60 percent increase in the statistical value of life. the value of life years). We begin with VL VLY LY and dif- One implication of this finding is that economic growth is ferentiate to obtain the finding that growth rates can be likely to result in increases in welfare through improved health, decomposed to find that gVL gVLY gLY, so that the growth in both because economic growth improves health and because the value of life is the sum of the growth in the value of a life the value of those health improvements increases as income year and growth in life expectancy. increases. This statement can be assessed formally by analyzing To understand the meaning of this relationship, one finds change over time in the value of life (VL) VLY LY, where it useful to consider the Viscusi and Aldy (2003) estimate of VLY is the value of a life year, and LY is life expectancy. Such the elasticity of the value of life with respect to income. numbers easily show the value of health gains that have been Substituting this elasticity income in the above equation, achieved. For example, India is the poorest country for which gVL 0.5 gY gLY, where gY is the growth rate in per capita calculations of revealed preference estimates of the statistical income. The relative magnitude of these two components is value of life have been done (Viscusi and Aldy 2003). This instructive. In developed countries such as the United States, study suggests a statistical value of life of US$0.6 million, or where growth in life expectancy has been on the order of two about US$10,000 to US$20,000 per life year saved. With months per year from a base of 75 or so years, the growth in Economic Approaches to Valuing Global Health Research | 159 life expectancy is (2 12) 75 0.2 percent per year. This rate is valuations clearly suggest that the potential value of health substantially smaller than the component of increased value of research for these countries is immense. That this discussion life related to growth in the valuation of gains in life, even with has focused only on the value of mortality reductions and neg- long-run real growth rates of only 2 to 3 percent, so the com- lected those of reductions in morbidity even further reinforces ponent attributable to growth in the value of life is 0.5 (2 to the potential gains from health research. 3 percent) 1 to 2 percent per year. In lower-income countries Not only is the magnitude of the value of these improve- in which growth in life expectancy is greater and baseline life ments in health immense in regard to how people value being expectancy is lower, life expectancy growth may be more healthier, but the total value of health also includes increased important but is still not likely to exceed the effects of growth productivity, decreased medical care costs, and increased ability in per capita income. For example, Mali experienced dramatic to plan for and invest in the future when health improves. growth in life expectancy of 1 percent per year (from 30 to 45 Although these effects are often not easy to quantify, they may years from 1950 to 1990), but growth in per capita income in prove to be important components of the total economic value Mali from 1980 to 2000 was about 3 percent annually, translat- of improvements in health. ing into effects on the value of life of 1.5 percent annually. More striking perhaps is China, where growth in life expec- The Connection between Gains in Health tancy from 40 to 70 years from 1950 to 1990 represented an and Health Research increase of about 2 percent per year, whereas growth in per capita income (from 1980 to 2000) was about 8 percent per To move from the finding that gains in health are highly valued year (translating into effects on the value of life of about 4 per- to the finding that health research is highly valued, gains in cent per year). Two implications are immediately obvious: health must be connected to health research. Making such asso- ciations is difficult even in high-income countries, but it has · First, the growth in the overall value of health over the been done with some success. For example, Cutler and Kadiyala period has been large--about 2 to 6 percent annually for (2003) argue that the major recent gains in life expectancy in the these initially low-income countries. United States have come from reductions in cardiovascular dis- · Second, the major driver of the value of health over time has ease, about two-thirds of which can be tied to advances that have not been an increase in health per se but, instead, increases resulted from medical research as opposed to secular trends in in income that produce increases in how health is valued. nonmedical factors that promote health, such as per capita This finding is likely to be even more true as life expectancy income and education. Similar calculations for lower-income increases so that continuing large increases in life expectancy countries have not been done. Nevertheless, although growth in become more difficult to achieve. income and education seem more likely to be important in pro- ducing increases in life expectancy for such countries, the major As noted previously, estimates such as these of the statistical role of reductions in childhood infectious disease in recent gains value of life can understandably be criticized on the basis of in life expectancy suggests the great value of the research that has associating a greater value with saving lives of wealthier persons produced such innovations as childhood immunizations, than saving lives of poorer persons. Nevertheless, it is striking improved sanitation, and oral rehydration therapy. Similarly, to note that in India--the only relatively low-income country the high burden of cardiovascular disease in lower-income for which data on the statistical value of life are available--the countries will be expected to decline in the coming years if the total value of recent health gains far exceeds even the immense benefits of research can be applied in those settings. gains in the United States. Data on the value of life for other Such conclusions speak to the value of health research as a low-income countries would help reinforce this finding, but whole for lower-income countries, and they reflect the gains the combination of the greater potential for large increases in that come both from research done initially to benefit higher- life expectancy and for greater growth in per capita income in income countries and from research done to benefit lower- low-income countries in the future suggests that the total value income countries. Although both of these classes of research of health gains for such countries may far exceed those of may produce benefits for lower-income countries, discussion higher-income countries. Indeed, it bears emphasis that, in of their differences is in order. In the case of research done ini- attempting to value health in monetary terms rather than tially with the needs of high-income countries in mind, clearly merely health terms, such as life years or QALYs or DALYs, an some types of research benefit low-income countries more important new avenue is opened by which the value of health than others. For example, research that has demonstrated the research can increase over time even as gains in life expectancy power of relatively inexpensive medications to reduce mortal- slow with increasing life expectancy, as they have in developed ity from cardiovascular disease (for example, aspirin in acute countries. Despite the concern that economic approaches myocardial infarction) has much greater potential to produce undervalue the health of persons with lower incomes, such large benefits for lower-income countries in the foreseeable 160 | Disease Control Priorities in Developing Countries | David Meltzer future than does research that advances techniques for expen- for Development. They include the essential national research sive acute treatments, such as cardiac catheterization. In this approach developed by the Council on Health Research for particular case of cardiovascular disease, estimates summarized Development and approaches developed by the Ad Hoc by Cutler and Kadiyala (2003) suggest that higher-income Committee on Health Research, the Advisory Committee on countries have also benefited more from research that Health Research, and the GFHR's own Global Forum combined advanced such low-technology treatments than from research matrix approach. The 2004 report compares these approaches on high-technology treatments. along several dimensions that include the following: Such patterns may not be maintained across other health conditions, but they do hold out the tantalizing thought that 1. Objective of priority setting health research initiated by high-income countries in their own 2. Focus at global or national level interest may sometimes be directed toward work of greater 3. Strategies or principles (especially relating to the process for value if issues of cost and generalizability to lower-income set- participation by stakeholders) tings are reflected more greatly in decisions about research pri- 4. Criteria for priority setting: orities. When research directed toward the needs of lower- · Burden of disease income countries is considered, cost and feasibility must, of · Analysis of determinants of disease burden course, receive great attention. However, the immense reduc- · Cost-effectiveness of interventions (resulting from tions in mortality that have come from advances in knowledge proposed research) about public health efforts--such as efficient and effective san- · Effect on equity and social justice itation, immunization, and oral rehydration programs, all of · Ethical, political, social, and cultural acceptability which were developed primarily with the needs of low-income · Probability of finding a solution countries in mind--suggest the value of research targeted · Scientific quality of research proposed toward the needs of lower-income countries. Rigorous studies · Feasibility (availability of human resources, funding, of the value of these innovations akin to those studies done to facilities) assess the value of health research for the United States are 5. Contribution to capacity strengthening. sorely needed to inform potential funders of the potential returns to such research. In addition to describing the dimensions considered in these approaches to priority setting, the 2004 report also sum- marizes several efforts to apply these approaches in individual METHODS OF ASSESSING THE VALUE OF HEALTH countries or for individual diseases or disease areas (see, for RESEARCH PROJECTS example, Remme and others 2002). The priorities identified include many of those identified in chapter 4 in this volume. Although the aggregate value of health research has unques- Examining these dimensions of priority setting, one finds tionably been large, not all health research projects are suffi- that some are intrinsically qualitative in nature but quantitative ciently valuable to justify their costs. Some of these latter analysis also sometimes plays an important role. For example, projects are inevitable because the outcome of research is intrin- several approaches, including the GFHR's combined matrix sically unpredictable. However, even if one abstracts from such approach, use DALYs to quantify the number of healthy years uncertainty by considering the expected returns from a project of life lost to the disease toward which the research is directed. before one embarks on it, research projects clearly vary in their Looking forward, one may find additional quantitative likely value relative to their cost. Such assessments are obvious- approaches helpful in assessing the value of research. For exam- ly difficult to make, yet they are routinely done as part of ple, new methods based on value-of-information (VOI) tech- the process of deciding the allocation of research funds. These niques have begun to be proposed to better inform such assess- assessments have traditionally been done rather informally, ments. In essence, VOI techniques model the uncertainty in the even in the most closely structured settings in which research outcomes of research and the value of research contingent on funding is sought, such as the NIH study sections in the United that uncertainty in order to assess the expected value of the States that review research proposals for NIH funding. research. For example, a research project that has a 5 percent An active discussion over the past decade has helped to iden- chance of success that it would produce health gains worth tify valuable strategies for priority setting in global health US$100 million and a 95 percent chance of failure, would have research and determine their implications for research priori- an expected value of (0.05 US$100 million) (0.95 ties in the field. The 2004 report on health research of the GFHR US$0) US$5 million. Although these techniques have rich summarized the major contributions to this discussion (GFHR roots in statistical and economic theory, they are often difficult 2003). The approaches that are described date back to as early to apply well because of the difficulty in valuing health out- as 1990 with the work of the Commission on Health Research comes and modeling the uncertain outcomes of research. As a Economic Approaches to Valuing Global Health Research | 161 result, these methods often can provide only loose bounds on how best to apply it in a given setting. With the limited funds the value of a research project (Meltzer 2001). Indeed, in the that have traditionally been available for research in lower- worst-case scenarios, in which the available information about income countries and the high value that appears to be placed the likely outcomes of research is essentially uninformative, on improvements in health in these countries, VOI studies, as VOI techniques only reproduce information on the burden and discussed previously, seem likely to frequently show that the costs of illness. Nevertheless, in settings where more structure benefits of research in lower-income countries far exceed can be put into the problem, these techniques have begun to be the costs. It is hoped that such findings may be used to increase applied successfully to specific research questions, such as the the pool of funds available for research in these settings. prioritization of research in Alzheimer's disease (Claxton and However, even if VOI techniques do not lead to increases in the others 2001). In general, the more remote the connection of the total funding available for research in low-income countries, research to health outcomes, the harder it is to gain meaningful they may be helpful in informing resource allocation decisions, information from a VOI approach. Thus VOI approaches are given available resources. In the context of a fixed research far better at providing information on the value of applied budget in which some research projects whose expected benefit research than they are at providing information on the value of exceeds their costs could still not be funded, rigorous data on basic research. the high rates of return on projects that did not make the fund- For lower-income countries, this ability to illustrate the ing line would then serve to highlight the returns to greater value of applied work may be especially valuable because often research in these settings. Because VOI calculations need not be a technology has already been developed and the question is complex (see box 7.1), the most important barriers to their Box 7.1 Value of Information Although VOI calculations can be quite complex (for suggests that a research program costing up to this example, Claxton and others 2001), they need not be. For amount would be worthwhile if it could be expected to a researcher seeking funding, much simpler calculations have this likelihood of success and percent efficacy. Even if may be sufficient to build a compelling case for funding. those probabilities could not be reliably estimated, mini- For example, quantifying the burden of illness in life years, mum values could be calculated at which the research pro- QALYs, or DALYs lost and converting this into a burden or gram would continue to be worthwhile. cost-of-illness measure in economic terms using an esti- As in their use in affluent countries, such calculations are mate of the statistical value of life may even be sufficient. likely to suggest that health research is of great value, often To take a simple example, imagine a research program in being predicted to return value many times its cost even a country with a statistical value of life of US$0.6 million accounting for uncertainty. In the preceding example, for (for example, the estimate for India) for a disease that cur- instance, even an immense US$100 million research pro- rently kills 1,000 people per year. If one abstracts from gram would have an expected return 20 times as large. Such issues such as the age of death and how one might differ- results would reflect the immense value in health research entially value deaths at different ages, this would have a but would also suggest caution: with limited research funds, potential value of US$0.6 billion per year. If the research not all programs whose benefits exceed their costs should study were viewed as having a 10 percent chance of pre- be funded if more valuable projects remain unfunded. In venting 50 percent of those deaths, its expected value this area, funding agencies can be helpful to the research would be P(success) percent efficacy statistical value community by highlighting how they have used calcula- of life 0.1 0.5 US$0.6 billion US$30 million per tions such as these and publicizing the ratio of research year. Potential annual productivity gains produced by a costs to returns for those projects they have been able to successful treatment and annual costs of the treatment fund. Even when a project is not funded despite a favorable could be added to, or subtracted from, these US$0.6 ratio of expected returns to research costs, reporting the billion annual benefits as well. One could go further, argu- missed opportunity will serve as a reminder to the world ing that the research would be valuable into the future, community of the potential of health research for lower- and discount an infinite stream of those annual returns. At income countries that remains unrealized and thereby, a discount rate (R) of 10 percent, this stream would have perhaps, promote greater investment in health research to a value of US$20 million/R US$200 million. This result benefit the people of these countries. 162 | Disease Control Priorities in Developing Countries | David Meltzer useful application to research in developing countries are the Alzheimer's Disease." International Journal of Technology Assessment in lack of awareness and acceptance of these methods on the part Health Care 17 (1): 38­55. of researchers and funders focused on these settings. Cutler, D. M., and S. Kadiyala. 2003. "The Return to Biomedical Research: Treatment and Behavioral Effects." In Measuring the Gains from Organizations concerned with international health can exercise Medical Research: An Economic Approach, ed. K. M. Murphy and valuable international leadership by exposing these researchers R. H. Topel, 110­62. Chicago: University of Chicago Press. and funders to these methods and encouraging their use. Cutler, D. M., E. Richardson, T. E. Keeler, and D. Staiger. 1997. "Measuring the Health of the U.S. Population." Brookings Papers on Economic Activity: Microeconomics 21782, Brookings Institution, Washington, DC. CONCLUSIONS GFHR (Global Forum for Health Research). 2003. The 10/90 Report on Health Research 2003­2004. Geneva: GFHR. Improvements in health in recent decades have been of great Gold, M. R., J. E. Siegel, L. B. Russell, and M. C. Weinstein. 1996. Cost- value in both high- and low-income countries. Furthermore, Effectiveness in Health and Medicine. New York: Oxford University Press. future improvements in health are likely to be highly valued as Meltzer, D. 1992. "Mortality Decline, the Demographic Transition, and Economic Growth." Ph.D. dissertation, University of Chicago, the value of health increases with continuing growth in Department of Economics. income. Health research has played a major role in the ------. 2001. "Addressing Uncertainty in Medical Cost-Effectiveness advances in health that have occurred, and the great potential Analysis: Implications of Expected Utility Maximization for Methods value of future gains in health suggests that health research to Perform Sensitivity Analysis and the Use of Cost-Effectiveness continues to merit increasing investment. New tools to Analysis to Set Priorities for Medical Research." Journal of Health Economics 20: 109­29. prospectively assess the value of research offer the promise of Murphy, K. M., and R. H. Topel. 2003. "The Economic Value of Medical even greater returns from health investment, especially for Research." In Measuring the Gains from Medical Research: An Economic more applied research that can be closely connected to meas- Approach, ed. K. M. Murphy and R. H. Topel, 41­73. Chicago: urable benefits at the population level. Because low-income University of Chicago Press. countries may often particularly benefit from applied research, Nordhaus, W. D. 2003. "The Health of Nations: The Contribution of Improved Health to Living Standards." In Measuring the Gains from these techniques to assess the value of research may be espe- Medical Research: An Economic Approach, ed. K. M. Murphy and cially helpful in ensuring that the value of applied research in R. H. Topel, 9­40. Chicago: University of Chicago Press. these settings is recognized. Providing researchers and policy Ram, R., and T. W. Schultz. 1979. "Life Span, Health, Savings, and makers in low- and high-income countries with the analytic Productivity." Economic Development and Cultural Change 27 (April): 399­421. tools needed to better identify and advocate for valuable Remme, J. H., E. Blas, L. Chitsulo, P. M. Desjeux, H. D. Engers, T. P. Kanyok, opportunities for health research may be an important avenue and others. 2002. "Strategic Emphases for Tropical Diseases Research: to increasing the level and effectiveness of spending for health A TDR Perspective." Trends in Parasitology 18 (10): 421­26. research. Rice, D. 1994. "Cost-of-Illness Studies: Fact or Fiction?" Lancet 344 (8936): 1519­20. Thaler, R., and S. Rosen. 1975. "The Value of Saving a Life: Evidence from REFERENCES the Labor Market." In Household Production and Consumption, ed. N. Terleckyj, 265­68. New York: Columbia University Press. Becker, G. S., T. J. Philipson, and R. R. Soares. 2003. "The Quantity and Viscusi, W. K., and J. E. Aldy. 2003. "The Value of a Statistical Life: A Quality of the Life and the Evolution of World Inequality." NBER Critical Review of Market Estimates throughout the World." Journal of Working Paper 9765, National Bureau of Economic Research, Risk and Uncertainty 27 (1): 5­76. Cambridge, MA. WHO (World Health Organization). 1994. "Global Comparative Bloom, D. E., D. Canning, and D. T. Jamison. 2004. "Health, Wealth, and Assessments in the Health Sector." In Disease Burden, Expenditures, and Welfare." Finance and Development 41 (1): 10­15. Intervention Packages, ed.C.J.L.Murray andA.D.Lopez.Geneva:WHO. Claxton, K., P. J. Neumann, S. Aranki, and M. C. Weinstein. 2001."Bayesian World Bank. 1993. World Development Report: Investing in Health. Value of Information Analysis: An Application to a Policy Model of Washington, DC: World Bank. Economic Approaches to Valuing Global Health Research | 163 Chapter 8 Improving the Health of Populations: Lessons of Experience Carol Ann Medlin, Mushtaque Chowdhury, Dean T. Jamison, and Anthony R. Measham In the past 50 years, the world has experienced enormous and countries than income growth. Furthermore, improvements in unprecedented gains in the health of human populations. health brought about by investments in technological progress Progress has been especially apparent in developing countries. generate an important and positive feedback loop favoring Average life expectancy has risen by more than 60 percent, from economic growth and development in these countries. 40 years in 1950 to 65 years today. In 1950, roughly 28 percent An important question that follows is what can be done to of children died before their fifth birthday, but by 1990, this further consolidate these gains and ensure that the fruits of number had fallen to 10 percent. Furthermore, many of the scientific and technology progress are placed in the hands of world's most deadly and debilitating diseases, including lep- the people in developing countries who stand to benefit most? rosy, measles, poliomyelitis (polio), and many childhood ill- Because the work of the Disease Control Priorities Project nesses, have been effectively contained in most areas and virtu- (DCPP) focuses primarily on identifying the most cost- ally eliminated in others. Smallpox, a highly contagious and effective interventions for diseases and conditions affecting the deadly disease that affected more than 50 million people a year health of populations in developing countries, this work pro- prior to 1950, has been completely eradicated. vides the starting point for analysis. The goal is to isolate the Researchers have identified economic growth, rising critical factors--in particular those "actionable" through spe- incomes, and better living conditions brought about by rapid cific public policies--that have contributed to the effective social and political transformations in many societies as major deployment and scaling up of proven cost-effective technolo- contributors to these impressive health gains. However, in gies and services in low-income settings. recent years, the role of scientific and technological progress To address this question, the DCPP joined forces with the has emerged as a crucial, but little understood, factor underly- What Works Working Group of the Global Health Policy ing these gains. As Davis (1956, 306­7) observes,"It seems clear Research Network, an initiative led by the Center for Global that the great reduction of mortality in underdeveloped areas Development in Washington, D.C., and funded by the Bill & since 1940 has been brought about mainly by the discovery of Melinda Gates Foundation. DCPP authors were asked to iden- new methods of disease treatment applicable at reasonable cost tify outstanding examples of successful implementation of pro- [and] by the diffusion of these new methods." grams and projects geared toward the deployment of proven New research has sought to validate, and indeed quantify, cost-effective interventions in their respective fields of interna- this basic intuition. For example, Jamison, Lau, and Wang tional health and to speculate on what kinds of programmatic (2005) show that technological progress (which is broadly aspects and broader public policy decisions might have con- defined as the generation or adoption of new technologies), tributed to their success. together with education, has been a far more important From an initial set of nominations, the What Works Working contributor to declining infant mortality rates in developing Group selected a subset of cases that conformed to strict 165 selection criteria, researched them thoroughly, and produced a The unabashed focus on success meant that the study report to be widely disseminated to policy makers and leading ignored potentially important information about factors that health experts in both developed and developing countries. In may be associated with programmatic failures or not-so- parallel, the DCPP initiated a systematic review of the case mate- successful cases that did not meet the strict criteria described rials to identify commonalities or factors that may have con- above. However, the inclusion of less-than-successful cases was tributed to the deployment and scaling up of those interven- not an option in light of time, resource constraints, and paucity tions. The objective was to identify a set of specific policy levers of available documentation. Thus, a significant limitation of and programmatic decisions that could facilitate the transplan- our study results from the lack of variance in the outcome tation of those and other cost-effective interventions to new and observed. This type of selection bias is a common problem that different settings. This chapter presents the results of that study. may result from the nonrandom selection of cases in qualita- tive research. Although bias cannot be eliminated without expanding the RESEARCH METHODS study to include unsuccessful examples, working skillfully with the presumption of bias to increase the level of confidence in The study consisted of a qualitative analysis of a set of case the findings is possible. First, counterfactual examples, even if studies selected to help illustrate how proven, cost-effective purely speculative (what would have happened if . . . ?), can be interventions have been successfully deployed and brought to used to further substantiate the hypothesis that circumstance scale with dramatic results in low- and middle-income coun- (B) has directly contributed to the observed phenomenon (A). tries in Africa, Asia, and Latin America and the Caribbean. We Second, theorizing in a constructive way about what the direc- examined evidence culled from interviews, peer-reviewed arti- tion of the bias might be and, therefore, minimizing its impact cles published in journals, and official project evaluations and on the results of the study are also possible. For example, any attempted to organize this information in a way that would potential bias more than likely results from overdetermining allow us to reach tentative conclusions about the most signifi- causality rather than overlooking or ignoring key factors cant elements associated with the interventions' success. related to success. In a related point, the study design makes The study thus followed one of Mill's (1843) five methods of discerning the relative contributions of the various factors dif- experimental reasoning: the method of agreement. Such a ficult, because weights cannot be assigned easily. method postulates that "if two or more instances of the phe- Remarkably few rigorous studies of this sort attempt to nomenon (A) . . . have only one circumstance (B) in common, track the implementation of proven, cost-effective interven- the circumstance (B) in which alone all the instances agree is the tions in the field. Although we have a good understanding of cause (or effect) of the given phenomenon (A)." For this study, the efficacy of the available arsenal of interventions for treating the phenomenon (A) is represented by success. Cases that qual- and preventing diseases specific to low-income countries, we ified as successes had to conform to the following five criteria: often know little about the range of programmatic and policy options that are needed to support these interventions in the · Scale. All cases selected for study involved a national, real world. This study represents the first major contribution regional, or global scale. Pilot projects or interventions toward the development of a body of knowledge in that area, implemented on a subnational scale were not considered. and the preliminary conclusions reached should be understood · Importance. Selected cases addressed a problem of major in that context (Collier and Brady 2004). public health significance that could be expressed, at the program's inception, in terms of disability-adjusted life years, a composite measure of mortality and morbidity CASES caused by the disease. · Health impact. Selected cases had documented evidence of From an initial set of nominations received from DCPP a clear and measurable effect on the health of the popula- authors as well as from other international health researchers, tion targeted by the intervention. Process indicators, includ- we selected a subset of 17 cases for study. We could not consider ing immunization coverage rates, were not considered an many cases that were nominated because of the absence acceptable substitute for health impact data. of reliable data. Thus, the 17 cases selected are merely a subset · Duration. All cases selected for study had a life span of at of the many successes in international health that have been least five consecutive years. achieved during the past 50 years, not the full universe. · Cost-effectiveness. Selected cases relied on interventions Nonetheless, the cases draw from all three continents of the that had been proven to be cost-effective at a threshold developing world--Africa, Asia, and Latin America and the of approximately US$100 per disability-adjusted life year Caribbean--and involve both communicable and noncommu- saved. nicable diseases as well as curative and preventive care. Most 166 | Disease Control Priorities in Developing Countries | Carol Ann Medlin, Mushtaque Chowdhury, Dean T. Jamison, and others cases are national-level programs, but a few involve regional dropped by 36 and 43 percent, respectively. Mortality attrib- initiatives, and one is global. uted to diarrhea fell 82 percent among infants and 62 per- Many near misses did not make the cut. The reasons for cent among children. The project closed in 1991. this exclusion varied. For example, a program in Costa Rica, · Guinea worm eradication. Twenty countries in Asia and El Salvador, and Guatemala to promote hand washing appears Sub-Saharan Africa began a global campaign to eradicate to have resulted in a dramatic decline in child morbidity guinea worm in the mid 1980s. Led by the Carter Center, and mortality but did not meet the duration criterion because the United Nations Children's Fund, the U.S. Centers for it was fully operational for only three years (1996­99). In Disease Control and Prevention, and the World Health another example, the evidence of a health impact was mixed: Organization, the campaign promoted improved water a successful schistosomiasis control program in the Arab safety through deep-well digging, environmental control, Republic of Egypt that included treatment of blood flukes in and the use of cloth filters for drinking water; health educa- infected individuals was later linked to high prevalence rates of tion programs; and case management, containment, and hepatitis C caused by the use of improperly sterilized syringes surveillance. Health impact--By 1998, 9 million to 13 mil- (Frank and others 2000). Nonetheless, the most common lion cases of guinea worm had been prevented and global rationale for excluding a case from this study had to do with a prevalence had dropped by 99 percent. The project is ongo- lack of consistent documentation and of analysis of the health ing in three countries. impact of the program in question. Thus, a reasonable conclu- · Family planning. In Bangladesh, family planning has been sion is that the true universe of cases is much larger than the promoted since the 1970s through a door-to-door outreach subset of cases we examined. program conducted by young, married women who provide Each case reviewed here illustrates how a discrete health information about limiting family size or spacing pregnan- intervention or combination of interventions was successfully cies along with products. An extensive media campaign brought to scale in a specific context. To gain insight into this accompanied the outreach program. Health impact-- process, we can distinguish between the intervention--for Contraceptive use among married women in Bangladesh is example, the tool or technology that has been proven to be cost- approximately 50 percent today, compared with only 8 per- effective for the treatment or prevention of a given disease-- cent in the mid 1970s, and the average number of children and the programmatic characteristics and policies that per family is 3.3, down from 7 in the mid 1970s. The project contributed to the successful delivery or deployment of the is ongoing. intervention through specially designed programs or projects. · Hib vaccination. Chile began to include the Hib vaccine as The following list of cases selected for review describes the part of its national immunization program in 1996. In The programs or projects that were scaled up, identifies the specific Gambia, a similar initiative was introduced in 1997. Health intervention or interventions deployed, and summarizes the impact--In Chile, the prevalence of Hib disease fell by existing evidence about health outcomes and impact:1 90 percent, and the incidence of pneumonia and other Hib-related illnesses fell by 80 percent. In The Gambia, the · Chagas disease control. In 1991, seven countries-- number of children developing Hib meningitis fell from Argentina, Bolivia, Brazil, Chile, Paraguay, Uruguay, and 200 per 100,000 to 21 per 100,000 only 12 months following later Peru--joined forces as part of an initiative for the the introduction of the vaccine. The projects are ongoing in Southern Cone countries led by the Pan American Health both countries. Organization to combat Chagas disease through a combina- · HIV/AIDS prevention. Thailand launched the 100 Percent tion of surveillance activities, house-to-house spraying, and Condom Program in 1991 to address the rising incidence other vector control methods. Health impact--Disease inci- of HIV/AIDS in the country. The program provided boxes dence had fallen by 94 percent by 2000. By 2001, disease of condoms to brothels free of charge, mandated the use of transmission had been halted in Chile, Uruguay, and large condoms by sex workers, and threatened brothels with parts of Brazil and Paraguay. The project is ongoing. penalties and closure for noncompliance. Health impact-- · Diarrheal treatment. In Egypt, the government launched a By 1992, condom use in brothels had risen to more than national program in the early 1980s to promote the use by 90 percent, up from 14 percent in 1989. The number of cases mothers of locally manufactured oral rehydration salts in a of new sexually transmitted infections fell from 200,000 in four-part strategy that included tailoring product design 1989 to 15,000 in 2001, and an estimated 200,000 new infec- and branding to accommodate local preferences and cus- tions were averted between 1993 and 2000. The project is toms; strengthening production and distribution channels, ongoing. both public and private; training health workers; and using · Health improvement of the poor using financial incen- social marketing and a mass media campaign. Health tives. In 1997, the Mexican government launched a new impact--Between 1982 and 1987, infant and child mortality social welfare program designed to help lift rural families Improving the Health of Populations: Lessons of Experience | 167 out of poverty by providing cash payments in exchange for Health impact--By 1995, the severity of caries in children their participation in nutrition and supplementation pro- between the ages of 6 and 12 had fallen by more than 80 per- grams, their use of prevention and basic health care servic- cent. The project is ongoing. es, and their children's school attendance. Health impact-- · Salt iodination. China launched the National Iodine After five years, the children of participating families were Deficiency Disorders Elimination Program in 1993. The gov- 12 percent less likely to experience illness than those of non- ernment requires producers to iodize salt and has stepped participating families, and their nutritional status had up its monitoring and enforcement capacity to ensure com- improved. Adult health indicators also improved. The proj- pliance. Health impact--Total goiter rates among children ect is ongoing. between the ages of 8 and 10 years fell from 20.4 percent in · Maternal health. The Sri Lankan government relied on pro- 1995 to 8.8 percent in 1999. The project is ongoing. fessional midwives and sustained investments in the coun- · Smallpox eradication. The campaign to eradicate small- try's health care system, including in rural areas, to improve pox, led by the World Health Organization and heavily maternal health. Health impact--The maternal mortality financed by the United States, was launched in the mid 1960s. ratio fell from approximately 500 per 100,000 live births in Strong leadership, dedication, and commitment on the part 1950 to 60 per 100,000 in 2003. The project is ongoing. of the international community and the timely discovery of · Measles elimination. In 1996, the seven southern African simple, new technologies--for example, the bifurcated nee- countries agreed to a coordinated immunization strategy, dle and the "ring" strategy of surveillance and contain- supported by improved surveillance and laboratory capacity, ment--characterized the effort. Health impact--The World to eliminate measles by including the vaccine as part of rou- Health Assembly declared smallpox eradicated in May 1980. tine immunization for all nine-month-old babies and · Tobacco control. Poland passed groundbreaking legisla- organizing nationwide catch-up and follow-up campaigns tion in 1995 imposing strong warning labels on cigarette for children age nine months to 14 years. Health impact-- packages, banning smoking from enclosed workplaces, and The number of measles cases reported annually in the prohibiting tobacco sales to minors. South Africa passed region fell from 60,000 in 1996 to 117 in 2000. The number similar legislation in 1999 to strengthen a previously of deaths attributed to measles fell from 166 to 0 during the imposed tax of 50 percent on the retail price of cigarettes. same period. The project is ongoing. Health impact--Cigarette consumption dropped 10 percent · Onchocerciasis control. The discovery of ivermectin between 1990 and 1998, resulting in a 30 percent decline in (Mectizan) in 1978 and Merck's decision to provide it free of lung cancer among men age 20 to 44, a nearly 7 percent charge to anyone who needed it allowed early successes decline in cardiovascular disease, and a decline in the num- based on weekly aerial spraying in 11 West African countries ber of babies with low birthweight. South Africa witnessed a to be further consolidated and later expanded to the other 30 percent decline in cigarette consumption in the 1990s, 19 endemic countries in Central and East Africa. Health especially among youths and the poor. The projects are impact--In West Africa, disease transmission has been vir- ongoing in both countries. tually halted, and 1.5 million previously infected people are · Trachoma control. The Moroccan National Blindness now symptom free. In Central and East Africa, the program Control Program, launched in 1991, promoted the use of has helped prevent an estimated 40,000 cases of blindness "SAFE" interventions (surgery, antibiotics to control the each year. The Onchocerciasis Control Program (OCP) infection, facial cleanliness, and environmental improve- ended in 2002. The African Programme for Onchocerciasis ments), with the goal of eliminating trachoma by 2005. Control is ongoing. Health impact--Overall prevalence rates have fallen by · Polio elimination. In 1985, the Pan American Health 75 percent since 1999, and the prevalence of active disease in Organization launched a campaign to eradicate polio from children under the age of 10 has seen a 90 percent reduction the Americas. National vaccine days were held twice a year since 1997. The project is ongoing. and were targeted at children under the age of five, regard- · Tuberculosis control. In 1991, China launched a 10-year less of their immunization status, to increase coverage in program in 13 of its 31 mainland provinces to apply the countries with weak routine immunization programs. An directly observed therapy short course (DOTS) strategy to extensive surveillance system and mop-up campaigns to turberculosis (TB) control. Peru, previously one of 23 high- address outbreaks were crucial during the campaign's final burden countries that collectively account for 80 percent of stages. Health impact--The last case of polio in the the world's new TB cases each year, launched a similar effort Americas was reported in 1991. the same year. Health impact--Within two years, China had · Salt fluoridation. In Jamaica, a formal agreement between achieved a 95 percent cure rate for new cases and a cure rate the Ministry of Health and the country's only salt producer of 90 percent for those patients who had previously com- introduced fluoridation to salt in 1987 to prevent caries. pleted treatment unsuccessfully. The number of people with 168 | Disease Control Priorities in Developing Countries | Carol Ann Medlin, Mushtaque Chowdhury, Dean T. Jamison, and others TB declined by more than 37 percent between 1999 and public sector's regulatory or legislative authority was critical. 2000. The project ended in 2001, but important elements Governments in Poland and South Africa passed strict laws, have been incorporated in the 10-year National Plan for the despite strong opposition from the tobacco industry, requiring Prevention and Control of TB (2001­10). In Peru, disease explicit health warnings on cigarette packs, banning smoking incidence declined each year by 6 percent. The program in enclosed public places, and prohibiting tobacco media achieved a case detection rate of 70 percent and an 85 per- advertisements, among other things. Governments also used cent cure rate. The project is ongoing. their authority creatively to encourage health-promoting behaviors and to discourage risky ones. In Mexico, the govern- GENERAL FINDINGS ment provided direct cash payments to poor families in exchange for visits to health care clinics and school attendance. Taken as a whole, the cases support four general findings. The In Thailand, local police worked in collaboration with health first two have special relevance because they serve to discon- officials to lend credibility to the government's threat to shut firm aspects of the prevailing wisdom about aid effectiveness-- down brothels that failed to comply with the no condom, no or at least present a serious challenge to such wisdom. sex policy, giving teeth to the national campaign. First, these cases demonstrate that a wide range of proven, Third, the cases reviewed for this study share a number of cost-effective interventions exists that can and have been common features or attributes that appear to have contributed brought to scale in developing countries, even in extremely to the successful outcomes. Without exception, they enjoyed low-income settings with limited health infrastructure and in and managed to reap the benefits of strong leadership, effective challenging macropolicy environments. In West Africa, aerial management, realistic financing arrangements, country owner- spraying of the blackflies' breeding sites, part of the strategy ship, and openness and receptivity to learning by doing, con- promoted by the OCP throughout the 1980s, "continued stantly improving on strategies and processes by incorporating unabated through wars between member countries and coups new research findings and technical innovation into program that grounded all other aircraft" (Eckholm 1989, 20). In Sudan, improvements. despite the difficulties created by the more than 20-year civil For example, successful projects appeared to benefit from a war, and in other areas of Sub-Saharan Africa, the campaign strong champion who could provide the necessary leadership to eradicate the guinea worm has made progress. The finding to bring relevant stakeholders together, encourage them to is significant in that it challenges a central tenet of the aid- focus and coordinate their activities, and instill in them a sense effectiveness literature: that only countries with a "good" policy of purpose and enthusiasm for their work. However, we did environment can benefit from external financial assistance find that leadership came packaged in many different shapes (Devarajan, Dollar, and Holmgren 2001). and sizes. In Jamaica, the curiosity and persistence of a Ministry The aid-effectiveness literature has tended to focus on a of Health dentist led to the identification of the island's only different set of outcomes--for example, macroeconomic and salt producer as the vehicle for fluoridation. In Mexico, structural reform--rather than on health outcomes, and this President Ernesto Zedillo Ponce de León seized on the innova- focus may partly explain the contradictory conclusions; how- tive proposal of a close adviser, Santiago Levy, then director- ever, an examination of whether such a conclusion is true goes general of social security, and launched a program linking well beyond the scope of this study. In any event, the cases education, health, and nutrition as part of an integrated strategy reviewed for this study displayed a striking degree of variation to lift rural families out of poverty, and the program was not in the political and economic contexts in which interventions abandoned when Zedillo left office. The new Vicente Fox were applied and brought to scale, and no clear pattern of administration, motivated by undeniable evidence of the pro- association was apparent between this variation and successful gram's effectiveness, instead sought to expand the program into outcomes in relation to health. urban areas and added an educational component. In a less vis- Second, the cases provided new evidence of the importance ible but nonetheless critical display of leadership and forward of the public sector to achieving successful health outcomes. thinking, the sustained investments of the Sri Lankan govern- This finding was a surprise, especially considering the strength ment over a nearly 50-year period to build a rural health net- of recent evidence documenting "weak links in the chain work emphasizing critical elements of maternal health have led between government spending for services to improve health to gains in the health of women unparalleled by countries at and actual improvements in health status" (Filmer, Hammer, similar, and higher, income levels. and Pritchett 2000, 199). The specific roles that the public sec- Strong program management was needed to ensure that tor played in achieving these outcomes varied tremendously. plans, once conceived, were implemented effectively. Successful In some instances, such as promoting maternal health in Sri cases had well-delineated goals that were clearly linked to Lanka and controlling TB in China and Peru, governments inputs, activities, outputs, and outcomes. This factor was espe- were involved in direct service provision. In other instances, the cially evident in the case of global or regional immunization Improving the Health of Populations: Lessons of Experience | 169 campaigns, given the many logistical challenges and the need ate to local conditions and vector behavior. Finally, adoption of for fluid and effective coordination of many countries and the ring vaccination strategy marked a crucial turning point stakeholder groups, often within a highly constrained time in the global campaign to eradicate smallpox, enabling rapid frame. However, similar management skills are needed for containment of the disease in remote parts of the world with- health service delivery systems, especially when patient referral, out vaccination of every child. tracking, and follow-up are essential components of the inter- In sum, a small number of features appear to be common to vention. In China, incentive schemes to motivate physicians, all the successful cases. A reasonable hypothesis suggested by extensive training and supervision of health care staff, and the evidence is that these five attributes represent the known set substantial investments in local TB dispensaries were all crucial of necessary, but not sufficient, conditions for successfully elements in improving management capacity for large-scale implementing cost-effective health interventions in the devel- rollout of the country's DOTS program, which covered a pop- oping world. ulation of 573 million in 1,208 counties in 13 provinces. Fourth, despite the obvious limitations of case study meth- A closely related requirement was having a realistic financ- ods in hypothesis testing and confirmation, the evidence from ing strategy that was compatible with a project's goals. Even the cases sheds important light on two important debates in when large sums of money were involved, deployment of the international health policy. First, the cases suggest that much intervention yielded tremendous returns at a relatively low cost more is involved than what is currently understood about per disability-adjusted life year. In the case of onchocerciasis whether weak policy environments can make good use of care- control, for which donors have invested US$560 million over a fully selected, strategic investments in health. As the next sec- period of 28 years, transmission has been virtually halted in 20 tion indicates, different types of programmatic characteristics West African countries, and nearly 600,000 cases of blindness and policies are needed for the deployment of different types have been averted at an annual cost of only US$1 per person. of interventions. How these characteristics interact with dif- In the case of guinea worm control, in which donors have ferent policy environments--whether strong, weak, or in invested approximately US$88 million over a 12-year period, between--deserves further scrutiny and exploration. Second, disease prevalence has fallen by 99 percent, and only 35,000 evidence from the cases of successful government action people remain affected, down from 3.5 million, at a cost of should call into question any premature and overly general US$5 to US$8 per person. conclusions about public sector ineffectiveness in developing Country ownership was another distinguishing feature of countries. Even though such a small sample of cases is surely successful programs. A government's willingness to commit insufficient to close the book on these important policy scarce funding to scaling up an intervention can be an impor- debates, it should at least encourage further study and refine- tant indication of this ownership, although not the sole predic- ment of the arguments. tor. Despite the extremely constrained budgets of the seven participating countries, the campaign to eliminate measles in southern Africa was almost entirely funded by their min- INTERVENTION TYPE, PROGRAMMATIC istries of health. The Thai government covered approximately CHARACTERISTICS, AND POLICIES 96 percent of the cost of the 100 Percent Condom Program. In Morocco, the government bore the bulk of the costs for imple- Programmatic characteristics and policies associated with menting the SAFE strategy to address blindness caused by tra- successful outcomes appear to vary by intervention type. choma, with contributions from the United Nations Children's The starting point for this discussion is the intervention, tech- Fund and the International Trachoma Initiative, an interna- nology, or tool in question. What allows for the widespread tional public-private partnership. deployment of a proven, cost-effective intervention? What are Most of the cases we reviewed benefited from new research the steps for converting a proven, cost-effective intervention findings and technical innovation. Successful cases appear to into a fully fledged health program that has been successfully display the openness and receptivity needed to make good use brought to scale, preferably at the national level? Is it possible of new knowledge and to support ongoing research when to distinguish between the specific health intervention and the appropriate or when gaps in knowledge prove to be a hin- programmatic characteristics and public policies associated drance to progress. In Bangladesh, a program to treat child- with its successful deployment? hood diarrhea trained mothers to make their own salt solution The cases under review were grouped according to the when the authorities determined that mass production and primary type of intervention deployed by the program or proj- distribution of prepackaged oral rehydration salts was unreal- ect in question. The types of interventions varied in terms of istic. Control of Chagas disease in the Southern Cone of Latin their emphasis on the delivery of standardized products to a America required public health officials in each country to population (product-intensive interventions), the delivery of devise and deploy environmental control strategies appropri- clinical services (service-intensive interventions), a personal 170 | Disease Control Priorities in Developing Countries | Carol Ann Medlin, Mushtaque Chowdhury, Dean T. Jamison, and others behavior change (behavioral change interventions), the control To the degree that product-intensive interventions place rel- of environmental hazards (environmental control interven- atively low technical demands on health care staff at the point tions), or some combination thereof. of delivery, they may be more easily deployed in low-resource Further scrutiny of subgroupings of cases revealed that settings than other types of interventions. Compared with certain programmatic characteristics, delivery modalities, and service-intensive interventions, they are less transaction inten- public policy instruments also appeared to vary by interven- sive, requiring fewer interactions between providers and clients. tion type. This finding appeared to substantiate the claim of Also, compared with behavioral-change interventions, they are the first edition of this volume (Jamison 1993, 11) that "com- less dependent on individual compliance, requiring simply that monalities of logistics, policy instruments, and approach" individuals make themselves available for treatment. Product- vary by intervention type and play a role in determining intensive interventions include mass drug administration whether the intervention or interventions will be deployed (chemotherapy), childhood immunizations, mineral fortifica- successfully. tion, and nutritional supplementation. (These specific inter- The typology presented here differs from the one elaborated ventions are addressed in detail in chapters 20, 22, and 28.) by World Development Report 2004: Making Services Work Product-intensive interventions are often, though not nec- for Poor People (World Bank 2004), in that the classification essarily, linked to vertical rather than horizontal delivery depends on characteristics inherent to the intervention in modalities. According to Gonzalez (cited in Mills 1983, 1972), question. By contrast, World Development Report 2004 identi- the vertical approach "calls for the solution of a given health fies three classes of service delivery arrangements: individual- problem through the application of specific measures through oriented clinical services, population-oriented outreach single-purpose machinery." By contrast, the horizontal services, and community- and family-oriented services that approach "seeks to tackle . . . health problems on a wide front support self-care. The focus is on differences in the relationship and on a long-term basis through the creation of a system of between provider and client and how these differences interact permanent institutions commonly known as `general health with market and public sector dynamics.2 However, our focus services.'" Where health systems are weak and poorly function- was on how characteristics of the interventions themselves give ing, vertical programs--in particular, mass campaigns--can be rise to certain programmatic or policy imperatives that may an effective means of rapidly providing coverage to a large pop- contribute, ultimately, to successful health outcomes. ulation. However, the same approach could result in an unfor- This section presents the five intervention types and tunate duplication of effort in countries where the health care explores the clusters of programmatic characteristics and poli- system is already strong and functioning properly. cies that appear to support the successful deployment of each intervention type, based on case study analysis. Mass Drug Administration. Of the product-intensive inter- ventions, those that can be delivered in pill or capsule form in Product-Intensive Interventions standardized doses, often through what is referred to as mass These types of interventions (box 8.1) involve the simple trans- drug administration, are perhaps the least complex to deliver fer of a standardized technology to an individual or to an entire and, as a result, may be the least costly. Onchocerciasis, also population. They can be targeted at either prevention or cure, known as river blindness, can be treated by a single dose of but the distinguishing feature is standardization. Unlike ivermectin administered annually to infected individuals. service-intensive interventions, product-intensive interven- Lymphatic filariasis can be treated in much the same way using tions need not be tailored to the unique health care needs of the a two-drug combination therapy of albendazole plus either individual receiving treatment. diethylcarbamazine or ivermectin administered annually in Box 8.1 Product-Intensive Interventions: Illustrations from Cases Hib vaccine Salt fluoridation Ivermectin Salt iodination Measles vaccine Smallpox vaccine Oral polio vaccine Source: Authors. Improving the Health of Populations: Lessons of Experience | 171 single doses for four to six years (the estimated productive life vaccine was available that was easy to manufacture, cost only span of the adult-stage parasite). about a penny a dose, protected for several years with a single The importance of product-intensive therapies can easily be inoculation, and was relatively stable in warm climates, reduc- illustrated by reference to the case of onchocerciasis control in ing the need for refrigeration. Whereas most vaccines took most parts of Africa. Although aerial spraying, an environmen- months to induce immunity, the smallpox vaccine acted with tal control intervention, had been used successfully to slow dis- remarkable speed, providing nearly total protection within ten ease transmission in 11 West African countries, it was not a to twelve days." This unique set of characteristics together viable option for 19 countries of East and Central Africa because meant that a "surveillance-containment" approach could of geographical differences. However, Merck scientists' discov- replace mass vaccination entirely, in effect reducing the number ery of ivermectin in 1978 and the company's generous commit- of distribution points required by the intervention, thereby ment to provide the drug free of charge to anyone who needed it permitting major strides in the global eradication campaign. changed the parameters of what was possible. Seizing the oppor- The more complex the intervention, the more challenging-- tunity, the African Programme for Onchocerciasis Control, an and probably costly--it will likely be to implement. international partnership led by the World Bank, the World Interventions that can be delivered in a single shot or that can Health Organization, the United Nations Development be easily incorporated into routine immunization (bundled) Programme, and the Food and Agriculture Organization of the are clearly the easiest to implement. As the case of Hib vaccina- United Nations, was created in 1995 with the goal of eliminating tion in Chile illustrates, delivering the intervention was rela- onchocerciasis as a disease of public health and socioeconomic tively straightforward after the government got past the hurdle importance in East and Central Africa. of evaluating its cost-effectiveness relative to other inter- It quickly became apparent that the weak and sometimes ventions. The government determined that the creation of a nonfunctioning health systems of many African countries were combined diphtheria-tetanus-pertussis and Hib vaccine was not up to the task; thus, a new approach was tried that took worthwhile and that the vaccine could be administered as part of advantage of the fact that the success of the intervention no an already well-functioning system of routine immunization. longer depended on a clinic-based delivery system. Under the Some of the common complexities associated specifically supervision of national public health ministries and non- with immunizations range from the need for multiple inocula- governmental organizations, community volunteers received tions administered at regular intervals, to the need to maintain training on organizing and managing the local ivermectin a reliable cold chain, or to the need for the large population campaigns. The community-directed approach of treatment coverage required to achieve "herd immunity"--whereby the with ivermectin has been so successful that it has been consid- likelihood of person-to-person transmission is drastically ered as a possible model for delivering other types of treat- reduced, even among the unimmunized population. In view of ments to remote areas. these potential complexities, polio elimination in the Americas represents a remarkable achievement. The oral polio vaccine Immunizations. Product-intensive interventions may vary in must be administered in three properly spaced doses, and cov- complexity, which has implications for their delivery or deploy- erage must be high to prevent "silent" epidemics. In the 1970s, ment. For vaccines, the need to maintain an effective cold chain before the campaign was launched, polio caused an estimated adds an additional layer of complexity to the delivery system. 15,000 cases of paralysis and 1,750 deaths each year (Musgrove Other pertinent factors are whether the intervention can be 1988). However, a carefully orchestrated campaign organized delivered as a single shot or iteratively, whether it can be bun- around achieving and maintaining high coverage through dled together with other products or must be delivered sepa- routine immunization and national vaccination days, the rately, and whether the number of distribution points is few or prompt identification of new cases, and the aggressive control many. In many cases, the characteristics of the disease or con- of outbreaks led to the elimination of polio from the Americas dition being addressed may affect the level of complexity. in 1991. The creation of the Inter-Agency Coordinating For example, the overwhelming success of the global effort Committee, made up of representatives from the Pan American to eradicate smallpox has been attributed, at least in part, to spe- Health Organization, the United Nations Children's Fund, cific characteristics of the variola virus. Unlike other infectious the U.S. Agency for International Development, the Inter- diseases, such as malaria or yellow fever, smallpox depends American Development Bank, Rotary International, and the solely on the human host and does not have an animal or insect Canadian Public Health Association, played a key role not only carrier. Unlike polio, smallpox does not produce silent or in generating political and financial support, but also in help- asymptomatic infection, thereby facilitating diagnosis and sur- ing address the logistical and managerial challenges inherent veillance of the disease (Tucker 2001). Other notable differences to the campaign. The Inter-Agency Coordinating Committee from other diseases have more to do with the vaccine than the model was so effective that it was quickly duplicated at the virus. Tucker (2001, 64) explains that "a freeze-dried smallpox country level. 172 | Disease Control Priorities in Developing Countries | Carol Ann Medlin, Mushtaque Chowdhury, Dean T. Jamison, and others The technical and logistical challenges associated with greater number of potential delivery points. Because salt measles elimination in southern Africa were no less complex. production is licensed at the provincial level in China, imple- Measles is one of the most contagious of all human diseases. The menting the change involved working with several layers of measles vaccine requires 90 percent coverage to achieve herd government bureaucracy; however, perhaps ironically, the sys- immunity and to stop the spread of the virus. Furthermore, it tem of central control eased the challenge. During a four-year often requires two doses to be effective and must be adminis- period, 55 salt factories were upgraded and 112 iodination cen- tered to infants no earlier than nine months of age, or about six ters were established throughout the country to support the months later than other recommended vaccines. If given earlier, initiative. The government also introduced changes in bulk and the vaccine will fail to trigger an active immune response, retail packaging to help consumers more easily recognize because infants are passively protected by their mothers' anti- iodized salt. The basic plan achieved nearly 90 percent cover- bodies until that age. Thus, the vaccination interval falls outside age, but the remaining challenge is to address the numerous of what most routine immunizations require. delivery points that function outside the national system, for To overcome this challenge, the southern African countries instance, in areas where people live near the sea and produce adopted a strategy known as catch up, keep up, and follow up. their own salt. In each country, beginning with the program's launch in 1996, the strategy involved organizing a national catch-up campaign Summary. In sum, product-intensive interventions can be in which mobile teams vaccinated all children, regardless extraordinarily complex even though they involve fewer trans- of their vaccination status, between the approximate ages of actions between providers and clients and lower technical 9 months and 14 years; sustained routine coverage; and ran at requirements at the point of delivery than other types of inter- least one follow-up campaign several years later. The countries ventions, particularly service-intensive interventions. However, also strengthened their surveillance and laboratory capabilities the relative simplicity of deployment when the scientific and to investigate all suspected measles cases. technical issues of development and production have been addressed may help explain why product-intensive interven- Mineral Fortification. A different type of product-intensive tions are perceived to be easier to implement than other types intervention, mineral fortification, requires fewer points of of interventions and why countries experiencing political and delivery and is far less labor intensive. However, potential chal- economic instability might prefer them. lenges include the need for a different set of technical capacities than is typical for health sector solutions; the possibility of a Service-Intensive Interventions need for significant initial investments to modify production Service-intensive interventions (box 8.2) include the full range processes or manufacturing capabilities; and the involvement of diagnostic and therapeutic health services usually provided of non­health sector entities, such as private industry. not only in the clinic setting, but also in the home or at school. Jamaica's salt fluoridation program beautifully illustrates Unlike product-intensive interventions, service-intensive inter- the simplicity of a single delivery point for an intervention. ventions cannot easily be standardized and may require With the agreement of the island's only salt producer, Alkali careful--and time-consuming--monitoring and reporting on Limited, in place, universal coverage was easy to achieve. All patients' progress. Thus service-intensive interventions are that was needed was a complementary legal and regulatory highly transaction intensive and typically place high technical framework to oversee the process. In this case, the start-up demands on the health staff at the point of delivery. Examples costs were small: only US$3,000 worth or so of new equipment range from primary care services, including essential obstetri- was needed, which the company was easily able to recoup with cal care, to surgical procedures, to treatment of communicable a slight increase in the price of salt. and noncommunicable diseases. In China, salt fortification with iodine was more difficult, The complexity of service-intensive interventions may vary, involving a larger investment in the production process and a just as in the case of product-intensive interventions. The more Box 8.2 Service-Intensive Interventions: Illustrations from Cases Bilamellar tarsal rotation procedure Maternal health care using midwives DOTS for TB Primary and basic health care services Source: Authors. Improving the Health of Populations: Lessons of Experience | 173 standardized the treatment protocol, the easier it will be to The cases we reviewed relied on a variety of strategies to administer on a large scale. However, standardized or not, the address this problem. Although the specific interventions var- transaction-intensive character of this type of intervention ied, the strategies ranged from traditional investments in pub- means that its successful deployment depends on the program's lic sector provision to improve access, to supply-side incentives or project's ability to overcome potential (and likely) con- to address quality concerns, to demand-side incentives to straints on human resources. A related concern is the overall strengthen the effective demand for health care services. For health system's capacity to effectively manage competing example, in an effort to improve maternal health even in demands on these resources. In contrast to product-intensive remote areas, Sri Lanka adopted the traditional model of pub- interventions, human capacity constraints for service-intensive lic sector provision, but with a twist. Instead of a physician- interventions are harder, but not impossible, to address based solution, which would have been extremely costly, through community mobilization or the use of volunteers, Sri Lanka relied instead on professional midwives to provide because of the need for specialized training. Chapter 71 inves- widespread access to maternal health care, building on a strong tigates developing countries' experiences with new types of health care system that provides free health care. Midwives professionals in service delivery settings that have traditionally serve a population of 3,000 to 5,000 each and live locally. They relied on physicians. visit pregnant women in their homes, register them for care, and encourage them to attend prenatal clinics (run by doctors). Single-Shot Surgical Services. If the intervention is relatively Midwives receive 18 months of training and are backed up standardized, it can sometimes be deployed using a vertical by supervision and a well-functioning referral network. modality in a manner similar to the more standardized, single- Established procedures for service delivery and supervision, shot, product-intensive interventions. This was the case of the along with frequent in-service training, help keep midwives surgical procedure used in Morocco as part of the broader current and delivering high-quality services. Health clinics are SAFE strategy to address trachoma. A relatively simple surgical supported by a network of cottage hospitals (clinics having procedure, the bilamellar tarsal rotation procedure, can be used doctors as well as nurses assigned to them), rural hospitals, and to halt corneal damage to prevent the onset of blindness caused maternity homes at the secondary level; tertiary provincial by repeated trachoma infections. Morocco's Ministry of Health hospitals with specialist services; teaching hospitals; and spe- organized mobile surgical teams of doctors and nurses to carry cialist maternity hospitals. out the corrective surgery in small towns and communities China faced a similar dilemma with regard to establishing throughout the country. In just eight years, the teams carried an effective system of outreach and referral to address a grow- out more than 26,000 surgeries. The effort required the ing TB problem by scaling up DOTS. Although DOTS is involvement of 43 physicians and 119 nurses working in relatively standardized as far as service interventions go (see 34 clinics. However, despite the relatively standardized nature chapter 16), the treatment protocol is highly transaction of this kind of service-intensive intervention, human capacity intensive, and complicated cases may require specialized treat- constraints may slow progress. Compared with product- ment. DOTS is thus also highly dependent on well-developed intensive interventions for which community volunteers can be systems of outreach and referral for its success. However, recruited to assist with distribution, service-intensive interven- China faced a challenging situation because most village doc- tions require a more specialized workforce. Even when nurses tors who were needed to conduct patient diagnosis, treatment, and other health workers with lower-level skills can substitute and surveillance in rural areas were in private practice and for more highly trained physicians, constraints on human had little incentive to treat patients for whom drugs were now resource capacity may persist. Morocco faces a backlog of provided free of charge. In response, the government created about 15,000 cases, many of which are urgent. a financial scheme to provide incentives for these doctors to participate. For each patient enrolled in the treatment pro- Strengthened Outreach and Referral Systems. If the inter- gram, village doctors received US$1. They received an vention is not easily standardized, or if it is highly transaction additional US$2 for each smear examination carried out in intensive, its successful deployment will also depend on well- the county TB dispensary during a two-month period and developed outreach and referral systems. Outreach systems are another US$4 for each patient who completed treatment. needed to ensure that those requiring care will have access to Simultaneously, the government made significant administra- care, and referral systems are needed to route patients requir- tive, managerial, and institutional investments. Tens of ing additional care toward specialized care and treatment facil- thousands of staff from TB dispensaries were trained, and ities. However, ensuring that such systems are in place for spe- supervisory systems were put into place. Furthermore, the cific programs and projects is a major challenge in countries government set up a national TB project office and a TB con- where the health care system is already weak and under consid- trol center to oversee and coordinate the various levels of erable strain. government involvement. 174 | Disease Control Priorities in Developing Countries | Carol Ann Medlin, Mushtaque Chowdhury, Dean T. Jamison, and others Demand-Side Incentives. Demand-side incentives can be received more on-the-job training, and clinics benefited from a designed to complement supply-side investments. Indeed, both steadier flow of pharmaceutical and other supplies. skeptics and supporters of governments' ability to translate public spending into effective service provision and positive Summary. In sum, service-intensive interventions are highly health outcomes encourage the use of demand-side incentives transaction intensive, especially compared with product- as a means of quality control to improve routine care (Filmer, intensive interventions. The more standardized the interven- Hammer, and Pritchett 2000). Economists consider demand- tion, the more likely that it can be delivered by means of side incentives to be valuable tools for stimulating weak military-like campaigns in the same manner as product- demand for services or for overcoming barriers to use that can intensive interventions, although the skill level of the health artificially dampen demand. This was how at least two of the workers involved will need to be higher. If complications of cases included in this study constructively used demand-side treatment are possible--or if the intervention cannot easily be incentives. However, in neither case was this use an either-or standardized--its deployment will likely require fairly elabo- proposition; that is, both demand-side and supply-side invest- rate systems of outreach and referral. Although recent ments were relied on to generate the successful outcomes that scholarship has strongly encouraged the use of demand-side qualified the cases for this study. mechanisms, particularly in an effort to address quality con- In Peru, the newly revised National Tuberculosis Control cerns in public sector service provision, evidence from the cases Program offered food packages, employment training, and highlighted in this chapter suggests that a variety of modalities stipends to patients to improve compliance with the drug treat- are possible and that a mix of supply- and demand-side ment regime. Simultaneously, the program was dramatically incentives may even be desirable. scaled up and the number of participating health centers rose from 977 to 6,539 over the next decade. In the clinics, nurses were the medical personnel responsible for administering Behavioral Change Interventions DOTS. In isolated rural areas, the program recruited local Behavioral change interventions (box 8.3) are designed to leaders to serve under the direction of the nursing staff to induce or encourage an individual behavior change or habit administer the treatment and follow up with patients. modification to achieve specific health goals. The focus is Mexico's Education, Health, and Nutrition Program (origi- usually on prevention, but need not be exclusively so--for nally known as PROGRESA, but now called "Oportunidades") example, the use of oral rehydration therapy to treat childhood also provides a compelling example of how a program can use diarrhea. Behavioral change interventions are often linked to demand-side incentives to stimulate demand for basic health the uptake of a specific product, as in the case of condoms and care services. The program offers cash transfers to families in insecticide-treated bednets. However, unlike product-intensive exchange for the attendance of mothers and children age five interventions, behavioral change interventions require active and under at nutrition monitoring clinics and to pregnant participation by the individual and cannot be passively women if they agree to prenatal care visits and nutritional sup- received in the form of an injection or supplement. Also, unlike plementation. The program also includes cash transfers to pro- service-intensive interventions, behavioral change interven- mote school attendance and performance. A rigorous evalua- tions do not depend on the involvement of a health care pro- tion provided evidence of the program's effect on the use of fessional on an ongoing basis. Illustrative examples include the health services: after just one year of implementation, atten- uptake of oral rehydration therapy for use by mothers in the dance at health care clinics was significantly higher in partici- home, not the clinic; face washing to prevent trachoma; and pating localities. However, the increased demand for services condoms to prevent HIV infection or other sexually transmit- was also met with significant improvements in the quality of ted infections. services available through public providers. Health care As Jamison (2002) notes, some changes in behavioral prac- providers in participating localities were paid more and tices associated with improved health, such as improved Box 8.3 Behavioral Change Interventions: Illustrations from Cases Stopping smoking Using nylon filters to purify water Using condoms to prevent HIV/AIDS Using oral rehydration salts Source: Authors. Improving the Health of Populations: Lessons of Experience | 175 hygiene and better nutrition, are linked to rising incomes, but In these cases, the targeted population not only was aware of the pathway for this link is not entirely clear, especially at the the health care problem, but also was eager to adopt solutions individual level, which makes it difficult to predict which mix to address it. Clearly this situation does not always prevail, but of programs and policies can be used to change behavior at the a reasonable conclusion is that a relatively simple technology population level. Information plays an important role, but the that addresses a recognizable, but as yet unmet, need will find a evidence suggests that its effect may be limited if it is not com- receptive audience. In Bangladesh, a mass media campaign to bined with other mechanisms to induce behavioral change. encourage families to have fewer children was backed up by a The cases under review typically used a mix of strategies to large cadre of female outreach workers, who went door to door induce behavioral change, including information, education, in rural areas to provide information to young married women and communication (IEC) campaigns; regulatory policies; and to make family-planning commodities available to them. taxation or subsidies; and financial incentives or disincentives. Market research indicates that almost all Bangladeshi women The evidence is suggestive, but hardly conclusive, that a relative were in favor of family planning but were unable to go against hierarchy exists among the available strategies and policy their husbands' objections if they were opposed to the use of instruments and that some are more effective than others at contraceptives. The fact that the campaign and the provision of altering how individuals perceive risk and weigh the costs and contraceptive commodities addressed an unmet need among benefits of behavioral change. Chapter 11, for example, pro- Bangladeshi women may provide a partial explanation for the vides an in-depth discussion of how and when fiscal instru- rapid acceptance of the program when it was launched on a ments may be used effectively to alter producers' and con- large scale. sumers' decisions in ways that encourage healthy behaviors. Regulatory Policies. By contrast, other behavioral changes, Information, Education, and Communication. Recent stud- such as using condoms to prevent HIV infection and other ies have challenged the effectiveness of mass IEC campaigns sexually transmitted infections and stopping smoking, have (Kremer and Miguel 2003), and evidence from the cases clearly been harder to induce, although the precise reasons for appears to support a degree of healthy skepticism. However, this difficulty remain elusive (see, in particular, chapters 18 and IEC campaigns did appear to have an effect in the cases under 46). Nevertheless, evidence from the cases suggests that gov- review when they were accompanied by the promotion of a ernments can put the right mix of policies in place to either new technology or a product. Also, acceptability appeared to discourage high-risk behaviors or encourage health-promoting increase if the product was adapted to fit the local circum- behaviors. For example, they may use regulatory policies to stances or cultural context. In Sub-Saharan Africa, educational ensure compliance through nonfinancial means and may com- campaigns to stop transmission of the guinea worm to humans plement these by using fines and sanctions as an enforcement have encouraged the construction and maintenance of safe mechanism. water sources (through deep-well digging and the application The Thai government's 100 Percent Condom Program is an of larvicide to contaminated ponds) and the use of cloth or excellent example of such a strategy. In 1991, the National AIDS nylon filters to purify drinking water, in addition to case iden- Committee launched a national program to be implemented at tification and containment. The primary IEC tool was so-called the provincial level requiring all workers in brothels and other worm weeks--that is, weeks of intensive health education and commercial sex establishments to refuse to have sex with any community mobilization. client not using a condom. The program had several compo- In Egypt, in addition to nurses and physicians, mothers were nents, including free distribution of condoms to health workers a primary target of the campaign to promote the use of oral during regular health checks and a media campaign to raise rehydration salts for the treatment of diarrhea in children, awareness of the risks of HIV and the dangers associated with especially those under three years of age. Television was the unprotected sex. What gave the program its unique character primary educational medium and proved to be an effective was its regulatory and enforcement component. Local govern- strategy for reaching a broad population base, including rural, ments, health authorities, and police officers were responsible illiterate households. Appropriate product design and branding for monitoring and enforcing condom use in the brothels. were essential. Oral rehydration salt packets were supplied in Those brothels that failed to comply with the strict policy a 200-milliliter size, not the standard 1-liter packs, because would be fined or forced to shut down. The results were impres- mothers did not have appropriate containers at home and sive. Condom use in brothels exceeded 90 percent in just the felt that a full liter was too much to give to a child to drink. first year of the program, up from 14 percent in 1989. The num- By contrast, in Bangladesh, where few households had access ber of new HIV infections fell by more than 80 percent, from to a radio, much less a television, health workers went door 142,819 cases in 1991 to 25,790 cases in 2001. Furthermore, the to door to teach mothers how to make the solution in their program appears to have generated important spillover effects, homes. or externalities, in unexpected places. For example, studies 176 | Disease Control Priorities in Developing Countries | Carol Ann Medlin, Mushtaque Chowdhury, Dean T. Jamison, and others indicate that indirect sex workers--a group that cannot be Environmental Control Interventions reached through similar enforcement strategies--have also As with behavioral change interventions, environmental con- begun to insist that their clients use condoms. trol interventions are geared toward prevention and are used in conjunction with other treatments or alone when effective vac- Taxation and Subsidies. Another option for discouraging cines or other prophylaxes are unavailable. However, rather high-risk behaviors is the adoption of taxation policies. A 1997 than focus on risk factors associated with individual behavior, study by the World Bank in partnership with the World Health environmental control interventions target risks associated Organization found that a price increase of 10 percent on cig- with the physical environment that are largely beyond the arettes would lower smoking rates by about 4 percent in high- individual's control. The physical environment refers to media income countries and about 8 percent in low-income countries in the natural (water, air, or soil) or man-made (housing, roads) (Jha and Chaloupka 2000, 358). In South Africa, a 50 percent environment. Examples of environmental control interventions tax on the retail price of cigarettes contributed to a 30 percent include many vector control strategies, such as aerial and decrease in consumption. In Poland, an increase in taxes on household spraying, water and sanitation projects, and air qual- cigarettes from 30 to 47 percent of the retail price, in conjunc- ity control measures. (See, in particular, chapters 41 and 43.) tion with other policies--including a ban on smoking in health Few of the cases reviewed for this study involved "pure" care establishments, schools, enclosed spaces in the workplace, environmental control interventions; although in several cases and elsewhere--contributed to a dramatic decline in smoking environmental control measures were coupled with other types rates. Before the fall of communism, Poland had the highest of interventions. Whenever activities involved in executing cigarette consumption in the world, but by the end of the the intervention fall outside the realm of typical health care 1990s, there were 4 million fewer smokers compared with services, nonhealth government agencies will be involved, and the previous decade, and cigarette consumption had fallen by strong political and technical leadership will be required of the 10 percent. These successes led to a 30 percent decrease in lung ministry of health to ensure that proper attention is given to cancer among men age 20 to 44 and a 19 percent decrease health care concerns during implementation. among men age 45 to 64 over the same period. Successful interventions in this category also appeared to be The section on service-intensive investments has already associated with strong multicountry partnerships, particularly discussed the remarkable power of explicit subsidies to encour- in relation to vector control activities. This factor raises an age health-promoting behaviors. The Mexican case of PRO- interesting question concerning governments' capacity to GRESA offers a by now familiar example of such a strategy. engage in these types of partnerships. Technical capacity, Although none of the cases included in this study dealt with although essential, is just the first hurdle, because what is ulti- subsidies explicitly directed at behavioral change interventions, mately required is a political and financial commitment at the evidence suggests that they have a significant effect. However, highest levels in participating countries. as Nugent and Knaul discuss in chapter 11, poorly targeted In the case of Chagas disease control, the Southern Cone subsidies can be costly, particularly relative to the result Initiative brought together the seven countries in the endemic achieved. Much more research is needed in this area to under- region under a coordinated, comprehensive, Chagas disease stand the specific contexts in which subsidies can achieve their control strategy. This multicountry approach was led by the desired effects and their cost-effectiveness relative to other Pan American Health Organization, following the recognition types of interventions. that disease transmission from neighboring countries was threatening the health gains achieved under Brazil's national Summary. In sum, behavioral change interventions differ eradication plan. A central aim of the Southern Cone Initiative from both product-intensive and environmental control was to eliminate the protozoan parasite, Trypanosoma cruzi, in interventions in that they require active participation by the the region by coordinating technical efforts to detect and elim- individual for the intervention to be fully effective. In some inate it. This coordination ensured consistent use of highly cases, this requirement is also true for service-intensive inter- effective control measures across the region and limited any ventions, but it is not as exclusively true as for behavioral possibility of reinvasion. Infested homes throughout the region change interventions. The degree to which incentives (or disin- were treated with long-lasting pyrethroid insecticides and centives) are needed to induce (or discourage) behavior structurally improved to eliminate hiding places for the blood- depends on the interaction between the usage characteristics of sucking insects that spread the parasite. These coordinated the intervention, the perceived risk of not using the interven- environmental control efforts, combined with blood screening, tion, and the perceived effectiveness of the intervention. To proved to be highly successful at interrupting Chagas disease have an effect, policy incentives (and disincentives) must either transmission in the region. succeed in changing the individual's risk-benefit assessment or Environmental control interventions also appear to be char- make ignoring the policy extremely costly. acterized by many consecutive years of sustained activity. In Improving the Health of Populations: Lessons of Experience | 177 some cases, this activity may take the form of ongoing mainte- have a critical role to play in providing technical assistance to nance of filters and distribution systems to prevent recontami- countries in formulating policies and developing strategic plans nation. In other cases, as in the case of vectorborne disease con- that are tailored toward their specific needs and capabilities. trol in which elimination or eradication is within reach, the Perhaps the most successful example of bundling interven- intervention must be sustained at least until transmission has tions included in our study is the Moroccan National Blindness been interrupted in humans. (The intervention may need to be Control Program, launched in 1991. The program was based sustained for much longer, or indefinitely, if the disease has an on the development in the mid 1980s of SAFE, a comprehen- animal host.) In the case of onchocerciasis, the lifetime of the sive strategy to treat and prevent trachoma. The philosophy of filarial load is 14 years, and vector spraying would need to be the new strategy, which was heavily researched and promoted sustained for at least 14 years to cut transmission. If reinvasions under the support and guidance of the Edna McConnell Clark or native vector population growth are allowed to take hold Foundation, was to augment the traditional medical approach before the human loads die off, transmission will occur and the to the treatment of trachoma with behavioral and environ- disease will not be effectively controlled. The OCP in West mental changes. The four main interventions the strategy rec- Africa has demonstrated the effectiveness of a long-term com- ommended were surgery (service-intensive intervention); mitment to vector control in reducing the burden of onchocer- antibiotics (product-intensive intervention); face washing ciasis. Since its inception in 1974, the OCP has used aerial (behavioral change intervention); and environmental activities, spraying of blackfly breeding sites to control vector popula- including water and sanitation programs (environmental con- tions and interrupt transmission. The original OCP area trol intervention). remains free of onchocerciasis transmission following the The program consisted of a wide-ranging partnership that withdrawal of regular spraying after 14 years. This sustained included five government divisions: the Ministry of Health, the attention proved to be highly successful in eliminating Ministry of National Education, the Ministry of Employment, onchocerciasis transmission and infection. Even after the intro- the Ministry of Equipment, and the National Office for Potable duction of ivermectin drug therapy to the program, vector con- Water. Targets were set--Morocco's political leaders were com- trol has remained a key strategy for maintaining health benefits mitted to eliminating trachoma by 2005--and the institutional in the expansion area of the OCP. and policy artillery to support the initiative was quickly put in In sum, environmental control interventions are associated place. Mobile surgical teams were deployed to small towns and with intersectoral collaboration, multicountry partnerships, villages to perform a simple, quick, and inexpensive procedure sustained activity, and attentiveness to context (for example, of the eyelid to halt corneal damage in infected patients; treat- epidemiological conditions, patterns of disease transmission, ment campaigns were organized to distribute the newly discov- and vector behavior). These factors place special capacity ered antibiotic, azithromycin, that could be administered in a requirements on governments that may differ markedly from single dose; IEC campaigns were launched at the community nonenvironmental interventions. Thus, to be effective, level with the participation of the Ministry of Education to ministries of health must be capable of advocacy and influence educate the population about the causes of the disease and how within the broader structures of government. to prevent it; and the National Office for Potable Water has expanded water and sanitation projects in many areas of the country. By 1999, prevalence levels had dropped 75 percent, Combination or Bundled Interventions from 28 to 6.5 percent, and acute infections in children had A two-front battle involving both prevention and treatment been reduced significantly. must be waged against most diseases and conditions to achieve the desired effect on morbidity and mortality. Under such cir- CONCLUSION cumstances, a combination or bundling of interventions is needed, adding an additional layer of complexity to the deploy- The accumulation of evidence presented in this study should ment of any particular intervention. An effective malaria con- help allay any remaining doubts about whether existing tech- trol strategy, for example, demands effective distribution and nologies and interventions, proven to be cost-effective in uptake of insecticide-treated bednets (a behavioral change randomized controlled trials, can be successfully deployed to intervention) and rapid treatment of malaria symptoms improve the lives and health of people throughout the devel- through strong outreach and referral systems (a service- oping world. The evidence suggests not only that it is possible, intensive intervention). Similarly, countries have adopted a but also that it has been achieved in many parts of the world, in dual approach to controlling the spread of HIV/AIDS: promot- many different socioeconomic and political settings. ing safe sex practices (a behavioral control intervention) and The study also found important commonalities among supporting the scale-up of antiretroviral therapy (a service- programs and projects that appear to have contributed to the intensive intervention). In this context, global partnerships successful deployment and rapid scale-up of cost-effective 178 | Disease Control Priorities in Developing Countries | Carol Ann Medlin, Mushtaque Chowdhury, Dean T. Jamison, and others interventions. Strong leadership, effective management, realis- increases, the quality of service delivery will improve, especially among tic financing, country ownership, and application of new poorer groups. research findings and technical innovation all played a role in implementation and appeared to have made major contribu- REFERENCES tions to the positive achievements of the cases under review. Collier, D., and H. E. Brady, eds. 2004. Rethinking Social Inquiry: Diverse In some respects, the study also presents a sobering view of Tools, Shared Standards. Lanham, MD: Rowman and Littlefield. the difficulties inherent in moving from a cost-effective inter- Davis, K. 1956. "The Amazing Decline of Mortality in Underdeveloped vention to a successful program or project. No single formula Areas." American Economic Review 46 (2): 305­18. is available, and identification of unique characteristics and Devarajan, S., D. Dollar, and T. Holmgren. 2001. Aid and Reform in Africa: attributes that will permit the large-scale, effective deployment Lessons from Ten Case Studies. Washington, DC: World Bank. of many known interventions is difficult. Eckholm, E. 1989. "River Blindness: Conquering an Ancient Scourge." In addition, evidence from the case studies suggests that the New York Times Magazine, January 8. programmatic characteristics and policies associated with suc- Filmer, D., J. S. Hammer, and L. H. Pritchett. 2000. "Weak Links in the Chain: A Diagnosis of Health Policy in Poor Countries." World Bank cessful outcomes vary depending on the type of intervention. Research Observer 15 (2): 199­224. Although no single formula exists, the implementation of the Frank, C., M. K. Mohamed, G. T. Strickland, D. Lavanchy, R. R. Arthur, programs and projects structured around various types of inter- L. S. Magder, and others. 2000. "The Role of Parenteral ventions appears to depend on certain types of organizational, Antischistosomal Therapy in the Spread of Hepatitis C Virus in Egypt." Lancet 355 (9207): 887­91. managerial, and financial capacities that can be anticipated and Jamison, D. T. 1993. "Disease Control Priorities in Developing Countries: specifically targeted for strengthening before the full-scale An Overview." In Disease Control Priorities in Developing Countries, ed. launch of a program or project. Thus, the findings of this study D. T. Jamison, W. H. Mosley, A. R. Measham, and J. L. Bobadilla, 3­34. may serve as pointers for future research seeking to understand New York: Oxford University Press. the range of government capacities that are needed to support ------. 2002. "Cost-Effectiveness Analysis: Concepts and Applications." the successful deployment and scaling up of interventions in In Oxford Textbook of Public Health, 4th ed., ed. R. Detels, J. McEwen, R. Beaglehole, and H. Tanaka. Oxford, U.K.: Oxford University Press. various contexts and in different parts of the world. Jamison, D. T., L. J. Lau, and J. Wang. 2005. "Health's Contribution to Economic Growth in an Environment of Partially Endogenous ACKNOWLEDGMENTS Technical Progress." In Health and Economic Growth: Findings and Policy Implications, ed. G. Lopez-Casasnovas, B. Rivera, and L. Currais, 67­91. Cambridge, MA: MIT Press. The authors gratefully acknowledge the helpful comments and Jha, P., and F. Chaloupka. 2000. "The Economics of Global Tobacco feedback from Gerald T. Keusch, Phil Musgrove, John Control." British Medical Journal 321: 358­61. Peabody, and members of the What Works Working Group. Kremer, M., and E. Miguel. 2003. "The Illusion of Sustainability." Thanks also to Carol Kolb for providing excellent research http:// emlab.berkeley.edu/users/emiguel/miguel_illusion.pdf. assistance. The basic idea for the What Works Working Group, Levine, R., and the What Works Working Group with Molly Kinder. 2004. and for the material in this chapter, came principally from Millions Saved: Proven Successes in Global Health. Washington, DC: Center for Global Development. Richard Klausner, and we owe him special acknowledgment. Mill, J. S. 1843. "Of the Four Methods of Experimental Inquiry." In A System of Logic, Raciocinative, and Inductive. Book 3. Reprint. Toronto: NOTES University of Toronto Press, 1974. Mills, A. 1983. "Vertical vs. Horizontal Health Programs in Africa: 1. Unless otherwise indicated, the background information and health Idealism, Pragmatism, Resources, and Efficiency." Social Science and impact data presented about the 17 cases reviewed for this study are drawn Medicine 17 (24): 1971­81. from Levine and What Works Working Group (2004). All materials are Musgrove, P. 1988. "Is Polio Eradication in the Americas Economically available at www.cgdev.org/publication/millionssaved. Justified?"Bulletin of the Pan American Health Organization 22 (1): 1­16. 2. According to World Development Report 2004 (World Bank 2004), because the relationship between provider and client differs, each of the Tucker, J. B. 2001. Scourge: The Once and Future Threat of Smallpox. New three types of service arrangements will experience a different con- York: Grove Press. stellation of market, government, and accountability failures. The report World Bank. 2004. World Development Report 2004: Making Services Work proposes that if these failures are properly addressed and client power for Poor People. New York: Oxford University Press. Improving the Health of Populations: Lessons of Experience | 179 Chapter 9 Millennium Development Goals for Health: What Will It Take to Accelerate Progress? Adam Wagstaff, Mariam Claeson, Robert M. Hecht, Pablo Gottret, and Qiu Fang The scale of the diseases and conditions that the Millennium for households. In Vietnam, they are estimated to have pushed Development Goals (MDGs) address is staggering: 3 million people into poverty in 1993 (Wagstaff and van Doorslaer 2003). · Almost 11 million children died before their fifth birthday Beyond the direct impact of ill health on households' living in 2000 (UNICEF 2001). Less than 1 percent of these 11 mil- standards through out-of-pocket expenditures, it indirectly lion deaths (79,000) occurred in high-income countries, affects labor income through productivity and the number of compared with 42 percent in Sub-Saharan Africa, 35 percent hours that people can work. The effects of illness on income, in South Asia, and 13 percent in East Asia. which may take time to appear, are often long lasting. Mal- · In 1998, an estimated 843 million people were considered nourished children are less likely to attend school and less likely undernourished on the basis of their food intake (FAO to learn when they do attend, reducing their productivity in 2000). Of the estimated 140 million children under the age later life. The devastating economic consequences of illness and of five who were underweight, almost half (65 million) were death are evident at the macroeconomic level as well. The AIDS in South Asia. epidemic alone has been estimated to reduce rates of eco- · Of the 3.1 million people who died from HIV/AIDS in nomic growth by 0.3 to 1.5 percentage points annually (Bell, 2003, almost all (99 percent) were in the developing Devarajan, and Gersbach 2003). world--74 percent in Sub-Saharan Africa alone (UNAIDS In the 1990s, the international community recognized the 2004). Tuberculosis and malaria together killed an equal importance of health in development. In a period when overall number; most of these deaths were among the poor. official development assistance declined, development assis- · In 1995, 515,000 women died during pregnancy or child- tance to health rose in real terms. World Bank lending for birth: 1,000 in the industrial world, contrasted with 252,000 health increased, with a doubling of the share of International in Sub-Saharan Africa (UNICEF 2001). Development Association disbursements going to health This burden of death and suffering is heavily concentrated in (OECD Development Assistance Committee 2000). The 1990s the world's poorest countries (Wagstaff and Claeson 2004). saw an increased global concern over the debt in the develop- Death and disease matter in their own right, but they also act as a ing world, fueled in part by a perception that interest payments brake on poverty reduction. Nobel laureate Amartya Sen (2002) were constraining government health expenditures in has described health as one of "the most important conditions developing countries. The enhanced Highly Indebted Poor of human life and a critically significant constituent of human Country Initiative, spearheaded by the International Monetary capabilities which we have reason to value." Health also matters Fund and World Bank in response to the unsustainable debt because it influences the living standards of both households burden of the poorest countries, was explicitly geared to chan- and countries. Health expenses can easily become burdensome nel freed resources into the health and other social sectors. The 181 Poverty Reduction Strategy Papers submitted by governments goals are indirectly related to health--for example, the goals on of developing countries seek debt relief or concessional education and gender. Gender equality is considered important (low-interest) International Development Association loans to to promoting good health among children. Other health out- set out their plans for fighting poverty on all fronts, including comes than those included in the MDGs measure progress on health. health--for example, targets related to noncommunicable The 1990s also saw the development of major new global diseases. These targets are referred to as the MDG plus and health initiatives and partnerships, including the Joint United are included in national priority setting, especially in many Nations Programme on HIV/AIDS (UNAIDS); the Global middle-income countries. Alliance for Vaccines and Immunization; the Stop TB Partnership; the Roll Back Malaria Partnership; the Global THE MILLENNIUM DEVELOPMENT GOALS Fund to Fight AIDS, Tuberculosis, and Malaria; and the Global FOR HEALTH: PROGRESS AND PROSPECTS Alliance for Improved Nutrition. A range of new not-for-profit organizations were set up to spur the accelerated discovery and Of the MDGs for which trend data are available or estimated, uptake in developing countries of low-cost health technologies the fastest progress has been on malnutrition, whereas overall to address the diseases of the poor; these organizations progress on under-five mortality and maternal mortality has included the International AIDS Vaccine Initiative, the been slower. Medicines for Malaria Venture, the Global Alliance for Tuberculosis, and the International Trachoma Initiative. In A Mixed Score at Halftime addition, the scale of philanthropic involvement in interna- In-depth analysis of the health-related MDGs shows a mixed tional health increased, with the launch of the Bill & Melinda score at halftime (Wagstaff and Claeson 2004): Gates Foundation and the Packard Foundation and the contin- ued attention to global health issues by such established entities · The number of people living in on-track countries (coun- as the Rockefeller Foundation. These initiatives brought not tries that will reach the MDGs if they maintain the rate of only new resources--funds, ideas, energy, and mechanisms-- progress they have already achieved during the period from but also new challenges to harmonization in the attempt to coor- 1990 to the present) matters. For the malnutrition target, dinate and link global goals with local actions in the fight against 77 percent of the developing world's people live in an on- disease, death, and malnutrition in the developing world. track country, but in Sub-Saharan Africa only 15 percent of As the 1990s closed, the international community decided the people live in an on-track country. that even more needed to be done. At the United Nations · Different indicators show different levels of improvement. Millennium Summit in September 2001, heads of 147 states For under-five mortality, the developing world was reduced endorsed the MDGs, nearly half of which concern different by an average of only a 2.5 percent in the 1990s, well short aspects of health--directly or indirectly (box 9.1). Several other of the target of 4.3 percent. Box 9.1 The Health-Related Millennium Development Goals Goal 1--eradicating extreme poverty and hunger. This goal mortality ratio, equivalent to an annual rate of reduction of includes as a target the halving between 1990 and 2015 of 5.4 percent. the proportion of people who suffer from hunger, with Goal 6--combating HIV/AIDS, malaria, and other diseases. progress to be measured in terms of the prevalence of The target is to halt and begin to reverse the spread of these underweight children under five years of age. The target diseases by 2015. implies an average annual rate of reduction of 2.7 percent. Goal 7--ensuring environmental sustainability. This goal Goal 4--reducing child mortality. The target is to reduce by includes as a target the halving by 2015 of the proportion two-thirds between 1990 and 2015 the under-five mortality of people without sustainable access to safe drinking water. rate, equivalent to an annual rate of reduction of 4.3 percent. Goal 8--developing a global partnership for development. Goal 5--improving maternal health. The target is to reduce This goal includes as a target the provision of access to by three-quarters between 1990 and 2015 the maternal affordable essential drugs in developing countries. Source: United Nations Millennium Declaration, the United Nations Millennium Summit 2000. 182 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others · Regional differences are also pronounced, with Sub-Saharan faster in 2000­15, especially in Sub-Saharan Africa (Wagstaff Africa faring worse than other regions. In Africa, trends in and Claeson 2004). Gender equality in school and access to reducing under-five mortality and underweight in children clean water will have a positive effect on progress toward the were barely above zero during the 1990s, and maternal mor- health MDGs. Even with economic growth and faster progress tality fell on average by just 1.6 percent a year compared on these nonhealth goals, however, many regions will still miss with the annual target rate of 5.4 percent. many of the health targets. The picture is bleakest for under- · Evidence on how the poor are faring within countries is five mortality--and for Sub-Saharan Africa. mixed. For malnutrition, the poorest 20 percent of the pop- The Goals Matter for All Countries. These goals need to be ulation within countries appears, on average, to have been taken seriously for three main reasons: experiencing broadly similar rates of reduction to the pop- ulation as a whole. However, for under-five mortality, the · Faster progress is important even if targets are missed. A key rate has been falling more slowly among the poor, while message of this chapter is that progress can be accelerated better-off families are seeing faster rates of progress. in all countries through a judicious mix of spending and pol- icy and institutional reform. Will the Second Half Go Better? · The goals facilitate benchmarking and monitoring of results. As a comparison of the child mortality experiences in the 1980s Because the goals focus on a limited set of outcomes, moni- and 1990s demonstrates, past performance is not necessarily a toring and evaluating progress toward the MDGs can show good predictor of future performance. The fact that a country what is achievable and where faster progress can be made. is on track on the basis of its performance in the 1990s does not · Focusing attention on national progress, as measured by guarantee that it will maintain the required annual rate of distributional analysis of the MDGs, forces countries to con- reduction of malnutrition or mortality during the second half sider how the benefits of progress are distributed among the of the MDG "window" from 2000 to 2015. Countries currently rich and poor within each country--the poor risk being left off track may possibly get on track in the second half if they can behind even in countries making progress overall. One limi- combine good policies with expanded funding for programs tation of the MDGs and targets is that they are national aver- that address both the direct and the underlying determinants ages. However, distributional analysis of MDG trends of the health-related goals. (Wagstaff and Claeson 2004) reminds us that progress needs to be for everyone, not just the better off. Progress has been Stimuli External to the Health Sector. The World Bank esti- uneven, with the poorer countries lagging behind the rest, mates that economic growth will fall somewhat in East Asia and for under-five mortality, the poor within countries are and the Pacific in 2000­15, turn from negative to positive in lagging behind the rest of the population. Europe and Central Asia as well as Sub-Saharan Africa, and increase somewhat in Latin America and the Caribbean, the SCALING UP: DEFINING INTERVENTIONS Middle East and North Africa, and South Asia (Jones and oth- AND REMOVING CONSTRAINTS ers 2003). Primary education completion rates will probably grow faster in the new millennium as a result of the global edu- A lack of interventions is not the primary obstacle to faster cation initiatives and partnerships on Education for All and progress toward the goals, although new interventions that can the Fast-Track Initiative. However, higher rates of educational be delivered by weak health systems could greatly improve attainment among women of childbearing age will not be progress--for example, malaria or HIV vaccines and effective achieved until 2005 or so, and even then the first full round of vaginal microbicides to block the spread of HIV and other effects on under-five mortality will not be felt until 2010. sexually transmitted infections. The main obstacle is the low More relevant is the fact that gender gaps in secondary edu- levels of use--especially among the poor--of existing effective cation may well narrow faster in the new millennium than in interventions. For example, if use of all the proven effective the 1990s as a result of the gender MDG (Goal 3: Eliminate preventive and treatment interventions for childhood illness gender disparity in primary and secondary education by 2005 were to rise from current levels to reach all, the number of and in all levels of education no later than 2015). To achieve under-five deaths worldwide could fall by as much as 63 per- parity with boys by 2015 in the proportion of the population cent (World Bank 2003b). who are age 15 and have completed secondary education, girls will have to achieve a faster growth in completion rates in the Array of Interventions, Programs, and Service Modalities new millennium than in the 1990s in most regions, especially The available interventions constitute a powerful arsenal for in South Asia and in East Asia and the Pacific. If the water preventing and treating the main causes of malnutrition and MDG (ensuring that households have access to safe drinking death (table 9.1).1 The major diseases and conditions that the water) is to be reached, access rates will need to grow much MDGs aim to prevent and control are discussed in several Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 183 Table 9.1 Effective Interventions to Reduce Illness, Deaths, and Malnutrition MDG Preventive interventions Treatment interventions Child mortality Breastfeeding; hand washing; safe disposal of stool; latrine use; safe prepara- Case management with oral rehydration therapy tion of weaning foods; use of insecticide-treated bednets; complementary feed- for diarrhea; antibiotics for dysentery, pneumo- ing; immunization; micronutrient supplementation (zinc and vitamin A); prenatal nia, and sepsis; antimalarials for malaria; care, including steroids and tetanus toxoid; antimalarial intermittent preventive newborn resuscitation; breastfeeding; comple- treatment in pregnancy; newborn temperature management; nevirapine and mentary feeding during illness; micronutrient replacement feeding; antibiotics for premature rupture of membranes; clean supplementation (zinc and vitamin A) delivery Maternal mortality Family planning (lifetime risk); intermittent malaria prophylaxis; use of Antibiotics for preterm rupture of membranes, insecticide-treated bednets; micronutrient supplementation (iron, folic acid, skilled attendants (especially active management calcium for those who are deficient) of third stage of labor), basic and emergency obstetric care Nutrition Exclusive breastfeeding for 6 months, appropriate complementary child feeding Appropriate feeding of sick child and oral rehy- for next 6­24 months, iron and folic acid supplementation for children, improved dration therapy, control and timely treatment of hygiene and sanitation, improved dietary intake of pregnant and lactating infectious and parasitic diseases, treatment and women, micronutrient supplementation for prevention of anemia and vitamin A monitoring of severely malnourished children, deficiency for mothers and children, anthelmintic treatment in school-age high-dose treatment of clinical signs of vitamin A children deficiency HIV/AIDS Safe sex, including condom use; unused needles for drug users; treatment of Treatment of opportunistic infections, sexually transmitted infections; safe, screened blood supplies; antiretrovirals co-trimoxazole prophylaxis, highly active in pregnancy to prevent maternal to child transmission and after occupational antiretroviral therapy, palliative care exposure Tuberculosis Directly observed treatment of infectious cases to prevent transmission and Directly observed treatment to cure, including emergence of drug-resistant strains and treatment of contacts, Bacillus early identification of tuberculosis symptomatic Calmette-Guérin immunization cases Malaria Use of insecticide-treated bednets, indoor residual spraying (in epidemic-prone Rapid detection and early treatment of uncompli- areas), intermittent presumptive treatment of pregnant women cated cases, treatment of complicated cases (such as cerebral malaria and severe anemia) Source: Authors. chapters (for example, see chapters 15, 19, 21­24, 28­31, 44, tive interventions that could save their lives or make them well and 45). The most cost-effective interventions and programs nourished. In middle- and high-income countries, 90 percent of are also discussed in several chapters (see chapters 59­62 and children are fully immunized, more than 90 percent of deliver- 65). The chapters dealing with health systems and service ies are assisted by a medically trained provider (that is, a doctor, delivery issues and other constraints related to the health nurse, or trained midwife, excluding traditional birth atten- MDGs are found in the latter part of the book (see chapters dants), and more than 90 percent of pregnant women have at 66­68, and 73). least one prenatal visit (UNICEF 2001). In South Asia, fewer In the case of child mortality, for example, diarrheal dis- than 50 percent of pregnant women receive a prenatal checkup, eases, pneumonia, and malaria account for 52 percent of deaths and only 20 percent of deliveries are assisted by a trained worldwide (World Bank 2003b). For each of these major causes provider. of childhood mortality, at least one proven effective preventive The story is similar for other childhood interventions-- intervention and at least one proven effective treatment inter- and for interventions for other goals. Condom use to prevent vention exist, capable of being delivered in a low-income set- transmission of HIV is low in much of Sub-Saharan Africa ting. In most cases, several proven effective interventions exist. and South Asia, and inexpensive one-time treatment with For diarrhea--the second-leading cause of child deaths--no antiretroviral medicine to prevent transmission from mother fewer than five proven preventive interventions and three to child covers only a small fraction of at-risk pregnant proven treatment interventions are available. women in most of the developing world. In Asia, where more than 7 million people are living with HIV/AIDS, no Effective Interventions Reaching Too Few People country has yet exceeded 5 percent antiretroviral therapy The high rates of malnutrition and death in the developing coverage among those who could benefit from it (World Bank world have several causes. First, people do not receive the effec- 2003c). 184 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others Just as shortfalls in coverage vary across countries, so do allocations. In practice, however, the amount of extra spending they vary within countries, with the poor and other deprived required would be difficult to attain on present trends and groups consistently lagging. These groups are less likely to would even be prohibitively expensive. In the case of East Asia receive full basic immunization coverage, to have their deliver- and the Pacific, for example, if economic growth proceeds as ies attended by a trained provider, and to have at least one pre- expected and the other relevant Millennium Development natal care visit to a medically trained provider. On the positive Targets are attained, the region would achieve the required side, the poor are often making fastest progress in coverage, rates of reduction of underweight and maternal mortality-- reflecting in part that the better off already have high coverage assuming that economic growth is accompanied by the devel- rates for many interventions. opment of appropriate human resources for health--even if government health spending continues to grow at its current rate. However, the region would miss the under-five mortality Underuse of Effective Interventions Costs Lives target. To reach that target, a minimum of 5 percentage points The low use of effective interventions--in the developing would need to be added to the annual rate of growth of the world in general and among the poor in particular--translates government health share of gross domestic product (GDP). into rates of mortality, morbidity, and malnutrition that are far That would take the projected share of GDP spent on govern- higher than necessary. If use of all the proven effective child- ment health programs to 3.7 percent in 2015--more than twice hood preventive and treatment interventions, for example, what it would be if the 1990s pattern of growth continued were to rise from their current levels to 99 percent--95 percent (Wagstaff and Claeson 2004). for breastfeeding--the number of under-five deaths worldwide In Sub-Saharan Africa, the situation is even starker. Even if could fall by as much as 63 percent (Jones and others 2003). faster economic growth materializes and the other targets are Deaths from malaria and measles could be all but eliminated, achieved, the share of government health spending in GDP and deaths from diarrhea, pneumonia, and HIV/AIDS could be would need to grow nearly sixfold over the coming decade, tak- reduced dramatically. If coverage rates of the key maternal ing the share to 12.2 percent of GDP in 2015. This percentage mortality interventions were increased from current levels to compares with a 2000 figure of 1.8 percent and a 2015 forecast 99 percent, an estimated 391,000 maternal deaths worldwide of 2.2 percent based on the 1990s annual growth in govern- (74 percent of current maternal deaths) might be averted ment spending for health. In conclusion, African countries will (Ramana 2003). One intervention stands out as especially not be able to reach the MDGs simply by multiplying their important: access to essential obstetric care, which accounts for health spending along the lines of historical expenditure pat- more than half the maternal deaths averted. terns, because the multiples required are beyond any realistic expectation of what these governments will be able to do dur- ing the next 10 years. WHAT DO COUNTRIES NEED TO DO? What Are the Implications? If the lack of interventions is not holding countries back from achieving the goals, what is? What do countries need to do to Poorly governed countries cannot expect to make much make progress toward the MDGs? progress toward the MDGs simply by scaling up their expendi- In countries with good governance, additional government tures on existing programs in proportion to current alloca- health spending does reduce child mortality (Rajkumar and tions. Although well-governed countries could, in principle, Swaroop 2002). Development assistance has a stronger effect in simply scale up existing spending to reach the targets, this countries with strong policies and institutions than in coun- option is unlikely to be affordable for them or their donors. tries with only average-quality policies and institutions--and This situation has two implications: an insignificant effect in countries where policies and institu- tions are weak. This assertion is also consistent with the find- · First, targeting additional government spending to activities ings of a study undertaken by the World Bank for the MDG that will have the largest effect on the MDGs is important report, The Millennium Development Goals for Health: Rising to for both sets of countries. the Challenges (Wagstaff and Claeson 2004). The study includes · Second, building good policies and institutions is important other outcomes with child mortality and uses the World Bank's for all countries: doing so increases the productivity not just Country Policy and Institutional Assessment index to measure of additional spending but also of existing spending com- the quality of policies and institutions. mitments. What do better policies and institutions entail in In principle, well-governed countries with good policies and the health sector? Health systems are very broad, and weak institutions could achieve the goals simply by scaling up their policies and institutions can arise at several points along expenditures on existing programs in proportion to current the pathway, from government health spending to health Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 185 outcomes (Claeson and others 2001). Countries can do a Targeting Specific Population Groups. Many countries subsi- number of things, with help from donors, to build stronger dize all government health services for everyone. These blanket policies and institutions. subsidy schemes not only fail to target interventions that give rise to externalities but also fail to disproportionately benefit the poor--despite the stronger equity case for subsidizing their Improving Expenditure Allocations and Targeting care and the fact that they tend to bear a disproportionate In most countries, government spending gets stuck in the cities burden of malnutrition as well as child and maternal mortality. and disproportionately accrues--in a financial sense--to There are many proven ways to target the poor--for example, people who are better off. by delivering essential services in clinics or health posts that only poor families attend or by promoting and delivering serv- Geographic Targeting. Resource allocation formulas can be ices in a way that segments the market and appeals to those in used to reduce government spending gaps across regions and low-income households. ideally to favor geographic zones that are furthest behind. These formulas have been used, for example, as part of Bolivia's Targeting Spending to Remove Bottlenecks. A planning and decentralization efforts since 1994 and have been associated budgeting approach is to assess--for a country--the health with some large--and pro-poor--improvements in maternal sector impediments to faster progress, to identify ways of and child health indicators. Targeting resources to poor regions removing them, and to estimate both the costs of removing and provinces may be most effectively implemented through them and the likely effects of their removal on MDG outcomes nontraditional mechanisms for priority setting and implemen- (Soucat and others 2002). MDG analysis along these lines-- tation, such as social investment funds. In Bolivia, a recent referred to sometimes as marginal budgeting for bottlenecks impact evaluation concluded that such funds were responsible (MBB)--has begun in several African countries and in some for a decline in under-five mortality from 88.5 to 65.6 per 1,000 states of India (UNICEF and World Bank 2003). In Mali, key live births over a five-year period (Newman and others 2002). bottlenecks were identified for supporting home-based prac- tices and delivering periodic and continual professional care. Changing the Allocation of Spending across Care Levels. They included low access to affordable commodities and the Spending on health in developing countries is characterized by need for community-based support for home-based care; low a high concentration of spending on secondary and tertiary geographical access to preventive professional care (immuniza- infrastructure and personnel. Some governments have tried to tion, vitamin A supplementation, and prenatal care); shortages scale back the share of hospital spending. Tanzania, for exam- of qualified nurses and midwives; and an absence of effective ple, reduced the share of hospital spending from 60 percent in third-party payment mechanisms for the poor for professional 2000 to 43 percent in 2002. Chapter 3 deals with the issue of continuous care. Important health systems bottlenecks, such as how to couple expenditure reallocations across levels of care human resources, drug availability, and health care manage- with measures to improve performance at each level of the ment, are discussed in chapters 71­73. health care system. Targeting Specific Programs. Programs such as those deliver- Improving Policies toward Households as Producers ing directly observed treatment short course (DOTS) for tuber- and Demanders of Care culosis or integrated management of infant and childhood illness (IMCI) for child health are good examples of programs Households are at the center of any efforts to scale up; they not that may yield high returns to government spending at the only demand and consume care, but they are also important margin. A recent World Bank study in India provides further producers of prevention and care. Policies to increase coverage support for the idea that the way government spending is allo- of cost-effective interventions to reach the health MDGs, there- cated across programs makes a difference to its effect on the fore, need to identify and influence the key constraints to both Millennium Development Indicators (World Bank 2003a). the production and the demand for those services at the Successful public health programs--large-scale programs with household and community levels. a measurable health effect over at least a five-year period--are further discussed in chapter 8. All successful programs have Lowering Financial Barriers. Low income is a barrier to the several factors in common: technical innovation and stake- use of most health interventions, and economic growth is an holder consensus, strong political leadership, coordination important weapon in the war against malnutrition and mor- across agencies and management, effective use of information tality. However, social protection programs are also important. and financial resources, and participation of the beneficiary Successful schemes aimed at households and communities are community. discussed in chapter 56. 186 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others One part of the affordability equation is price. User charges numeracy and literacy skills learned at school to acquire health- for MDG interventions are to be discouraged. Why? Many of specific knowledge later in life. Although better-educated girls those interventions involve benefits that spill over to people will mean healthier women and healthier children in years to who do not receive the intervention; high coverage of immu- come, a shorter and more direct route to increasing health- nization is a classic example. However, an equity case also can specific knowledge and skills is through information dissemi- be made for reducing prices facing the poor and near poor, nation, health promotion, and counseling in the health sector. even where no spillovers occur. Subsidies should be targeted to Several success stories exist. In Brazil, after health workers services with spillovers and to the poor. In practice, subsidies trained by IMCI provided information and counseling at are often badly targeted in at least one respect if not both. health facilities and in the community, health knowledge Exceptions exist. In Ifakara, Tanzania, a voucher program for among mothers improved, as did feeding practices (Santos and mosquito nets was launched successfully for pregnant women others 2001). After only 18 months, the nutritional status of and children under five (Schellenberg and others 2001). children in the area improved as well. Social marketing and Some recent programs, especially in Latin America, have not media campaigns--for example, malaria and social marketing only made health care affordable for the poor but have also of insecticide-treated nets (see chapter 21)--have also proved made it profitable. Rather than simply reducing the cost of effective in some circumstances. using specific interventions, these programs provide users with cash payments, which are linked to specific interventions and Reducing Time Costs Transportation systems, road infrastruc- restricted to certain groups--often poor mothers and their ture, and geography influence the demand for care delivered by children. The experience with these programs in targeting and formal providers through their effect on time costs, which can achieving results is encouraging (Mesoamerica Nutrition be substantial. In rural communities, where the roads are poor Program Targeting Study Group 2002; Morris and others 2003; and the transportation unreliable, the time spent waiting for Palmer and others 2004). the transportation is also a major cost. Time costs tend to be a Risk aversion coupled with the unpredictability of illness major issue for maternal mortality: health centers are unable to provides a motivation for pooling risks through an insurance provide essential obstetric care for a complicated delivery, and scheme. The Arab Republic of Egypt, for example, introduced women would have to travel to distant hospitals to get a school health insurance program for all children attending such services. Road rehabilitation and other transportation school. The program resulted in larger increases in coverage projects are important here, but so are subsidies linked to among the poor and achieved considerable effect on use and the use of health services. Malaysia and Sri Lanka provide free out-of-pocket expenditures (Yip and Berman 2001). However, or subsidized transportation to hospitals in emergencies insurance in the developing world is very limited, and those (Pathmanathan and others 2003). Other options for tackling who are least able to smooth consumption without insurance inaccessibility include using outreach and establishing partner- are the least likely to have insurance coverage (Musgrove, ships between government and nongovernmental organiza- Zeramdini, and Carrin 2002). Another problem is that many of tions (NGOs), private providers, or community organizations. the schemes are small scale, and evaluations of these schemes do not generally measure health effect or effect on equity, thus Providing Access to Water and Sanitation The availability of resulting in limited evidence (Palmer and others 2004). adequate supplies of water and improved sanitation is associ- ated with better maternal and child health outcomes, at least Providing Information--Enhancing Knowledge. Lack of among the better educated, even after controlling for other knowledge is a major factor behind poor health. It results in influences. This result is not altogether surprising. Hand wash- people not seeking care when needed, despite the absence of ing is easier if the household has piped water that provides price barriers, and it also results in people--especially poor readily available quantities of safe water. The safe disposal of people--wasting limited resources on inappropriate care. feces is easier if the household has an improved form of sanita- Ignorance may also result in people not getting the maximum tion. The developing world lags well behind the industrial health gain out of inputs they have available to them and use. world in both; the poorer people fare especially badly. They are Many people do not know that hand washing confers much of less likely to be connected to a network, and the sources they the health benefit of piped water (see chapter 41). Not surpris- rely on tend to be more costly per liter than the networked serv- ingly, piped water has a much greater effect on the prevalence ices used by the better off. of diarrhea among the children of the better off and better The challenge from a health perspective is to get maximum educated. Better-educated women--especially those with a health benefits from investments in access to water and sanita- secondary education--achieve better health outcomes for tion infrastructure. Efforts to work across sectors on water and themselves and their children not by using health-specific health, in order to influence the health MDGs, are under way in knowledge that they acquire at school, but by using general Ethiopia, Peru, and Rwanda. Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 187 Improving Health Service Delivery women with two or more prenatal visits. In Peru, comparisons Health providers--in the public and private sectors, as well as of primary health care clinics with and without community in both formal and informal sectors--should deliver interven- participation in governance suggested decreases in staff absen- tions of relevance to the MDGs. Many are efficient, deliver high teeism and waiting times and suggested increases in perceived quality services, and are responsive to their patients. Many, quality by patients (Cotlear 1999). The approach probably however, are not; many are not even there to deliver any serv- works best for primary care and in situations in which strong ices at all. As a result, resources--public and private--are often technical and advisory support is provided to community rep- nonexistent, underused, or wasted. resentatives who are close to the service being delivered. Two things can make a difference. One is the quality of management. Better management means a clearer delineation Contracting. Evidence on the effect of contracting within the of responsibilities and accountabilities inside organizations, public sector is mixed, and the experiences are mainly based a clearer link between performance and reward, and so on. on lessons learned from middle-income countries. In several Management means getting accountabilities right within an countries in Europe and Central Asia, evidence shows a positive organization. The other thing that can make a difference is get- effect from performance-based payment, but that is not neces- ting accountabilities right between the organization and the sarily the same as contracting, which can occur without public (World Bank 2003d). performance-related pay. The best evidence relates to the use of target payments for the attainment of a given level of Improving Management--Increasing Accountability within coverage--for example, for immunization or cervical cytology Provider Organizations. Management styles in government- at the primary care level (Langenbrunner 2003). In Argentina funded and government-implemented health schemes have and Nicaragua, social security institutes have increased pro- recently begun to change, focusing on performance--that is, ductivity by establishing capitation-based payments for an on outputs and outcomes--rather than on inputs and pro- integrated package of inpatient and ambulatory services cesses. Good performance is rewarded, financially or in some (Bitran 2001). Key influences on the success of contracts with- other way. The focus is on clients and on the belief that an in the public sector include whether the provider has the abil- organization is ultimately accountable to its clients. A client- ity to respond, whether service commitments are congruent oriented strategy emphasizes customer choice and satisfaction. with funding levels, whether output and key components of Business techniques enhance performance and are a standard performance expectations are easily measurable, and how far part of strategic planning. capacity strengthening of the payer or funder is addressed. This new approach is evident in several countries, and ele- Contracting with nonprofit organizations is most common ments of the approach are visible in successful nutrition and in low-income countries (see chapter 12, which contains a child health programs (see chapter 56). For example, in Tamil longer discussion of contracting with NGOs). Most cases have Nadu's Integrated Nutrition Program, community nutrition had positive effects on target outcome or output variables. workers were given clearly defined duties. Information on out- In Bangladesh, contracts with nonprofit organizations for puts not only enabled the community to keep the workers planning and implementing an expanded program on immu- accountable but also enabled the nutrition workers to see how nization project were credited with a dramatic increase in their program was working. In Ceara's Programa de Agentes de immunization. In Haiti, contracting for a primary health care Saude, which is credited with a substantial reduction in child package also significantly increased immunization coverage mortality (Victora and others 2000), health agents and nurse- (Eichler, Auxilia, and Pollock 2001). In Bangladesh, supervisors were assigned clear tasks and given clear responsi- Madagascar, and Senegal, significant reductions in nutrition bilities. The intended outcomes of the program were empha- rates were attributed to contracting initiatives (Marek and oth- sized to health workers and members of the public, and the ers 1999). Only a few cases assess efficiency. Contracting with health agents were held accountable through community- nonprofits works best when the contractors have well- based monitoring and rewarded for good performance. functioning accountability arrangements and strong intrinsic motivation and when the government makes timely payments Governance. The accountability of provider organizations to to the NGOs. The government needs to be capable of assessing, the public can be improved through enhanced governance or selecting, and managing the ongoing relationship with con- contracting. Having community representatives participate in tractors. The methodological quality of evaluating contracting the governance and oversight of providers can improve the is often poor and needs to be improved. An exception is the productivity and quality of public sector providers. In Burkina Cambodian contracting trial that used a rigorous cluster ran- Faso, participation of community representatives in public domized design, but the intervention groups had greater input primary health care clinics increased immunization coverage, of resources than the control communities, which may have the availability of essential drugs, and the percentage of been partly responsible for the difference in performance. 188 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others Results on contracting with for-profit private service proxies for the goals that can help monitor progress, test the providers are also mixed. Experience from the hospital sector impact of policies, and adjust programs going forward (World warns that weak government contracting capacity often allows Bank 2001). Such indicators should be simple, easily measura- the provider to capture efficiency gains or to expand volume-- ble, representative, easy to understand, scientifically robust, and not necessarily of cost-effective services--to generate more ethical. They need to be assessed regularly because the MDGs income. In Zimbabwe, the cost per service decreased, but the themselves are difficult to collect, thus entail delays, and are lack of volume control led to an increase in total cost (McPake therefore not useful for regular monitoring of progress. Greater and Hongoro 1995). Other adverse outcomes are possible. In investments are needed in systems to monitor these intermedi- Brazil, contracting with for-profit hospitals led to increases ate indicators and to track expenditures on public health. in access, but also increases in fraud (false billing) and cream- Although some good practices in surveillance are being skimming to avoid costly patients (Slack and Savedoff 2001). developed--for example, in Brazil, China, and India--few low- These problems seem less pronounced in primary health care. income developing countries can afford to invest in the infra- In Peru and El Salvador, contracting with private primary structure required for strong surveillance systems. Most rely on health care providers increased access, choice, and consumer alternative short- to medium-term solutions for data gather- satisfaction (Fiedler 1996). Contracting with for-profit ing, such as intermittent household surveys, health facility providers seems to work best when the government invests in surveys, and simplified facility-based routine reporting. A few the development of capacity to manage the contracting process countries have made special efforts to improve the surveillance (Mills, Bennett, and Russell 2001); when quality is at least as of a specific intervention, such as AIDS and tuberculosis treat- high in the private sector as in the public sector; and when the ment or childhood immunization, whereas others have services involve primary care or other relatively observable attempted to monitor progress toward a specific MDG. services, such as diagnostic services. INDEPTH (International Network of Field Sites with Continuous Demographic Evaluation of Populations and Their Health in Developing Countries), which is supported by Strengthening Core Public Health Functions the Rockefeller Foundation with help from other donors, coor- Vulnerable populations need to be protected from risks and dinates a range of surveillance sites, many of them in Africa, damages, informed, and educated. Public health regulations and the Health Metrics network aims at improving the quality need to be established and enforced. Infrastructure needs to be of surveillance data. Some governments are explicitly develop- in place to reduce the impact of emergencies and disasters ing or modifying their monitoring and evaluation framework on health. All this action needs to be implemented through a to focus on the MDGs. public health system that is transparent and accountable. Governments in developing countries generally recognize that Intersectoral Actions--Going Beyond the Ministry of these public health functions are important, but they often lack Health. A review of the evidence base for the key determinants the capacity and financial resources to implement them. of the health and nutrition MDGs identifies significant poten- Indeed, few low-income countries invest in these public health tial for intersectoral synergies (Wagstaff and Claeson 2004). functions. By employing public health professionals with core public Transportation Although roads and transport are vital for health competencies, the government can develop and enforce health services, especially for reducing maternal mortality, it is standards; can monitor the health of communities and popula- not just the physical infrastructure that matters. Also impor- tions; and can emphasize health education, public information, tant are the availability of transportation and the affordability health promotion, and disease prevention. Public action can of its use, as shown in a study in Nigeria (Eissen, Efenne, and help improve consumer knowledge and change attitudes so Sabitu 1997). Transportation and roads complement health that private markets can operate effectively to meet the needs of services. A 10-year study in Rajasthan, India, found that better the poor, for example, through social marketing of insecticide- roads and transportation helped women reach referral treated bednets to reduce malaria transmission or of condoms facilities, but many women still died because no corresponding for protection against HIV/AIDS. improvements took place at household and facility levels. Working with the transportation sector is also important for Government-Led Monitoring and Evaluation. Integrated reducing HIV transmission in many settings and making disease surveillance, program assessment, and collection and progress on the HIV/AIDS-related MDG. analysis of demographic and vital registration data are essen- tial if governments and donors are to ascertain whether poli- Hygiene Improved hygiene (use of hand washing) and sanita- cies and programs are positively affecting health goals. tion (use of latrines and safe disposal of children's stools) are at Governments can use a list of intermediate indicators and least as important as drinking water quality in shaping health Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 189 outcomes, specifically in reducing diarrhea and associated child Cost of Achieving the MDGs Globally mortality (Esrey and others 1991). Constructing water supply The global estimates of what it would cost to achieve the MDGs and sanitation facilities is not enough to improve health out- range from an additional US$20 billion to US$70 billion a year. comes; sustained human behavior change must accompany the A World Bank study (http://www.worldbank.org/html/extdr/ infrastructure investment. By collaborating with other sectors, mdgassessment.pdf) estimates that the additional official the health sector can develop public health promotion and development assistance required to meet the health goals is in education strategies and implement them in partnership with the range of US$20 billion to US$25 billion per year, which is agencies that plan, develop, and manage water resources. The roughly four times the current amount of official development health sector can also work with the private sector to manufac- assistance spending for health in 2002 (US$6.5 billion) and ture, distribute, and promote affordable in-home water purifi- three times all external financing, including that of foundations cation solutions and safe storage vessels--and advocate for and loans from multilateral sources (see chapter 13). The dra- water, sanitation, and hygiene interventions in strategies to matic shortfalls in resources required to achieve the MDGs reduce poverty. were emphasized during the 2002 Monterrey Conference on Financing for Development, which brought significant atten- Indoor Air Quality Indoor air pollution is caused by use of tion to issues concerning the estimation of the cost of achiev- low-cost, traditional energy sources, such as coal and biomass ing the health MDGs. for cooking and heating, the main source of energy for 3.5 bil- Another analysis conducted by the Commission on lion people. Indoor air pollution is a major risk factor for pneu- Macroeconomics and Health (2001) of the World Health monia and associated deaths in children and for lung cancer in Organization estimated that an additional US$40 billion to women who risk exposure during cooking (see chapter 42). US$52 billion annually would be required until 2015 to scale up Studies in China, Guatemala, and India are under way to the coverage for malaria, tuberculosis, HIV/AIDS, childhood improve access to efficient and affordable energy sources mortality, and maternal mortality (Kumaranayake, Kurowski, through local design, manufacturing, and dissemination of and Conteh 2001). A third study using the production frontiers low-cost technologies, modern fuel alternatives, and renewable approach estimated that between US$25 billion and US$70 bil- energy solutions. The community-based project in China was lion of additional spending was needed to bring poorly per- initiated by the Ministry of Health, which was troubled by the forming countries up to the level of high performers (Preker and leveling off of child mortality reductions among the rural poor others 2003). A fourth study prepared by the World Bank for the and was seeking ways to influence major environmental deter- Development Committee estimated at least US$30 billion annu- minants of child mortality. The program combines appropri- ally in additional aid was needed to accelerate all the MDGs, ately improved stoves and ventilation with behavior-change including health (Development Committee 2003).Whatever the modification; it is in an early stage of implementation, and method of analysis, all global estimates show that reaching the results on outcomes are not yet available. Agricultural policies MDGs will require significant additional resources compared and practices influence food prices, farm incomes, diet diver- with the current levels of funding for health. sity and quality, and household food security. Policies that focus on women's access to land, training, and agricultural inputs; on their roles in production; and on their income from Cost of Achieving the MDGs in Countries agriculture are more likely to have a positive effect on nutrition Global estimates of what it costs to achieve the health MDGs than interventions without a focus on women, particularly if are not very useful for countries wanting to plan and budget in combined with other strategies, such as women's education and order to reach the MDGs. The substantial range of estimates behavior change (Johnson-Welch 1999; Quisumbing 1995). between US$20 billion and US$75 billion per year to achieve The MDG agenda highlights the need not only to prioritize the MDGs at a global level has led to debates over the most within health to achieve better health outcomes, but also to appropriate costing method for country-specific analysis and better inform priority setting in resource allocations between to the development of new costing methodologies for obtain- sectors, identifying intersectoral synergies and finding ways to ing consistent and reliable estimates to use for policy dialogue maximize benefits for health. and decision making at the country levels. Some of the meth- ods are summarized in box 9.2. COSTING AND FINANCING ADDITIONAL SPENDING FOR THE MDGS Preliminary Country Cost Estimates. Table 9.2 provides a set of preliminary country-level estimates for the cost of Additional health spending will be required in many countries removing bottlenecks and accelerating progress toward the to accelerate progress toward the health goals (see chapter 12). health MDGs (MBB method) and for the cost of achieving What will it cost, and how will extra spending be financed? the health MDGs (Millennium Project tools) in selected 190 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others countries. The estimates are presented for illustration of Encouraging Risk Pooling Rather Than Out-of-Pocket orders of magnitude and should not be used for intercountry Spending. Health spending can be broken down into three comparison. categories: Financing Extra Health Spending The additional resources needed to reach the MDGs are large · private (out-of-pocket expenditures and private insurance) at both country and global levels, as discussed in the previous · public (financing from general revenues and social insur- section. The key question is how to finance the extra spending ance contributions) that is needed. · external sources (development assistance). Box 9.2 Estimating the Cost of Scaling Up to Achieve the MDGs The following are the country-specific models for MDG removal of bottlenecks are discussed, and the inputs are cost analysis: identified for improving coverage, for example, in a vil- lage. Cost estimates are based on these inputs by scaling · The MDG Needs Assessments Model developed by the up the cost to cover the district, province, or nation. United Nations Millennium Project, (Millennium · Elasticity estimates through econometric modeling devel- Project 2004). The Millennium Project model yields oped by the World Bank staff (Wagstaff and Claeson total cost estimates for full coverage of the needs of a 2004). A few studies have used econometric techniques defined population with a comprehensive set of health to analyze the effect on MDG outcomes of certain cross- interventions in a given year. It uses unit cost of cover- sector determinants (such as economic growth, water ing one person multiplied by the total population in and sanitation, education, and road infrastructure) as need in a given year to yield the direct health cost. well as government expenditures on health. Eco- Additional resource requirements are added (on the nometric analysis has been used mostly to analyze the basis of assumptions rather than actual inputs) for, effect of changes in government health expenditures on among other items, health system improvement, salary outcomes using cross-sectional or panel data at a global increases for human resources, administration and scale. But in one particular study in India, the method- management, promotion of community demand, and ology was used to estimate the marginal costs of avert- research and development. ing a child's death at the state level. The estimates could · The Marginal Budgeting for Bottlenecks Model devel- vary from as low as US$2.40 per child death in a low- oped by the United Nations Children's Fund, the World income state to US$160 in a middle-income state in Bank, and the World Health Organization (Soucat and India. others 2002, 2004; UNICEF and World Bank 2003). The · The Maquette for Multisectoral Analysis of MDGs is MBB model yields additional resources required for under development by the World Bank (Bourguignon removing a set of health system bottlenecks that are and others 2004). The thesis for this new approach is considered to hinder the delivery of health services to that development aid is a key ingredient of a country's the population through three delivery modes: family- development process, but its effectiveness has to be community, outreach, and clinical levels. The MBB assessed at the country level within each country's local method also estimates the effect on outcomes (for implementation and macroeconomic constraints. The instance, child and maternal mortality) of increased objective of the model is to calculate the financial needs coverage and use of the health services provided. First, a to attain a targeted path to 2015 and determine an opti- set of high-impact services are selected on the basis of a mal allocation of additional funding toward different country's epidemiological needs. These services are the social sectors for the MDGs. This modeling framework same as those cost-effective priority interventions iden- is still at an early stage of development and will be tified in the relevant disease control priorities chapters. applied later to countries. This model is anticipated to Second, health system bottlenecks hindering delivery of draw extensively from results of other models, such as these services are identified. Then, strategies for the elasticity analysis and MBB models. Source: Millennium Project 2003, 2004; Soucat and others 2004; Bourguignon and others 2004. Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 191 Table 9.2 Alternative Cost Estimates Using Millennium Project and Marginal Budgeting for Bottlenecks Models Country Model used Cost estimate (US$ per capita per year) Ethiopia MBB 3.56 Madagascar (Toamasina) MBB 2.38 Mali (one region) MBB 3.97 Ethiopia Millennium Project 32.00 Bangladesh Millennium Project 20.60 Cambodia Millennium Project 22.50 Ghana Millennium Project 24.70 Tanzania Millennium Project 34.70 Uganda Millennium Project 32.10 Source: Authors. Private spending absorbs a larger share of income in poorer than they are in richer countries. However, differences exist countries. In low-income countries, it absorbs a larger share of across countries that cannot be explained by per capita income GDP, on average, than domestically financed public spending. alone. In low-income and lower-middle-income countries, it invari- Countries need to ascertain whether their low spending ably means out-of-pocket expenditures rather than private is caused by unduly low general revenues or by unduly low allo- insurance (Musgrove, Zeramdini, and Carrin 2002). This situ- cations to health and explore ways of making appropriate ation leaves many near-poor households heavily exposed to the adjustments. Bolivia managed to raise its general revenue share risk of impoverishing health expenses. The risk is clearly consistently in the 1990s as the result of a sustained reform greater the poorer the country, because poorer countries tend, process begun in 1983. The health sector there has been one of on average, to have larger shares of poor people (World Bank the beneficiaries of this growth of tax revenues: government 2000). Governments thus have a major role to play in helping health spending as a share of GDP grew at an annual rate of shape effective risk-pooling mechanisms, in addition to nearly 10 percent in the 1990s. increasing their own spending and targeting it to services for Although raising domestic resources takes time, countries the poor that will have a large positive effect on the MDGs. that can apparently afford to spend more out of their own resources should be encouraged to start the process. Development agencies have a role here--in providing technical Getting Governments to Spend What They Can Afford support of tax reform, in helping develop government com- Government spending is an important part of the picture, and mitment to health in public expenditure allocations, and in the issue is how much they can afford. Unlike private spending, giving financial assistance, both to ease the adjustment costs government spending as a share of GDP is higher in richer and to provide support while the gap is being closed between countries. However, at any given per capita income, a surpris- current and affordable spending. ing amount of variation occurs across countries in the share of GDP allocated to government health programs. Countries that Recognizing the Limits of Development Assistance. Official appear able to spend similar shares of GDP on government development assistance tends to account for a larger share of health programs end up spending quite different amounts. government health spending in poorer countries. Development How can extra domestic resources be mobilized if countries assistance for health is especially important in Sub-Saharan are spending less than they can afford? Domestically financed Africa. Twelve countries in Sub-Saharan Africa had external government health spending comes from general revenues, funding exceeding 35 percent of total health expenditures in social insurance contributions, or both. The amount of general 2000 (World Bank 1998). revenues flowing into the health sector is the product of the Increased development assistance is needed to achieve the amount of general (tax and nontax) revenues collected by MDGs. Development assistance, however, is not without its the government (the general revenue share) and the share of drawbacks. Many donors require that assistance be kept in general revenues allocated to the health sector (the health share parallel budgets outside the ministry of finance, which risks of government spending) (Hay 2003). Low government health undermining government efforts to appropriately plan and spending could be attributable to either share or both shares target expenditures. Such off-budget expenditures make it being low. In poorer countries, both shares are typically lower difficult in some countries to properly target resources to 192 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others particular interventions, geographic locations, or population Bourguignon, F., M. Bussolo, H. Lofgren, H. Timmer, and D. van der groups, even though such targeting may be essential for Mensbrugghe. 2004. "Towards Achieving the Millennium Development Goals in Ethiopia: An Economywide Analysis of improving the effect of expenditures on outcomes and the Alternative Scenarios." World Bank, Washington, DC. probability of reaching the health goals. Donors often require Claeson, M., C. Griffin, T. Johnston, M. McLachan, A. Soucat, A. Wagstaff, recipient governments to maintain separate accounts and to and A. Yazbeck 2001. "Poverty-Reduction and the Health-Sector." In provide separate progress reports, thereby increasing the Poverty Reduction Strategy Sourcebook. Washington, DC: World Bank. administrative burden on weak health ministries. Most impor- Claeson, M., A. Wagstaff, E. Bos, P. Hay, and J. Baudouy. 2004. "The Case for Mobilizing New Research behind the Health Millennium tant, donor commitments of expenditures in health are short Development Goals." In Global Forum Update on Research for Health term, whereas the needs are permanent. Thus, any external 2005, 73­75. Geneva: Global Forum. financing must at some point be substituted with additional Commission on Macroeconomics and Health. 2001. "Macroeconomics domestic revenues or expenditure reallocations. This substitu- and Health: Investing in Health for Economic Development--Report tion or transition to domestic sources of funding has typically of the Commission on Macroeconomics and Health." December 20, World Health Organization, Geneva. been difficult to achieve, leading to a dropoff in effort in Cotlear, D. 1999. "Peru: Improving Health Care for the Poor." Human important health programs, such as immunizations and repro- Development Department (Latin America and the Caribbean Human ductive health services. Development) Paper 57, World Bank, Washington, DC. Consensus on how to improve aid effectiveness is grow- Development Committee. 2003. "Supporting Sound Policies with ing among development partners, and partners at the High Adequate and Appropriate Financing." Discussion paper, World Bank, Level Forum on Health MDGs (http://www.hlfhealthmdgs. Washington DC. org). This agenda includes supporting countries in developing Eichler, R. 2001. "Improving Immunization Coverage in an Innovative Primary Health Care Delivery Model: Lessons from Burkina Faso's more MDG-responsive Poverty Reduction Strategy Papers, Bottom up Planning, Oversight, and Resource Control Approach That tracking resource flows, strengthening monitoring and evalua- Holds Providers Accountable for Results." Discussion paper, World tion, and more effectively dealing with the human resources Bank, Washington, DC. crisis in health. Effective monitoring can help ensure that Eissen, E., D. Efenne, and K. Sabitu. 1997. "Community Loan Funds and Transport Services for Obstetric Emergencies in Northern Nigeria." increased external funds do not simply lead to reduced domes- International Journal of Gynecology and Obstetrics 59 (Suppl. 2): tic financing (the fungibility problem) but actually boost overall S237­44). spending for health. In concert with moves affecting all devel- Esrey, S. A., J. B. Potash, L. Roberts, and C. Shiff. 1991."Effects of Improved opment assistance, donors and governments are trying to see Water Supply and Sanitation on Ascariasis, Diarrhea, Dracunculiasis, that in the health area external funds are pooled and that min- Hookworm Infection, Schistosomiasis, and Trachoma." Bulletin of the World Health Organization 69 (5): 609­21. istries can use a common management and reporting format. FAO (Food and Agriculture Organization of the United Nations). 2000. In addition, a research agenda to support acceleration toward The State of Food Security in the World. Rome: FAO. the health MDGs is being proposed; it needs to focus on how Fiedler,J.L.1996."The Privatization of Health Care in Three LatinAmerican to translate knowledge into action and on how to remove Social Security Systems." Health Policy and Planning 11 (4): 406­17. health systems constraints to scaling up coverage of cost- Filmer, D., and L. Pritchett. 1999."The Impact of Public Spending on Health: effective interventions that are available but do not reach those Does Money Matter?" Social Science and Medicine 49, pp. 1309­23. who need them (Claeson and others 2004; Task Force on Gwatkin, D., S. Rutstein, K. Johnson, and R. P. Pande. 2000. Socio-economic Health Systems Research 2004). Differences in Health, Nutrition, and Population. Washington, DC: World Bank. Hay, R. 2003. "The `Fiscal Space' for Publicly Financed Health Care." NOTE Oxford Policy Institute Policy Brief, Washington, DC. Johnson-Welch, C. 1999. "Focusing on Women Works: Research on 1. Intervention in this chapter refers to the direct action that leads to Improving Micronutrient Status through Food-Based Interventions." prevention or cure. International Center for Research on Women, Washington, DC. Jones, G., R. W. Steketee, R. E. Black, Z. A. Bhutta, and S. S. Morris. 2003. "How Many Child Deaths Can We Prevent this Year?' Lancet 362 REFERENCES (9377): 65­71. Bell, C., S. Devarajan, and H. Gersbach. 2003. The Long-Run Economic Cost Kumaranayake, L., Christoph Kurowski, and Lesong Conteh. 2001. "Costs of AIDS: Theory and an Application to South Africa. Washington, DC: of Scaling Up Priority Health Interventions in Low-Income and World Bank. Selected Middle-Income Countries: Methodology and Estimates." Commission for Macroeconomics and Health Working Paper WG5:19, Bitran, R. 2001. "Paying Health Providers through Capitation in World Health Organization, Geneva. Argentina, Nicaragua, and Thailand: Output, Spending, Organizational Impact, and Market Structure." USAID Partners for Langenbrunner, J. 2003. "Resource Allocation and Purchasing in ECA Health Reform Project, Washington, DC. Region: A Review." Discussion paper, World Bank, Washington, DC. Bokhari, F., P. Gottret, and Y. Gai. Forthcoming. "Government Health Marek, T., I. Diallo, B. Ndiaye, and J. Rakotosalama. 1999. "Successful Expenditures, Donor Funding and Health Outcomes." World Bank, Contracting of Prevention Services: Fighting Malnutrition in Senegal Washington, DC. and Madagascar." Health Policy and Planning 14 (4): 382­89. Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 193 McPake, B, and C. Hongoro. 1995."Contracting Out of Clinical Services in Insecticide-Treated Nets on Child Survival in Rural Tanzania." Lancet Zimbabwe." Social Science and Medicine 41 (1): 13­24. 357 (9264): 1241­47. Mesoamerica Nutrition Program Targeting Study Group. 2002. "Targeting Sen, A. 2002. "Why Health Equity?" Health Economics 11 (8): 659­66. Performance of Three Large-Scale, Nutrition-Oriented Programs in Slack, K., and W. D. Savedoff. 2001. "Public Purchaser­Private Provider Central America and Mexico." Food and Nutrition Bulletin 232 (2): Contracting for Health Services: Examples from Latin America and the 162­74. Caribbean." Sustainable Development Department Technical Paper Millennium Project. 2003. "Millennium Development Goal Country Case 111, Inter-American Development Bank, Washington, DC. Studies: Methodology and Preliminary Results." October. United Soucat,A.,W.Van Lerberghe, F. Diop, S. Nguyen, and R. Knippenberg. 2002. Nations, NY. "Marginal Budgeting for Bottlenecks: A New Costing and Reallocation ------. 2004. "Millennium Development Goals Needs Assessments: Practice to Buy Health Results." World Bank, Washington, DC. Country Case Studies of Bangladesh, Cambodia, Ghana, Tanzania, and ------. 2004. "Marginal Budgeting for Bottlenecks: A New Costing and Uganda." Unpublished working paper for the Millennium Project. Resource Allocation Practice to Buy Health Results--Using Health United Nations, NY. Sector's Budget Expansion to Progress toward the Millennium Mills, A., S. Bennett, and S. Russell. 2001. The Challenge of Health Development Goals in Sub-Saharan Africa." Unpublished paper. World Sector Reform: What Must Governments Do? New York: St. Martin's Bank, Washington, DC. Press. Task Force on Health Systems Research. 2004. "Informed Choices for Morris, S., E. Flores, P. Olinto, and J. Medina. 2003. "A Randomized Trial Attaining the Millennium Development Goals: Towards a Cooperative of Conditional Cash Transfers to Household and Peripheral Health Agenda for Health Systems Research." Lancet 364: 997­1003. Centres: Impact on Child Health and Demand for Health Services." Presented at Fourth International Health Economics Association UNAIDS (United Nations Joint Programme on HIV/AIDS). 2004. Report World Congress, San Francisco, CA, June. on the Global AIDS Epidemic. Geneva: UNAIDS. Musgrove, P., R. Zeramdini, and G. Carrin. 2002. "Basic Patterns in UNICEF (United Nations Children's Fund). 2001. Progress since the World National Health Expenditure." Bulletin of the World Health Summit for Children: A Statistical Review. New York: UNICEF. Organization 80 (2): 134­42. Victora, C., F. Barros, J. Vaughan, A. Silva, and E. Tomasi. 2000."Explaining Newman, J., M. Pradhan, L. Rawlings, G. Ridder, R. Coa, and J. Evia. 2002. Trends in Inequities: Evidence from Brazilian Child Health Studies." "An Impact Evaluation of Education, Health, and Water Supply Lancet 356 (9235):1093­38. Investments of the Bolivian Social Investment." World Bank Economic Wagstaff, A., and M. Claeson. 2004. The Millennium Development Goals for Review 6 (2): 241­74. Health: Rising to the Challenges. Washington, DC: World Bank. OECD (Organisation for Economic Co-operation and Development) ------. 2005. "The Millennium Development Goals for Health: Rising to Development Assistance Committee. 2000. Recent Trends in Official the Challenges: Appendix A, pp. 169­174." World Bank. Washington, Development Assistance to Health. Paris: OECD. DC. Palmer, N., D. Mueller, L. Gilson, A. Mills, and A. Haines. 2004. "Health Wagstaff, A., and E. van Doorslaer. 2003. "Catastrophe and Impover- Financing to Promote Access in Low Income Settings--How Much Do ishment in Paying for Health Care: with Applications to Vietnam We Know?" Lancet 364: 1365­70. 1993­1998." Health Economics 12 (11): 921­34. Pathmanathan, I., J. Liljestrand, J. M. Martins, L. C. Rajapaksa, C. Lissner, World Bank. 1998. Assessing Aid: What Works, What Doesn't, and Why. A. de Silva, and others. 2003. "Investing in Maternal Health: Learning Oxford, U.K.: Oxford University Press. from Malaysia and Sri Lanka." Health, Nutrition, and Population ------. 2000. World Development Report 2000/2001: Attacking Poverty. Department, World Bank, Washington, DC. New York: Oxford University Press. Preker, A. S., E. Suzuki, F. Bustero, A. Soucat, and J. Langenbrunner. 2003. "Costing the Millennium Development Goals." Background paper to ------. 2001. "`Health, Nutrition, and Population Development Goals: The Millennium Development Goals for Health: Rising to the Challenges, Measuring Progress Using the Poverty Reduction Strategy World Bank, Washington, DC. Framework." Report of a World Bank Consultation, World Bank, Washington, DC. Quisumbing, A. R. 1995."Gender Differences in Agricultural Productivity: A Survey of Empirical Evidence." IFPRI Discussion Paper 5. ------. 2003a. Attaining the Millennium Development Goals in India: How International Food Policy Research Institute, Washington, DC. Likely and What Will It Take? Washington, DC: World Bank. Rajkumar, A., and V. Swaroop. 2002. "Public Spending and Outcomes: ------. 2003b. Global Economic Prospects and the Developing Countries. Does Governance Matter?" Policy Research Working Paper 2840, Washington, DC: World Bank. World Bank, Washington, DC. ------. 2003c. World Development Indicators 2003. Washington, DC: Ramana, G. 2003. Background paper for The Millennium Development World Bank. Goals for Health: Rising to the Challenges. World Bank, Washington, DC. ------. 2003d. World Development Report 2004: Making Services Work for Santos, I., C. G. Victora, J. Martines, H. Goncalves, D. P. Gigante, N. J. Valle, Poor People. Washington, DC: World Bank. and G. Pelto. 2001. "Nutrition Counselling Increases Weight Gain Yip, W., and P. Berman. 2001."Targeted Health Insurance in a Low Income among Brazilian Children." Journal of Nutrition 131 (11): 2966­73. Country and Its Impact on Access and Equity in Access: Egypt's School Schellenberg, J. R., S. Abdulla, R. Nathan, O. Mukasa, T. J. Marchant, N. Health Insurance." Health Economics 10 (3): 207­20. Kikumbih, and others. 2001. "Effect of Large-Scale Social Marketing of 194 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others Chapter 10 Gender Differentials in Health Mayra Buvini´c, André Médici, Elisa Fernández, and Ana Cristina Torres In health, more than in other social sectors, sex (biological) and from trials with male subjects only (male norm) to both sexes, gender (behavioral and social) variables are acknowledged use- whereas female biology can affect the onset and progression of ful parameters for research and action because biological differ- disease, and women's lower position in society can affect their ences between the sexes determine male-specific and female- health-seeking behaviors (Pinn 2003; Sen, George, and Ostlin specific diseases and because behavioral differences between the 2002). genders assign a critical role to women in relation to family As A. K. Sen (1990) and others have indicated, gender bias health. Until recently, however, the importance of sex and gen- results in the neglect of female children and in selective abor- der informed work on female-specific diseases but did not carry tion and excess female mortality in China, India, and other over to diseases shared by men and women. As a result, the South Asian countries, explaining the "missing" women in pop- literature contained comparatively little about which diseases ulation counts. In addition, such bias can have intergenera- affect men and women differently, why that difference might be tional health effects, starting with maternal undernutrition the case, and how to structure prevention and treatment in and leading to fetal growth retardation, low birthweight, child response to these differences. This situation has changed, how- undernutrition, and ailments in adult children of disadvan- ever, and interest in measuring, understanding, and responding taged mothers (Osmania and Sen 2003). to sex and gender differentials in disease has surged, nurtured by This chapter only partially addresses women's health needs. breakthroughs in science and advances in advocacy.1 It omits important disease conditions for women, such as lung In line with this interest and using global burden-of-disease cancer and HIV/AIDS, where men and women currently have data for 2001, this chapter reviews worldwide gender differen- similar disease burdens. (In the case of HIV/AIDS this balance tials in mortality and morbidity that result in excess disease is changing, and women's disease burden is rising over men's, burdens for women and examines cost-effective interventions especially for the 18­25 age group and for specific world drawn from chapters 17 (on sexually transmitted infections), regions.) It also does not cover important sources of the disease 26 (on maternal and perinatal conditions), 29 (on health serv- burden for women that are not measured in disability-adjusted ice interventions for cancer control in developing countries), life years (DALYs), such as burden from female genital mutila- 31 (on mental disorders), 32 (on neurological disorders), 51 tion (FGM). Last, the emphasis on disease underplays women's (on musculoskeletal disability and rehabilitation), and 57 (on reproductive and other health needs. contraception) to address them. The focus on women's excess disease burden is justified to ANTECEDENTS fill gaps in knowledge regarding women's health that are in part a product of male bias and male norms in clinical studies. In The chapter's emphasis on gender differentials and inequalities the past, medical research often wrongly assumed that women in health rather than on women's absolute health conditions were biologically weaker (male bias) and extrapolated findings reflects the evolution of thinking on women and health issues. 195 (Annex 10.A charts these advances in the past two decades, ferences between men's health and women's health that arise highlighting milestones and influential publications.) In the from this lower position and the consequent unequal power 1960s and 1970s, the field of international women's health issues relationship between the sexes. Sex and gender can act alone, emerged from and was influenced by an interest in women's fer- independently, or interactively in determining differentials in tility behavior as a means of curbing population growth and by the burden of disease (Krieger 2003). Some women's excess an interest in maternal and child health to improve child wel- health burdens, such as uterine cancer, are based almost solely fare, with little or no attention paid to mothers (McNamara on biology. At the other end of the continuum, some women's 1981; Rosenfield and Maine 1985). Much of the work in the excess health burdens, such as injuries from domestic fires or 1980s sought to bring a woman-centered perspective into pop- domestic abuse, are solely gender based. ulation and maternal and child health programs. This focus However, in most cases sex and gender interact to determine included awareness of how women's lower status in society women's disease burdens. Two salient examples are depressive affected health delivery and health-seeking behaviors and how disorders and HIV/AIDS. Women are twice as likely as men to women's time burdens in poor households affected child health. become depressed, and genetics and hormones influence the The issues raised included the quality of care in health and risk of depression. However, genes and sex hormones cannot family-planning programs and the nature of women's work and entirely explain women's excess burdens, and gender factors its impact on child survival and health (Bruce 1990; Leslie play an important role (WHO 2000). HIV infection rates 1988). Reducing maternal mortality became a major develop- among teenage girls are 5 to 16 times higher than among ment objective (Herz and Measham 1987). teenage boys in Sub-Saharan Africa. This earlier age of HIV The 1994 United Nations International Conference on exposure for girls is partly explained by the greater biological Population and Development in Cairo placed women's repro- efficiency of male-to-female transmission and partly by girls' ductive health and rights at the center of the population and lack of knowledge, opportunities, and bargaining power in sex- development debate, and the United Nations Women's ual relations that make them prime victims of the rapid spread Conference in Beijing the following year reinforced the impor- of the disease. tance of women's empowerment and of a gender perspective Existing knowledge about the interplay between sex and in health. Along with the global burden-of-disease effort, gender in determining disease is imperfect and evolving researchers estimated the loss of women's healthy years of life (Krieger 2003; Pinn 2003). This chapter groups women's excess caused by gender violence (Heise, Pitanguy, and Germain health burdens from diseases into the following four broad 1994), and gender was identified as central to women's risk of categories: and treatment for HIV/AIDS (Gupta 2000; Mann 1993). The World Health Organization analyzed how differences between · diseases specific to women (that is, where biology plays a women and men in access to and control over resources deter- major role in the disease) mine differential exposure to risk and access to the benefits of · diseases related to women's average greater longevity (where health technology and care (WHO 1998). After more than two both sex and gender tend to play important roles) decades almost solely devoted to maternal and reproductive · diseases that result from the interaction of sex and gender health issues, attention expanded to cover a range of women's · diseases that are predominantly gender based (that is, that health issues unrelated to reproduction and to identify and cor- result from specific behavioral, social, and cultural factors rect gender differentials and inequities in health (Sen, George, associated with women's condition). and Ostlin 2002). These new emphases complemented renewed interest in health inequities and their reduction in the field of Sex and gender have a much wider influence on disease than international health (see, for instance, Evans and others 2001). is usually acknowledged. They influence the etiology, diagnosis, progression, prevention, treatment, and health outcomes of disease as well as health-seeking behaviors and exposure to risk. Framework Whereas sex plays a bigger role in the etiology, onset, and pro- Both sex and gender matter in health. We use the term sex to gression of disease, gender and its consequences influence dif- describe differences between men and women that are primar- ferential risks, symptom recognition, severity of disease, access ily biological in origin and that may be genetic or phenotypic. to and quality of care, and compliance with care. In addition, By contrast, we use the term gender to describe differences that poverty and social exclusion because of race and ethnicity are primarily caused by social conditions or cultural and reli- interact with sex and gender and contribute to women's excess gious beliefs and norms regarding the sexes. Structural gender disease burdens in ways that are largely unexplored to date inequalities that place women in a subordinate position to men (Breen 2002). underlie and contribute to gender differentials in disease (Sen, Factors that influence gender differentials in relation to the George, and Ostlin 2002). A gender perspective addresses dif- risk of disease include (a) biological (genetic, physiological, and 196 | Disease Control Priorities in Developing Countries | Mayra Buvini´c, André Médici, Elisa Fernández, and others hormonal) differences between the sexes; (b) women's longer males live shorter but healthier lives than females. Even though life expectancy; (c) nature and rate of change of women's more boys are born than girls, gender differences in mortality labor force participation compared with men's participation; eventually change the sex balance in populations so that by age (d) women's differential access to social protection mecha- 30 or so women start outliving men, and the absolute female nisms (health and social insurance); (e) cultural norms, advantage in survivability increases with age (Kinsella and Gist religious beliefs, and family arrangements and behaviors 1998). Therefore, differences in life expectancy at birth by gen- determining gender roles and gender hierarchy in society; der, using 2000 data, vary in favor of females, ranging from one (f) gender differences in educational attainment; (g) income year in the low-income countries of South Asia and Sub- differences between the genders resulting from the interaction Saharan Africa to seven years in Europe and Central Asia and of all the previous factors; and (h) interactions between race, nine years in the middle-income countries of Latin America ethnicity, income, and gender. and the Caribbean. Women's overall underutilization of health services has been Overall, however, women have higher morbidity than men. well documented. For instance, even though women in India Murray and Evans (2003) find that in relation to expected lost report more illness than men, hospital records show that men healthy years at birth, whereas men lose 7.8 years over their receive more treatment (World Bank 1996); in Thailand, men lifetimes as a result of poor health, women lose 10.2 years. In are six times more likely than women to seek clinical treatment other words, women spend about 15 percent of their lives in for malaria, a disease that affects women and men similarly unhealthy conditions and men spend just 12 percent. Therefore, (Hanson 2002); and in Brazil, the Dominican Republic, living longer lives should not be taken to indicate better health Jamaica, Paraguay, and Peru, low-income women underuse for women. Women live less healthy lives and are saddled with health services (Levine, Glassman, and Schneidman 2001). higher morbidity in part because they outlive men (Verbrugge Three groups of factors influence this underuse of health 1989). Supporting the less healthy lives assertion, women up to services. The first group is service factors, such as accessibility; age 65 reported worse health status in virtually all 64 household affordability (money and time costs); and appropriateness or surveys from 46 countries (Sadana and others 2000). Because of adequacy, including friendliness, of the health and social infra- these differences in morbidity, the concept of healthy adjusted structure for meeting women's needs. The second group is user life expectancy at birth describes differences in health conditions factors, which include social constraints, such as restrictions on between males and females better than the concept of life women's mobility and women's average lower incomes and expectancy at birth. greater time burdens than men's; asymmetric information about health needs and rights and the availability of services, which disproportionately affects poor women; and marital sta- GENDER DIFFERENTIALS IN DISEASE BURDENS tus, family roles, and work conditions affecting access and use. The third group is institutional factors, including men's The global burden of disease for 2001 proportionally affects decision-making power and control over health budgets and males slightly more than females. About 52 percent of DALY facilities, which affect local perceptions of illness and norms losses are attributed to males, but this proportion varies concerning treatment, and stigmatization and discrimination between 54.8 percent in Europe and Central Asia and 49.9 per- in health settings, which affect women among the poor and cent in South Asia. The only region where the global burden of women of minority ethnic and racial groups. disease affects females more than males is South Asia. Table 10.1 shows the burden of disease by region, gender, and age group. The burden of disease during early childhood (age 0 Context through 4) is somewhat smaller for girls than boys; however, The global demographic dynamic, a product of the interplay of from age 5 through 29, females lose more DALYs than males, nature and nurture, biology and society, helps determine gen- but only in developing countries. Larger differences favoring der differentials in health. In 2001, the world's population, an women appear starting at age 30 and continue until after the age estimated 6.2 billion, was 50.3 percent male and 49.7 percent of 70, when women, because of their greater longevity, lose female. The surplus male population was concentrated in the more DALYs than men. However, when DALYs are estimated developing countries, whereas the developed countries had a per 1,000 males and females as in table 10.1, women in the higher proportion of women, primarily in the older age groups older age groups lose fewer DALYs than men both in low- and (WHO 2001). middle-income countries (LMICs) and in high-income In the developed countries, the number of women age 80 countries (HICs). and older was more than double the number of men in the Communicable diseases and maternal conditions con- same age group. This female advantage in longevity helps shape tribute significantly to females' burden of disease in develop- a gender paradox in health outcomes worldwide: on average, ing countries and add little to their burden in HICs, where Gender Differentials in Health | 197 Table 10.1 DALYs by Region, Gender, and Age Group, 2001 (DALYs per 1,000 population) Age (years) Region 0­4 5­14 15­29 30­44 45­59 60­69 70­79 80+ Total Males Low- and middle-income countries 754.6 75.4 137.4 194.5 349.3 600.1 799.4 950.8 271.5 East Asia and the Pacific 408.0 46.5 99.6 124.2 276.4 522.8 735.3 965.1 192.6 Europe and Central Asia 345.4 50.1 137.9 220.8 412.8 695.6 855.3 952.5 278.3 Latin America and the Caribbean 449.1 56.7 155.0 173.2 290.4 493.2 682.6 870.8 217.8 Middle East and North Africa 585.6 67.5 109.3 151.4 337.8 608.5 825.8 1,019.8 219.8 South Asia 805.2 83.0 131.9 196.0 383.0 661.7 861.7 899.6 285.1 Sub-Saharan Africa 1,480.7 142.2 248.2 532.2 609.0 778.6 1,002.5 1,194.3 528.6 High-income countries 128.2 28.6 84.4 97.0 189.6 355.7 541.5 721.5 168.2 World 698.5 70.9 131.1 178.5 315.4 540.4 721.0 856.7 256.3 Females Low- and middle-income countries 753.4 78.5 142.5 162.1 275.3 501.3 745.4 946.7 259.8 East Asia and the Pacific 434.9 47.1 86.2 103.4 225.7 442.9 699.5 966.8 182.1 Europe and Central Asia 303.6 39.1 87.0 117.7 223.6 423.7 663.2 909.6 212.5 Latin America and the Caribbean 395.8 51.1 101.8 120.5 222.3 423.8 604.9 827.6 179.0 Middle East and North Africa 538.0 62.7 103.6 140.6 278.3 509.2 767.6 1,069.1 203.3 South Asia 865.5 98.3 168.7 188.5 334.0 616.6 901.0 938.3 304.4 Sub-Saharan Africa 1,367.4 138.7 317.3 448.9 478.8 701.4 1,009.8 1,196.5 504.3 High-income countries 117.5 28.9 72.5 77.7 144.8 258.4 414.6 625.9 153.3 World 696.5 73.7 134.2 148.1 247.2 441.6 641.0 800.0 243.4 Source: WHO 2001. noncommunicable diseases prevail in both women's and men's the prevalence of asymptomatic illness, such as sexually disease burden. Injuries weigh more heavily in males' than in transmitted infections among women; the differences in females' burden of disease across regions. In summary, in both health-seeking behaviors that favor males accessing formal developed and developing countries, the overall burden of dis- health care, which is the main source for health statistics; and ease is higher for males than for females; however, this situation the exclusion of some conditions that affect only women, reverses in developing countries for young girls and women in such as FGM, from global burden-of-disease estimations their prime childbearing years. In addition, in LMICs, females (Hanson 2002). Thus, the findings in the preceding section are more affected than males by highly preventable communi- may be affected by a quality problem, and the estimates in this cable diseases. section are probably conservative. Table 10.2 breaks down 11 conditions specific to women WOMEN'S EXCESS DISEASE BURDENS by region. In 2001, conditions specific to women accounted for 5.3 percent of women's total DALY losses, compared with The 2001 global burden-of-disease data underestimate both 0.7 percent for two conditions (prostate cancer and benign women's and men's disease burdens because of the incomplete- prostate hypertrophy) specific to men.2 Most causes of mortal- ness of health statistics, especially in the developing world. This ity or morbidity specific to women are related to maternal underestimation is probably more pronounced for women conditions and malignant neoplasms. The DALY losses associ- because they experience more disability--which is less well ated with conditions specific to women are around 6 percent in recorded than mortality--than men. This underestimation is both HICs and LMICs, but maternal conditions are more aggravated by underreporting resulting from the stigma associ- prevalent in LMICs, whereas neoplasms cause more DALY ated with certain diseases in women, such as sexual infections; losses in HICs. 198 | Disease Control Priorities in Developing Countries | Mayra Buvini´c, André Médici, Elisa Fernández, and others Table 10.2 Percentage of DALYs Resulting from Conditions Specific to Women by Region, 2001 East Asia Europe and Latin America Middle East and the Central and the and Sub-Saharan Condition HICs LMICs Pacific Asia Caribbean North Africa South Asia Africa World Maternal 0.5 4.0 2.0 0.8 2.5 4.1 4.9 5.7 2.8 Maternal hemorrhage 0.0 0.6 0.2 0.0 0.2 0.4 0.8 1.0 0.5 Maternal sepsis 0.1 0.8 0.5 0.2 0.7 0.7 0.9 1.1 0.7 Hypertensive disordersa 0.0 0.3 0.1 0.0 0.2 0.2 0.4 0.5 0.3 Obstructed labor 0.0 0.4 0.1 0.0 0.1 0.3 0.6 0.5 0.3 Abortion 0.0 0.5 0.1 0.0 0.2 0.5 0.7 0.9 0.5 Other 0.4 1.4 1.0 0.6 1.3 2.0 1.5 1.7 0.5 Neoplasms 5.5 1.7 1.8 4.1 3.1 1.4 1.5 0.8 2.2 Breast cancerb 3.4 0.8 1.0 2.0 1.3 0.9 0.6 0.3 1.1 Cervix uteri cancer 0.4 0.6 0.5 0.7 1.0 0.3 0.7 0.4 0.6 Corpus uteri cancer 0.8 0.1 0.1 0.7 0.5 0.1 0.0 0.0 0.2 Ovarian cancer 0.9 0.2 0.3 0.7 0.3 0.1 0.2 0.1 0.3 Chlamydiab 0.1 0.3 0.2 0.2 0.4 0.5 0.4 0.3 0.3 Total 6.1 6.0 4.0 5.1 6.0 6.0 6.8 6.8 5.3 Source: WHO 2001. a. Related to maternal conditions only. b. Even though these conditions are not specific to women, women account for more than 90 percent of the DALY losses associated with these conditions. Table 10.3 shows the gender ratio and burden of disease of We determined priority conditions affecting females in dif- eight conditions, by region, that are more prevalent among ferent age groups using a method that took into account both women than among men. The selection of diseases was done the gender ratio and the weight of specific conditions in using as the threshold the mean plus one standard deviation females' total DALYs lost. The results indicate that women are (SD) of the distribution of gender ratio scores for each disease. affected by communicable diseases and maternal conditions The diseases selected were then screened for their importance until age 29 and by noncommunicable diseases after age 30, in women's disease burden, using the same criterion of the with chronic diseases having a heavy weight during the last mean plus one SD in the distribution of DALY scores for stages of the life cycle. Conditions for which females' burden of women. Although some diseases, such as unipolar depressive disease is more or less double that of males at a specific stage in disorders and osteoarthritis, are priorities for both HICs and the life cycle are migraine at age 5 to 14, fires and panic disor- LMICs, others, such as Alzheimer's disease, are more relevant der at age 15 to 29, and unipolar depressive disorders at age 60 in HICs, reflecting women's longer life expectancy. Conditions to 69. such as age-related vision disorders, migraine, fires, and cere- brovascular diseases have particular relevance in specific regions. PRIORITY DISEASE GROUPS FOR WOMEN Combining gender-specific conditions and shared condi- tions that disproportionately affect women gives a total of 19 Table 10.4 presents conditions with excess burdens for women priority conditions for women. Taken together, these condi- divided into the four groups defined earlier. tions represent about one-fifth of women's total DALY losses and indicate priorities for research and the search for cost- effective methods of promotion, prevention, and treatment. Conditions Specific to Women Note that some important contributors to females' health In developed countries, advances in medical technology have burdens, such as malaria and HIV/AIDS, have been omitted almost eliminated the burden of disease resulting from mater- because females do not currently suffer disproportionately nal conditions. Three types of cost-effective intervention pack- from these diseases. However, the growing feminization of the ages for maternal conditions are the prevention of pregnancy by HIV/AIDS epidemic in developing countries should result in means of effective family-planning methods; the prevention of excess disease burdens for women in the near future. complications (for example, hemorrhage); and the prevention Gender Differentials in Health | 199 200 | Disease Control Priorities in Developing Table 10.3 Gender Ratio and Women's Excess Burden of Disease for Top Priority Conditions by Region, 2001 Countries Latin American Middle East East Asia and Europe and and the and North Sub-Saharan | Mayra World LMICs HICs the Pacific Central Asia Caribbean Africa South Asia Africa Buvini´ BOD BOD BOD BOD BOD BOD BOD BOD BOD Condition GR (percent) GR (percent) GR (percent) GR (percent) GR (percent) GR (percent) GR (percent) GR (percent) GR (percent) c, André Alzheimer's disease 1.81 1.48 n.a. n.a. 2.13 7.02 1.55 1.51 2.40 2.16 1.63 1.58 n.a. n.a. n.a. n.a. n.a. n.a. and other dementias Médici, Osteoarthritis 1.64 1.46 1.63 1.26 1.71 3.30 1.64 2.15 1.76 2.76 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. Elisa Unipolar depressive disorders 1.53 4.22 1.51 3.89 1.69 7.30 n.a. n.a. 1.73 5.15 1.72 6.94 1.44 3.86 1.56 4.33 1.54 1.17 Age-related vision disorders n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1.63 2.10 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. Fernández, Migraine n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 2.81 1.14 n.a. n.a. n.a. n.a. n.a. n.a. Other cardiovascular diseases n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1.45 3.20 n.a. n.a. n.a. n.a. and Fires n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 2.43 2.04 n.a. n.a. others Cerebrovascular diseases n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1.47 1.80 Source: WHO 2001. n.a. not applicable; BOD burden of disease; GR gender ratio. Note: The gender ratio is the ratio of female to male DALYs. For each world region or classification, data are shown only for those diseases that are classified as "priority" given that they meet the selection criteria according to the methodology used in this study (that is, the gender ratio was equal to or higher than the mean plus one SD of the distribution of gender ratio scores, and in a second screening, the BOD was equal to or higher than the mean plus one SD of the distribution of female DALY scores). Table 10.4 Priority Conditions by Category, 2001 to women include reaching poor and socially excluded women with basic maternal and reproductive health services; Region where strengthening the adoption of preventive health behaviors in Gender gender ratio is developing countries; extending the quality of care to other Category and condition ratio apparent conditions specific to women, including neoplasms; and edu- Conditions specific to women cating and empowering women to promote their own healthy Maternal hemorrhage n.a. n.a. behaviors. Maternal sepsis n.a. n.a. Hypertensive disorders related n.a. n.a. to maternal conditions Conditions Associated with Women's Greater Longevity Obstructed labor n.a. n.a. The main group of diseases with excess burdens for women Abortion n.a. n.a. associated with women's greater longevity are Alzheimer's Other maternal conditions n.a. n.a. disease; musculoskeletal disorders, such as osteoarthritis, Breast cancer 244.62 World rheumatoid arthritis, and osteoporosis; and cardiovascular dis- Cervix uteri cancer n.a. n.a. eases,3 which together account for 12 percent of total DALY Corpus uteri cancer n.a. n.a. losses for women worldwide. LMICs account for about 80 per- Ovarian cancer n.a. n.a. cent of the DALYs resulting from these conditions, likely because of the lack of medical care during the early stages of Chlamydia 9.76 World these diseases. Conditions associated with women's greater longevity Alzheimer's disease and other dementias account for 1.5 per- Alzheimer's disease and 1.81 World cent of total female DALYs, and this burden is almost twice as other dementias high as that for men. About 46 percent of this burden is Osteoarthritis 1.64 World concentrated in HICs and 54 percent in LMICs, but DALYs lost Cerebrovascular diseases 1.47 Sub-Saharan Africa per capita are much greater in the HICs than in more densely Other cardiovascular diseases 1.45 Middle East and populated LMICs. Because of population aging, during the North Africa next 50 years the number of people with Alzheimer's disease is Age-related vision disorders 1.63 Europe and expected to more than double, with more women affected than Central Asia men (McCann and others 1997). Studies of the effects of estro- Conditions arising from the interaction of sex and gender gen therapy on Alzheimer's disease have been inconsistent: Unipolar depressive disorders 1.53 World estrogen may increase the risk of both dementia and other dis- Migraine 2.81 Latin America and eases such as stroke in postmenopausal women (Shumaker and the Caribbean others 2003). Gender-based conditions As concerns musculoskeletal disorders, osteoarthritis affects Fires 2.43 South Asia 9.6 percent of men and 18.0 percent of women age 60 or older worldwide and accounts for 1.5 percent of total female DALYs. Source: WHO 2001. Osteoarthritis is related to aging and is most common in over- n.a. not applicable. weight women over the age of 45. Demographic changes in developing countries, especially middle-income countries, of death or disability resulting from complications through indicate that osteoarthritis prevention and treatment needs will emergency obstetric care. Proven technologies also exist for increase during the next decade. Most prevention and treat- screening and early detection of some neoplasms, although ment are linked to regular exercise, healthy weight manage- their implementation has been uneven. Problems pertaining to ment, physical and occupational therapy, and pain manage- conditions in this category are not simply related to the avail- ment with over-the-counter medications. ability of medical technologies, but also to the behavioral and Because cardiovascular diseases are generally thought of social factors that influence women's exposure to risks and by society as "men's diseases," women tend to delay seeking underuse of services as well as the economic and institutional treatment for cardiac-related events (Seils, Friedman, and factors that influence the availability and quality of services, Schulman 2001). However, cardiovascular and cerebrovascular especially in developing countries. diseases account for about 8.2 percent of total female DALYs, Even though the international women's health movement more than half of which is caused by cerebrovascular has promoted significant advances in the quality of care diseases. for reproductive health and maternal conditions in the past In addition to age, smoking, and obesity, another risk factor 25 years, the main challenges in relation to conditions specific that exposes women to a greater burden of cardiovascular Gender Differentials in Health | 201 diseases is depression, which is associated with increased depression in developing countries by creating systems that morbidity and mortality from heart diseases and is highly help them overcome social stigma and economic and social prevalent in women (Linfante and others 2003). Also, women's barriers. symptoms of heart disease tend to be different from men's, In addition to treatment, prevention of depression and increasing the difficulties of diagnosis (Seils, Friedman, and other mental illness needs to address women's role in society Schulman 2001). Finally, evidence suggests that physical and the control they have over their lives and circumstances. activity significantly reduces the risk of cardiovascular events; According to the World Health Organization, pertinent factors however, women tend to exercise less than men (Manson and are related to having sufficient autonomy to exercise some others 2002). Whether this observation can be generalized to all control in response to severe events, access to adequate age groups and whether it occurs because of biological factors resources to be able to make choices, and social supports or social norms deserve further attention. (WHO 2000). The promotion of healthy behaviors, including exercise, is also crucial for prevention. Conditions Arising from the Interaction of Sex and Gender In the group of conditions resulting from the interaction of Gender-Based Conditions biological and social factors, unipolar depressive disorders have The main characteristic of gender-based conditions is that they the most significant gender ratio and most unequal burden have no biological referent and can, therefore, be prevented by of disease in every region except East Asia and the Pacific. means of behavioral change. The role of social components in Unipolar depressive disorders account for 4.2 percent of this category explains excessive health burdens for women in women's global burden of disease. Even though the DALYs lost particular world regions. For example, women are dispropor- per capita are similar in LMICs and HICs, these disorders tionately affected by fires in South Asia, an outcome of the vio- represent a higher share of women's total burden of disease lence caused by dowries. Too often women die in what are in HICs (7.3 percent) than in LMICs (3.9 percent). Another called "cooking fire accidents," whereas in reality they are mur- important consideration is the high comorbidity between dered so that their husbands may remarry and obtain another depression and other psychiatric disorders (for example, dowry. anxiety disorders). Neuropsychiatric disorders account for Another characteristic of the diseases and injuries in this 11.8 percent of women's total global burden of disease and category is that they are often underreported because of 23.5 percent in HICs. stigma and social pressures. As a result, the data probably The exact contributions of biology and society in the etiol- underreport the true extent of the problem. FGM and domes- ogy of depression are unknown. Some believe that genetic tic violence are two examples. In 2000, estimates indicated causes account for about half of the risk for mood disorders that 100 million to 140 million girls and women had under- (Zubenko and others 2002), whereas others suggest that gender gone FGM and that more than 2 million girls were at risk. At roles, stressors, social relationships, and personality traits least 28 African and Middle Eastern countries practice FGM may play a larger role than hormones and neurotransmitters for social, cultural, or religious reasons (WHO 2000). (Bromberger 2004). As concerns biology, aside from genetic Unfortunately, global burden-of-disease data do not report predisposition, the fluctuation of sex hormones, especially the resulting DALYs lost. Regarding violence against women, estrogen, during women's reproductive life is believed to be an 10 to 50 percent of women report having been physically important risk factor for depression (Bromberger 2004). As abused by an intimate partner, and 12 to 25 percent report concerns social factors, poverty; lack of proper nutrition and attempted or completed forced sex. Although men experience education; stressful and insecure life circumstances; and more absolute DALY losses from violence, women are also domestic and sexual violence and the concomitant feelings of seriously affected. In 1998, interpersonal violence was the loss, entrapment, and lack of control are likely at the root of 10th-leading cause of death for women age 15 to 44 depression (Bromberger 2004; WHO 2000). worldwide. Treatment for depression includes medication and psy- Given the complexity of these health problems, comprehen- chotherapy or counseling, and instruments are available to sive interventions are necessary. Changes are needed in the assess the severity of depression, including prenatal and post- following areas: partum depression. In developing countries, severe depression and anxiety disorders go mostly untreated (76 to 85 percent of · legislation and law enforcement serious cases receive no treatment), partly because of igno- · public policies and programs in areas such as health, educa- rance, social barriers, and stigmatization, which may affect tion, and police and legal services women's access to treatment (WHO 2004). A main challenge, · training of service providers and creation of gender- therefore, is increasing women's access to treatment for sensitive services, especially at the community level 202 | Disease Control Priorities in Developing Countries | Mayra Buvini´c, André Médici, Elisa Fernández, and others · education of the general public to create awareness, behav- episode vary from US$73 (public hospital) to US$86 (mission ioral change, and promotion of advocacy groups hospital) (Levin and others 1999). In Bolivia, the cost of a · better data collection, research, and understanding of cesarean section ranges from US$56 to US$104 (Rosenthal and the individual and social mechanisms sustaining these Percy 1991), and the cost of a normal delivery varies from problems. US$11 to US$16 (Dmytraczenko and others 1998). In developing countries, 61 percent of maternal deaths occur 23 to 48 hours after delivery because of such problems as COST-EFFECTIVENESS OF INTERVENTIONS postpartum hemorrhage and hypertensive disorders or after 48 hours because of sepsis. Complications from unsafe abor- This section summarizes the costs and health benefits of strate- tions account for 13 percent of maternal deaths, though this gies to address conditions that are specific to women and figure is probably an underestimate because of the scarcity of conditions that affect women disproportionately. Table 10.5 data. Little information is available on costs related to postna- presents cost-effectiveness estimates for recognized effective tal care given the different kind of interventions and the sever- interventions. ity of cases, but the literature generally agrees that emergency obstetric care can reduce costs. As concerns postabortion care, costs per case in LMICs could vary from US$4.40 to US$17.19 Conditions Specific to Women (Dayaratna and others 2000). Conditions specific to women include the cluster of diseases Millions of premature deaths, illnesses, and injuries could be related to women's maternal function plus chlamydia (which is avoided by helping women prevent unwanted pregnancies and predominantly but not exclusively a female disease) and obtain prompt treatment for reproductive health problems. The female-specific cancers. contraception costs per couple-year of protection could vary, depending on the method used, from US$6 (intrauterine Maternal Conditions. The analysis developed in chapter 26 device) to US$20 (condoms or injections) (Dayaratna and uses a model for maternal and perinatal conditions that gener- others 2000). A 1999 experiment by the Planned Parenthood ates 128 potential scenarios. According to the findings of this Association of South Africa considered total health planning theoretical exercise, the cost per DALY averted of mother and costs per couple-year of protection, including travel expenses to baby packages could vary from US$77 to US$151 in Sub- health clinics. Comparing these costs with total health planning Saharan Africa and from US$143 to US$278 in South Asia, costs in services provided by community-based doctors, the depending on the complexity of the intervention. study found that the former cost US$44 per couple-year of pro- Prenatal care prevents almost a quarter of maternal deaths, tection and the latter cost US$42. especially when backed by essential and emergency obstetric care to deal with conditions detected during the course of Chlamydia. Although not specific to women, chlamydia is pregnancy care. Good prenatal care includes information, edu- nine times more prevalent among women than among men, cation, and communication activities and behavior-change and its consequences and treatment are much more compli- communication to increase women's skills in relation to the cated and severe for women, affecting women's and infants' identification of danger signs and potential complications and health during pregnancy and the postnatal period. Chlamydia where to seek care in these cases (Dayaratna and others 2000). is widespread in low-income countries. Chlamydia, as well as In Uganda, for example, integral prenatal care ranged from other sexually transmitted diseases, could be prevented by US$2.26 (public services) to US$6.43 (religious mission using condoms, with an average cost per DALY averted in services) per pregnant woman per year (Levin and others 1999). developing countries estimated at US$3.40 in noncore target Another important service is supplementation with iron groups and US$12.60 in core target groups (Mumford and and folic acid. Iron deficiency accounts for 1.8 percent of others 1998). Detecting chlamydia in pregnant women could women's deaths and 2.6 percent of female DALY losses. Iron cost $4.38 per case, with treatment, at $3.82 per case, being less and folic acid supplements administered to highly anemic expensive than detection (Shultz, Schulte and Berman 1992). pregnant women can save lives at a cost of US$13 per DALY Chlamydia's adverse effects are trachoma (chronic conjunctivi- averted (Berman and others 1991), demonstrating that this tis, endemic in Africa and Asia), reproductive tract infections, intervention is very cost-effective. genital ulcer disease in tropical countries, and infertility. The Good maternal health services can strengthen the entire cost of each adverse outcome averted varies from about US$85 health system. A health facility that is equipped to provide to US$308 (Shultz, Schulte, and Berman 1992). essential obstetric care can also treat accidents, trauma, and other medical emergencies. The costs of emergency obstetric Neoplasms. Cancers specific to women are responsible for care vary depending on the country. In Uganda, costs per high levels of female morbidity and mortality, with cervical Gender Differentials in Health | 203 Table 10.5 Cost-Effectiveness of Selected Interventions Aimed at Conditions Specific to Women or That Affect Women Disproportionately Intervention Region or country Lower end of range Upper end of range Source Maternal conditions Mother and baby package LMICs US$2 per capita US$4 per capita WHO 1997 Mother and baby package LMICs US$18 per DALY averted US$58 per DALY averted Raviez, Griffin, and Follmer 1995; World Bank 1993 Mother and baby package Sub-Saharan Africa US$77 per DALY averted US$151 per DALY averted Chapter 26 of this publication Mother and baby package South Asia US$143 per DALY averted US$278 per DALY averted Chapter 26 of this publication Integral prenatal care Uganda US$2.26 per pregnant woman US$6.43 per pregnant woman Levin and others 1999 Iron and folic acid nutritional Low-income countries US$13 per DALY averted in highly n.a. Berman and others 1991 supplementation anemic pregnant women Emergency obstetric care Uganda US$73 per episode US$86 per episode Levin and others 1999 Cesarean delivery Bolivia US$56 per case US$104 per case Rosenthal and Percy 1991 Normal delivery Bolivia US$11 per case US$16 per case Dmytraczenko and others 1998 Postabortion care (dilation LMICs US$4.40 per case US$17.19 per case Dayaratna and others 2000 and curettage) Total health-planning costs South Africa US$42 per couple-year of US$44 per couple-year of Dayaratna and others 2000 protection protection Chlamydia Prevention Developing countries, US$47.98 per DALY averted US$651.82 per DALY averted Over and Piot 1993 1990 (noncore target groups) (core target groups) Prevention (use of condoms) Developing countries, US$3.40 per DALY averted US$12.60 per DALY averted Mumford and others 1998 (noncore target groups) (core target groups) Detection Various sites, 1992 US$4.38 per pregnant woman n.a. Shultz, Schulte, and Berman 1992 Treatment Various sites, 1992 US$3.82 per pregnant woman n.a. Shultz, Schulte, and Berman 1992 Detection and treatment Various sites, 1992 US$84.92 per adverse outcome US$307.88 per adverse Shultz, Schulte, and Berman averted in low-prevalence context outcome averted in high- 1992 (5 percent) prevalence context (20 percent) Neoplasms Detection of cervical cancer Ecuador, 1996 US$2.95 per visit US$3.51 per visit Mumford and others 1998 (laboratory costs) Honduras, 1991 US$5.60 per visit US$12.89 per visit Mumford and others 1998 Zimbabwe, 1995 US$2.99 per visit US$3.89 per visit Mumford and others 1998 Cervical cytology screening Vietnam, 2000 US$725 per discounted DALY n.a. Suba and others 2001 averted Cervical cytology screening South Africa, 2000 US$39 per DALY averted US$81 per DALY averted Goldie and others 2001 Treatment of cervical cancer Zimbabwe, 1994 US$12.35 per visit US$95.82 per visit Mitchell, Littlefield, and Mexico, 1994 US$52.51 per visit US$432.42 per visit Gutter 1997 Treatment of cervical cancer Developed countries US$2,384 to US$28,770 per DALY n.a. Rose and Lappas 2000 averted based on actual survival Treatment of cervical cancer Developed countries US$308 to US$3,712 per DALY n.a. Rose and Lappas 2000 averted based on estimated survival Treatment of ovarian cancer Thailand, 1995 US$234.25 per case US$472.27 per case Tintara and Leetanapon 1995 Management of breast Brazil, 1995 US$1,667.88 per case n.a. Arredondo, Lockett, and Icaza cancer 1995 204 | Disease Control Priorities in Developing Countries | Mayra Buvini´c, André Médici, Elisa Fernández, and others Table 10.5 Continued Intervention Region or country Lower end of range Upper end of range Source Alzheimer's disease and other dementias Acetylcholinesterase Developed countries US$13 per hour of caregiver time n.a. Marin and others 2003 inhibitors saved Developing countries US$10 per hour of caregiver time n.a. Marin and others 2003 saved Osteoarthritis Celecoxib monotherapy Sweden US$1,394 per QALY n.a. Haglund and Svarvar 2000 Rofecoxib monotherapy United Kingdom US$2,184 per life year saved (result n.a. Moore and others 2001 was sensitive to the use of gas- trointestinal protective agents) Different packages of drugs, United States US$2,001 to US$2,140 per QALY, n.a. Sigal 2002 including acetaminophen, depending on the drug naproxen, misoprostol, combination celecoxib, and rofecoxib Total hip arthroplasty United States US$6,893 per QALY for n.a. Chang, Pellissier, and Hazen 85-year-old men 1996 Knee replacement Australia US$6,000 per QALY Sigal and others 2004 Source: Authors. QALY quality-adjusted life year. Note: Costs are based in current U.S. dollars as presented in each study. Many studies do not present well-documented data regarding reference period of costs. To avoid mistakes on interpretation, we kept the costs in the currency informed by the authors. n.a. indicates that information is not available. cancer being one of the most important. Recommended strate- in Zimbabwe. Differences in treatment costs are associated gies involve early detection and treatment. The following are with the kinds of procedures used. How the results of cost- the main strategies to prevent cervical cancer: effectiveness studies for cervical cancer prevention and screen- ing interventions in developed countries might translate to · screening and treatment performed during the same visit health care delivery settings in developing countries is not clear, · screening and treatment performed at two separate visits but prevention could clearly play an important role. · traditional three-visit intervention, in which a cytology sam- Many studies of breast cancer prevention view diet as an ple is obtained during the first visit, a diagnostic colposcopy important condition explaining the predisposition for breast is performed for those who screened positive during the sec- cancer. Ministries of health in many developing countries invest ond visit, and treatment is provided at the third visit. in promotion and prevention, issuing communications and The data on costs associated with cervical cancer detection guidelines for early detection using self-testing as a cost-effective and treatment in developing countries are limited. In way to provide information.Few studies of the cost-effectiveness Honduras in 1991, costs per visit for cervical cancer detection of different breast cancer treatments are available, especially in varied from US$5.60 (small clinics) to US$12.90 (larger clinics) developing countries. One on the management of breast cancer (Mumford and others 1998). Lower detection costs were found in Brazil in 1995 showed extremely high costs of US$1,678 per in Ecuador in 1996 (US$2.95 to US$3.51 per visit) and in death averted (Arredondo, Lockett, and Icaza 1995). Zimbabwe in 1995 (US$3.00 to US$3.90 per visit). Recent stud- ies on cervical cancer screening in South Africa show that the Conditions That Affect Women Disproportionately two-visit method is more cost-effective than the traditional Few studies on shared diseases that affect women dispropor- three-visit method, US$39 per DALY averted compared with tionately include gender-related considerations, especially in US$81 (Goldie and others 2001). Studies in Vietnam found developing countries. Most literature on Alzheimer's disease, costs equivalent to US$725 per DALY averted with cytology unipolar depressive disorders, and osteoarthritis presented in screening (Suba and others 2001). this section is based on studies in developed countries with Regarding treatment of cervical cancer, Rose and Lappas's no specific analysis of gender differences in relation to cost- (2000) studies in developed countries find costs varying from effectiveness. US$2,384 to US$28,770 per DALY averted. Costs are lower in developing countries, ranging from US$52.51 to US$432.42 Alzheimer's Disease and Other Dementias. Alzheimer's dis- per visit in Mexico and from US$12.35 to US$95.82 per visit ease is linked to genetic and other risk factors, including Gender Differentials in Health | 205 increasing age, positive family history of dementia, and lower measures of using several kinds of medicines (acetaminophen, levels of education. Treatment is based mostly on drugs, and naproxen, misoprostol, celecoxib, and rofecoxib) under differ- the practical benefits of treatment translate mainly into ent conditions are available. Sigal (2002) shows that by using reduced caregiver hours. different combinations of medicines, treatment costs can vary Some studies have found that interventions aimed at reduc- from US$2,001 to US$2,140 per quality-adjusted life year, but ing caregiver stress, even providing low-dose antipsychotic these costs are prohibitive for developing countries. medication, can be effective. However, the costs of undertaking Another treatment for osteoarthritis is synovial fluid replace- such interventions have not been quantified; thus, their cost- ment, but given the costs of this intervention, it is not currently effectiveness cannot be calculated. Institutional care for recommended for developing regions. Surgical interventions patients with any form of dementia is extremely limited in for osteoarthritis,such as joint replacement,are most commonly LMICs. The costs of setting up institutions for those with performed in developed countries. Sigal and others (2004) Alzheimer's disease and the costs of care are prohibitive. In this review a number of interventions for osteoarthritis and suggest context, inexpensive, home-based care appears to be the only a cost per quality-adjusted life year of US$6,000 for knee viable option for Alzheimer's disease patients in developing replacements. In developing countries, however, the availability countries. These countries will therefore have to face the chal- of surgical interventions is constrained by its costs and by the lenges of addressing families' needs in relation to financial and availability of surgeons qualified to perform the operation. social support and caregiver training. Another issue is the tradeoff between women's income-earning opportunities and RESEARCH AGENDA their traditional primary role as family caregivers. Health research and practice should give priority attention to Unipolar Depressive Disorders. Depression is among the the 19 conditions with excess burdens for women that this most disabling and costly illnesses in the world, especially for chapter has identified and clustered into four main groups, women. Despite good short-term treatment outcomes, long- according to the interplay of sex and gender in their etiology, to term outcomes remain disappointing. Costs associated with improve women's health status worldwide. A number of cost- depression affect not just the sufferers themselves, but also their effective health technologies are available, but women are dying families and friends (time dedicated to caregiving); employers because these technologies are not available and accessible to (payment for treatment and care, as well as for reduced pro- all. Thus, a priority need is to deploy them more widely in ductivity); and society (provision of mental health care developing countries. The research challenge is also urgent, financed by taxpayers). Most of these costs are difficult to given the unacceptably high disease burdens for women in obtain, but the consensus is that the indirect costs of depression developing countries and the rising numbers of older women are larger than the direct costs. worldwide. Research on sex and gender factors affecting The treatment setting for depression is usually primary women's disease burdens should give physicians new informa- health care, with many kinds of episodic treatments combining tion that will increase their options in relation to diagnostic old and new generations of antidepressants and psychosocial practices and drug approaches, not only for women, but procedures. Averting depressive episodes results in average also for men, thereby improving the provision of gender- gains of up to 50 disability days per treated case per year. Studies appropriate health care for all. of the factors influencing women's access to screening, preven- Key general items in a research agenda to reduce women's tion, and treatment for depression and the cost-effectiveness of excess disease burdens include the following: treatment options should be a priority in developing countries. · Expand and sharpen analyses of sex and gender and their Osteoarthritis. Despite clear evidence of a reduction in symp- interaction in the etiology, onset, progression, prevention, toms and delayed progression of osteoarthritis with weight and treatment of diseases that women and men share, but reduction, no formal studies of cost-effectiveness are available. where women face an excess health burden, and in the Education and exercise programs for osteoarthritis are avail- assessment of cost-effective interventions. able in developed countries, but such programs are unknown · Increase research to identify the determinants of women's in developing countries. Studies of the effect of diet and physi- underuse of health services, paying special attention to the cal exercise in preventing osteoarthritis in women are a priority accessibility, affordability, and appropriateness of services. not just in developed countries, but especially in developing Identify and analyze best practices in health service delivery countries with fiscally strapped health systems and growing that incorporate gender variables to inform training and elderly populations. human resource development programs for health sector Acetaminophen is thought to be the most cost-effective ini- providers in developing countries for both services specific tial treatment with drugs. In addition, some cost-effectiveness to women and general services. 206 | Disease Control Priorities in Developing Countries | Mayra Buvini´c, André Médici, Elisa Fernández, and others · Investigate how interactions between sex and gender, race, demiology studies to assess the importance of different ethnicity, and poverty affect the etiology, onset, and pro- genetic risks. gression of disease as well as access to and compliance with · Expand research on the biological and behavioral determi- prevention and treatment. Investigate the effects of stigma- nants of depression in women, on screening alternatives, tization and discrimination on service quality and use. and on cost-effective prevention and treatment options · Use demographic accounting methods to project and plan adapted to specific contexts in developing countries. for the demands that women's overall greater longevity Explore ways to reduce the stigmatization and discrimina- imposes on health systems and investigate cost-effective tion associated with depression in developing countries. treatment options, effect of gender variables on prevention · Continue research to document the risks and benefits of and treatment, and viable options for increasing older common, but unproven, approaches of preventing and women's access to health and to social insurance and pro- treating diseases specific to women, such as the finding that tection mechanisms. Investigate affordable options for fam- long-term hormone replacement therapy does not reduce ily care of elderly patients that take into account the time cardiovascular disease in postmenopausal women as had and the physical and emotional burdens on women, the long been thought but instead increases risks for cardiovas- traditional family caretakers. cular disease and breast cancer. · Improve the methodology used for disease classification, and · Establish a research program to document the prevalence expand data collection efforts to address the largely unre- and disease burdens associated with FGM, and seek preven- ported causes of women's disease burdens, emphasizing con- tion and treatment alternatives. ditions specific to women, including FGM and other domes- tic-, social-, and religious-based violence against women. · Promote research, health promotion activities, health CONCLUSIONS services, and advocacy efforts that will help women adopt desirable nutrition and physical exercise practices for opti- The purpose of this chapter has been to move beyond the mal health. Exercise is a preventive measure that can help traditional international health focus on women's diseases reduce women's excess disease burdens throughout the life related to their reproductive and maternal functions and to cycle and has numerous indirect psychological and social highlight those conditions for which sex and gender consider- benefits, but it will often require changing deeply rooted cul- ations, if adequately incorporated into prevention and treat- tural mores. ment, could reduce women's excess health burdens and, as a · Support research on the costs and effectiveness of treat- result, increase health equity. Because of the approach taken, ments for diseases that affect both men and women but the chapter has excluded a set of important conditions for affect women disproportionately, especially in developing which the gender ratio is similar or is unfavorable for men, countries. including HIV/AIDS, injuries resulting from domestic vio- lence, and malaria. Therefore, this chapter should not be Research needs pertaining to specific conditions include the viewed as covering all diseases and conditions important to following: women or all conditions that result from the interaction of sex and gender. · Carry out further testing of innovative technologies for The chapter has two other main limitations. First, the dataset diagnosing neoplasms specific to women and explore new (the global burden of disease) understates certain disease options for preventing their growth and proliferation. burdens, especially for women, because it does not estimate Regarding treatment, the literature describes few best prac- disability weights for some gender-based conditions, such as tices and cost-effective measures, especially for breast and FGM. Second, the information available for cost-effectiveness ovarian cancer. analysis is inadequate. Following this chapter's emphasis on the · Evaluate the viability of applying the results of cost- importance of sex and gender in explaining women's excess effectiveness analysis carried out in developed countries disease burdens, assuming that sex and gender considerations in on the prevention and treatment of cancers specific to prevention and treatment would affect the analysis of cost- women in developing countries to help adapt viable effectiveness seemed reasonable; however, general information procedures. for assessing cost-effectiveness was deficient, and gender-related · Promote and study the cost-effectiveness of approaches to information was entirely lacking. This limitation was additional preventing Alzheimer's disease, osteoarthritis, and other to the common limitation of cost-effectiveness analysis in terms chronic conditions related to aging. of underestimating the value of prevention for conditions spe- · Compare the prevalence of Alzheimer's disease in cific to women and non­health sector interventions important different populations, including carrying out genetic epi- to women. Gender Differentials in Health | 207 Annex 10.A Milestones and Influential Works in International Women's Health (1980­2003) Year Milestones and Influential Work 1980 · Maternal and child health · Infant feeding practices · Women's reproductive and fertility behaviors 1984 · Child survival research (Mosley and Chen 1984) · Women's issues for child survival 1985 UN Women's Conference in Nairobi · Gender perspective/framework (DAWN) 1987 Safe Motherhood Initiative (World Bank) · Barber Conable speech (Conable 1986) · Herz and Measham 1987 Reproductive Health and Dignity: Choices by · Germain 1987 Third World Women (International Women's Health Coalition) 1988 Women's Work and Child Nutrition (Leslie 1988) · Debunks myth of incompatibility between breastfeeding and work 1989 Quality of Care for Women in Family Planning · Quality of care (Bruce 1990) (Population Council) 1990 1990­1994: 10 Working Papers (World Bank) · Life-stage focus: adolescents, postreproductive age, and so on · Burden focus: cervical cancer, abortion, HIV/AIDS, reproductive tract infections, violence, and so on 1992 Women's Health: Across Age and Frontier (WHO) · Socioeconomic, cultural, and legal factors affecting women's health · Policy directions for women's nutrition · Access and empowerment 1993 Disease Control Priorities in Developing · Cost-effectiveness issues in women's health Countries (World Bank) · Missing women in India (A. K. Sen 1990) · Sector study of women's health in India (1992) 1994 New Agenda for Women's Health and · Lifecycle approach to account for specific and cumulative effects of nutrition Nutrition (World Bank) · Cost-effective intervention packages advocated · Women and AIDS (Mann 1993) · Stigma and lack of empowerment as risk factors International Conference on Population and · Agenda setting, policy making, and programming for reproductive health since Development, Cairo the 1994 Cairo conference 1995 UN Beijing Women's Conference · Reproductive health and rights 1996 In Her Lifetime: Female Morbidity and Mortality · Burdens exclusive to, greater for, and of particular significance to women in Sub-Saharan Africa (Institute of Medicine) reviewed · Burdens tracked across life span 1998 Women, Aging and Health (WHO) · Social, cultural, political, economic determinants of major health issues facing aging and postmenopausal women · Life-cycle approach Gender and Health: A Technical Paper (WHO) · Role of social and cultural factors and power relations between men and women in promoting and protecting health 1999 Safe Motherhood and the World Bank · Effect of safe motherhood on labor supply, productive capacity, community economic well-being Gender and HIV/AIDS (Joint United Nations · Gender-specific personal and societal vulnerability to HIV/AIDS Program on HIV, International Center for Research on Women) 2000 Women of South East Asia: Health Profile (WHO) · Life-cycle approach to review gender-specific and disproportionate burdens Improving Women's Health: Issues and · Role of biological and social factors in women's exposure risk and disease Interventions (World Bank) progression · Life cycle approach 208 | Disease Control Priorities in Developing Countries | Mayra Buvini´c, André Médici, Elisa Fernández, and others Annex 10.A Continued Year Milestones and Influential work 2000 2000: Investing in the Best Buys (World Bank) · Emphasis on identifying and funding most cost-effective programs and factors that lead to program success 2002 2002: Reproductive Health Outlook: Older Women · Health conditions and interventions for older women 2002: International Position Paper on Women's · Best clinical practices to address conditions associated with menopause Health and Menopause (National Institutes of Health) Source: Adapted from an original project proposal for this volume by S. Goldie, R. Anhang, and M. Buvinic´. 2004. The Evolving Agenda for Women's Health. ACKNOWLEDGMENTS One Works and What Do They Cost? Policy Project. Washington, DC: Futures Group. The authors would like to acknowledge the useful comments of Dmytraczenko, T., I. Aitken, S. E. Carrasco, J. Holley, W. Abramson, A. E. Barrett-Connor, Mahmoud F. Fathalla, and Philip Musgrove Valle, and M. Effen. 1998. Evaluación del seguro de maternidad y niñez en Bolivia. Bethesda, MD: Abt Associates. and the collaboration of Sue Goldie, Anthony Measham, and Evans, T., M. Whitehead, F. Diderichsen, A. Bhuiya, and M. Wirth. 2001. Sonbol A. Shahid-Salles in the preparation of this chapter. "Challenging Inequities in Health." In Challenging Inequities in Health: From Ethics to Action. New York: Oxford University Press. NOTES Germain, A. 1987. "Reproductive Health and Dignity: Choices by Third World Women." Paper commissioned for the International Conference 1. See, for instance, the documentation of efforts to understand gender on Better Health for Women and Children through Family Planning, differentials in infectious diseases (Altman 2004) and in unipolar depres- Nairobi, Kenya, October 1987, by the International Women's Health sive disorders (Gilbert 2004). Coalition, New York. 2. Males suffer the weight of some external causes in the burden of dis- Gilbert, S. 2004. "New Clues to Women Veiled in Black." New York Times, eases more than females. For example, in 2001, mortality and morbidity March 16. caused by war and violence worldwide accounted for about 2.7 percent of males' total DALY losses and 0.6 percent of females' losses. Goldie, S. J., L. Kuhn, L. Denny, A. Pollack, and T. C. Wright. 2001. "Policy 3. Cardiovascular diseases include cerebrovascular diseases but do not Analysis of Cervical Cancer Screening Strategies in Low-Resource include ischemic heart disease, which affects men more than women, or Settings: Clinical Benefits and Cost-Effectiveness." Journal of the coronary artery disease. American Medical Association 285 (24): 3107­15. Gupta, R. G. 2000."Approaches for Empowering Women in the HIV/AIDS Pandemic: A Gender Perspective." Paper prepared for the Expert REFERENCES Group Meeting on the HIV/AIDS Pandemic and Its Gender Implications, Windhoek, Namibia, November 13­17. Altman, L. K. 2004. "Action Urged on Diseases with Dangers for Women." New York Times, February 28. Haglund, U., and P. Svarvar. 2000. "The Swedish ACCES Model: Predicting the Health Economic Impact of Celecoxib in Patients with Osteoarthritis Arredondo, A., L. I. Lockett, and E. de Icaza. 1995. "Cost of Diseases or Rheumatoid Arthritis." Rheumatology 39 (Suppl. 2): 51­56. in Brazil: Breast Cancer, Enteritis, Cardiac Valve Disease, and Bronchopneumonia." Revista de Saude Publica 29 (5): 349­54. Hanson, K. 2002. "Measuring Up: Gender, Burden of Disease, and Priority Setting." In Engendering International Health, ed. G. Sen, A. George, Berman, P., J. Quinley, B. Yusuf, S. Anwar, U. Mustaini, A. Azof, and and P. Ostlin, 313­45. Cambridge, MA: Massachusetts Institute of I. Iskandar. 1991. "Maternal Tetanus Immunization in Aceh Province, Technology Press. Sumatra: The Cost-Effectiveness of Alternative Strategies." Social Science and Medicine 33 (2): 185­92. Heise, L., with J. Pitanguy and A. Germain. 1994. Violence against Women: The Hidden Health Burden. Washington, DC: World Bank. Breen, N. 2002. Social Discrimination and Health: Gender, Race, and Class in the United States. In Engendering International Health, ed. G. Sen, A. Herz, B., and A. R. Measham. 1987. The Safe Motherhood Initiative: George, and P. Ostlin, 223­55. Cambridge, MA: Massachusetts Proposals for Action. Washington, DC: World Bank. Institute of Technology Press. Kinsella, K., and Y. J. Gist. 1998. Gender and Aging: Mortality and Health. Bromberger, J. T. 2004. "A Psychosocial Understanding of Depression in International Programs Center Report IB/98-2. Washington, DC: U.S Women: For the Primary Care Physician." Journal of the American Department of Commerce, Economics and Statistics Administration. Medical Women's Association 59 (3): 198­206. Krieger, N. 2003. "Gender, Sexes, and Health: What Are the Connections Bruce, J. 1990. "Fundamental Elements of the Quality of Care: A Simple and Why Does It Matter?" International Journal of Epidemiology 32 (4): Framework." Studies in Family Planning 21 (2): 61­91. 652­57. Chang, R. W., J. M. Pellissier, and G. B. Hazen. 1996. "A Cost-Effectiveness Leslie, J. 1988. "Women's Work and Child Nutrition in the Third World." Analysis of Total Hip Arthroplasty for Osteoarthritis of the Hip." World Development 16 (11): 1341­62. Journal of the American Medical Association 275 (11): 858­65. Levin, A., T. Dmyatraczenko, F. Ssengooba, M. McEuen, F. Mirembe, M. Conable, B. 1987. "Population Growth and Policies in Sub-Saharan Nakakeeto, and others. 1999. Costs of Maternal Care Services in Masaka Africa." Address to the annual meetings of World Bank and District, Uganda. Bethesda, MD: Abt Associates. International Monetary Fund, Washington, DC. September. Levine, R., A. Glassman, and M. Schneidman. 2001. La Salud de la Mujer Dayaratna, V., W. Winfrey, W. McGreevey, K. Hardee, J. Smith, E. en América Latina y el Caribe. Washington, DC: Inter-American Mumford, and others. 2000. Reproductive Health Interventions: Which Development Bank. Gender Differentials in Health | 209 Linfante, A. H., R. Allan, S. C. Smith, and L. Mosca. 2003. "Psychosocial Status." Global Programme on Evidence for Health Policy Discussion Factors Predict Coronary Heart Disease, but What Predicts Paper Series 15, World Health Organization, Geneva. Psychosocial Risk in Women?" Journal of the American Medical Seils, D. M., J. Y. Friedman, and K. A. Schulman. 2001. "Sex Differences in Women's Association 58 (4): 248­53. the Referral Process for Invasive Cardiac Procedures." Journal of the Mann, J. M. 1993. "Women and AIDS: Critical Issues." Speech given at American Medical Women's Association 56 (4): 151­55. the International Center for Research on Women Meeting, New York, Sen, A. K. 1990. "More Than 100 Million Women Are Missing." New York April 16. Review of Books 37 (20): December 20. Manson, J. E., P. Greenland, A. Z. LaCroix, M. L. Stefanick, C. P. Mouton, Sen, G., A. George, and P. Ostlin. 2002. "Engendering Health Equity: A A. Oberman, and others. 2002. "Walking Compared with Vigorous Review of Research and Policy." In Engendering International Health, Exercise for the Prevention of Cardiovascular Events in Women." New ed. G. Sen, A. George, and P. Ostlin, 1­33. Cambridge, MA: England Journal of Medicine 347(10): 716­25. Massachusetts Institute of Technology Press. Marin, D., K. Amaya, R. Casciano, K. L. Puder, J. Casciano, S. Chang, and Shultz K., J. Schulte, and S. Berman. 1992. "Maternal Health and Child others. 2003. "Impact of Rivastigmine on Costs and on Time Spent in Survival: Opportunities to Protect Both Women and Children from Caregiving for Families of Patients with Alzheimer's Disease." the Adverse Consequences of Reproductive Tract Infections." In International Psychogeriatrics 15 (4): 385­98. Reproductive Tract Infections: Global Impact and Priorities for Women's McCann, J., L. Hebert, D. Bennett, V. Skul, and D. Evans. 1997. "Why Reproductive Health, ed. A. Germain, K. Homes, P. Piot, and J. Alzheimer's Disease Is a Women's Health Issue." Journal of the Wasserheit, 145­82. New York: Plenum Press. American Medical Women's Association 52 (3): 132­37. Shumaker, S., C. Legault, S. Rapp, L. Thal, R. B. Wallace, J. K. Ockene, and McNamara, R. 1981. "To the Massachusetts Institute of Technology: An others. 2003. "Estrogen Plus Progestin and the Incidence of Dementia Address of the Population Problem," April 28, 1977. In The McNamara and Mild Cognitive Impairment in Postmenopausal Women." Journal Years at the World Bank: Major Policy Addresses, 415­16. Baltimore and of the American Medical Association 289 (20): 2651­62. London: John Hopkins University Press. Sigal, L. 2002. "Cost-Effectiveness Analysis of Osteoarthritis Treatment Mitchell, M. D., J. Littlefield, and S. Gutter. 1997. "Reproductive Health: Options." Data presented at the American College of Rheumatology From Policy to Practice." Draft manuscript. Management Sciences for 66th Annual Scientific Meeting, New Orleans, December 16. Health, Boston, MA. Sigal, L., S. E. Day, A. B. Chapman, and R. H. Osbourne. 2004. "Can We Moore, R. A., C. J. Phillips, J. M. Pellissier, and S. X. Kong. 2001. "Health Reduce Disease Burden from Osteoarthritis." Medical Journal of Economic Comparisons of Rofecoxib versus Conventional Australia 180 (5 Suppl.): S11­17. Nonsteroidal Anti-inflammatory Drugs for Osteoarthritis in the Suba, E. J., C. H. Nguyen, B. D. Nguyen, and S. S. Raab. 2001. "De Novo United Kingdom." Journal of Medical Economics 4: 1­17. Establishment and Cost-Effectiveness of Papanicolaou Cytology Mosley, W. H., and L. C. Chen, eds. 1984. Child Survival: Strategies for Screening Services in the Socialist Republic of Vietnam: Viet/American Research. Cambridge, U.K.: Cambridge University Press. Cervical Cancer Prevention Project." Cancer 91 (5): 928­39. Mumford, E., V. Dayaratna, W. Winfrey, J. Sine, and W. McGreevey. 1998. Tintara, H., and R. Leetanapon. 1995. "Cost-Benefit Analysis of Reproductive Health Costs: Literature Review. Washington, DC: Futures Laparoscopic Adnexectomy." International Journal of Gynecology and Group. Obstetrics 50 (1): 21­25. Murray, C. J. L., and D. B. Evans, eds. 2003. Health Systems Performance Verbrugge, L. M. 1989. "Gender, Aging, and Health." In Aging and Health: Assessment: Debates, Methods, and Empiricism. Geneva: World Health Perspectives on Gender, Race, Ethnicity and Class, ed. K. S. Markides. Organization. Newbury Park, CA: Sage. Osmania, S., and A. Sen. 2003. "The Hidden Penalties of Gender WHO (World Health Organization). 1997."Mother-Baby Package Costing Inequality: Fetal Origins of Ill-Health." Economics and Human Biology Spreadsheet." Unpublished spreadsheet, WHO, Geneva. 1 (1): 91­104. ------. 1998. World Health Report 1998. Geneva: WHO. Over, M., and P. Piot. 1993. "HIV Infection and Sexually Transmitted ------. 2000. Women's Mental Health: An Evidence Based Review. Geneva: Diseases." In Disease Control Priorities in Developing Countries, ed. D. T. WHO. Jamison, W. H. Mosley, A. R. Measham, and J. L., Bobadilla, 455­527. New York: Oxford University Press for the World Bank. ------. 2001. Global Burden of Disease data. Accessible at www.fic. nih.gov/dcpp/gbd.html. Pinn, V. W. 2003."Sex and Gender Factors in Medical Studies. Implications for Health and Clinical Practice." Journal of the American Medical ------. 2004. "Prevalence, Severity, and Unmet Need for Treatment of Association 289 (4): 397­400. Mental Disorders in the World Health Organization World Mental Health Surveys." Journal of the American Medical Association 291 (21): Raviez, M., C. Griffin, and A. Follmer. 1995. "Health Policy in Eastern 2581­90. Africa: A Structured Approach to Resource Allocation." Unpublished report, World Bank, Washington, DC. World Bank. 1993. World Development Report: Investing in Health. New Rose, P. G., and P. T. Lappas. 2000. "Analysis of the Cost-Effectiveness York: Oxford University Press. of Concurrent Cisplatin-Based Chemoradiation in Cervical Cancer: ------. 1996. Development in Practice: Improving Women's Health in Implications from Five Randomized Trials." Gynecologic Oncology 78 India. Washington, DC: World Bank. (1): 3­6. Zubenko, G. S., H. B. Hughes III, B. S. Maher, J. S. Stiffler, W. N. Zubenko, Rosenfield, A., and D. Maine. 1985. "Maternal Mortality: A Neglected and M. L. Marazita. 2002. "Genetic Linkage of Region Containing the Tragedy--Where Is the M in MCH?" Lancet 2 (8446): 83­85. CREB1 Gene to Depressive Disorders in Women from Families with Rosenthal, G., and A. Percy. 1991. "Maternal Services in Cochabamba, Recurrent, Early-Onset, Major Depression." American Journal of Bolivia: Costs, Cost-Recovery, and Changing Markets." Arlington, VA: Medical Genetics 114 (8): 980­87. MotherCare, John Snow. Sadana, R., C. D. Mather, A. D. Lopez, C. J. L. Murray, and K. Iburg. 2000. "Comparative Analysis of More Than 50 Household Surveys on Health 210 | Disease Control Priorities in Developing Countries | Mayra Buvini´c, André Médici, Elisa Fernández, and others Chapter 11 Fiscal Policies for Health Promotion and Disease Prevention Rachel Nugent and Felicia Knaul Governments use fiscal policy to encourage healthy behavior. The chapter is divided into five sections: The instruments of government for this purpose are taxes and subsidies, and direct provision of certain health services for free · The first section provides a general framework through or at subsidized rates. Examples of fiscal policies for health are which fiscal policy options can be considered in terms of taxes on tobacco and alcohol, subsidies on certain foods, and their impact on health and the health sector. tax incentives for health care purchases. · The next section examines the experiences in developing Government intervention through fiscal policy works best countries of using subsidies to achieve health-related objec- when public institutions and credibility are strong, the design tives (columns 1 and 2 in table 11.1). and application of the fiscal instruments are appropriate, and · The third section presents examples of how taxes are used consumers' and producers' responsiveness to a price signal is in a number of countries to promote health (column 3 in high. When these conditions are not present, direct provision, table 11.1). information and education campaigns, or legislation may be · The next section discusses nonhealth goods where fiscal preferable in conjunction with fiscal policy. policies are often used and have important indirect health The purpose of this chapter is to review country experiences benefits (lower part of column 2 in table 11.1). with promoting health through fiscal policies and to examine · The final section presents conclusions and suggestions for the usefulness and success of these policies. The chapter consid- further research and policy development. ers both the role of fiscal policies in the production of health and the effect of these policies on the well-being of the economy-- fiscal policy for health and healthy fiscal policy.1 Little exists in USE OF FISCAL POLICY FOR HEALTH the literature linking fiscal policy and health promotion except IN DEVELOPING COUNTRIES in relation to tobacco. This work contributes to filling that gap. The chapter deals specifically with experiences at the coun- Fiscal policies come in a wide range of designs, but the main try level with tax policies affecting some goods related to effect is either to alter the price of health-related goods or to health, such as food, tobacco, alcohol, and condoms; subsidized alter the quantity available. Table 11.2 summarizes the health provision of workplace promotion of healthy behavior and interventions subject to fiscal policies. The behaviors that caregiving; and direct subsidies affecting food provision and require these interventions are divided into the following fortification, cooking fuels, water purification and soap, con- categories: doms, bednets, vaccines, and medical research. The chapter only touches on health care provision and does not discuss its · unhealthy consumption (foods, tobacco, and alcohol), for financing, either directly by governments or through insur- which the most salient fiscal policies are taxes on consumers ance, because other chapters deal with those topics. and producers, and fines 211 Table 11.1 Fiscal Policies for Health Promotion Covered in Chapter 11 Subsidy Health-related Subsidized Health-related Government for or tax products receiving provision products directly financing of imposed on direct subsidies of health taxed health care Consumer Medicine Caregiving (partially covered) Tobacco Not covered Food Alcohol Cooking fuel Food Water purification Imported medicine and supplies Soap Condoms Bednets Producer Vaccinations Workplace promotion of healthy Fuel usage (partially covered) Food additives behavior (partially covered) Medical research Source: Authors. · health promotion and disease and accident prevention SUBSIDIES FOR HEALTH AND (hygiene, pollution, safety, public health, maternal and child HEALTH-RELATED PRODUCTS and reproductive health, infectious disease, and healthy lifestyles), for which the most important fiscal instruments Using examples primarily from developing countries, this sec- are subsidies, but which may also be affected by tax policy tion of the chapter analyzes the range of subsidies that are · health care goods and inputs, including insurance and available to promote healthy behavior and the consumption of human resources, that may be exempted from taxation, health-related goods. The first sections deal with consumer subsidized, or guaranteed as a constitutional right subsidies both to promote the consumption of health- · other goods that indirectly promote health (education, producing goods and of health care. The second section dis- housing, agriculture, energy, charitable giving, charities that cusses producer subsidies. provide targeted subsidies, and so on), which are often sub- ject to their own particular tax regime or sets of subsidies Consumer Subsidies that affect their production or consumption and, therefore, also affect health behavior Governments use consumer subsidies to encourage the use of a · research and development initiatives that can be applied to beneficial product by lowering the price consumers pay-- health and health care goods and are sensitive to tax exemp- usually in situations where the consumers are too poor, the tions and subsidies. market prices of the good are too high, or both situations apply--to otherwise achieve a socially optimal consumption Fiscal interventions can have various rationales, such as level. Examples include subsidies for staple foods, condoms, macroeconomic benefits, equity, or efficiency--and promoting soap, insecticide-treated bednets, cooking fuels, and medicines. health may or may not be the primary goal. A fiscal policy may be designed to affect some other sphere of behavior or a good Staple Foods. Ample evidence indicates that food subsidies other than health--for instance, education--and the effects on are effective in improving nutrition; however, appropriate tar- health or the use of health care may be indirect. The shaded geting is often a problem (Alderman 2002). Subsidies may be boxes in table 11.3 indicate the possible rationale behind each targeted to specific foods, specific delivery locales or geo- type of fiscal policy. graphic areas, or specific populations. Often targeting includes A fiscal policy should be effective, efficient, and cost- all three. effective and should promote or maintain equity goals. An Food-specific subsidies, whether in the form of general effective tax or subsidy reaches the intended target and alters subsidies, ration cards, quotas, or food stamps, increase food health-related behavior in the desired manner. An efficient consumption. They will have a positive effect on health if this policy minimizes resource distortions and involves minimal consumption occurs in undernourished populations that administrative costs. A cost-effective policy has the lowest cost require increased caloric or nutrient intake. In some cases, food relative to the desired health goal. subsidy programs have had unintended macroeconomic and 212 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul Table 11.2 Use of Taxes and Subsidies to Promote Health by Type of Intervention Tax Taxes preferences Subsidies Consumer (sales and Director- Targeted value Credits and subsidized Subcontract consumer Intervention Payroll added tax) Excise Production Input Fines exemptions provision provision subsidies Producers Rights Unhealthy consumption Foods Alcohol Tobacco Health promotion and disease and accident prevention Hygiene (soap) Pollution (for example, fuels) Safety (for example, seat belts) Public health (vaccines, clean water, supplementation, education, and information) Fiscal Child, maternal, and Policies reproductive health (information and for education, Health supplementation, and medical attention) Promotion Infectious disease (condoms and healthy and workplaces) Disease Healthy lifestyles (food, exercise, and Prevention healthy workplaces) (Continues on the following page.) | 213 214 Table 11.2 Continued | Disease Tax Taxes preferences Subsidies Control Consumer Priorities (sales and Director- Targeted value Credits and subsidized Subcontract consumer Intervention Payroll added tax) Excise Production Input Fines exemptions provision provision subsidies Producers Rights in Developing Health care goods Insurance Countries Medical attention Medicines | Rachel Human resources for health Nugent Other goods that indirectly promote health and Personal leave Felicia Caregiving and family leave (maternity, paternity, and Knaul chronic illness) Education (for mothers or women, early childhood, or special needs) Housing (provision, flooring, and roofing) Agriculture (type of products produced or imported) Energy (types of fuels, heating, cars, and gasoline) Charitable giving Research and development Source: Authors. Note: Shaded boxes indicate possible fiscal policies for each type of intervention. Table 11.3 Taxes and Subsidies by Policy Rationale Rationale Macroeconomic or Promotion of another Policy Healthy behavior fiscal benefits Equity Efficiency good or type of behavior Taxes Payroll Consumer (value added tax/sales) Levy Excise Producer Input Fines Subsidies Tax credits Direct provision Fiscal Subsidized provision Policies Subcontracted provision for Targeted subsidies Health Subsidies to producers Promotion Rights Tax exemptions and Source: Authors. Disease Note: The shaded boxes indicate the possible rationale behind each type of fiscal policy. Prevention | 215 microeconomic consequences (Adams 2000; del Ninno and These kinds of issues underscore the importance of design Dorosh 2002; Pinstrup-Anderson 1988; Siamwalla 1988). They considerations and country conditions in creating effective and become expensive if they are too widely available, can create efficient food subsidy programs. incentives for black market activities, and can affect prices and In sum, many food subsidy programs avoid the political and volumes in agricultural and trade markets. administrative challenges of explicit targeting by allowing uni- Indonesia switched from a general rice support system to a versal access to the subsidies on the assumption that the needy limited subsidy during the 1997 macroeconomic crisis. The will self-select into the programs. However, Adams (2000) earlier system had successfully reduced food insecurity to low shows that countries with targeted food programs--for exam- levels, but higher prices increased the cost of maintaining the ple, Chile, Jamaica, and Peru--provide much higher income subsidy and led to food being smuggled out of the country transfers to the poor than do self-targeted programs of the kind (Tabor and Sawit 2001). The government targeted the new rice used in Egypt, Morocco, and Tunisia. subsidy to the poor and issued ration cards. Within roughly a year of implementation, the subsidy was reaching an estimated Condoms. Preliminary investigation indicates that subsidies 85 percent of the poor. Only about 10 percent of the subsidy on condoms can be effective in increasing their use in both appeared to be reaching nontarget population groups. general and high-risk populations, but whether price reduc- India has subsidized essential consumer goods for decades, tion, increased access, or education leads to greater use is not including health-related goods such as food grains, edible oils, clear (Price 2001) because information campaigns about the sugar, and fuels (S. Jha 1992). The government rationed certain health benefits of condoms usually accompany price subsidies. goods in the belief that only the truly needy would endure Recent surges in social-marketing schemes to distribute con- waiting in lines and purchasing the poorer quality products doms as part of the fight against HIV/AIDS, especially in that were involved in the subsidy schemes. This is called self- Africa, have increased condom use.2 targeting. However, Jha shows that 40 percent of the population Few researchers have compared HIV infection rates--or purchased subsidized rice in 1990, only half of whom were even condom use rates--before and after the introduction of a poor. The government recently modified the program to better subsidy on condoms. Cohen and others (1999) conclude that in target the subsidy to the poor and removed such barriers as a particular jurisdiction in Louisiana, free distribution through bulk purchasing (Rao 2000). public clinics and 1,000 small businesses in areas with high lev- The Arab Republic of Egypt's generalized program also illus- els of HIV and other sexually transmitted diseases achieved trates the problems that beset broad food subsidy programs. significantly higher condom distribution than a fee-based sys- The program reached its zenith in 1980, when it subsidized 20 tem (77 percent use during the last sexual encounter compared food products and accounted for 15 percent of government with 64 percent) and that the revenues from cost recovery were expenditures (Adams 2000). The program has been scaled back insufficient to justify imposition of the fee. The dropoff in to cover four staple foods and now accounts for 6 percent of condom use during the cost-recovery period persuaded the government expenditures. Nevertheless, about 75 percent of jurisdiction to reinstate the free distribution program. the population holds ration cards entitling them to purchase Another example suggests that promotion and information the subsidized foods. The program is intended to achieve self- are also effective. A social-marketing effort in Turkey in the targeting, but the nonpoor purchase many of the subsidized early 1990s offered condoms at a commercial price but foods. The program accounts for 44 percent of the total calorie included intensive advertising and other promotional efforts. It supply of the poorest quintile group, but in rural areas, the rich achieved sales well beyond original expectations and gained obtain more calories from subsidized food than the poor do. 41 percent of the market share (Yaser 1993). Musgrove (1993) reviews 104 supplementary feeding pro- grams in 19 countries in Latin America and the Caribbean. The Water Purification. The U.S. Centers for Disease Control and review covers a range of program sizes, from those serving Prevention and the Pan American Health Organization 1,000 individuals to those supplying 28 million people; of types designed the Safe Water System Initiative to improve the qual- of subsidies (namely, food distribution, direct feeding, and ity of drinking water for households that draw their water from direct payments); of levels of coverage of the targeted popula- sources outside the home. The principle underlying the initia- tion, ranging from 1.9 to 100.0 percent; and of extent of tive is to subsidize storage containers, disinfectant, and educa- coverage of the poor, varying from 5.8 to 88.0 percent. The per tion on proper handling to avoid contamination (Quick and capita costs of reaching beneficiaries differed widely. The most others 1999). Numerous countries have implemented similar common reasons for program ineffectiveness were spreading initiatives, including Bangladesh, Bolivia, Burkina Faso, Kenya, resources too thinly across beneficiaries, targeting foods with and Zambia. The government provides containers and chemi- minor health benefits, choosing inappropriate beneficiaries, cals at subsidized prices, but the costs are still higher than the and encountering excessive costs in distributing resources. cost of boiling water (Quick and others 2002). 216 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul Soap. Another proven method for reducing the incidence of grams have been designed to promote the use of liquid petro- diarrhea and other hygiene-related diseases is hand washing, leum gas, natural gas, or kerosene, which burn more cleanly with or without soap. Whether the key factor is education or and emit a low amount of smoke and particulates, but none has the subsidized provision of soap is unclear. Some investigators been efficacious or efficient (UNDP 2003). claim that small-scale programs that subsidize soap and edu- Liquid petroleum gas subsidies have been shown to benefit cate households about the benefits of hand washing are self- middle- and higher-income families in urban areas rather than financing because of the consequent reduction in disease the poor (UNDP 2003). In attempting to target the poor more (Borghi and others 2002). accurately, Côte d'Ivoire and Senegal focused subsidies on Luby and others' (2001) results from Pakistan suggest that smaller liquid petroleum gas cylinders but found that poor education alone may be just as effective as education accompa- consumers still preferred charcoal (UNDP 2003). Electricity nied by soap provision in reducing diarrheal disease. By con- subsidies in low-income countries are also often skewed toward trast, Hoque (2003) and other researchers suggest that the cost the well off, who are more likely than the poor to be connected of soap is a barrier to its widespread use among extremely poor to the electricity grid (Alderman 2002). populations and that behavioral change may be difficult to achieve without a subsidy. Medicines and Medical Supplies. In relation to the direct pro- vision of health-related goods, including drugs, supplies, and Insecticide-Treated Bednets. The degree of subsidization of services of medical personnel, governments may subsidize and bednets has become a controversial issue, with some arguing regulate drug prices, make bulk purchases from manufacturers for full subsidization and others for partial subsidization. Most for distribution at reduced prices, and distribute certain drugs long-term studies indicate that consumers resist purchasing with complete or partial subsidies to target populations. bednets even at subsidized prices after they have had access to Specific interventions--for instance, antiretrovirals, vaccines, free bednets (Snow and others 1999). or reproductive health care--are often more heavily subsidized A number of researchers have undertaken studies in various or may be targeted by population group or disease--for exam- locations in Africa to assess the effect of selling bednets rather ple, child and maternal health, tuberculosis, and malaria. With than providing them free to vulnerable populations the exception of antiretroviral drugs, the health benefits and (Armstrong-Schellenberg and others 2001; Kolaczinski and low costs of these medicinal interventions make them good others 2004; Snow and others 1999). The key issue is con- targets for subsidization. sumers' responsiveness to changes in the prices of bednets, through either subsidies or a reduction in taxes and tariffs. General Health Care. In developing countries, where informal Many households do not own a bednet because they cannot sectors tend to be large, providing subsidized health care is an afford it, while other reasons are lack of information, poor important tool for health promotion. Some countries have access to markets, and cultural preferences (Hanson and chosen direct provision of health goods, whereas others com- Worrall 2002; Simon and others 2002). The evidence suggests bine the public provision of services with subsidized health that responsiveness to price changes alone may be modest, but insurance for families below a certain income cutoff. Both in combination with removing some of the other barriers, models require identifying the families that are unable to afford demand for bednets could increase substantially in malaria- health care and the types of services that are considered public affected regions (Simon and others 2002). goods. Nigeria removed tariffs and taxes on bednet insecticide in One example of subsidizing the production and provision 2001, and the 18 percent price drop resulted in an estimated 9 of health care is the Mexican program originally called to 27 percent increase in purchases (Simon and others 2002). PROGRESA and now known as Oportunidades. This program Another study that reviewed a public sector subsidy for bednets is also an example of how income transfers for other goods combined with private sector marketing and distribution by can affect health and how cross-subsidies can be used to means of a social-marketing scheme concluded that the pro- strengthen the incentive effects of a fiscal policy to promote gram was successful because 18 percent of children slept under healthy behavior. The government launched the program in bednets as a result; however, the low insecticide retreatment 1997 to provide subsidized health, nutrition, and education to rate led the authors to conclude that subsidies were needed on poor families. By mid 2004, it was serving the majority of both bednets and insecticide (Armstrong-Schellenberg and those living below the poverty line. Oportunidades combines others 2001). a cash transfer equivalent to 20 to 30 percent of families' incomes that includes incentives for positive behaviors in rela- Clean Cooking Fuels. High rates of respiratory illness occur as tion to health, nutrition, and schooling with subsidized basic a result of exposure to smoke and particle emissions from bio- health interventions. The program is largely financed from mass burning in many developing countries. Fuel subsidy pro- federal budgets. Fiscal Policies for Health Promotion and Disease Prevention | 217 Oportunidades is successful both in terms of targeting the tutions to grant and monitor the tax benefits, and the ability to poorest households and in terms of achieving measurable gains forgo alternative public investments. in health, health care use, nutritional status and growth, school attendance, and school achievement. Gertler (2004), for exam- TAXES AND TAX EXPENDITURES: ple, finds significant and cumulative reduction in illness rates DESIGN AND OUTCOMES among children, lower prevalence of anemia, and an addition- al centimeter of growth in the first year of the program. The following section describes various examples of the use of The program's success is attributable to many factors, taxation directed at both consumers and producers. This sec- including a rigorous longitudinal evaluation process; an inte- tion of the chapter also analyzes the design issues that are grated package of services; and the presence of financial stimuli important in order to guarantee that these instruments con- tied to school attendance, visits to health clinics, and participa- tribute to achieving healthy fiscal policy. tion in health education initiatives. Furthermore, the program incorporates several targeting methods. Taxes on Consumers Sales taxes--including excise taxes and value added taxes-- and exemptions from those taxes are the most common fiscal Producer Subsidies policy tools used to influence consumers' health purchases. Governments use producer subsidies to encourage production Examples are exempting medicines and foods from sales tax that improves health by lowering manufacturers' costs in situa- and imposing an excise tax on cigarettes and alcohol. tions in which the private market supply is inadequate to meet Developed countries often use income tax incentives to provide social needs. Examples include medical supplies, vaccines, food deductions and credits for specific health care purchases. additives, and medical research. Box 11.1 discusses issues surrounding use of taxes for health. Food Fortification. Governments sometimes subsidize the "Sin" Taxes on Tobacco and Alcohol. A wide range of coun- fortification of staple foods through the addition of selected tries and local jurisdictions have taxed tobacco, with acknowl- micronutrients as a way of achieving broadly based nutrition edged success in reducing consumption (P. Jha 1999). The improvements. Challenges involve maintaining a relationship health benefits of curbing the demand for cigarettes may go between the public sector, which initiates and funds the pro- beyond eliminating the health consequences of smoking and gram, and the private sector, which implements the fortifica- secondhand smoke if consumer expenditures are diverted from tion. Incentives for private providers are often needed in the cigarettes to healthier alternatives (for example, food). form of tax exemptions, import preferences, subsidies for start- Taxes on alcohol are widespread and are used primarily to up costs, quality control, and training. Illegal markets selling raise revenue. Governments typically impose taxes at the pro- nonfortified products at a lower price often arise in response ducer, wholesale, and retail levels that are levied as a percentage (Alderman 2002; Dorosh, del Ninno, and Sahn 1996; Rao of the sale price or are based on a flat amount per unit. 2000). Harmful alcohol consumption is controlled through prohibi- tion, government monopolization of sales, "dry" days, restric- Health Research. Government support for health research tions on hours when sales are legal, restrictions on age and consists of the provision of direct subsidies for private sector locations for sales and consumption, laws against drinking and investment, the granting of tax benefits for private research and driving, limits to alcohol content, laws against the sale of cer- development (R&D) investment, the establishment of property tain types of alcohol, and licensing. rights and a system to protect them, and the promotion of pri- Alcohol taxes do contribute revenue to government coffers vate goods by other means (OECD 2003). Despite the strong in developing countries, generally in higher proportions than evidence from developed countries that the private sector will in developed countries (WHO 2002a), but smuggling and tax underinvest in R&D and that tax incentives increase R&D evasion are common. For example, Zimbabwe raised taxes on investment, developing countries should be cautious in apply- certain beers in 1995 but repealed the increase within months ing those results to their own situations. Empirical investiga- when tax revenues dropped significantly (WHO 2002a). Some tions tend to conclude that producer subsidies for R&D in developing countries have lowered alcohol taxes with conse- developing countries are not effective (Shah 1995; Zee, Stotsky, quent negative results. Mauritius experienced a dramatic and Ley 2002). Many conditions need to be in place to realize increase in drunk-driving arrests, alcohol-related fatalities, high social returns and to minimize rent seeking and profiteer- and hospital admissions after it reduced taxes on alcohol ing, including a strong private sector research effort that is (WHO 2002a). In sum, alcohol taxes do reduce drinking, but stimulated by the public investment, the presence of appropri- the evidence that such taxes are well targeted to those most at ate targeting, a transparent and fair set of public laws and insti- risk of problem drinking is not strong. 218 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul Box 11.1 Using Taxes to Influence Consumption and Production Behavior Taxes as a tool for health policy face significant imple- informal or illegal markets. Smuggled or contraband mentation obstacles. First, targeting can be difficult. A products that cannot be regulated or certified for quality close link must exist between the consumption of the and safety, such as alcohol or tobacco in particular, may product or behavior to be taxed and a specific population be more harmful to health than goods that are legally with a health risk. For instance, all consumers would pay a produced and sold. tax on "junk" food, even though it would only present a Any tax should be efficient in terms of both its admin- health threat to a small percentage of them. The taxed istration and its effect on resource allocation. Tax author- good must also be appropriately defined in relation to ities need a well-functioning system for imposing, collect- close substitutes; for example, taxing only certain forms of ing, and monitoring taxes and taxed products, and the tobacco such as cigarettes, but not chewing tobacco, may public should perceive the system as fair and credible in increase consumption of the latter. Governments may also order to achieve a high degree of compliance. distinguish between locally produced goods and imported Finally, a tax should be cost-effective in achieving its goods, often because of lobby groups. If governments stated goal of improving health outcomes. The net costs of place a higher tax on the good that is less harmful, this imposing the tax should compare favorably with the net action will encourage greater consumption of the more costs of using another policy instrument, such as regula- harmful good. tion or direct government provision. Depending on the Key weaknesses in using taxes for health policy include characteristics of the tax base, the health goal and the the feasibility of smuggling and the existence of large revenue goal may even be at odds. Source: Authors. Food Taxes. The issue of taxing unhealthy foods has received in a setting where overnutrition and undernutrition coexist increasing attention in the wake of the Global Strategy on Diet, may have mixed outcomes. Physical Activity, and Health, which was approved by member A natural experiment in Poland during the economic down- countries of the World Health Organization (WHO 2004). The turn of the 1990s suggests a beneficial role for price policy in a global strategy points to the rising prevalence of obesity and consumer switch from animal fats to vegetable fats with lower overweight in developing countries, along with that of nutri- amounts of trans fatty acids (Zatonski, McMichael, and Powles tion-related noncommunicable diseases, and recommends that 1998). A dramatic decline in ischemic heart disease and related countries consider fiscal policies and other measures to reduce circulatory system diseases during the first half of the 1990s is those problems. most easily explained by the removal of consumer subsidies Governments can use excise taxes to reduce the consump- from foods of animal origin, the aggressive marketing of mar- tion of unhealthy foods only if tax rates are sufficient to garines, and a general decline in food purchasing power. The change consumption in a way that improves health outcomes, major change in the food supply appeared to be a reduction in if they tax enough harmful foods or food ingredients, and if foods containing animal fats; however, no direct relationship they levy the taxes in an effective manner. Guo and others' can be conclusively attributed without further study of the (1999) study in China demonstrates significant potential for Polish experience and the experiences of other countries price changes to affect consumption. The researchers studied undergoing similar transitions. dietary intake in a sample of urban and rural Chinese house- Governments may choose to address food-related health holds and show that a 10 percent increase in the price of pork problems by taxing imports of high-fat or high-sugar food; potentially reduces fat consumption by 8 percent. Energy and however, such efforts conflict with rules governing interna- protein intake would both drop by 2 percent. The overall tional trade. Fiji, for example, tried to ban the import of mut- effect may be different for the poor and the rich. The poten- ton flaps, an extremely fatty food that was contributing to the tially harmful effects on the poor of increasing the price of country's obesity problem. To comply with its World Trade pork would be buffered by substitutions from other food Organization obligations, Fiji had to ban the sale of all mutton groups, such as oil, wheat flour, and coarse grains, but con- flaps, not just imports (Evans and others 2001). One analysis cerns remain that overall nutrition would worsen. These suggests that the same kind of broad treatment would be nec- results suggest that using price changes to alter dietary intake essary to grant subsidies to healthy foods, but taxing unhealthy Fiscal Policies for Health Promotion and Disease Prevention | 219 domestic foods alone would probably not pose a problem continue to impose tariffs on imports. For example, Malaysia is under World Trade Organization rules (WHO 2003). Further- a major producer and imposes a 25 percent tax on imports. more, avenues for using other regulatory and economic Brazil used to impose both an import tax and a distribution tax policies to improve the consumption of healthy foods may be that amounted to a total of 45 percent of the original condom acceptable under the World Trade Organization Agreement on price, but it granted a permanent sales tax exemption when Technical Barriers to Trade and the Agreement on Agriculture condom sales increased following a temporary tax holiday. if countries can justify them as contributing to legitimate national health objectives. Taxes on Producers Agricultural policies affect food prices, food choices, and farm incomes in addition to the food security of both rural and Producer taxes are usually aimed at discouraging socially urban populations. Each country must assess the potential for harmful products or processes. They can be imposed either on reorienting its agricultural policies so as to produce a healthier the use of certain inputs, such as more heavily polluting fuels, food supply. Developing countries are generally more likely to or on their outputs, such as emissions of air pollutants. directly subsidize food consumption than food production; Theoretical and simulation models have examined the use however, they frequently make indirect subsidies available of taxes on fuels and emissions taxes to control air pollution through the provision of cheap fuel, chemical inputs, water, (World Bank 1994a, 1994b), but empirical data are lacking. and loans to the agriculture sector. These policies may be envi- Models of taxes suggest potential to induce substitution by ronmentally and fiscally costly and rarely contribute to cleaner fuels and reductions in overall energy use, but actual improved population health. results will depend on the availability of fuel substitutes within Research is needed on individual countries' agricultural countries. Data on Chilean manufacturing support the possi- policies and food supply needs to make them more compatible bility of clean fuels substitution but indicate the likelihood of (Nugent 2004). At the same time, the dynamics of food choice uneven sectoral incidence of the emissions tax. For example, and the effects of price manipulation need to be better under- bakeries were responsive to changes in relative prices, whereas stood before tax and subsidy systems can be designed to effec- metal products plants were unresponsive, and meat packers tively promote healthy food choices. were unable to adjust their electricity demand but could reduce energy from other sources (World Bank 1994b). If this poten- Sales Tax Exemptions on Healthy and Staple Foods and tial were realized on a global or regional basis--for example, Medicines and Other Health Care Goods. Governments may through agreements to the Kyoto Protocol--a double benefit of set tax policies to ensure that certain expenditures on health- reducing harmful externalities and raising significant revenues related behaviors and health goods are tax deductible or tax might be achieved. exempt for firms, employers, or individuals. Exemptions should apply to a limited number of goods that are easily dif- ferentiated from goods that are not exempted. Note that in FISCAL POLICY TO PROMOTE HEALTH countries with large informal sectors, income tax systems are The fiscal policies discussed in this chapter in relation to health weak, and fiscal policies for the deductibility of credits are and health care goods can be applied to other goods and mar- unlikely to be effective. kets, such as housing and education, some of which may have South Africa provided value added tax exemptions for a important effects on health. This chapter does not provide an short list of essential foods and found a varied consumption exhaustive discussion of the goods that indirectly promote pattern by commodity, with the poor receiving most of the health, but it does briefly consider some of these policies in benefits of the maize exemption, but few of the benefits of the relation to workplaces, employment leave policies, and day milk exemption (Alderman and del Ninno 1999). care. Note that policies focused on formal labor markets will Mexico imposes a 15 percent value added tax on almost all not be effective in reaching large segments of the population in goods. Exemptions include medicines, physician's services, and many countries. Policies that provide health-related services, some foods. Recently, a government proposal to make drug and such as day care, that are not based on formal labor market food purchases eligible for value added tax and to channel the participation may have a broader effect. resulting revenues into financing programs targeted to the poor has given rise to extensive debate. Those in favor have argued that the existing subsidy is regressive because most drug Workplace Health and food purchases are by the wealthy (Fundación Mexicana Governments can use tax relief and financial support to pro- para la Salud 2001). ducers to encourage firm-specific actions to promote health. Many developing countries concerned about the spread of Many countries mandate safeguards in the workplace and HIV/AIDS have dropped import taxes on condoms, but others levy penalties against occupational health violations. Most 220 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul government actions are mandates rather than fiscal policies, may have to use low-quality care or leave children unattended. but a combination of approaches may also be used. Van der Gaag and Tan (1997) argue for public subsidies based A growing area for workplace health promotion is on cost-benefit analysis of early childhood development pro- HIV/AIDS. Bloom and others (2004) suggest that the failure on grams. They conclude that the greatest payoff comes from tar- the part of most firms to act--even if they correctly perceive geting the most deprived families and that the private benefits the business, human, and social challenges HIV/AIDS poses-- are sufficient to expect better-off parents to pay. is attributable to a lack of incentives. Significant externalities Two large-scale, home-based day care programs targeted to (benefits to society and firms) are likely to result from promot- poor families are Community Well-Being Homes (Hogares ing greater action by firms. Some private firms have begun Comunitarios de Bienestar) run by the Colombian Institute for providing HIV/AIDS prevention and treatment services to Family Well-Being (Instituto Colombiano de Bienestar employees, families, and their communities ("Face Value: AIDS Familiar) in Colombia (Myers 1995) and the Integrated Child and Business" 2004). Sometimes government support is Development Program in Bolivia. The former is an interesting involved, but little information is available to evaluate the case of a targeted cross-subsidy because the financing comes potential of fiscal policy. from the wealthier formal sector by means of a payroll tax, whereas the services are targeted to the poorest families. The program was 85 percent subsidized in the early 1990s. Parents Maternity Leave, Sick Leave, and Family Care Leave paid a proportion of the caregivers' wages on a sliding-scale Government policy can alter choices regarding different types user fee (Young 1996). The Bolivian program includes nutri- of worker leave. Many countries have financial or legislative tion, health, and cognitive development interventions and is support for caregiving, although most focus on children. one of the few early childhood programs in developing coun- Rhum (1998) cites evidence that more than 100 countries-- tries that has been formally evaluated (Behrman, Cheng, and and almost all the industrial countries--have some legislation Todd 2000). The evaluation shows that the program signifi- about parental leave, although in several countries it is unpaid. cantly increases cognitive achievement, although the results Caregiving policies that allow people to take time off work depend on age and the duration of exposure to the program. to care for aged and chronically ill family members are less common than policies for child care, particularly in devel- oping countries, but tax benefits and allowances for these types CONCLUSIONS of caregiving are becoming increasingly available in the indus- trial countries (Brodsky, Habib, and Mizrahi 2000; Pijl 2003; A broad range of experiences cited in this chapter demonstrates Wiener 2003). Although the provision of services in kind by the that fiscal and health policies interact in a number of areas. government is still an important mechanism, the trend is Although substantial research has focused on tobacco and alco- toward empowering consumers by offering subsidies or tax hol, other links--for example the promotion of health in the deductions that allow them to choose among caregiving workplace--have been much less recognized or studied, partic- options. Countries tend to use a combined approach to financ- ularly in developing countries. ing that relies on payroll taxes imposed on employees and The research presented in this chapter suggests that fiscal employers, general taxation, and copayments. Important issues policy can be a useful tool for influencing health in developing that developing countries need to address in this respect countries. Nevertheless, budgetary limitations to withstand include targeting compared with universal provision, the pressure for program expansion, leakages to unintended bene- mechanisms to pay for or to insure care, and the extent to ficiaries, public compliance with the tax system, and corruption which long-term care should be integrated into the health care among both government officials and the public are important and social service systems (Brodsky, Habib, and Mizrahi 2000; factors to take into account in design and implementation. WHO 2003). Table 11.4 summarizes some lessons learned on the use of fis- cal policy to promote health. Governments may find it worthwhile to examine their use Day Care and Early Childhood Education of fiscal policies to identify the entire range of effects and have Some countries use targeted fiscal policies, such as income tax health ministries participate in this exercise. More generally, the deductions or direct provision, to increase the use and quality chapter indicates an area for increased interaction between of early childhood education and child care services. Important ministries of health and finance. Healthy fiscal policy and fiscal health, labor market efficiency, growth, and equity arguments policy for health should be topics that are debated, agreed on, support subsidizing these services, particularly for low-income and formalized between the two areas of policy making to families, because without subsidies women may be forced to guarantee that those developing fiscal policy take both its eco- limit their work or to leave the labor market, and families nomic and its health implications into account. Fiscal Policies for Health Promotion and Disease Prevention | 221 Table 11.4 Lessons in Using Fiscal Policy for Health Promotion Intervention choice Program design Instrument design Policy regime · Select interventions that directly · Ensure that the health benefits of · Choose the appropriate recipients · Ensure that policy is consistent address the health objective. the desired change are apparent for a subsidy or tax preference. and predictable. · Ensure that interventions are and significant. · Do not spread the benefits across · Ensure that institutions carrying sufficient to effect the health · Make sure that the tax base is too large a group. out a policy are open, account- change, but not excessive. adequate and stable and that no · Note that targeting by demo- able, and uncorrupted. · Choose interventions with high untaxed close substitutes are graphic, geographic, or need cate- · Consider tradeoffs between effi- health returns and low costs available. gories is more efficient than no ciency and distributional goals. relative to alternatives. · Be aware that a large informal targeting or self-targeting. · Seek non­health sector opportuni- labor sector will limit the · Be aware of the price elasticities ties to effect health goals. effectiveness and equity of of a taxed good so that its benefit delivery. incidence is clear. · Avoid programs whose expenses may become unsustainable because of uncontrollable factors. Source: Authors. Rigorous evaluation studies are needed of most of the fiscal REFERENCES policy interventions discussed in this chapter. Such studies Adams, R. 2000. "Self-Targeted Subsidies: The Political and Distributional should address the health, fiscal, macroeconomic, and distrib- Impact of the Egyptian Food Subsidy System." Economic Development utional effects of using fiscal policy to achieve health goals and and Cultural Change 49 (1): 115­36. should be performed in a range of countries with mixed pub- Alderman, H. C. 2002. Price and Tax Subsidization of Consumer Goods. lic and private sector capacity to deliver health services. The Social Safety Net Primer Series. Washington, DC: World Bank Institute. studies should also examine the differing effects of policies in Alderman, H. C., and C. del Ninno. 1999. "Poverty Issues for Zero Rating VAT in South Africa." Journal of African Economies 8 (2): 182­208. urban and rural settings and across income quintiles. Of par- Armstrong-Schellenberg, J., S. Abdulla, R. Nathan, O. Mukasa, T. J. ticularly high priority are further studies of the results of sub- Marchant, N. Kikumbih, and others. 2001."Effect of Large-Scale Social sidizing drugs, medical supplies, and hygiene interventions Marketing of Insecticide-Treated Nets on Child Survival in Rural with or without education campaigns. Those areas may reveal Tanzania." Lancet 357 (9264): 1241­47. new fiscal approaches for addressing the disease burden in Behrman, J. R., Y. Cheng, and P. Todd. 2000. The Impact of the Bolivian Integrated "PIDI" Preschool Program. Philadelphia: University of developing countries. Pennsylvania. Bloom, D., L. R. Bloom, D. Steven, and M. Weston. 2004. "Business and HIV/AIDS: Who Me? A Global Review of the Business Response to DISCLAIMER AND ACKNOWLEDGMENT HIV/AIDS." Paper prepared for the World Economic Forum's Global Health Initiative in partnership with the Harvard School of Public The results and conclusions in this chapter are those of the Health and Joint United Nations Programme on HIV/AIDS, World authors and do not necessarily reflect the opinions of the insti- Economic Forum, Geneva. tutions for which they work. The authors are grateful to Ana Borghi, J., L. Guinness, J. Ouedraogo, and V. Curtis. 2002. "Is Hygiene Mylena Aguilar, Hector Arreola, Ania Burczysnka, Esperanza Promotion Cost-Effective? A Case Study in Burkina Faso." Tropical Medicine and International Health 7 (11): 960­69. Calleja, Marissa Courey, Monica Hurtado, Vanesa Leyva, Brodsky, J., J. Habib, and I. Mizrahi. 2000. Long-Term Care in Five Eugenia Rocha, Swathi Sista, and Sinaia Urrusti for excellent Developed Countries. Geneva: World Health Organization. research assistance and to the Mexican Health Foundation's Cohen, D., R. Scribner, R. Bedimo, and T. Farley. 1999. "Cost as a Barrier Council on Health and Competitiveness and the National to Condom Use: The Evidence for Condom Subsidies in the United Council for Science and Technology of Mexico for support for States." American Journal of Public Health 89 (4): 567­68. Cook, E., and M. Vlaisavljevich. 1994. "Implications of Health Reform for this research. State and Local Fiscal Policy." National Tax Journal 47 (3): 639­54. del Ninno, C., and P. Dorosh. 2002. "In-Kind Transfers and Household NOTES Food Consumption: Implications for Targeted Food Programs in Bangladesh." Discussion Paper 134, International Food Policy Research Institute, Washington, DC. 1. The idea of healthy fiscal policy is discussed in Cook and Vlaisavljevich (1994), Joffe and Mindell (2004), and Secretaría de Salud Dorosh, P., C. del Ninno, and D. Sahn. 1996. "Market Liberalization and (2001). the Role of Food Aid in Mozambique." In Economic Reform and the 2. Social marketing is defined as the use of marketing principles to Poor in Africa, ed. D. Sahn. Oxford, U.K.: Clarendon Press. influence behavior for a socially desirable outcome. It provides a desirable Evans, M., R. Sinclair, C. Fusimalohi, and V. Liav'a. 2001. "Globalization, product at an affordable price with adequate promotion and placement Diet, and Health: An Example from Tonga." Bulletin of the World (that is, access). Health Organization 79 (9): 856­62. 222 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul "Face Value: AIDS and Business." 2004. Economist 373 (8404): 68. Water Treatment and Safe Storage: A Promising New Strategy." Fundación Mexicana para la Salud. 2001. "Política Fiscal Saludable: Epidemiology Infection 122 (1): 83­90. Propuestas a Considerar sobre la Reforma Fiscal." Unpublished paper, Rao, V. 2000. "Price Heterogeneity and Real Inequality: A Case Study of Fundación Mexicana para la Salud, México, D.F. Poverty and Prices in Rural South India." Review of Income and Wealth Gertler, P. 2004. "Do Conditional Cash Transfers Improve Child Health? 46 (2): 201­11. Evidence from PROGRESA's Control Randomized Experiment." Rhum, C. 1998. "Parental Leave and Child Health." NBER Working Paper American Economic Review 94 (2): 336­41. 6554, National Bureau of Economic Research, Cambridge, MA. Guo, X., B. Popkin, T. Mroz, and F. Zhai. 1999. "Food Price Policy Can Secretaría de Salud. 2001. "Programa Nacional de Salud: 2001­2006-- Favorably Alter Macronutrient Intake in China." Journal of Nutrition La democratización de la salud en México: Hacia un sistema universal 129 (5): 994­1001. de salud." Secretaría de Salud, México, D.F. Hanson, K., and E. Worrall. 2002. Report on the Analysis of SMITN2 End- Shah, A., ed. 1995. Fiscal Incentives for Investment and Innovation. New of-Project Survey. London: London School of Hygiene and Tropical York: Oxford University Press. Medicine. Siamwalla, A. 1988. "Some Macroeconomic Policy Implications of Hoque, B. 2003. "Handwashing Practices and Challenges in Bangladesh." Consumer-Oriented Food Subsidies." In Food Subsidies in Developing International Journal of Environmental Health Research 13 (Suppl. 1): Countries: Costs, Benefits and Policy Options, ed. P. Anderson. S81­87. Baltimore: Johns Hopkins University Press, 323­30. Jha, P. 1999. Curbing the Epidemic: Governments and the Economics of Simon, J., B. Larson, A. Zusman, and S. Rosen. 2002. "How Will the Tobacco Control. Washington, DC: World Bank. Reduction of Tariffs and Taxes on Insecticide-Treated Bednets Affect Jha, S. 1992. "Consumer Subsidies in India: Is Targeting Effective?" Household Purchases?" Bulletin of the World Health Organization Development and Change 23 (4): 101­28. 80 (11): 892­99. Joffe, M., and Mindell, J. 2004. "A Tentative Step towards Healthy Public Snow, R. W., E. McCabe, D. Mbogo, C. Molyneux, V. Mung'ala, and Policy." Journal of Epidemiology and Community Health 58: 966­68. C. Nevill. 1999. "The Effect of Delivery Mechanisms on the Uptake of Bed Net Re-impregnation in Kilifi District, Kenya." Health Policy and Kolaczinski, J., N. Muhammad, Q. Khan, Z. Jan, N. Rehman, T. Leslie, and Planning 14 (1): 18­25. others. 2004. "Subsidized Sales of Insecticide-Treated Nets in Afghan Refugee Camps Demonstrate the Feasibility of a Transition from Tabor, R., and M. Sawit. 2001. "Social Protection via Rice: The OPK Rice Humanitarian Aid Towards Sustainability." Malaria Journal 3 (15): Subsidy Program in Indonesia." Developing Economies 39 (3): 267­94. 1­11. UNDP (United Nations Development Programme). 2003. World Bank Luby, S. P., M. Agboatwalla, A. Raza, J. Sobel, E. Mintz, K. Baier, and Energy Sector Management Assistance Program--India: Access of the others. 2001. "Microbiologic Effectiveness of Hand Washing with Soap Poor to Clean Household Fuels. New York: UNDP. http://www. in an Urban Squatter Settlement, Karachi, Pakistan." Epidemiological worldbank.org/esmap/. Infection 127 (2): 237­44. Van der Gaag, J., and J. P. Tan. 1997. The Benefits of Early Child Musgrove, P. 1993. "Feeding Latin America's Children." World Bank Development Programs: An Economic Analysis. Washington, DC: World Research Observer 8 (1): 23­45. Bank. WHO (World Health Organization). 2002. Alcohol in Developing Societies: Myers, R. 1995. Early Childhood Care and Development Programs in Latin A Public Health Approach. Geneva: WHO. America and the Caribbean: A Review of Experience. Washington, DC: Inter-American Development Bank. http://www.ecdgroup.com/ ------. 2003. Using Domestic Law in the Fight against Obesity: An download/gw1eccdl.pdf. Introductory Guide for the Pacific. Geneva: WHO. Nugent, R. 2004. "Food and Agricultural Policy in the Prevention of ------. 2004. Strategic Plan for Diet, Nutrition, and Physical Activity. Non-communicable Diseases." Food and Nutrition Bulletin 25 (2): Geneva: World Health Assembly. 200­8. Wiener, J. 2003. "The Role of Informal Support in Long-Term Care." In OECD (Organisation for Economic Co-operation and Development). Key Policy Issues in Long-Term Care. World Health Organization 2003. Tax Incentives for Research and Development: Trends and Issues. Collection on Long-Term Care, ed. J. Brodsky, J. Habib, M. J. Hirschfeld. Paris: Science Technology Industry. Geneva: World Health Organization. Pijl, M. 2003. "The Support of Carers and Their Organizations in Some World Bank. 1994a. Energy Pricing and Air Pollution: Econometric Evidence Northern and Western European Countries." In Key Policy Issues in from Manufacturing in Chile and Indonesia. Washington, DC: World Long-Term Care: World Health Organization Collection on Long-Term Bank. Care, ed. J. Brodsky, J. Habib, M. J. Hirschfeld. Geneva: World Health ------. 1994b. How Relative Prices Affect Fuel Use Patterns in Organization. Manufacturing: Plant-Level Experience from Chile. Washington, DC: Pinstrup-Anderson, P. 1988. Food Subsidies in Developing Countries: Costs, World Bank. Benefits, and Policy Options. Washington, DC: International Food Yaser, Y. 1993. "Extensive Advertising." Integration 32: 32­33. Policy Research Institute; Baltimore, MD: Johns Hopkins University Young, M. E. 1996. Desarrollo del Niño en la Primera Infancia: Una Press. Inversión en el Futuro. Washington, DC: World Bank, Human Resour- Price, N. 2001. "The Performance of Social Marketing in Reaching the ces Department. http://web.worldbank.org/WBSITE/EXTERNAL/ Poor and Vulnerable in AIDS Control Programs." Health Policy and TOPICS/EXTEDUCATION/EXTECD/. Planning 16 (3): 231­39. Zatonski, W. A., A. J. McMichael, and J. W. Powles. 1998. "Ecological Study Quick, R., A. Kimura, A. Thevos, M. Tembo, I. Shamputa, L. Hutwanger, of Reasons for Sharp Decline in Mortality from Ischaemic Heart and E. Mintz. 2002. "Diarrhea Prevention through Household-Level Disease in Poland since 1991." British Medical Journal 316 (7137): Water Disinfection and Safe Storage in Zambia." American Journal of 1047­51. Tropical Medicine 66 (5): 584­89. Zee, H., J. Stotsky, and E. Ley. 2002. "Tax Incentives for Business Quick, R., L. Venczel, E. Mintz, L. Soleto, J. Aparicio, M. Gironaz, and Investment: A Primer for Policymakers in Developing Countries." others. 1999. "Diarrhea Prevention in Bolivia through Point-of-Use World Development 30 (9): 1497­516. Fiscal Policies for Health Promotion and Disease Prevention | 223 Chapter 12 Financing Health Systems in the 21st Century George Schieber, Cristian Baeza, Daniel Kress, and Margaret Maier This chapter assesses health financing policy in low- and cost-effective health services to figuring out how to finance and middle-income countries (LMICs). It discusses the basic func- deliver those services equitably and efficiently, to recognizing tions of health financing systems and the various mechanisms the need to scale up health systems to meet basic service needs for effective revenue collection, pooling of resources, and pur- and achieve the MDGs, which will require large amounts of chase of interventions (WHO 2000). It analyzes the basic DAH for poor countries (see, for example, WHO 2000, 2001; financing challenges facing LMICs as a result of revenue gener- World Bank 1993, 2004b). ation and collection constraints, increasing flows of develop- This chapter updates and reviews the global evidence on ment assistance for health (DAH) coupled with donors' health spending, health needs, revenue-raising capacity, organi- concerns about aid effectiveness, and the difficult economic zation of health financing, and trends in DAH. It discusses situation facing many LMICs as a result of globalization and the key challenges that country policy makers face in ensuring poor economic management. access to services and financial protection while dealing with a In 2001, about US$3.059 trillion--approximately 9 percent new health policy world defined by new instruments such as sec- of global gross domestic product (GDP)--was spent on health torwide approaches (SWAps) and Poverty Reduction Strategy care worldwide (WHO 2004b; World Bank 2004e); however, Papers (PRSPs). The chapter also discusses the scope and poten- only 12 percent of this amount was spent in LMICs, which tial effects of new and relatively large global funding sources, account for 84 percent of the global population and 92 percent such as the Bill & Melinda Gates Foundation; the Global Fund to of the global disease burden (Mathers and others 2002). Fight AIDS, Tuberculosis, and Malaria; and the Global Alliance Ongoing epidemiological, demographic, and nutrition transi- for Vaccines and Immunization (GAVI) Vaccine Fund. tions will pose significant challenges for health financing sys- tems in LMICs in the near future as the communicable disease burden lessens and the noncommunicable disease and injury HEALTH FINANCING SYSTEMS burdens expand. At the same time, the current communicable Health financing provides the resources and economic incen- disease burden in low-income countries (LICs) and in many tives for the operation of health systems and is a key deter- middle-income countries (MICs), especially that caused by minant of health system performance in terms of equity, malaria, tuberculosis, and HIV/AIDS, poses a serious threat to efficiency, and health outcomes. public health, health systems, and economic growth. As a result of the international focus on poverty reduction, the HIV/AIDS pandemic, and the Millennium Development Health Financing Functions Goals (MDGs), international health financing policy has Health financing involves the basic functions of revenue collec- evolved over the past decade from defining a basic package of tion, pooling of resources, and purchase of interventions. 225 Revenue Pooling Resource allocation Service collection and purchasing (RAP) provision Government Taxes agency Public charges/ resource sales Social insurance Public or sickness funds Mandates Public providers Private insurance Grants organizations Private providers Loans Employers Private insurance Private Individuals Communities and households Out of pocket Source: Authors. Figure 12.1 Interactions among Revenue Raising, Risk Pooling, Resource Allocation, and Service Provision · Revenue collection is how health systems raise money from · managing these revenues to equitably and efficiently pool households, businesses, and external sources. health risks; and, · Pooling deals with the accumulation and management of · ensuring the purchase of health services in an allocatively revenues so that members of the pool share collective health and technically efficient manner. risks, thereby protecting individual pool members from large, unpredictable health expenditures. Prepayment allows These financing functions are generally embodied in the fol- pool members to pay for average expected costs in advance, lowing three stylized health financing models: relieves them of uncertainty, and ensures compensation should a loss occur. Pooling coupled with prepayment · national health service (NHS): compulsory universal cover- enables the establishment of insurance and the redistribu- age, national general revenue financing, and national own- tion of health spending between high- and low-risk individ- ership of health sector inputs uals and high- and low-income individuals. · social insurance: compulsory universal coverage under a · Purchasing refers to the mechanisms used to purchase serv- social security (publicly mandated) system financed by ices from public and private providers. Figure 12.1 illus- employee and employer contributions to nonprofit insurance trates these functions and their interactions. funds with public and private ownership of sector inputs · private insurance: employer-based or individual purchase of In terms of health policy at the country level, these three private health insurance and private ownership of health financing functions translate into the following: sector inputs. · raising sufficient and sustainable revenues in an efficient Although these models provide a general framework for and equitable manner to provide individuals with both a classifying health systems and financing functions, they are not basic package of essential services and financial protection useful from a micropolicy perspective because all health sys- against unpredictable catastrophic financial losses caused by tems embody features of the different models. The key health illness or injury. policy issues are not whether a government uses general 226 | Disease Control Priorities in Developing Countries | George Schieber, Cristian Baeza, Daniel Kress, and others revenues or payroll taxes, but the amounts of revenues raised Nations Millennium Project (2005) and the Commission for and the extent to which they are raised in an efficient, equitable, Africa (2005). and sustainable manner. Similarly, nothing is intrinsically good Others have argued, however, that absent resources to fund or bad about public versus private ownership and provision. drug purchases, provide facilities with some discretionary The important issue is whether the systems in place ensure funding, and motivate providers, use of primary health care by access, equity, and efficiency. the poor will remain low because of both poor quality and lack of drugs, and the poor will purchase these essential services on the private market. The Bamako Initiative shows that user fees Revenue Collection may be an important revenue source where institutions are Governments use a variety of financial and nonfinancial mech- weak, resources are limited, and the choice is between having anisms to carry out their functions, including directly providing drugs or not having them (World Bank 2003, 76­77). services; financing, regulating, and mandating service provi- Furthermore, studies indicate that user fees can improve bene- sion; and providing information (Musgrove 1996). A substan- fit incidence if user fee and waiver policies have been well tial literature is devoted to the various sources for financing designed and implemented and if providers are compensated health services and the economic and institutional effects of for forgone revenues. Indeed, proponents of user fees argue using these sources in terms of efficiency, equity, revenue- that as long as the fees are set below private market levels, this raising potential, revenue administration, and sustainability "savings" may result in a net reduction in overall out-of-pocket (Schieber 1997; Tait 2001; Tanzi and Zee 2000; WHO 2004b; spending for the poor (Bitrán and Giedion 2003). These diverse World Bank 1993). An additional source of revenue receiving experiences demonstrate the difficultly involved in making increasing attention is efficiency gains (Hensher 2001). LICs blanket statements regarding user fees. As the World Bank rarely use tax credits as a financing source (Tanzi and Zee 2000). (2003, 71) points out, "user fees, as with other public policy The key fiscal issue for LMICs is for their financing systems, decisions, must balance protection of the poor, efficiency in both public and private, to mobilize enough resources to allocation, and the ability to guarantee that services can be finance expenditures for basic public and personal health serv- implemented and sustained." ices without resorting to excessive public sector borrowing (and creation of excessive external debt); to raise revenues Risk Pooling and Financial Protection equitably and efficiently; and to conform with international Preventing individuals from falling into poverty because of cat- standards (Tanzi and Zee 2000). Institutional constraints are astrophic medical expenses and protecting and improving the particularly important, including a country's economic health status of individuals and populations by ensuring finan- structure--for example, large rural populations and limited cial access to essential public and personal health services pro- formal sector employment; ineffective tax administration; and vide a strong basis for public intervention in financing health lack of data, all of which tend to preclude LMICs from using systems. Public intervention may be needed because of market the most efficient and equitable revenue-raising instruments failures in private financing and provision (for instance, infor- (Schieber and Maeda 1997; Tait 2001). The high level of mation asymmetries) and instabilities in insurance markets inequality in most LMICs means that governments face the dif- (such as favorable risk selection by insurers and moral hazard). ficult situation of needing to tax the politically powerful and Indeed, in virtually all Organisation for Economic Co- wealthy elites to raise significant revenues in an equitable operation and Development (OECD) countries except the manner but of being unable to do so easily. As Tanzi and Zee United States, governments have decided to publicly finance or (2000, 4) point out, "tax policy is often the art of the possible require private financing of the bulk of health services. rather than the pursuit of the optimal." However, given both low income levels and limits on possibili- Another area of health financing that continues to generate ties for domestic resource mobilization in LICs and some heated debate is user fees (that is, charges to individuals for MICs, these countries face severe challenges in publicly financ- publicly provided services). The need to significantly scale up ing essential public and personal health services. They also resources to meet the MDGs in LICs has pushed the user fee often confront difficult tradeoffs with respect to financing issue to the forefront of this debate. Arhin-Tenkorang (2000) these basic essential services and providing financial protection and Palmer and others (2004) suggest that the overall effect is against the costs of catastrophic illness. negative: use decreases, particularly among the poor, and fre- quently, administrative costs of collecting the fees are higher Ensuring Financial Protection. Ensuring financial protection than the revenue generated. Further, Kivumbi and Kintu means that no household spends so much on health that it falls (2002) suggest that granting waivers and exemptions for the into and cannot overcome poverty (ILO and STEP 2002b). poor is difficult, if not impossible. Given those findings, many Achieving adequate levels of financial protection requires max- have called for the abolition of user fees, including the United imizing prepayment for insurable health risks; achieving the Financing Health Systems in the 21st Century | 227 largest possible pooling of health risks within a population, At least four alternative organizational arrangements exist thereby facilitating redistribution among high- and low-risk for risk pooling and prepayment: ministries of health (MOHs) individuals; ensuring equity through prepayment mechanisms or NHSs, social security organizations (SSOs), voluntary pri- that redistribute costs from low- to high-income individuals; vate health insurance, and community-based health insurance and developing purchasing arrangements that promote effi- (CBHI). Each of these is linked to distinctive instruments for cient delivery of good-quality services. revenue collection (for example, general revenues, payroll Meeting those requirements depends on how health systems taxes, risk-rated premiums, and voluntary contributions) and arrange the three key health financing functions of revenue col- for purchase of health services. lection, risk pooling, and purchasing. Although all health Within these organizational structures, three alternatives financing functions play an important role in ensuring finan- often coexist for generating revenues and financing equity sub- cial protection, risk pooling and prepayment--whether sidies: subsidies within a risk pool, subsidies across different through taxes or individual premiums--play the central and risk pools, and direct public subsidies through transfers from often the most poorly understood roles. the government. Although medical savings accounts (with or Risk pooling refers to the collection and management of without public subsidization) are also sometimes referred to as financial resources so that large individual and unpredictable a risk pooling mechanism, they do not pool risks over groups financial risks become predictable and are distributed among all and, therefore, are far more limited in terms of predictability members of the pool. The pooling of financial risks is at the core and equity subsidization. They are simply intertemporal mech- of traditional insurance mechanisms. Whereas pooling ensures anisms for smoothing health risks over an individual's or predictability and the potential for redistribution across individ- household's life cycle. ual health risk categories, prepayment provides various options Subsidies within a risk pool, whether financed through gen- for financing those risks equitably and efficiently across high- eral revenues or payroll taxes, are prerequisites for pooling risks and low-income pool members. The health financing models in traditional NHSs and SSOs. The goal of collecting revenues described earlier embody different means for creating risk pools through an income-related or general revenue­based contribu- and financing such pools through prepaid contributions. tion (in contrast to a risk-related contribution, as is generally In most LMICs, multiple public and limited private the case with private insurance) is to generate subsidies from arrangements coexist, making system fragmentation the norm high- to low-income individuals. These systems are effective rather than the exception. This situation increases administra- when payroll contributions are feasible, when the general rev- tive costs; creates potential equity and risk selection problems, enue base is sufficient and a large proportion of the population for example, when the wealthy are all in one pool; and limits participates in the same risk pool, or when both conditions pool sizes. Moreover, health care risks change over the life cycle exist. Moreover, in a system with multiple, competing, public of an individual or household, but because generally little cor- and private insurers and a fragmented risk pool, payroll contri- relation exists between life cycle needs and capacity to pay, sub- butions may increase incentives for risk selection. In the case of sidies are often necessary and are facilitated by risk pooling. a NHS or SSO, financial resources might be insufficient or Risk pooling and prepayment functions are central to the inappropriate for spreading the financial risks or for creating creation of cross-subsidies between high-risk and low-risk an equity subsidy, particularly if the general revenue or payroll (that is, a risk subsidy) and rich and poor (that is, an equity sub- contribution base is regressive. sidy) individuals. The larger the pool, the greater the potential Subsidies across different risk pools involve the creation of for spreading risks and the greater the accuracy in predicting funds, often called solidarity or equalization funds, financed by average and total pool costs. Placing all participants in a single a portion of contributions to each risk pool. This mechanism pool and requiring contributions according to capacity to pay is found in systems with multiple insurers in, for example, rather than individual or average pool risk facilitates cross- Argentina, Colombia, Germany, and the Netherlands. A key subsidization and, depending on the level of pooled resources, element of this mechanism's success is the implementation of can significantly increase financial protection. adequate systems of compensation among different risk and However, spreading risks through insurance schemes is not income groups. enough to ensure financial protection, because it can result in Finally, in many OECD countries, direct public transfers low-risk, low-income individuals subsidizing high-income, funded through general taxation are made to insurers for sub- high-risk individuals. Furthermore, significant portions of the sidizing health care for certain groups or for the entire popula- population may not be able to afford insurance. For this reason, tion. They are also used in some LMICs, although at a limited most health care systems aim not only at spreading risk, but level because of low revenue collection capacity. also at ensuring equity in financing of health care services In most LMICs, risk pool fragmentation significantly through subsidies from high- to low-income individuals. impedes effective risk pooling, while limited revenue-raising Equity subsidies are the result of such redistribution policies. capacity precludes the use of broad public subsidies as the main 228 | Disease Control Priorities in Developing Countries | George Schieber, Cristian Baeza, Daniel Kress, and others source of finance. Therefore, targeting scarce public subsidies that follows on the need for additional funds from external across different risk pooling schemes is probably the most financing sources. feasible way to finance equity subsidies for the poor and those As also discussed by Hecht and Shah in this book outside formal pooling arrangements. However, this method (chapter 13), external funds--development assistance for has important transaction costs. Because a significant portion health--have become an increasingly importance source of of the population is excluded from the formal sector, using this health financing in LICs, supporting some 20 percent of LIC mechanism for ensuring universal financial protection is spending. Specifically, DAH from governments, multilateral limited, particularly in LICs. Even if significant subsidies are and bilateral agencies, and private foundations increased from available from general taxation, the lack of insurance portabil- an average of US$6.7 billion between 1997 and 1999 to US$9.3 ity restricts its usefulness as a subsidization mechanism among billion in 2002. Sub-Saharan Africa received 36 percent of DAH risk pools because individuals may lose their coverage when funds in 2002, and in 13 extremely poor countries, DAH they change jobs. LICs and certain MICs will be challenged accounted for more than 30 percent of health spending (WHO both to publicly finance essential public and personal health 2004b). services and to ensure financial protection through equity sub- The relationship between health expenditures and health sidies. Thus, LMICs should strive to achieve the best value for outcomes is not always clear. Higher spending does not neces- publicly financed health services in terms of health outcomes sarily translate to better health outcomes. Although the evi- and equity and should try to facilitate effective risk pooling for dence tenuously demonstrates a positive relationship between privately financed services. Providing public financing for cost- public spending on health and selected health indicators, it falls effective interventions is one critical aspect of determining far short of a definitive statement (Bidani and Ravallion 1997; which services to finance publicly. Filmer and Pritchett 1999; Gupta, Verhoeven, and Tiongson 2001; World Bank 1993, 2003). Health outcomes also vary Distributing and Sourcing Health Expenditures. As table 12.1 across income groups, with the poor generally receiving fewer shows, health spending is derived from three broad sources: services and having worse health outcomes. As in the case of public sector (expenditures financed out of general revenues health services and health outcomes, health spending is often and social insurance contributions), private sector (expendi- not pro-poor (Gwatkin and others 2003). The quality of a tures financed out of pocket and by private insurance), and country's institutions also plays a key role in determining the external sources (grants or loans from international funding effectiveness of health spending (Devarajan, Swaroop, and agencies). In 2001, high-income countries spent an average of Heng-Fu 1996; Rajkumar and Swaroop 2002; Wagstaff and 7.7 percent of their GDP on health (country weighted), MICs Claeson 2004; World Bank 1993). spent 5.8 percent, and LICs spent 4.7 percent. Even though a clear upward trend between a country's Mobilizing Government Revenues. Governments of LICs income level and the level of public and total health spending is have recognized the need for greater domestic investments apparent in terms of both absolute spending and share of GDP, in health. In the 2001 Abuja Declaration on HIV/AIDS, spending for any given income level varies a great deal, particu- Tuberculosis, and Other Related Infectious Diseases, African larly at lower income levels (Musgrove, Zeramdini, and Carrin leaders pledged to increase health spending to 15 percent of 2002). The composition of health spending also exhibits major their government's budgets (Haines and Cassels 2004; UNECA differences.As incomes increase, both private and out-of-pocket 2001). Yet LICs' ability to raise enough revenue to meet needs shares of total health spending decrease. In LICs, private and and demands for publicly financed health services is highly out-of-pocket spending and external assistance account for the constrained (Gupta and others 2004; Schieber and Maeda bulk of all health spending. As countries move up the income 1997). Even though revenue mobilization is directly correlated scale, public spending predominates and both out-of-pocket with income, wide cross-country variation in revenue mobi- spending and external assistance decrease drastically. lization within income groups is apparent. For example, LMICs with high levels of out-of-pocket spending have Myanmar's tax revenues amounted to only 4 percent of its limited opportunities for risk pooling, which hinders alloca- GDP, whereas Lesotho's were 36 percent (WHO 2002). tive efficiency and financial protection efforts.1 Moreover, low As table 12.2 shows, during the early 2000s, LICs collected overall spending levels in many LICs and some MICs result in the equivalent of about 18 percent of their GDP as revenues, limited access to essential services and limited financial pro- whereas high-income countries collected almost 32 percent. tection, particularly for the poor. As Musgrove (personal com- Given projected future economic growth on the order of 4 per- munication with G. Schieber, April 2004) indicates, if GDP is cent for LMICs during 2006­15, they will face difficulties in adjusted for basic subsistence needs, poor households in LICs mobilizing additional domestic revenues (World Bank 2004b). appear to be spending a substantial share of their postsubsis- In other words, even though economic growth is a necessary tence income on health, reinforcing much of the discussion condition for progress, it is unlikely to provide the financing Financing Health Systems in the 21st Century | 229 230 | Disease Control Priorities in Developing Table 12.1 Composition of Health Financing by Region and Country Income Level, 2001 Countries (Averages) Public Social security Private Out-of-pocket External | George expenditures expenditures expenditures expenditures Private prepaid provision for Per capita Total health on health as on health as on health as on health as a plans as a health as a Schieber health expenditures a percentage a percentage of a percentage percentage of total percentage of percentage of Region and country Per capita expenditures as a percentage of total health total public health of total health private health private health total health income level GDP (US$) (US$)a of GDP expenditures expenditures expenditures expenditures expenditures expenditures ,Cristian East Asia and the Pacific 1,387 84 (46) 5.6 59.3 11.1 40.7 83.4 3.5 11.9 Baeza, Eastern Europe and Central Asia 2,053 132 (131) 5.5 67.1 42.1 32.9 94.9 3.5 2.6 Latin America and the Caribbean 3,705 237 (264) 6.4 56.2 28.5 43.8 81.5 13.7 4.0 Daniel Middle East and North Africa 2,834 102 (82) 5.6 52.7 15.6 47.3 79.1 8.1 3.1 Kress, South Asia 737 38 (21) 4.6 49.0 6.2 51.0 97.7 0.2 9.9 Sub-Saharan Africa 868 42 (29) 4.5 54.0 1.0 46.0 83.3 6.9 21.7 and High-income countries 21,198 1,527 (2,860) 7.7 70.1 33.1 29.9 74.0 16.2 0.1 others MICs 3,026 176 (106) 5.8 61.7 28.5 38.3 86.4 8.9 3.4 LICs 576 25 (19) 4.7 51.7 2.2 48.3 84.4 4.0 20.0 Sources: WHO 2004b; World Bank 2004e. Note: All figures are weighted by country. a. Per capita health expenditures include population-weighted averages (in parentheses). Table 12.2 Average Central Government Revenues, Early 2000s Total revenue as a Tax revenue as a Social security taxes as Region and country income level percentage of GDP percentage of GDP a percentage of GDP Americas 20.0 16.3 2.3 Asia and the Pacific 16.6 13.2 0.5 Central Europe, Baltic states, Russian Fed., 26.7 23.4 8.1 and other former Soviet republics Middle East and North Africa 26.2 17.1 0.8 Sub-Saharan Africa 19.7 15.9 0.3 Small islands (population less than 1 million) 32.0 24.5 2.8 LICs 17.7 14.5 0.7 Lower-middle-income countries 21.4 16.3 1.4 Upper-middle-income countries 26.9 21.9 4.3 High-income countries 31.9 26.5 7.2 Source: IMF 2004b. base needed to deal with the HIV/AIDS pandemic or to achieve privately provided health services and pharmaceuticals. The the health MDGs. government, through the MOH, generally operates like a NHS. It provides basic public health and other services, including some tertiary-level hospital care, generally in major urban Trends in Health System Financing areas, to the entire population within an extremely limited As countries move to different stages of the income spectrum, budget. In general, because of the small size of formal sector their health financing profiles transition as well. The following employment, social insurance is limited, except perhaps for discussion compares countries at different stages of the income government employees. Community-based health insurance spectrum. Given health systems' variability across time periods, may be available to varying degrees but is unlikely to play a countries, and income levels, the analysis provides only a snap- major role. Private health insurance, if any, is extremely limited shot. Figure 12.2 illustrates transitions in general health systems because of people's inability to pay and institutional constraints as countries move from low- to middle- to high-income status. to the industry's development, including the lack of well- In LICs, almost half of health spending is private, virtually developed financial markets and regulatory environments. all out of pocket, and usually in the form of payments for As countries' economies improve, government revenues tend to increase because of the expansion of the more readily taxable formal sector. Other institutions, such as financial Low-income Middle-income High-income markets, legal systems, and regulatory capabilities, are able to countries countries countries develop. Although private spending still accounts for some 40 percent of all health spending in MICs, the out-of-pocket Private insurance Patient out of pocket share declines as private health insurance markets develop. Patient out of Patient out of The MOH generally continues to provide basic public health pocket National health pocket service model services and to serve as the insurer of last resort for the poor or for the entire population for specific chronic conditions as National health Social insurance social health insurance mechanisms develop. Social insurance insurance model Countries move into the high-income group with improved Government Government Private health institutions, more efficient governments, and greater revenue- budget budget insurance model raising capacity and spend a relatively small share on basic pub- Community lic health. With few exceptions, publicly financed universal financing coverage--or, in some cases, publicly mandated private coverage--becomes the goal. MOHs maintain responsibility Source: Maeda 1998. for public health and surveillance and for the general regulato- Figure 12.2 Health Care Financing System Trends by Country ry environment but generally do not directly provide services. Income Level Risks are pooled either through a NHS, as in Italy and the Financing Health Systems in the 21st Century | 231 United Kingdom, or through single or multiple insurance revenue mobilization efforts, and future recurrent costs. Most mechanisms, as in France and Germany. The Netherlands studies indicate that the macroeconomic saturation point for requires wealthier individuals to be insured through a private aid lies somewhere between 15 and 45 percent of GDP, depend- system. Private spending declines to 30 percent, and out-of- ing on the country's policy environment (Clemens, Radelet, pocket spending represents about 20 percent of total health and Bhavnani 2004; Collier and Dollar 1999; Collier and spending. Although health financing systems are highly coun- Hoeffler 2002; Foster 2003). try specific, available information on sources of health spend- Aid can have a number of negative effects. If aid flows are ing and government revenues supports these stylized models. not included in the recipient country's budget, they can result in corruption. Aid may substitute donors' priorities for countries' priorities. A country may have insufficient human ABSORPTION, EFFECTIVENESS, AND resources, physical infrastructure, or managerial capacity to SUSTAINABILITY OF DONOR FUNDS use funds effectively. Resources that may already be in short supply and that are critical for effective service delivery may In recent years, several new dimensions have emerged in the be diverted from other important activities. New resources debate on international health financing, namely the effective- may overwhelm the system, and the donors' reporting and ness of large increases in DAH and the enormous costs of scal- administrative requirements may impose additional burdens ing up health and other social systems to meet the MDGs. Both on countries. the donor community and recipient countries have raised con- Absorptive capacity problems may also result from cerns pertaining to countries' absorptive capacity, aid effective- demand-side constraints at the individual, household, or com- ness, and sustainability. munity levels, including lack of education, limited informa- tion, travel costs, and income loss (Ensor and Cooper 2004). Conditional cash transfers are among the demand-side innova- Countries' Absorptive Capacity tions developed to improve the use of essential public health Large increases in DAH channeled to LICs have raised ques- services by the poor that have been receiving increased atten- tions about whether countries can make effective use of these tion. Such programs were initially developed in Latin America new aid flows. As table 12.3 shows, absorptive capacity has as part of social safety-net programs and provide direct cash macroeconomic, budgetary management, and service delivery payments to poor households contingent on certain behavior, dimensions. such as completing a full set of prenatal visits or attending Increased aid has important macroeconomic implications health education classes (Rawlings 2004). Conditional cash given its potential effect on exchange rates, inflation, balance transfers are in effect negative user fees. Even though investiga- of trade, overall competitiveness, aid dependency, domestic tors have found that such programs are quite successful in Table 12.3 Major Constraints to Countries' Absorption of Additional External Resources Macroeconomic Institutional Physical and human Social, cultural, and political National government Debt sustainability Monetary and fiscal policy Administrative, management, and Stable national political institutions Competitiveness instruments planning skills Power-sharing mechanisms Dutch disease Exchange rate management Training technicians and Social stability sector specialists Fiscal instruments and n.a. Public expenditure management: Sector management skills Cultural norms allocative mechanisms Budget preparation and execution Connectivity and communications Weak institutions Accounting and auditing networks Power-sharing mechanisms Service delivery and n.a. Local government institutions Accessibility Cultural norms local governments Private sector capacity Sanitation and water Ethnic, caste, and class relations Roads Geography Local government skills and capacity Source: World Bank 2004a. n.a. not applicable. 232 | Disease Control Priorities in Developing Countries | George Schieber, Cristian Baeza, Daniel Kress, and others MICs and have the potential to improve human capital and countries' budgetary and financial management capacity, and health outcomes and reduce poverty with relatively modest fostering a more transparent and predictable implementation administrative costs, their applicability in LICs is still structure (World Bank 2004a). unresolved. The effect of the composition of aid on countries' efforts to Health sector supply and demand constraints can also hin- mobilize domestic resources is also critical given the strong der countries' effective employment of large increases in health push by heavily indebted countries, several Group of Seven resources. As Mills, Rasheed, and Tollman point out in this (G-7) countries, and the United Nations Millennium Project book (chapter 3) and elsewhere, these constraints can occur at for grant assistance. Gupta and others (2004) find that all levels of service delivery and governances (Oliviera-Cruz, increases in overall aid (net loans plus grants) result in a decline Hanson, and Mills 2003). Additional funding alone does not in total domestic revenues; however, the effects of loans were create sufficient conditions for overcoming structural weak- quite different from those of grants. Each 10.0 percent increase nesses, particularly in the short run. If aid is targeted to specif- in loans was associated with a 2.3 percent increase in domestic ic diseases or interventions, effective use of such aid may revenues, whereas a 10.0 percent increase in grants was associ- "consume" different amounts of a country's administrative ated with a 2.8 percent decrease in domestic revenues. The capacity. Increased public funds may supplant private spending same study also finds higher levels of corruption result in not only by the poor, but also by the nonpoor, resulting in lim- reduced domestic revenue-raising efforts. ited marginal effects on the poor (Filmer and Pritchett 1999). Fiscal Sustainability Aid Effectiveness Fiscal sustainability is an often used but rarely defined term, Given calls for increases in aid of anywhere between US$25 bil- though it has generally been defined in terms of self- lion and US$75 billion a year, the question of aid effectiveness sufficiency. In its broadest context, achieving sustainability has taken on increased importance.2 A protracted debate has means that, over a specific period, the managing entity will generated the following findings concerning aid (Burnside and generate sufficient resources to fund the full costs of a particu- Dollar 1997; Clemens and Radelet 2003; Clemens, Radelet, and lar program, sector, or economy, including the incremental Bhavnani 2004; Collier and Dollar 1999; Collier and Hoeffler service costs associated with new investments and the servicing 2002; Foster 2003; United Nations Millennium Project 2005; and repayment of external debt. WHO 2001; World Bank 2004a): Knowles, Leighton, and Stinson (1997, 39) define health system sustainability as the "capacity of the health system to · Aid has diminishing returns. replace withdrawn donor funds with funds from other, usually · Countries' absorptive capacity is limited. domestic, sources" and sustainability of an individual program · Aid is fungible overall and among sectors. as the "capacity of the grantee to mobilize the resources to · Aid achieves better results in good policy environments.3 fund the recurrent costs of a project once it has terminated." · Aid requires ownership by countries; for example, donor- However, given the enormous unmet needs in the poorest imposed conditions rarely work. countries, coupled with stagnant economic performance, some · Aid is related to increased investment and growth. donors are now defining sustainability on the basis of the man- · Debt repayments have a negative effect on economic aging entity's commitment of a stable and fixed share of pro- growth. gram costs (Brenzel and Rajkotia 2004; Kaddar, Lydon, and · Aid has high transaction costs for countries. Levine 2003). · Aid makes governments accountable to donors as opposed In light of criticisms leveled at the International Monetary to their citizens. Fund (IMF) regarding its structural adjustment programs and fiscal ceilings, IMF has recently paid increased attention to fiscal Serious overall and health sector­specific questions pertain sustainability. However, evaluating a country's fiscal situation to the levels, predictability, variability, fungibility, and sustain- and defining sustainability are not easy matters (Croce and ability of aid flows, and debate continues between those argu- Juan-Ramon 2003; Dunaway and N'Diaye 2004; Hemming, ing for vertical disease­specific program assistance and those Kell, and Schimmelpfennig 2003; Tanzi and Zee 2000). Work supporting broader health system reform changes (WHO is under way to develop operational indicators of debt and fiscal 2001). As Mills, Rasheed, and Tollman show in this book (chap- sustainability and to define the concept of fiscal space (Dunaway ter 3), evidence on the effectiveness of both approaches is and N'Diaye 2004; Heller 2005). Understanding the details of mixed. Aid unpredictability and uncertainty need to be IMF fiscal programs and ensuring stable and predictable long- addressed by aligning donors' disbursement and commitment term DAH are important conditions for avoiding the macroeco- cycles with those of recipient countries, strengthening nomic distortions discussed earlier. Financing Health Systems in the 21st Century | 233 HEALTH FINANCING ISSUES IN LICs Although assessments of global initiatives and alliances are generally positive, some observers have concerns about their This section discusses the severe challenges LICs face in mobi- effects on health systems and prioritization (Travis and others lizing sufficient revenues, both domestically and externally, to 2004). Increasing concerns are being expressed about the meet even the basic health needs of their populations. "verticalization" of DAH and the development of separate health system "silos," each dedicated to specific diseases and The Needs Gap activities. This strategy is especially problematic in light of the scarce human resources available for health in many LICs Since the release of the 1993 World Bank World Development (Global Health Trust 2004; Joint Learning Initiative 2004). Report: Investing in Health, researchers have undertaken As a result of these concerns, the G-7 countries are currently numerous efforts to estimate the costs of a basic package of discussing a number of new, broad-based, global financing essential health services. The Commission on Macroeconomics mechanisms to mobilize and facilitate the transfer of resources and Health (WHO 2001) estimated that, in 1997, the 48 poorest from developed countries to LICs, and significant progress developing countries were spending on average US$11 per has been made in relation to the International Finance Facility capita (US$6 per year in public funds) and that the level of (IFF), a proposal advanced by the U.K. government. The IFF spending would have to rise to US$34 per capita to ensure delivery of an essential package.4 On the basis of these data, the will frontload development assistance by issuing bonds on international markets that would be secured based on legally Commission on Macroeconomics and Health estimated that binding, long-term donor commitments. The IFF would repay total DAH should rise to US$27 billion in 2007 and to US$38 bondholders using future donor payments. Depending on the billion by 2015 to scale up coverage (WHO 2001). number of donors involved, the IFF could raise an additional Within the framework of the MDGs, a number of other US$50 billion a year in development assistance between now studies have been undertaken to determine the financial and 2015. One of the many advantages of this kind of mecha- resources needed to meet the goals. In this book, Wagstaff and nism is that a portion of funding for international develop- others (chapter 9) review the estimates and methodologies ment is effectively taken out of the annual budgetary process in from these various studies, finding that the annual cost of scal- participating countries. In this way, the hope is that the revenue ing up to meet the MDGs is between US$25 billion and US$75 streams available to fund development can be rationalized, billion. The United Nations Millennium Project (2005) both in terms of the total volume of assistance and in terms of estimates that the additional overall development assistance the stability of annual flows. needed for scaling up to meet all the MDGs will be US$74 bil- These global funds have added a major new dynamic to lion by 2015. All the studies indicate that most LICs will face global health policy and a new level of influence over LICs. enormous constraints in raising additional resources through Large grants are approaching the World Bank's financing levels domestic resource-mobilization efforts and that the interna- for the health sector. Moreover, such funding is often targeted tional community must essentially finance most of the gap. to specific diseases or interventions, frequently outside the basic broadly based financing instruments required by the New Global Alliances and Funds World Bank and the IMF. This factor raises important issues of Recent years have witnessed a marked increase in the number donor coordination and harmonization of procedures and has of global alliances and institutions aimed at alleviating specific implications for IMF fiscal ceilings. health sector deficiencies, a number of which owe their exis- tence to resources made available by philanthropic organiza- tions.5 The GAVI Vaccine Fund and the Global Fund to Fight Financing Instruments AIDS, Tuberculosis, and Malaria are perhaps the largest and During the past decade, a new reform instrument known as the most well known. While the GAVI Vaccine Fund is both a fun- SWAp has heavily influenced health financing, particularly for der and an implementer, Roll Back Malaria is an example of an LICs. Concomitantly, the World Bank and the IMF have alliance that is a global partnership without a funding mecha- imposed a series of requirements and instruments to ensure nism. Some entities like the Global Fund are purely financial that external assistance is targeted to the poor through PRSPs. vehicles with little alliance structure. The effect of these new These new policy blueprints and requirements are radically dif- alliances and funds is significant. Since its inception in 2000, ferent from previous DAH mechanisms, which were largely the GAVI Vaccine Fund has raised and spent more than US$1 funded on a bilateral basis through projects. billion for immunization, and the Global Fund has commit- ments of more than US$5 billion and has signed grant SWAps. Starting in the mid-1990s, donors and recipient coun- agreements with more than 70 countries worth in excess of tries established the SWAp to address the limitations of project- US$3 billion. based forms of donor assistance, to ensure that overall health 234 | Disease Control Priorities in Developing Countries | George Schieber, Cristian Baeza, Daniel Kress, and others reform goals were met, to reduce large transaction costs for · provide a comprehensive coordination framework for the countries,and to establish genuine partnerships between donors World Bank, the IMF, and other development partners and countries in which both had rights and responsibilities. The · improve public governance and accountability core elements of a SWAp follow (McLaughlin 2003, 2004): · improve priority setting. · The government is "in the driver's seat." The PRSP process is country driven, involves broadly based · The partnership results in a shared vision and agreed-upon participation, is results oriented and focused on outcomes that priorities for the sector. benefit the poor, is comprehensive in recognizing the multidi- · A comprehensive sector development strategy that reflects mensional nature of poverty,is partnership oriented,and is based all development activities to identify gaps, overlaps, or on a long-term perspective (IMF 2004a; World Bank 2004d). inconsistencies. The PRSP process has made poverty reduction the priority · An expenditure framework that clarifies sectoral priorities issue for development (SHC Development Consulting 2001). and guides all sectoral financing and investment. Because macroeconomic and sectoral strategies need to be for- · A partnership across development assistance agencies that mulated around the PRSP, health reform strategies must be reduces governments' transaction costs. included and focus on the poor. As of September 2004, about 42 LICs had developed PRSPs that are serving as the basis for SWAps explicitly recognize the need to tie health sector World Bank and IMF financing in those countries. Extensive changes to new aid instruments, to macroeconomic and public evaluations of PRSPs by the World Bank and the IMF, by bilat- sector management, to poverty reduction, and to achievement eral donors, and by other development partners have painted of the MDGs (Cassels 1997). A key aspect of this approach is to the following mixed picture of their success (IMF 2004a; IMF improve countries' policy-making processes, including budget and World Bank 2002, 2003; World Bank 2004d): and public expenditure management, by capturing all funding · PRSPs have the potential to encourage the development of sources and expenditures and by putting resource allocation country-owned, long-term strategies for poverty reduction decisions into a medium-term budget and expenditure frame- and growth, but tensions concerning ownership among work that is based on national priorities (Foster 1999). To date, countries, the World Bank and the IMF, and other donors SWAps are in various stages of development and implementa- remain. External partners have not adapted their assistance tion, and few conform fully to the specifications listed above programs to PRSP processes in a coordinated manner, and (Institute for Health Sector Development 2003). At this point better frameworks for accountability of both countries and in their evolution, SWAps should be viewed as a way of coordi- partners are needed. nating development assistance and creating country owner- · Country participation has improved; however, greater ship. They should be judged on how well they do these things inclusiveness is still needed. Moreover, the process has not compared with the previous environment characterized by strengthened domestic institutional policy-making pro- multiple, stand-alone projects. cesses or accountability. · PRSPs are an improvement over previous processes in terms PRSPs. Starting in the mid 1990s, the World Bank and the IMF of results orientation, poverty reduction focus, and long- began to radically change both the focus and the tools for pro- term perspective. They have fallen short in terms of being a viding development assistance to poor countries. In response strategic reform road map, especially in relation to under- to criticisms about the ineffectiveness of previous development taking structural reforms, boosting economic growth, link- assistance efforts and the high level of indebtedness in some of ing with medium-term expenditure frameworks and budg- the world's poorest countries, the two organizations focused on ets, integrating sectoral strategies into the macroeconomic debt forgiveness for heavily indebted poor counties, poverty framework, assessing the social effects of macroeconomic reduction, and improved economic growth. Debt forgiveness strategies, understanding links between macroeconomics required countries to reprogram the bulk of the savings from and microeconomics, integrating strategy components, and forgiven debt into social sectors such as health and education. linking medium- and long-term operational targets. In 1999, the World Bank and the IMF stipulated that all of · Capacity constraints have been serious impediments to their concessionary assistance to 81 eligible poor countries effective implementation, but little attention has focused on would need to be based on a Poverty Reduction Strategy Paper capacity building. (IMF and World Bank 2002, 2003). This new approach was · Monitoring and evaluation is still a significant weakness. intended to the following: Evaluations of the health sector components of PRSPs raise · strengthen country ownership many of these issues (DFID Health Systems Resource Centre · enhance the poverty focus of country programs 2003; WHO 2004a). As more and more partners buy into this Financing Health Systems in the 21st Century | 235 process and as increased amounts of development assistance community-based health insurance provides some are funneled through PRSPs, their effectiveness will ultimately financial protection by reducing out-of-pocket spending. depend on country commitment, capacity, and processes; part- There is evidence of moderate strength that such ner flexibility; and funding availability. At this stage, PRSPs schemes improve cost-recovery. There is weak or no evi- still seem to be a work in progress. dence that schemes have an effect on the quality of care or the efficiency with which care is produced. In absolute Community-Based Health Insurance terms, the effects are small and schemes serve only a lim- As noted earlier, private and out-of-pocket spending accounts ited section of the population. The main policy implica- for almost half of total health spending in LICs. Given LIC gov- tion of this review is that these types of community ernments' limited abilities to mobilize revenues, country and financing arrangements are, at best, complementary to donor attention has turned to informal sector insurance mech- other more effective systems of health financing. anisms as a way to improve financial protection, mobilize rev- enues, and improve the efficiency of out-of-pocket spending. The evidence from these reviews suggests that, even though Community-based health insurance is an umbrella term for the CBHI provides financial protection for those enrolled and some various types of community financing arrangements that have degree of resource mobilization, the overall effect is relatively emerged because of high out-of-pocket spending, uncertainty small and schemes are less effective in reaching the very poor. surrounding anticipated financial flows from donors, and large Thus, CBHI is unlikely to be a panacea for substantially improv- and unregulated private sectors. Here, CBHI refers to prepay- ing risk pooling and mobilizing resources in LICs, and for MICs, ment plans that attempt to pool risks to reduce the financial CBHI is less relevant given higher incomes and levels of formal risk an individual faces because of illness (Atim and others sector employment. This finding does not suggest that CBHI 1998; Bennett, Creese, and Monash 1998; Bennett, Kelley, and should not be part of an overall solution to financing health Silvers 2004). care, but it indicates that CBHI is unlikely to play a major role. CBHI is found throughout the world but is particularly The most critical challenge facing LICs is raising sufficient prevalent in Sub-Saharan Africa (Bennett, Kelley, and Silvers revenues to meet their basic health needs and the health MDGs. 2004). CBHI plans are relatively heterogeneous in terms of Although increased grant funding is badly needed, the large populations covered, services offered, regulation, management amounts of funds often targeted to a few countries and for spe- function, and objectives. The Commission on Macroeconomics cific diseases and interventions raise questions of country and Health found that CBHI plans provided significant finan- absorptive capacity, potential distortions of health systems' pri- cial protection and extended access to a large number of rural orities, and interactions with IMF fiscal ceilings. Concomitantly, and low-income populations (WHO 2001), but that affordabil- LICs must improve their institutions in order to increase ity impeded access for the very poor. As a result, the commis- absorptive capacity and increase the effectiveness of all official sion called for increased support for CBHI and for the estab- development assistance. It is also critical for the international lishment of a cofinancing scheme that would match dollar for community to reassess the entire official development assistance dollar the premiums individuals paid toward their health insur- and DAH structure; to develop country-compatible mecha- ance with a government or donor dollar (WHO 2001). nisms to reinforce promised international redistribution; and to One recent review of the CBHI experience found less positive improve the targeting, levels, predictability, and timeliness of results, noting "no evidence from the documents reviewed that external assistance. [CBHI schemes] positively impact health status or at least the utilization of services and financial protection for their mem- HEALTH FINANCING ISSUES IN MICs bers and/or for society at large, particularly the poor" (ILO and STEP 2002a, 54). The review finds that most CBHI schemes MICs benefit from higher levels of domestic funding, higher "tend to be small organizations (70 percent covering less than initial levels of risk pooling and prepayment, and stronger 200 members) with community participation in key decisions at health systems than LICs. Many MICs are now focused on one point or another in their history but with limited legal or de ensuring access and financial protection through universal facto ownership by the community and with significant depend- health coverage. Chile, Colombia, the Republic of Korea, ence from other health subsystems or subsidies as reflected by Mexico, Poland, and Thailand are implementing universal cov- their low market exposure" (ILO and STEP 2002a, 54). erage reforms or have already done so. However, they and many In his assessment of CBHI, Ekman (2004, 249) notes the other MICs still face challenges similar to those facing LICs. following: Alternative Risk Pooling Arrangements Overall, the evidence base is limited in scope and Country experience shows that the critical factors for increas- questionable in quality. There is strong evidence that ing coverage--that is, the number of individuals covered and 236 | Disease Control Priorities in Developing Countries | George Schieber, Cristian Baeza, Daniel Kress, and others Table 12.4 MIC Reforms and Innovations for Achieving Universal Coverage Organizational arrangement Reforms or innovations Social security Opening affiliation to self-employed and informal sector workers Mandating universal participation Providing direct public subsidies to the organization for including the poor Subsidizing premiums for the poor, self-employed, and workers in the informal sector MOHs and NHSs Separating the purchase and provision of care Using public and private purchasing Reforming provider payments Private health insurance Regulating voluntary health insurance Making private insurers eligible for mandatory social security for health Providing demand-side subsidies for health insurance Integration reforms (reforms that allow synergic Using public and private purchasing interaction of multiple organizational Providing demand-side subsidies for health arrangements) insurance for the poor and for high-risk groups Setting up risk equalization and solidarity funds Providing health education to stimulate demand Having a virtual single pool Source: Authors. the extensiveness of the benefit package--are increased risk face risks of adverse selection because of the voluntary nature pooling and prepayment and better access to equity subsidies. of enrollment and the exclusion of the poorest (Bitrán and As discussed earlier, most MICs face fragmented risk pools others 2000; Instituto Mexicano de Seguro Social 2003). Chile, (ILO and STEP 2002b). Table 12.4 presents MIC approaches to Colombia, Costa Rica, and the Philippines have addressed reforming risk pooling arrangements for achieving universal exclusion either by subsidizing the SSO directly or by subsidiz- coverage. ing premiums for the poor and informal and self-employed Most MICs face an additional strategic decision: whether to workers who join. The Republic of Korea and Taiwan, China, pursue aggregation of all pools in a single organization (a sin- have implemented mandatory universal participation, includ- gle pool) or whether to allow for the existence of multiple risk ing gradual expansion to the whole population, whereas pooling organizations, which would explicitly or implicitly Panama has expanded coverage to dependents of contributing compete for members and would be subject to the same rules members. regarding benefit packages, revenue collection mechanisms, Some of the most important advantages underlying SSO and portability of benefits (that is, a virtual single pool). innovations include the existence of organizational capacity Colombia and Turkey have opted for the virtual single pool and of pools of funds (or sometimes a single large fund) that reform, whereas Costa Rica has chosen a single risk pool. In the allow newly enrolled individuals and groups to take advantage OECD context, Germany and the Netherlands have virtual sin- of the risk and income cross-subsidization mechanisms and gle pool arrangements, whereas New Zealand and the United purchasing arrangements that are already in place. This Kingdom have single pool arrangements. approach results in an immediate enlargement of the risk pool in contrast to creating other pooling organizations as interme- Reforms of Social Security. Because SSOs traditionally cover diate steps for the future merging of schemes. However, SSOs salaried formal sector workers from whom payroll contribu- usually cover only a relatively small portion of the total popu- tions can be collected, requiring informal sector workers or lation, and their focus on formal sector workers and use of pay- self-employed workers to join is difficult. The reforms for con- roll contributions as their main revenue collection mechanisms fronting this issue range from voluntary enrollment to various might be an insurmountable obstacle for reaching the informal types of subsidization, as detailed in table 12.4. sector and the poor, particularly those in rural areas. Country experiences are also illuminating. For instance, In countries where a SSO is well established and covers a Chile and Mexico have opened SSOs to the informal sector and large population, it might face problems in including informal the self-employed through voluntary affiliation, yet they still sector workers in the absence or even the presence of public Financing Health Systems in the 21st Century | 237 subsidies if the incentive structure is not well designed. For Single Pool versus Virtual Single Pool. Most MICs must also instance, in the case of Mexico, the SSO operates a scheme that decide whether to aggregate all pools into a single organization is partially subsidized by the central government, yet the or to aim for a virtual single pool (Baeza and Packard 2005). scheme still has experienced severe adverse selection and has The implementation of more effective and efficient cross- few participants, which has also discouraged actively promot- subsidies between groups with different income and health ing enrollment (World Bank 2004c). risks is facilitated by merging smaller pools into large pools-- in some cases, national pools. Indeed, the main preliminary les- Reforms of MOHs and NHSs. MIC reform approaches for sons emerging from Costa Rica, the Republic of Korea, and MOHs and NHSs include introducing internal markets, Taiwan (China), all of which have achieved universal coverage, including separating the purchasing function from the provi- suggest that the combination of a clearly defined benefit pack- sion of health services; using public-private purchasing; age and reforms for enlarging risk pools plays a paramount role reforming provider payment systems; and decentralizing. In in achieving greater inclusion through solidarity in financing theory, efficiency gains in the system could be used to provide and increasing access. access to new enrollees, to increase the number and quality of Yet for most MICs, the reality is that multiple pooling services to all participants in the system, or to do both. Success arrangements exist, leading to a fragmented, inefficient, and in these areas has been limited (Baeza and Packard 2005). inequitable health financing situation overall. Given that devel- Effective modernization of public sector management and civil oping proper regulations and incentive systems for counterbal- services statutes has to date been missing from most provider ancing such problems is complicated, both institutionally and payment and health sector reform efforts. cost-wise, fostering a virtual single pool is likely the most feasi- ble option for these countries. Private Health Insurance Reforms in MICs. Since the 1980s, MICs have seen two main reforms related to private insurance: (a) the facilitation and promotion of voluntary health insurance, Sources of Health System Financing in MICs including formal recognition of competing private health insur- As discussed earlier in this chapter, health systems use different ance, and (b) the integration of regulated private insurance as sources of financing and revenue collection, including general one component of mandatory social security schemes for formal taxation, payroll contributions, risk-rated premiums, and user workers. Many MICs, such as Indonesia, Mexico, and the fees. However, concern is increasing about the use of payroll Philippines,now recognize and regulate voluntary private health contributions as a mechanism for collecting revenue. In a insurance.In Chile and Colombia,private insurers participate in recent study on financial protection in Latin America, Baeza the provision of mandatory risk pooling for social security. and Packard (2005) argue that to extend effective risk pooling The literature provides some evidence of the potential ben- to the informal and nonsalaried sectors and to achieve univer- efits and problems resulting from the introduction of private sal participation in risk pooling arrangements, policy makers health insurance and competition in the insurance market need to delink health insurance financing and eligibility from (Londoño and Frenk Mora 1997; Sheshinski and López-Calva labor market status or employment sector, by gradually reduc- 1998). However, an ample literature also deals with the equity ing and eventually eliminating payroll contribution financing. and efficiency problems of private health insurance competi- In addition to extending protection against health shocks, this tion, including risk selection (insurers seeking to enroll low- delinking might also have a positive effect on labor market risk individuals) and underservice (insurers setting barriers mobility and formalization. This delinking can be achieved to the use of services, for instance, by not contracting with through shifting health financing toward general taxation, providers of expensive interventions or in low-income areas) which is likely preferred on equity and efficiency grounds, or (Arrow 1963; Hsiao 1994, 1995; Laffont 1990; Milgrom and through risk-rating premiums as a transition if fiscal con- Roberts 1992). straints do not permit full fiscal financing. As to whether harnessing private health insurance con- tributes to or damages MICs' chances for achieving universal coverage, the question is whether MICs can take advantage of Donor Disengagement from MICs the benefits of introducing health insurance competition and As demonstrated both by the composition and recipients of avoid the related efficiency and equity problems. MICs must DAH and by new health financing and policy agendas, global confront the feasibility of introducing specific financial, regula- health financing policy is currently focused on LICs, leaving tory, and organizational reforms at a level of transaction costs most MICs under the radar. The MDG agenda is predomi- that would not offset the benefits of competition and privati- nantly a LIC agenda, or at least most MICs perceive it as such. zation (Baeza and Cabezas 1998; Coase 1937; Newhouse 1998; Thus, the question is what to do with the MIC policy dialogue. Williamson 1985). Is the status quo tantamount to disengagement? If so, the 238 | Disease Control Priorities in Developing Countries | George Schieber, Cristian Baeza, Daniel Kress, and others international community is in danger of losing important Global health financing policy makers face the following financing lessons that would most likely be of great use for challenges: LICs. In addition, a more concerted effort is needed to analyze what the MDGs mean for MICs--particularly in light of their · The architecture for formulating, coordinating, and imple- increased noncommunicable disease and injury burdens, areas menting global health financing policy at the international that the MDGs do not address--and to invest more in the and country levels needs to be improved. evidence base for MIC-relevant reforms. It is important to · The donor community needs to harmonize procedures, maintain broad goals, but also important to develop new MIC- ensure aid predictability, and guarantee longer-term assis- specific indicators, especially for financial protection, which is tance. at the core of poverty alleviation in MICs as well as LICs but is · Donors need to meet their development assistance obliga- not explicitly reflected in the MDGs. tions as well as provide more assistance to help countries The PRSP process also sends a clear signal to the interna- improve their domestic resource mobilization efforts. tional community to focus on the LICs. Unfortunately, few · The IMF needs to improve understanding of its fiscal pro- PRSPs consider the role of the health system in ensuring finan- grams and be more flexible in reconciling fiscal constraints cial protection and reforms of risk pooling arrangements, with increased official development assistance and DAH. which are at the core of most LIC and MIC health sector · The global community needs to improve the knowledge financing strategies. A new approach is needed to support base in terms of good (and bad) international practice with MICs' efforts to improve public subsidy management and respect to health financing. In this context, absorptive health system performance to ensure financial protection. Such capacity constraints on both the demand and the supply approaches are also critical in assisting LICs with their poverty sides must be removed. Better use of existing tools, includ- reduction and health financing reform efforts in the future. ing cost-effectiveness analysis, and development of new tools are needed to help poor countries realistically priori- tize their financing and spending options and deal with the CONCLUSIONS tradeoffs between financing essential services and providing financial protection. Global health financing policy is in transition. Infusions of large · The potential for verticalization as a result of increased lev- amounts of grant money from new financing entities have els of DAH needs to be assessed rigorously and empirically, changed the players involved in shaping global health policy. taking into account the benefits of such assistance as well as Decisions made by the World Bank and the IMF in 1999 requir- its potential distortionary effects on other programs and on ing PRSPs as the basis for concessionary financing have pushed health systems as a whole. By focusing limited resources on LICs to develop their health policies in the context of an overall a few targeted areas, countries can achieve impressive results strategy framework for poverty reduction that considers intra- in terms of disease control efforts; however, many disease sectoral, intersectoral, and macroeconomic tradeoffs. eradication efforts have succeeded because such efforts Clearly, neither increased domestic resource mobilization enhanced overall system capacity. nor future economic growth will provide the resources neces- · The existing assistance instruments need to be objectively sary for LICs to finance their health needs, whether defined in and fully analyzed. Examples of potential inconsistencies, terms of a basic package of essential health services or whether such as disease-specific program grants versus PRSPs, need identified within the framework of the MDGs. Increasing offi- to be highlighted and addressed. cial development assistance is thus critical for LICs to make · The issue of financial sustainability needs to be assessed progress in either respect. However, the projected magnitude objectively and apolitically. The international donor com- and speed of scaling up raises serious questions about coun- munity needs to face up to the realities of those poor coun- tries' absorptive capacity, aid effectiveness, predictability, and tries whose economies are not sustainable in the medium stability and about new investments' financial sustainability at term and to consider redistributional policies to assist the country and donor levels. Even though empirical evidence them. is still lacking, concerns have arisen that new sources and · The donor community needs to put MICs on the agenda increased levels of funding for disease-specific programs will both in terms of their economic and social development and lead to verticalization and could distort health systems. The in terms of their use as good practice examples for LICs as donor community and countries urgently need to reform the they transition to MIC status. current system of DAH, to improve institutions in developing countries, and to develop mechanisms to ensure that donors Because of the different accountabilities of the various mul- meet their DAH commitments. Finally, MIC issues need to tilateral and bilateral organizations, global funds and alliances, receive greater attention. and private foundations, coordinating global health financing Financing Health Systems in the 21st Century | 239 policy has become increasingly complex. Given that interna- 5. One of the main funding organizations is the Bill & Melinda Gates tional redistribution of wealth is central to meeting basic needs Foundation, which is investing approximately US$1.35 billion per year, with a considerable portion of that allocated to global health issues. in poor countries, the lack of an effective international mecha- nism to enforce agreed-on transfers of wealth is problematic. Under these circumstances, the global community must help REFERENCES countries prioritize on the basis of realistic expectations of promised donor assistance and harmonization. Arhin-Tenkorang, D. 2000."Mobilizing Resources for Health: The Case for User Fees Revisited." Working Paper 81, Center for International Providing countries with advice on good practice and assist- Development, Cambridge, MA. ing both LICs and MICs to develop equitable and efficient Arrow, K. J. 1963. "Uncertainty and the Welfare Economics of Medical institutional structures, revenue-raising mechanisms, and Care." American Economic Review 53 (5): 851­83. spending prioritizations are important areas worthy of more Atim, C., F. Diop, J. Etté, D. Evrard, P. Marcadent, and N. Massiot. 1998. international focus and collaboration. Assessments of the costs The Contribution of Mutual Organizations to Financing, Delivery, and and constraints in reaching the health MDGs, taking into Access to Health Care: Synthesis and Research in Nine West and Central African Countries. Bethesda, MD: Abt Associates, Partnerships for account the large increases in marginal costs to cover the most Health Reform Project. difficult-to-reach 5 or 10 percent of the population, are impor- Baeza, C., and M. Cabezas. 1998. "Is There a Need for Risk Adjustment in tant knowledge products in a resource-constrained world. Health Insurance Competition in Latin America?" Discussion paper Making better use of cost-effectiveness information and devel- prepared for the World Bank, Latin America and the Caribbean Region, oping better-costing tools are necessary for assisting countries, Human and Social Development Department, Washington, DC. and donors could help by providing better information on Baeza, C., and T. G. Packard. 2005. Beyond Survival: Protecting Households from the Impoverishing Effects of Health Shocks in Latin America. where to focus policies to remove bottlenecks to the absorption Washington, DC: World Bank. of additional resources, particularly in terms of achieving the Bennett, S., A. Creese, and R. Monash. 1998. "Health Insurance Schemes MDGs. A needed step for assisting LIC and MIC governments for People outside Formal Sector Employment." Discussion of is developing and disseminating evidence about effective health Analysis, Research, and Assessment Paper 16, World Health Organization, Geneva. financing polices, both in severely resource-constrained LICs Bennett, S., A. G. Kelley, and B. Silvers. 2004. 21 Questions on CBHF: An that have achieved good health outcomes and in MICs that Overview of Community-Based Health Financing. Bethesda, MD: Abt have achieved universal coverage with good health outcomes at Associates, Partnerships for Health Reform Project. reasonable spending levels. Last, the donor community must Bidani, B., and M. Ravallion. 1997. "Decomposing Social Indicators Using harmonize its procedures, simplify aid instruments, ensure the Distributional Data." Journal of Econometrics 77 (1): 125­39. predictability of assistance, and create a more effective global Bitrán, R., and U. Giedion. 2003. "Waivers and Exemptions for Health policy environment. Services in Developing Countries." Social Protection Discussion Paper 308, World Bank, Washington, DC. Bitrán, R., J. Muñoz, P. Aguad, M. Navarrete, and G. Ubilla. 2000. "Equity in the Financing of Social Security for Health in Chile." Health Policy DISCLAIMER 50 (3): 171­96. Brenzel, L., and Y. Rajkotia. 2004. "Vaccine Financing Report." Paper pre- The findings, interpretations, and conclusions expressed in this pared for the World Bank Human Development Network, Health paper are entirely those of the authors and should not be Nutrition and Population Unit, Washington, DC. attributed in any manner to the Bill & Melinda Gates Burnside, C., and D. Dollar. 1997. "Aid, Policies, and Growth." Policy Foundation, the RAND corporation, or to the World Bank, its Research Working Paper 1777, World Bank, Policy Research Department, Macroeconomics and Growth Division, Washington, DC. affiliated organizations, the members of its Board of Directors, Cassels, A. 1997. A Guide to Sectorwide Approaches for Health Development: or the countries it represents. Concepts, Issues, and Working Arrangements. Geneva: World Health Organization. Clemens, M. A., and S. Radelet. 2003. "The Millennium Challenge NOTES Account: How Much Is Too Much, How Long Is Long Enough?" Working Paper 23, Center for Global Development, Washington, DC. 1. For a detailed analysis of country-specific and global health expen- Clemens, M. A., S. Radelet, and R. Bhavnani. 2004. "Counting Chickens diture trends, see Musgrove, Zeramdini, and Carrin (2002). When They Hatch: The Short-Term Effect of Aid on Growth." Working 2. In addition to aid, countries receive significant financial inflows Paper 44, Center for Global Development, Washington, DC. through foreign direct investment, expatriate workers' remittances, special Coase, R. H. 1937. "The Nature of the Firm." Economica 4 (16): 386­405. targeted assistance, South-South support, and so on, and these inflows must also be taken into account (World Bank 2004b). Collier, P., and D. Dollar. 1999. "Aid Allocation and Poverty Reduction." 3. Clemens, Radelet, and Bhavani's (2004) study shows that aid can be Unpublished manuscript, World Bank, Development Research Group, somewhat effective in countries with weaker policy environments. Washington, DC. 4. More recent data for all LICs indicate per capita spending of US$19 Collier, P., and A. Hoeffler. 2002. "Aid, Policy, and Growth in Post-Conflict if the data are population weighted and US$25 if they are country Societies." Policy Research Working Paper 2902, World Bank, weighted. The public share is 52 percent (country weighted). Washington, DC. 240 | Disease Control Priorities in Developing Countries | George Schieber, Cristian Baeza, Daniel Kress, and others Commission for Africa. 2005. Our Common Interest: Report of the Com- ILO and STEP (International Labour Organization and Strategies and mission for Africa. London: Commission for Africa. http://www. Tools against Exclusion and Poverty). 2002a. "Extending Social commissionforafrica.org/english/report/thereport/cfafullreport_1. Protection in Health through Community Based Health Or- pdf. ganizations." Discussion paper, ILO and STEP, Geneva. Croce, E., and V. H. Juan-Ramon. 2003. "Assessing Fiscal Sustainability: ------. 2002b. "Toward Decent Work: Social Protection for Health for All A Cross-Country Comparison." Working Paper 03/145, International Workers and Their Families." Working paper, ILO and STEP, Geneva. Monetary Fund, Washington, DC. IMF (International Monetary Fund). 2004a. Evaluation of the IMF's Role Devarajan, S., V. Swaroop, and Z. Heng-Fu. 1996. "The Composition of in Poverty Reduction Strategy Papers and Poverty Reduction and Public Expenditures and Economic Growth." Journal of Monetary Growth Facility. Washington, DC: IMF, Independent Evaluation Economics 37 (2­3): 313­44. Office. DFID (Department for International Development) Health Systems ------. 2004b. 2004 Government Finance Statistics. Washington, DC: IMF. Resource Centre.2003.A Review of Human Resource Content of PRSP and IMF and World Bank. 2002. "Review of the Poverty Reduction Strategy HIPC Documentation in 6 Selected African Countries. London: DFID. Paper (PRSP) Approach: Early Experience with Interim PRSPs and Dunaway, S., and P. N'Diaye. 2004. "An Approach to Long-Term Fiscal Full PRSPs." Paper prepared for the International Development Policy Analysis." Working Paper 04/113, International Monetary Fund, Association and the IMF, Washington, DC. Washington, DC. ------. 2003. "Poverty Reduction Strategy Papers: Detailed Analysis and Ekman, B. 2004. "Community-Based Health Insurance in Low-Income Progress to Date." Paper prepared for the International Development Countries: A Systematic Review of the Evidence." Health Policy and Association and the IMF, Washington, DC. Planning 19 (5): 249­70. Institute for Health Sector Development. 2003. "Mapping of Sector-Wide Ensor, T., and S. Cooper. 2004. "Overcoming Barriers to Health Service Approaches in Health." Report prepared for the Swedish International Access: Influencing the Demand Side." Health Policy and Planning Development Cooperation Authority Sectorwide Approach Seminar, 19 (2): 69­79. San Francisco, CA, June 19. Filmer, D., and L. Pritchett. 1999. "The Impact of Public Spending on Instituto Mexicano de Seguro Social. 2003. Informe al Ejecutivo Federal y al Health: Does Money Matter?" Social Science and Medicine 49 (10): Congreso de la Unión sobre la situación financiera y los riesgos del 1309­23. Instituto Mexicano del Seguro Social. Mexico City: Instituto Mexicano de Seguro Social. Foster, M. 1999. "Lessons of Experience from Sectorwide Approaches in Joint Learning Initiative. 2004. "Human Resources for Health: Health." Paper prepared for the World Health Organization and the Overcoming the Crisis." Harvard University Press Global Equity Inter-Agency Working Group on Sector-wide Approaches and Initiative, Cambridge, Mass. Development Cooperation, Geneva. Kaddar, M., P. Lydon, and R. Levine. 2003. "A Critical Review of Financial ------. 2003. The Case for Increased Aid: Final Report to the Department for Sustainability: The GAVI Experience." Discussion paper prepared for International Development. Chelmsford, U.K: Mick Foster Economics. London School of Hygiene and Tropical Medicine, Workshop on the Global Health Trust. 2004. Specific Programs and Human Resources: Economics of Immunization, London, October 29­30. Addressing a Key Implementation Constraint. Cambridge, MA: Harvard Kivumbi, G. W., and F. Kintu. 2002. "Exemption and Waivers from Cost University. Sharing: Ineffective Safety Nets in Decentralized Districts in Uganda." Gupta, S., B. J. Clements, A. Pivovarsky, and E. R. Tiongson. 2004. "Foreign Health Policy and Planning 17 (Suppl. 1): 64­71. Aid and Revenue Response: Does the Composition of Aid Matter?" In Knowles, J. C., C. Leighton, and W. Stinson. 1997. Measuring Results of Helping Countries Develop: The Role of Fiscal Policy, ed. S. Gupta, B. J. Health Sector Reform for System Performance: A Handbook of Indicators. Clements, and G. Inchauste, 385­406. Washington, DC: International Special Initiatives Report 1. Bethesda, MD: Abt Associates, Monetary Fund. Partnerships for Health Reform Project. Gupta, S., M. Verhoeven, and E. R. Tiongson. 2001. "Public Spending Laffont, J-J. 1990. The Economics of Uncertainty and Information. on Health Care and the Poor." Working Paper 01/127, International Cambridge, MA: Massachusetts Institute of Technology Press. Monetary Fund, Washington, DC. Londoño, J-L., and J. Frenk Mora. 1997. "Structured Pluralism: Toward an Gwatkin, D., S. Rutstein, K. Johnson, E. A. Suliman, and A. Wagstaff. 2003. Innovative Model for Health System Reform in Latin America." Health Initial Country-Level Information about Socioeconomic Differences in Policy 41 (1): 1­36. Health, Nutrition, and Population. Vols. I and II. Washington, DC: Maeda, A. 1998. "A Model for the Evolution of Health Systems." World Bank. Presentation prepared for the World Bank, Washington, DC. Haines, A., and A. Cassels. 2004. "Can the Millennium Development Goals Mathers, C. D., C. Stein, D. M. Fat, C. Rao, M. Inoue, N. Tomijima, and Be Attained?" British Medical Journal 329: 394­97. others. 2002. "Global Burden of Disease 2000: Version 2 Methods and Heller, P. 2005. "Fiscal Space-What It Is and How To Get It." Finance and Results." Global Programme on Evidence for Health Policy Discussion Development 42 (2). Paper 50, World Health Organization, Geneva. Hemming, R., M. Kell, and A. Schimmelpfennig. 2003. "Fiscal McLaughlin, J. 2003. "Accelerating Progress toward the Health MDGs: Vulnerability and Financial Crisis in Emerging Market Economies." Important Lessons Learned from Development Assistance." Paper Occasional Paper 218, International Monetary Fund, Washington, prepared for the World Bank, Washington, DC. DC. ------. 2004. "The Evolution of the Sectorwide Approach (SWAp) and Hensher, M. 2001. "Financing Health Systems through Efficiency Gains." Explaining the Correlation between SWAps and Reform Initiatives." Commission on Macroeconomics and Health Working Paper Series Paper prepared for the World Bank, Washington, D.C. WG3:2, World Health Organization, Geneva. Milgrom, P., and J. Roberts. 1992. Economics, Organization, and Manage- Hsiao, W. C. 1994. "`Marketization': The Illusory Magic Pill." Health ment. Englewood Cliffs, NJ: Prentice-Hall. Economics 3 (6): 351­57. Musgrove, P. 1996. "Public and Private Roles in Health: Theory and ------. 1995. "Abnormal Economics in the Health Sector." Health Policy Financing Patterns." Health, Nutrition, and Population Discussion 32: 125­39. Paper, World Bank, Washington, DC. Financing Health Systems in the 21st Century | 241 Musgrove, P., R. Zeramdini, and G. Carrin. 2002. "Basic Patterns in Diseases." Paper presented at UNECA Special Summit on HIV/AIDS, National Health Expenditures." Bulletin of the World Health Tuberculosis, and Other Related Infectious Diseases, Abuja, Nigeria, Organization 80 (2): 134­46. April 27. http://www.uneca.org/adf2000/Abuja%20Declaration.htm. Newhouse, J. 1998."Risk Adjustment: Where Are We Now?" Inquiry 35 (2): United Nations Millennium Project. 2005. UN Millennium Project 2005: 122­31. Investing in Development--A Practical Plan to Achieve the Millennium Oliviera-Cruz, V., K. Hanson, and A. Mills. 2003. "Approaches to Development Goals. New York: United Nations. Overcoming Constraints to Effective Health Service Delivery: A Wagstaff, A., and M. Claeson. 2004. The Millennium Development Goals for Review of the Evidence." Journal of International Development 15 (1): Health: Rising to the Challenges. Washington, DC: World Bank. 41­65. WHO (World Health Organization). 2000. The World Health Report 2000: Palmer, N., D. H. Mueller, L. Gilson, A. Mills, and A. Haines. 2004. "Health Health Systems--Improving Performance. Geneva: WHO. Financing to Promote Access in Low Income Settings--How Much ------. 2001. Report of the Commission on Macroeconomics and Health: Do We Know?" Lancet 364 (9442): 1365­70. Investing in Health for Economic Development. Geneva: WHO. Rajkumar, A. S., and V. Swaroop. 2002. "Public Spending and Outcomes: ------. 2002. The Report of Working Group 3 of the Commission on Does Governance Matter?" Policy Research Working Paper 2840, Macroeconomics and Health. Geneva: WHO. World Bank, Washington, DC. ------. 2004a. Poverty Reduction Strategy Papers: Their Significance for Rawlings, L. 2004. "A New Approach to Social Assistance: Latin America's Health--Second Synthesis Report. Geneva: WHO. Experience with Conditional Cash Transfer Programs." Social Protection Discussion Paper 0146, World Bank, Washington, DC. ------. 2004b. World Health Report 2004: Changing History. Geneva: WHO. Schieber, G., ed. 1997. Innovations in Health Care Financing. Washington, DC: World Bank. Williamson, O. E. 1985. The Economic Institutions of Capitalism. New York: Free Press. Schieber, G., and A. Maeda. 1997. "A Curmudgeon's Guide to Financing Health in Developing Countries." In Innovations in Health Care World Bank. 1993. World Development Report 1993: Investing in Health. Financing, ed. G. Schieber, 1­40. Washington, DC: World Bank. New York: Oxford University Press. SHC Development Consulting. 2001. Sector Programmes and PRSP ------. 2003. World Development Report 2004: Making Services Work for Implementation: Chances and Challenges. Kassel, Germany: SHC Poor People. New York: Oxford University Press. Development Consulting. ------. 2004a. "Aid Effectiveness and Innovative Financing Mechanisms." Sheshinski, E., and L. F. López-Calva. 1998. "Privatization and Its Benefits: Report prepared for the 2004 Annual Meetings of the International Theory and Evidence." Consulting Assistance on Economic Reform II Monetary Fund and the World Bank, Washington, DC, October 2­3. Discussion Paper 35, Harvard University, Cambridge, MA. ------. 2004b. Global Development Finance: Harnessing Cyclical Gains for Tait, A. 2001. "Mobilization of Domestic Resources for Health through Development. Washington, DC: World Bank. Taxation: A Summary Survey." Commission on Macroeconomics and ------. 2004c. Poverty in Mexico: An Assessment of Conditions, Trends, and Health Working Paper WG3:14, World Health Organization, Geneva. Government Strategy. Washington, DC: World Bank. Tanzi, V., and H. H. Zee. 2000. "Tax Policy for Emerging Markets: ------. 2004d. The Poverty Reduction Strategy Initiative: An Independent Developing Countries." Working Paper 00/35, International Monetary Evaluation of the World Bank's Support through 2003. Washington, DC: Fund, Washington, DC. World Bank, Operations Evaluation Department. Travis, P., S. Bennet, A. Haines, T. Pang, Z. Q. Bhutta, A. A. Hyder, and ------. 2004e. 2004 World Development Indicators. Washington, DC: others. 2004. "Overcoming Health Systems Constraints to Achieve World Bank. the Millennium Development Goals." Lancet 364: 900­6. UNECA (United Nations Economic Commission for Africa). 2001. "Abuja Declaration on HIV/AIDS, Tuberculosis, and Other Related Infectious 242 | Disease Control Priorities in Developing Countries | George Schieber, Cristian Baeza, Daniel Kress, and others Chapter 13 Recent Trends and Innovations in Development Assistance for Health Robert Hecht and Raj Shah After nearly a decade during which levels of external develop- Fight AIDS, Tuberculosis, and Malaria and special U.S. financ- ment assistance for health (DAH) stagnated, an encouraging ing for HIV/AIDS, plus rapid growth in grant awards from the rise has occurred in the volume of such assistance. Donors and Bill & Melinda Gates Foundation and in World Bank developing countries are testing and implementing innovative International Development Association (IDA) grants. Com- approaches to the use of DAH, while simultaneously seeking mitments from all external sources, including foundations, rose ways to raise the effectiveness of existing streams of aid and from an annual average of US$6.7 billion in 1997­99 to about more traditional financing mechanisms. In short, DAH has US$9.3 billion in 2002. entered a dynamic phase that holds considerable promise. Total DAH is the sum of external financing for health from Nevertheless, it continues to suffer from a broad range of several different sources: bilateral agencies as reported through disappointments: misuse and inefficiency in the deployment of the creditor reporting system of the Organisation for Economic funds, gaps in essential areas that require financing support, Co-operation and Development (OECD); multilateral agen- and weaknesses in institutional and management arrange- cies, including the United Nations (UN) system--especially the ments. Substantial room for improvement exists. World Health Organization (WHO), the United Nations This chapter documents those recent trends, analyzes the Children's Fund, the United Nations Population Fund, and the effects and assesses the performance of DAH, and points to global and regional development banks; the European Union; areas that require priority attention. In the first part, we present philanthropic organizations; and the Global Fund to Fight statistics on DAH, updating the World Development Report AIDS, Tuberculosis, and Malaria. Because no central repository 1993 (World Bank 1993) and the report of the Commission on of data on all the sources of DAH is currently available and Macroeconomics and Health (CMH 2001). In the second part, comprehensive information is not published on any regular we assess the performance of DAH. In the third part, we present basis, painstaking and time-consuming efforts are required to recent innovations to underscore the current dynamic nature assemble accurate, comparable data about all these sources. of such assistance. The chapter concludes with some sugges- After a long period of decline in official development assis- tions on future directions. tance (grants from bilateral government channels and UN agencies plus net flows from development banks) during the TRENDS AND GAPS IN DEVELOPMENT 1990s, the OECD reported a real increase of 7 percent from ASSISTANCE FOR HEALTH 2001 to 2002 and a further increase of 4 percent from 2002 to 2003. Those increases took official development assistance to an Despite a decline in overall official development assistance in all-time high, in both nominal and real terms, of US$68.5 bil- the 1990s, DAH rose in real terms and as a proportion of offi- lion. As a percentage of gross national income, this represents cial development assistance (table 13.1). New funding sources an increase from the all-time low of 0.22 percent recorded became available in 2000­2, including the Global Fund to during most years from 1997 to 2001 to about 0.25 percent in 243 Table 13.1 Development Assistance for Health, Selected Years health in 2003, about US$1 billion took the form of IDA cred- (US$ millions) its, and most of the rest was in the form of loans that reflected the costs of borrowing. Because the face value of the financial Annual average, commitment can be considered to be reduced by repayments in Source 1997­99 2002 the case of subsidized and market rate loans, some argue that Bilateral agencies 2,560 2,875 the net financial value of such loans, rather than their face Multilateral agencies 3,402 4,649 value, should be used in calculating DAH. This calculation fur- European Commission 304 244 ther complicates the task of monitoring DAH. Global Fund to Fight AIDS, 0 962 Despite these various cautions and qualifications, it is clear Tuberculosis, and Malaria that DAH has grown in recent years. This upward trend has Bill & Melinda Gates Foundation 458 600 been driven by several factors, including (a) donors' increasing Total 6,724 9,330 attention to the challenges presented by the Millennium Development Goals (MDGs), which are heavily centered on Sources: Michaud 2003; OECD 2004a. maternal and child health and control of communicable dis- 2003, still well below the target of 0.7 percent set by the OECD's eases; (b) strong global mobilization to confront the AIDS pan- member states in 1970. Only five countries--Denmark, demic in developing countries since 1998­99, especially in Luxembourg, the Netherlands, Norway, and Sweden--currently Africa; and (c) donors' expanding interest in research and achieve this target, and six others have now set prospective dates development (R&D) in relation to new health technologies to for its achievement--namely, Belgium (2010), Finland (2010), address the major diseases prevalent in poor countries. In con- France (2012), Ireland (2007), Spain (2012), and the United trast, external funding for health system development, human Kingdom (2013). resources, and noncommunicable diseases has increased more Bilateral assistance for health rose from an annual average of slowly. US$2.2 billion (3.8 percent of the total) during 1997­99 to In terms of the areas that have benefited from the growing US$2.9 billion (6.8 percent) in 2002. Among the bilateral volume of DAH, three stand out: HIV/AIDS, immunization, arrangements, the United States accounted for about 40 per- and new health product development. According to Michaud cent of the total, even though as a percentage of gross domestic (2003), in 2002 about US$900 million in DAH was for product (GDP), its allocation to international development HIV/AIDS, followed by US$210 million for tuberculosis (TB), was among the lowest of all the high-income countries. and US$160 million for malaria control. The Joint United Within the UN system, DAH rose from an average of Nations Programme on HIV/AIDS (UNAIDS) also reports a US$1.6 billion per year during 1997­99 to US$2 billion in substantial rise in external financing for AIDS prevention, treat- 2002. Commitments from the development banks remained ment, and social mitigation activities over the past four years stationary at about US$1.4 billion. However, changes in (UNAIDS 2004). Most of the increase in assistance for immu- accounting by the World Bank to include financing for health nization has taken place through the Global Alliance for activities contained in projects managed by other sectors (such Vaccines and Immunization (GAVI), which has amassed com- as urban, water and sanitation, transportation, and social mitments of about US$1.3 billion to finance the expansion of development), suggest that its new commitments for health existing childhood immunization programs and the accelerated actually rose from about US$1 billion in 2001 to US$1.3 billion introduction of hepatitis B and Haemophilus influenzae type B in 2002 and US$1.7 billion in 2003. vaccines. Assistance for health technologies directed at diseases In the future, consensus will need to be reached on whether that are prevalent in the developing world has been channeled allocations by the multilateral development banks to projects in through new public-private partnerships. Examples include the other sectors or to projects that are classified as multisectoral-- International AIDS Vaccine Initiative (IAVI), the Medicines for especially broad budget support to governments, which may be Malaria Venture, and the International Partnership for specifically tied to domestic spending and policy reforms in Microbicides. Estimates indicate that the 10 largest public- health--should be counted as DAH. Another issue in DAH private partnerships have raised more than US$1 billion over accounting involves distinguishing between allocations for the past five years (IPPPH 2004). health from the multilateral banks that take different forms-- In terms of the sources of the expanded volume of DAH, a namely, outright grants (a recent innovation for the World small number of institutions account for much of the recent Bank and the regional banks); subsidized loans for the poorest increase. Among the traditional donors, these institutions countries, which at the World Bank are IDA credits; and loans include the World Bank and the governments of Canada and for the middle-income developing countries that reflect the the United Kingdom. At the same time, as a share of GDP, actual costs of borrowing by the development banks. For exam- contributions from Ireland, the Netherlands, Norway, and ple, of the US$1.7 billion in World Bank commitments for Sweden remain high. Among nontraditional sources, the Bill & 244 | Disease Control Priorities in Developing Countries | Robert Hecht and Raj Shah Melinda Gates Foundation stands out as a major new player as come from external sources (UNAIDS 2001). At the 2004 of the late 1990s. With a focus on the development of new International AIDS Conference in Bangkok, UNAIDS raised its drugs, vaccines, and diagnostics for the developing world, the estimate of resources needed to more than US$15 billion a year Gates Foundation's commitments for health started in 1998 by 2010 (UNAIDS 2004). and rose rapidly to some US$600 million in 2002, with annual These global calculations have been followed by more commitments expected to approach US$1 billion in 2004 and detailed costing exercises at the country level, which hold the beyond. promise of yielding more accurate and meaningful figures than Development assistance for health is channeled to a large the global estimates. Donors working with government special- number of low- and middle-income developing countries, but ists in developing countries have tested a variety of methods. the largest recipient region is Africa. In 2002, about 35 percent of The UN Millennium Project has used a bottom-up approach, all such assistance went to Africa, followed by Latin America and which is based on expanded coverage of key interventions and the Caribbean with around 14 percent, East Asia and South Asia fixed unit costs, assuming no shared costs or benefits among with 11 percent each, and the Middle East with 7 percent. The different interventions, omitting the possibility of private remaining 22 percent was for global programs (Michaud 2003). financing, and adding a rough amount for system improve- This growth in funding for the control of communicable ments (UN Millennium Project 2004). The World Bank has diseases and new health technologies to address them is impor- followed two other approaches. One, in India, is based on tant, given the high burden of illness and premature death observed elasticities of change in children's health and nutri- these diseases cause. Nevertheless, the focus on AIDS, TB, and tion outcomes in relation to public expenditures on health, malaria should be matched by similar increases in investments primary education, water, and so on (World Bank 2003a). in broader health system improvements. Relatively little DAH is Another, in Ethiopia, Mali, and other countries, is based on being channeled to address the serious problems of shortages detailed modeling of the costs of removing bottlenecks in health in health workforces in poor countries and their low produc- service delivery to enhance the coverage, utilization, and quali- tivity or to deal with weaknesses in health management ty of key health interventions proven to have a positive effect on information systems, in supply chain logistics for drugs and maternal and child health outcomes (Soucat and others 2003; commodities, and so on. Even though focused spending on World Bank and Ministry of Health, Ethiopia 2005). AIDS, TB, and malaria will clearly touch on these weaknesses, A comparison of the Millennium Project's and the World it will not on its own go to the core of the problem or lead to Bank's results for East Africa is interesting. The Millennium sustainable solutions. For example, GAVI allocates resources to Project calculates that nearly US$30 per capita are needed in strengthen immunization infrastructure such as cold chains additional spending for health, whereas the World Bank calcu- and to train health workers to deliver vaccinations more lates that about US$4 per capita are needed for Ethiopia to effectively. The Global Fund to Fight AIDS, Tuberculosis, and reduce child and maternal death rates by 30 to 40 percent by Malaria provides funds to prepare health workers to deliver and 2015. The large difference between the two sets of results sug- monitor compliance with antiretroviral treatments. Useful as gests that more work needs to be done to move toward con- those activities are, they will not address the underlying weak- sensus on the best methodology for countries to use. nesses in human resources for health in poor countries, such Part of the difference is due to technical factors. The as low levels of pay, unattractive conditions of service, and Millennium Project approach covers all the health MDGs, uncertain prospects for career advancement. whereas the bottlenecks method has focused on the MDGs The recent rise in DAH is encouraging, but it is still far short pertaining to child and maternal health only. The Millennium of the volume of external financing for health that is needed, Project also calculates costs to achieve the MDGs in their according to recent estimates and political pronouncements. entirety, whereas the bottlenecks method addresses incremental On a global level, estimates of what donors need to provide to improvements. For example, in relation to the child mortality help countries reach the MDGs for health have typically ranged goal, the bottlenecks analysis for Ethiopia considers a substantial from US$15 billion to US$35 billion per year. The Commission decline to be from 176 to 107 deaths per 1,000 live births, but the on Macroeconomics and Health suggested a figure of about MDG is 59 deaths per 1,000 births. In addition, the Millennium US$30 billion a year. While preparing for the Monterrey Project multiplies additional units of service by a standard cost Summit on Finance for Development, the World Bank calcu- per unit, whereas the bottlenecks method estimates the cost of lated a funding gap of US$15 billion to US$25 billion a year system improvements and then divides this amount by the addi- (Devarajan, Miller, and Swanson 2002). For the United Nations tional services rendered to derive incremental unit costs. General Assembly special session on AIDS in June 2001, The two approaches also have important differences in polit- UNAIDS suggested that spending on HIV/AIDS alone in the ical philosophy. The Millennium Project approach sets high developing countries needed to rise to about US$9 billion targets for DAH and health spending, which are based on full annually by 2005, with about two-thirds of this amount to achievement of the MDGs, regardless of the starting points and Recent Trends and Innovations in Development Assistance for Health | 245 gaps and without addressing the feasibility of reaching the tar- · collaboration across governments, donors, and nongovern- gets. The World Bank's approach is less ambitious but may be mental organizations (NGOs) in program design and seen as more realistic and as pointing the way to implementation implementation based on gradual improvements--improvements that countries · consistent, predictable funding support, even after success can pursue as additional financial resources and capacity to has been achieved manage them effectively are combined on the ground. · simple and flexible technologies that can be adapted to local conditions and do not require complex skills to operate and maintain MAKING DEVELOPMENT ASSISTANCE FOR · programmatic approaches that recognize and address the HEALTH MORE EFFECTIVE: LESSONS LEARNED need to help build health system infrastructure, especially in human resources More assistance is part of the answer to helping developing · household or community participation in the design, exe- countries achieve more rapidly the improved health outcomes cution, and monitoring of program activities. they seek and that are enshrined in the MDGs. To this end, we need to know how effective DAH has been and what can be done to make it more effective. Policy Environment Despite valid criticisms of DAH, some health programs-- Development assistance for health supports a vast array of inspired and supported by donors--have worked at scale and activities and services, some focused on specific diseases (polio, contributed to more than four decades of steady improvements TB, HIV/AIDS), some on strengthening health systems (disease in health, as measured by under-five mortality and overall life surveillance, nurse and midwife training), and some on partic- expectancy. The record of public health successes in developing ular services (reproductive and child health services). But has countries is becoming increasingly clear, as noted in a recent DAH actually changed health outcomes? Recent work from the review of four decades of experience (Levine and What Works World Bank, the Commission on Macroeconomics and Health, Working Group 2004). The success stories cover a broad spec- and others suggests that it has (Rajkumar and Swaroop 2002). trum of circumstances. They are found in all regions and cover However, DAH does not work as effectively in countries where both communicable and noncommunicable diseases. They the policy environment is poor, even though some carefully have been driven by new technologies, including vaccines, targeted disease control activities can confer limited benefits. drugs, and diagnostics; community- and clinic-based care; and With good policies and institutions (strong property rights, knowledge for behavior change. reduced corruption, an efficient bureaucracy), an extra 1 per- Significant gains have occurred even in the poorest coun- cent of GDP in aid is estimated to reduce infant mortality by tries and in those with weak institutional environments. 0.9 percent. By contrast, where policies are average, the decline Consider, for example, the high levels of TB case detection in is estimated at only 0.4 percent, and where policies are poor, aid the Democratic Republic of Congo and in Myanmar (Stop TB is estimated to have no significant effect on infant mortality Partnership 2003), the successes in polio eradication in African (World Bank 2004b). countries experiencing civil wars, and the growing availability The issue is not black and white: there are gradations of of antiretroviral therapy in Haiti. good policy, and as policies improve, the productivity of aid Some of this progress can be attributed to the general effects increases. For example, Bangladesh has made large strides in of economic growth and improvements in education, water, reducing under-five mortality in recent years, relying on NGOs and sanitation. However, specific, compelling examples of the to deliver many services. If Bangladesh were able to raise the success of DAH-backed initiatives are available. For example, quality of its governance from below average to above average, programs to immunize against measles, to control river even at its same public spending levels, it would realize more blindness and guinea worm, and to fortify salt with iodine have rapid gains. An additional dollar of government health spend- had sustained and widespread effects (Levine and What Works ing would reduce under-five mortality by 14 percent, compared Working Group 2004). These programs have been successful at with 9 percent without such improvements (World Bank scale, have generated sustainable health improvements at the 2003c). population level, and have succeeded in a broad range of insti- Tradeoffs may be necessary between targeting assistance tutional environments. toward the neediest countries and achieving the greatest effect Recent analyses and reviews of donor-supported successes from DAH, but even in needy countries with weak policies, in international health have noted a set of factors that tend to some kinds of carefully targeted assistance for health (for contribute to positive outcomes: example, immunizations delivered by reputable NGOs) can · strong internal (governments) and external (donors) politi- have a positive effect. In addition, in countries with weak cal leadership policies, a focus by donors on policy dialogue and technical 246 | Disease Control Priorities in Developing Countries | Robert Hecht and Raj Shah assistance to improve the environment for DAH can set the This effort set the stage for a large inflow of donor financing, stage for a larger infusion of financial support down the road. starting around 1995 and continuing to the present. Conditionality Fungibility of Development Assistance for Health Conditionality is making the availability of funding dependent Much aid is earmarked, both across sectors and within them. on a government completing an agreed task, such as enacting a One part of a development agency gives a grant to the ministry new health law or spending a certain share of its budget on of health for a health sector reform, while another does the health activities. It can work, but only if the government is same for a primary education project. An agency makes a loan committed to making such a change. Tying aid to policy to the ministry of health for a TB project, while another makes changes is a common practice, but recent studies have cast a loan for a malaria control project. The donors' intent is that doubt on the ability of such conditionality to bring about these activities remain tightly sealed: the funds for health sector reform (World Bank 1998, 2003a). If governments are not reform are to be kept separate from the funds for the primary committed to reform, conditionality will not make them education project; the TB project funds are to be kept separate reform. Donors themselves often undermine the rigor and from the malaria control project funds. The idea is to ensure credibility of conditions because they usually face strong inter- that the government makes a certain spending choice. It is nal pressures to continue disbursement of the funds anyway, based on the assumption that the choice would not be made if even when governments do not adhere to the agreed-on condi- the government had been handed a blank check for the same tions. On the other hand, if governments are committed to amount. reform, conditions can help by enabling governments to com- The implied view of such aid is that what you see is what mit publicly to certain reforms and persuade private investors you get; that is, a government receives US$1 million for a water of their seriousness. For example, the government of Uganda's project and the net effect is US$1 million worth of extra commitment to decentralizing the management of basic health spending on the water sector. This view has recently been chal- services and to making local authorities accountable to com- lenged, with the alternative view being that aid is at least part- munities was reinforced by conditions in the Uganda poverty ly fungible (Burnside and Dollar 2000; World Bank 1998). reduction support credit, which several donors financed. Hence, as a result of the inflow of development assistance for a Similarly, the Chinese government's commitment to reaching specific health activity, the government changes the way in the poor with TB control services was reinforced by stipula- which it spends the rest of its resources, both in the health sec- tions in donor-funded TB projects that they target the coun- tor and in terms of allocations between health and other sec- try's poorest provinces and reach out to deprived households tors. As a result, for each dollar earmarked for a specific health (World Bank 2003c). project, spending on health rises by less than a dollar but by Donors cannot force policies on governments, but they can more than would have been the case had the government help with policy design. Donors can alert governments to the received an extra dollar in its overall budget. Similarly, spend- reasons for reform and help nurture commitment, but at the ing for the specified purpose in the health sector rises by less end of the day, it is governments that have to sustain any than a dollar, but by more than would have been the case had reforms (box 13.1). Undertaking analytical work, providing the government allocated an extra dollar of its own resources training and technical assistance, disseminating ideas about to health generally. policy reform and development, and stimulating debate in civil Assessing whether aid is indeed fungible is not straightfor- society can all be valuable activities for donors to support while ward. The difficulty is knowing how the government would a government's commitment to reform is growing. have responded had its own resources increased by the amount Vietnam in the late 1980s and early 1990s is a good example. of the aid or had it received a blank check for the same amount. At a landmark meeting in 1986, the ruling Communist Party Recent research suggests that, despite considerable variation decided to break with the past and introduce sweeping eco- across countries, on average only 29 cents of each additional nomic reforms. In the health sector, the reforms included dollar of aid goes into government development programs, introducing user fees at public facilities, legalizing private med- with the rest leaking out into nondevelopment programs such icine, deregulating the pharmaceutical industry, and opening as military spending (World Bank 2004b). the pharmaceuticals and medical equipment subsectors to One important implication of those findings is that donors international trade. Initially, Vietnam saw no increase in aid, should spend less time and effort trying to channel their exter- but such agencies as the United Nations Development nal funding to specific programs for priority diseases and Programme and the World Bank helped facilitate the reform populations. Instead, a more useful exercise would be to engage process by organizing international workshops for the in a dialogue with the government on basic changes in the Vietnamese to exchange ideas on policy with their neighbors. overall patterns of public spending for health--that is, the total Recent Trends and Innovations in Development Assistance for Health | 247 Box 13.1 Donors and Commitment: Nutrition in Bangladesh and Thailand In most countries, nutrition has not become a visible issue between the government and the NGOs involved in on the national political agenda, because nutrition advo- community activities have complicated the situation. cates have not succeeded in linking improved nutrition with Contradictory messages from donors and frequent changes political and economic goals or in creating popular demand in leadership within the government have added to the to eliminate malnutrition. In Bangladesh in the early 1990s, challenge. the United Nations Children's Fund and the World Bank More recently, donors and advocates for nutrition joined forces to present a case to the government showing within the government have proposed that nutrition how the country could not achieve its economic goals activities that build on earlier successes be included in the unless it reduced malnutrition. This effort persuaded the Health, Nutrition, and Population Sector Program, which government's financial planners that funding a national is scheduled to be finalized and approved in early 2005. nutrition program was a good investment, and the govern- In contrast, in Thailand, building commitment for ment approved a new nutrition project in 1995. nutrition was achieved during the 1980s and nurtured However, the issue is not just how to build initial com- with little external support. Government-sponsored mitment, often the main focus of organizations such as the efforts through studies, workshops, and media outreach World Bank. Commitment can be fragile, and the issue is generated commitment for nutrition by building broad how to broaden and maintain commitment and comple- consensus (in the government, NGOs, and the private ment it with systematic investments in institutional capac- sector) on the benefits of nutrition--not as a welfare issue, ity development. The first nutrition investment in but as a human development issue. This initial commit- Bangladesh was completed in 2001. Children's nutritional ment was sustained by ensuring that policy statements status and households' health-seeking behaviors improved were closely linked to national investment plans, by build- substantially in project areas, and malnutrition rates ing strong technical and managerial capacity for nutrition declined. A follow-on nutrition investment was approved, in the country (often by means of external aid), and by but because of weak government commitment, it is strug- linking those actions with a strong buy-in and demand gling in the challenging policy environment in which the from communities. Malnutrition rates in Thailand social sectors operate. The Ministry of Health has not declined from 51 percent in the early 1980s to 18 percent assigned high priority to the program, and conflicts in 1990 and continue to fall. Source: Heaver 2002. allocation and the amounts for, say, providing child health and donors' ability to earmark their funding effectively, many high- communicable disease control services and for improving impact health services can be delivered to the population on a community and primary-level health delivery systems. If the targeted basis even when national policies and institutions are government followed through on those basic changes, then weak--and this aid would not occur in the absence of DAH. donors would transfer their financial assistance to the health This argument applies especially to services with simple sector as a whole. technologies--for instance, basic childhood vaccinations that The finding that aid is indeed fungible has encouraged some can be delivered on a single occasion through annual cam- donors to search for broader development assistance mecha- paigns or disease treatment programs (such as short-course nisms that recognize the importance of the entire expenditure drug therapy for TB) that can be provided through tightly program and explicitly avoid earmarking. Such mechanisms managed top-down efforts (Jamison 2004). The high coverage include the Multi-Country AIDS Program in Africa, which and treatment success using the directly observed short-course supports national HIV/AIDS strategies; sectorwide approaches therapy approach for TB in the Democratic Republic of in health; and poverty reduction support credits (PRSCs) that Congo in recent years, even during the civil war, can be cited as back a broad public spending agenda. an example of how a well-protected enclave project with An opposing viewpoint is that although, in general, good strong donor backing can be successful (Stop TB Partnership policies matter and the fungibility of aid tends to undermine 2003). 248 | Disease Control Priorities in Developing Countries | Robert Hecht and Raj Shah Another argument against broad budget support of health funding will respect the government's overall spending sector funding or in favor of the more traditional earmarking plans and limits of DAH is that it helps maintain governments' and donors' · pooling donor funds in a single account and untying aid so focus on implementing and monitoring specific health serv- that the government can procure goods and services from ice interventions and on the necessary technical and mana- the lowest-cost source and not just from the donor gerial improvements to ensure the achievement of targets in countries those areas. On the basis of experience in countries such as · limiting the number of country coordination bodies that Bangladesh, Ghana, Mozambique, and Zambia, donors are can bring together national and international actors increasingly of the view that broad support to a national health involved in health. sector program leads to superficial oversight of bureaucratic processes and a corresponding loss of technical focus and At the same time, some of the experiments in country-level depth (Foster, Brown, and Conway 2000). Although a sector- coordination of DAH reveal the difficulties of implementing wide approach is theoretically fully compatible with careful the principles of better donor harmonization. In some monitoring of key outputs and health outcomes and with in- instances, persuading donors to pool their funds and sever the depth technical improvements, in practice, achieving this mix links between funding and procurement has proven difficult. of objectives may prove difficult. Some donors face pressure from their legislatures to maintain In summary, the debate on earmarked versus broad DAH separate accounting for the use of their DAH allocations so support for national health programs continues. More analysis they can claim credit for achieving progress and thus "plant is needed to produce clearer conclusions on the advantages and flags" on individual health projects. In addition, monitoring disadvantages of the two approaches. and evaluation systems have frequently not been strong enough to yield timely and meaningful data on progress, a critical fail- ure if disbursements are linked to performance rather than to Transaction Costs of Aid spending on specific inputs. Multiple national coordination In a single low-income country, more than 20 donors-- bodies for government, donors, and NGOs for different dis- including bilaterals, multilaterals, global programs, founda- eases (AIDS, TB) and services (immunization, polio eradica- tions, and large NGOs--may be involved in the health sector. tion) also persist in many settings. In short, a number of polit- The demands placed on recipient countries can be huge, and ical and technical changes are needed to ensure the successful donors are starting to acknowledge this burden. They are rec- implementation of a harmonized donor agenda at the country ognizing that their individual procedures for reporting, level in the poorest developing nations. accounting, and managing funds--which often encompass dif- ferent budget structures; different ways of measuring progress toward objectives; different regulations for the procurement of Unpredictability of Development Assistance for Health goods, services, and works; and different approaches toward Some donors have taken steps to put in place instruments for and cycles for disbursing funds--place heavy and unreasonable DAH that extend the length of their financial commitments. demands on recipient countries. Demands are particularly The development of multiphase funding "slices" of 3 to 5 years heavy in poor countries that are forced to allocate limited embedded in a 10- or 15-year program of support is one way human resources away from service delivery to manage donor to lengthen commitments. Nevertheless, donor financing for funding. health is not yet as reliable or sustained as is often claimed or The donor community is working to harmonize and hoped for, even under these new, long-term arrangements. In simplify its procedures to reduce these transaction costs. In the some developing countries, cuts in DAH have been sharp. health sector, experiments are taking place in several develop- Donors' budgets are subject to the usual business and political ing countries, including Bangladesh, the Kyrgyz Republic, cycles and may go up or down during their annual budgetary Mozambique, and Zambia, to determine how best to lower the processes. For such countries as the Comoros and Eritrea, costs (OECD 2004b). Some of the principles of improved where the year-to-year changes in external funding can amount donor action include the following: to as much as a fifth of all public spending for health, the fluctuations are so great that they make planning and imple- · ensuring that countries, not donors, drive the coordination menting coherent national health programs nearly impossible · fostering strategic coherence through a poverty reduction (figure 13.1). strategy and the health, nutrition, and population analyses Further work is needed to design mechanisms for DAH that feed into it that provide greater assurance of sustained financial support. · promoting financial coherence through a medium-term The challenges are to overcome the factors that result in inter- expenditure framework and an agreement that all donor ruptions to long-term DAH. Those factors include changes in Recent Trends and Innovations in Development Assistance for Health | 249 Percentage of total health expenditures large scale--a series of innovative approaches and instruments 60 to improve the effectiveness of DAH. Those innovations include the use of broad budgetary support to countries with 50 strong governance and institutions, the implementation of sec- torwide approaches in health, the use of performance-based 40 financing mechanisms, a shift to direct engagement with the private sector, and the implementation of programs designed 30 to move resources expeditiously to the frontlines of the battle for improved health (that is, to communities). Evidence on the 20 effectiveness of those innovative approaches and the conditions under which they tend to work is starting to accumulate. 10 Budget Support in Strong Policy Environments 0 1995 1996 1997 1998 1999 2000 In low-income settings where policies, governance, and institu- Burundi Comoros Eritrea Ethiopia tions are sound, donors have increasingly sought to provide Somalia Tanzania Uganda broad, untied, and flexible budget support to governments to help support a full public expenditure program aimed at rais- Source: World Bank 2004b. ing the level of spending and the effectiveness of resource use Figure 13.1 External Financing as a Percentage of Total Health for health. Frequently, this support has taken the form of PRSC Expenditures, Selected Countries, 1995­2000 operations, including grants and credits. The PRSC is typically built on the foundations of a national poverty reduction strat- egy that analyzes the links between poor health outcomes and political leadership and aid agency management that lead to income poverty and identifies policies that can improve the modifications of earlier agreements and end up reducing health of the general population, especially that of poor house- external funding levels or reallocating those funds to other holds. The policies are then used to design a medium-term activities. public expenditure program or framework that, in turn, is One example of an innovative approach is the recently pro- backed by external funding from donors in the form of a PRSC. posed establishment of an international finance facility. This This approach draws on three of the key lessons from facility would use financial commitments from governments of decades of experience with DAH: developed countries to tap funds in capital markets and would use those funds to frontload development assistance so as to · A good policy environment improves the use of external accelerate progress toward the MDGs. The International financing. Finance Facility for Immunization, which the United Kingdom · The fungibility of DAH makes it logical to allocate external formally announced at the World Economic Forum in January funds to a general budget that prioritizes health rather than 2005 and is aiming to launch later in the year, will expand to narrow projects in the health sector. external financing for childhood vaccinations. The pilot project · An integrated system for managing public finance for health could be used to pledge funding against multiyear advance- improves national ownership of policies and programs to purchase contracts for new vaccines (such as rotavirus vaccine) improve the health of the poor and raises the chances that that may not reach the market for several years. Such an effort such funding will be sustained over a long period and could help create a more assured market and reduce the risks eventually will use domestic resources. for vaccine companies, thereby speeding up the introduction of these new health commodities to low-income countries (GAVI Plenty of examples are now available of the use of DAH for 2004b). PRSCs that focus on improvements in health. The earliest health-oriented PRSCs were in Mozambique and Uganda, fol- lowed since by similar operations in Benin, Mauritania, and RECENT INNOVATIONS TO IMPROVE THE other low-income African countries. In Mauritania, for exam- EFFECTIVENESS OF DEVELOPMENT ASSISTANCE ple, the country received a transfer of US$25 million in exter- FOR HEALTH nal financing to back a public spending plan that doubled health spending from about US$8 per capita in 2000 to US$16 In recent years, donor agencies working with developing coun- per capita in 2004. The plan also emphasized increasing health tries have been testing--and in some cases rolling out on a investments that are designed to lower maternal and child 250 | Disease Control Priorities in Developing Countries | Robert Hecht and Raj Shah deaths and combat communicable diseases by expanding rural months of program execution, more than 100,000 eligible health facilities, by providing higher pay and other incentives women and children joined the insurance scheme. for health personnel working in rural areas, and by improving the availability of drug supplies at lower-level facilities. External Pooling and Donor Harmonization financing was not earmarked for these actions in the health sec- As mentioned earlier, another innovation in recent years has tor. Instead, the government committed itself to spending for been the use of sectorwide approaches as a way for multiple these purposes from a consolidated national budget that was donors to pool their funds for a commonly agreed-on program closely monitored by civil society, government officials, and and to use similar, streamlined procedures for procurement, donor representatives (World Bank 2004a). The early results monitoring and evaluation, and reporting. Sectorwide from Mauritania are encouraging. Spending for health has approaches grew out of sector investment programs for health risen, with most of the increase going to those parts of the that were launched in the early 1990s as a way to bring donors country and for those kinds of services likely to have the largest together to support broader government objectives in health. effect on the health of the poorest households. The main features of sectorwide approaches are as follows: In middle-income countries, an analogous shift of DAH in strong policy environments has been the increasing use of a. a partnership among a broad coalition of donors, with the single-tranche, programmatic, sector adjustment loans. These government taking the lead; loans have emerged as a favored instrument for DAH in certain b. a comprehensive sector policy framework to achieve goals Latin American countries with sound management of public over the short and medium terms; finances and internationally accepted procurement practices. c. an agreed-on expenditure program; Unlike the PRSCs, the programmatic sector loans have tended d. the improvement of management systems and capacity to target a single sector (such as health) or, occasionally, two building (Swedish International Development Cooperation sectors (such as health and education in the case of Brazil or Agency 2003). health insurance and pensions in the case of Ecuador). Whereas the PRSCs have their analytical roots in poverty stud- The main difference between sectorwide approaches and ies, the programmatic sector loans tend to be based on sector PRSCs is that, in the former, pooled donor funding is disbursed assessments. After the government has taken key legal, institu- against specific expenditure items--for example, construction tional, and spending actions to improve the efficiency of health of health facilities or purchase of drugs--whereas in PRSCs, spending or to target services for poor households, donor donor funds are transferred to the general budget, with dis- funds are transferred in a block or tranche to the government. bursements triggered by policy actions. In 2003, the World Bank approved four programmatic sector A prime example of a sectorwide approach is the Ghana loans--for Brazil, Colombia, Ecuador, and Peru--totaling health sector support program, in which 17 donor organiza- US$900 million (D. Cotlear, personal communication, tions have committed US$442 million over a five-year period to December 12, 2003). improve the health status of the population while focusing Another recent example is the World Bank's US$750 million efforts on reducing inequalities in health. The program includes Maternal and Child Insurance Program sector adjustment loan the following main spheres of action aimed at strengthening to Argentina, which followed decisions made by that govern- priority health interventions: developing human resources for ment in 2003 to create a mother and child health insurance health services, enhancing infrastructure and support services, scheme for poor provinces, to increase spending for communi- fostering partnerships for health, improving regulation, cable diseases, and to establish a national health council to set reforming organizational arrangements, improving health policies on the sharing of revenues to be used for health sector financing, enhancing financial management systems, between the central government and the provinces. The central strengthening management information systems and perform- pillar of this project, as well as the follow-on operation in 2004, ance monitoring, and linking with traditional medicine. is the implementation of the Maternal and Child Insurance Program. It delivers a publicly financed package of essential Performance-Based Financing services to uninsured mothers and children at the provincial Developing countries and their international partners are level. The donor funding is used in an innovative way to provide increasingly adopting methods for financing health care matching grants from the national to the provincial level, on the activities that link the availability of funding to concrete, basis of a capitated payment per mother and child enrolled plus measurable results on the ground. Such performance-based additional transfers to the province for performance. That per- financing was advocated a decade ago in the 1993 World formance is measured in terms of key health service goals (for Development Report (World Bank 1993) and in other policy example, coverage of vaccine programs, incidence of low birth- documents in the early 1990s, although relatively little practical weight, and number of prenatal consultations). In the first four knowledge of this type of financing was available at the time. Recent Trends and Innovations in Development Assistance for Health | 251 Since then, much more experimentation has taken place, a performance basis (Hecht 2004). Many of the earliest experi- and the important potential--as well as the challenges--of ments are from Latin America and the Caribbean. In Haiti, for performance-based financing for achieving national and global example, the government contracted NGOs to provide child health goals is becoming apparent. health and family planning services. The government gave the Performance-based financing is now being widely and NGOs an advance each year and then a quarterly sum based on actively tested at several levels of the health care system. These a negotiated budget. At the end of the year, performance was tests include situations in which the following occurs: measured against various indicators, including the extent of immunization coverage, the percentage of families using oral · Governments of developing countries pay health care rehydration to treat acute diarrhea, the share of pregnant providers in NGOs and the private sector to deliver essential women attending prenatal care, and the average waiting times health services to poor households. in clinics. The NGOs' performance determined the bonus they · Central governments determine the transfer of funds to received, which could be up to 10 percent of the original nego- local governments on the basis of their performance in tiated budget. As a result, the Haitian NGOs made changes in strengthening health services. their service delivery schemes and improved their performance · Donors release funding to recipients in developing in immunization and oral rehydration in particular (Eichler, countries as and when they achieve certain key health Auxilia, and Pollack 2001). In Guatemala, the government is targets. implementing a large performance-based program with NGOs that currently covers nearly 4 million people, mostly among the Performance-Based Contracts with Nongovernmental country's indigenous population (box 13.2). Similar schemes Organizations. A number of governments in low-income have been implemented in Argentina, El Salvador, and countries are funding NGOs to deliver basic health services on Nicaragua. Box 13.2 Large-Scale Performance Contracting with Nongovernmental Organizations in Guatemala Guatemala has successfully implemented contracting on a The contracting system under the program appears to large scale with NGOs to deliver health services. The gov- have resulted in important gains in health service delivery. ernment started the Program to Extend Coverage of Basic Immunization rates rose from 69 percent in 1997 to 87 per- Health Services in 1997, soon after the end of a long civil cent in 2001. Household surveys currently under way will war. The program has continued under successive admin- be able to assess the program's effect on mothers' and chil- istrations. By 2000, a total of 89 NGOs were involved in dren's health outcomes. providing health care to about 3.7 million people under During the program's early years, the government had 137 contracts. to overcome a number of obstacles. Government health The contracts specify a range of maternal and child workers resisted the scheme because they feared that con- health services, as well as the prevention and treatment of tracting with NGOs was a hidden form of privatization of a number of diseases, including malaria. The NGOs are government health services. The NGOs were initially paid about US$8 per person served, mostly in cash, but reluctant to get involved, because they thought that the also in kind in the form of such items as vaccines and government was demanding too much in the way of medicines. Payments are released quarterly after the improved performance and also doubted that the govern- NGOs' performance has been checked and verified. ment would pay them in a timely manner. Given the Performance is measured according to a series of indi- financial fragility of many local NGOs in Guatemala, the cators, including coverage of immunization and prenatal government had to make advance payments to the NGOs care, distribution of iron sulfate tablets to pregnant and release quarterly payments without delay to build women and to children, and frequency of home visits by confidence in the relationship between the public and pri- NGO outreach staff. The government has hired private vate sectors. firms to develop the monitoring system, which also looks at the NGOs' accounting practices. Source: Hecht 2004; World Bank 2000. 252 | Disease Control Priorities in Developing Countries | Robert Hecht and Raj Shah More recently, countries in South Asia have begun to enter of the activities run by the underperforming municipalities into performance-based health programs with NGOs. In (G. M. LaForgia, personal communication, October 21, 2003). Afghanistan, under a recently approved World Bank­financed project for health service rehabilitation, the government is con- Donor Disbursements to National Governments and Other tracting with NGOs to run health centers. NGOs that achieve Recipients. A number of innovative approaches are in place specified targets will be eligible to receive additional payments of that make donor financing of health programs conditional on up to 10 percent of their baseline subsidies from the government. successful performance on the ground. One example is the In a similar vein, the central and state governments in India World Bank's credit buy-down program for polio eradication have started to reimburse NGOs and private providers on the (box 13.3). basis of performance. The national TB program reimburses GAVI has also been a pioneer in the performance-based private laboratories for testing sputum samples to detect TB; it approach to grant assistance. Through its sister organization, also pays NGOs and private doctors a fixed sum per infected the Vaccine Fund, which raises and disburses funds for the patient who is cured using the directly observed short-course alliance, GAVI provides commodity assistance to countries in therapy approach. In one district of Kerala state where this the form of new and underused vaccines (hepatitis B, scheme is well advanced, NGOs and private providers have Haemophilus influenzae type B, and yellow fever, with new helped boost coverage from some 55 percent of those infected products for rotavirus and pneumococcus to follow); safe with TB to 78 percent. injection supplies; and support for strengthening national In Cambodia, government agreements and funding to immunization systems. In addition, GAVI allocates grant funds NGOs to operate district health services showed impressive to countries on the basis of their performance in increasing results compared with the standard approach, whereby the coverage rates for diptheria-pertussis-tetanus immunizations. government ran district services. The NGOs operated in one of Countries' applications to GAVI specify current coverage levels. two ways: (a) on a fully contracted-out basis, with complete On the basis of these data, their performance is assessed responsibility for service delivery, including hiring and firing annually, and US$20 per child is given to the country staff members and setting wages and procuring and distribut- for each additional child immunized with the diphtheria- ing essential drugs and supplies, or (b) through a pure man- pertussis-tetanus vaccine. agement contract, in which the NGOs worked within the In 2004, GAVI made its first payment for performance veri- Ministry of Health system and had to strengthen the existing fied by means of externally audited health data. Eight countries district structure. The NGOs that were fully contracted out received US$15 million in performance-based payments for raised immunization rates by 40 percentage points between their achievements in increasing immunization rates to reach 1997 and 2001, twice the rate of improvement produced by the an additional 750,000 children. Sierra Leone, for example, qual- government-run districts. The rate of growth in prenatal care ified for these payments on the basis of its performance in rais- in the contracted-out districts was more than triple that in the ing coverage from 44 percent of children in 2000 to 62 percent government-run districts, and the use of modern contracep- in 2002, as the country emerged from civil war (GAVI 2004a). tion methods expanded 50 percent more in the contracted-out districts (Bhushan, Keller, and Schwartz 2002). Stronger Engagement with the Private Sector Central Government Transfers to Local Authorities. In As donors have increasingly become aware of the extent of pri- Brazil's Family Health Project, the central government is mak- vate sector involvement in the health sector in developing ing per capita transfers to local municipalities on the basis of countries--that is, both the share of health services delivered planned increases in certain services, such as safe deliveries for by private providers and the share of total health spending low-income women and poor children treated for various ill- coming from private sources, including out-of-pocket nesses and monitored for their nutritional status and growth. payments--they have sought to use DAH to engage the private For example, at least 40 percent of babies are to be delivered in sector in pursuit of basic health goals. maternity facilities managed under the government's family Innovative approaches include both the transfer of develop- health program. Participating outreach workers are to provide ment assistance to the private sector through government an average of at least nine home visits to targeted low-income channels in developing countries and the provision of direct families each year. All doctors enrolled in the program are to financial support to private institutions (World Bank 2003b). undergo special training. If the municipalities reach those tar- In the former category, social investment funds have been gets and several others, they will continue to be eligible for established in many regions as a way to channel DAH to com- future financial transfers; otherwise, the level of central gov- munity groups and NGOs involved in running health centers ernment support will be reduced, and remedial measures will and disease control programs (Jorgensen and Domelen 2001), be put in place to try to improve the targeting and effectiveness especially in Africa and Latin America. In a similar vein, donors Recent Trends and Innovations in Development Assistance for Health | 253 Box 13.3 IDA Credit Buy-Downs for Polio Eradication To ensure financing for the MDGs, governments, founda- Nations Foundation, the World Bank implemented two tions, agencies, and development banks are all exploring projects in fiscal 2003, one in Pakistan and the other in new financing approaches that have the potential to Nigeria. The partnerships will buy down a country's IDA increase resource flows, adjust the concessionality of fund- loans on successful completion of the country's polio ing (that is, reduce interest rates and thus increase the eradication program. Because of the generous loan terms, grant element) where appropriate, and help focus more each grant dollar unlocks roughly US$2.50 for countries attention on effects. to fight polio. To fund the buy-downs, the partnership has The IDA credit buy-down mechanism was recently established a trust fund with US$25 million from the Bill piloted in several projects supporting polio eradication, & Melinda Gates Foundation and US$25 million from clearly a global public good. The mechanism enhances the Rotary International and the United Nations Foundation. concessionality of IDA's assistance in priority areas, mobi- This US$50 million investment has the potential to buy lizes additional resources from external partners, and down roughly US$125 million in IDA loans. In this way, focuses the attention of governments, partners, and World developing countries can mobilize what ultimately Bank staff on clearly defined performance objectives. becomes grant funding to eradicate polio and to con- Working in partnership with the Bill & Melinda Gates tribute to the global campaign to eliminate the transmis- Foundation, Rotary International, and the United sion of polio. Source: World Bank 2004d. have been prime movers behind schemes to encourage govern- financial and in-kind support from major Indian companies ments to contract with NGOs and private hospitals and labo- that can then be used to support a range of HIV prevention ratories for basic services targeted to the poor, such as cataract programs, such as condom promotion, peer education, and surgery and TB case detection and treatment in India (Central voluntary counseling and testing targeted at truck drivers, TB Division 2002; World Bank 2002). commercial sex workers, and others at high risk (Sengupta and In terms of direct DAH financing to the private sector in Sinha 2004). developing countries, the most common and longstanding The other area in which donor funds are increasingly being examples are in the social marketing of health-related personal used to stimulate private sector action and leverage private products, such as contraceptives, kits for treating sexually funding is through public-private partnerships for new health transmitted infections, insecticide-impregnated bednets to technologies, including vaccines, drugs, and diagnostics. prevent malaria, and point-of-use water purification kits. Private financing, technical input, and management make Donors are currently providing millions of dollars each year to sense in this area, because typically it is the private sector that subsidize the purchase of these items by poor families in devel- has the technical knowledge and the manufacturing and distri- oping countries. More recently, other donor engagements with bution capacity to create and market new health products, but the private sector have included the Global Alliance for major scientific risks and the lack of an attractive market in Improved Nutrition, in which a consortium of donors that poor countries are barriers to investment. The public-private includes the Bill & Melinda Gates Foundation and the govern- partnerships aim to overcome those barriers through a combi- ments of Canada, the Netherlands, and the United States have nation of up-front financing for R&D (so-called push funding) pooled funds that can be used to expand the fortification of and market guarantees for effective products (so-called pull basic foods with micronutrients by private manufacturers. The financing). The 20 largest partnerships for new products have Global Alliance for Improved Nutrition is helping to fortify raised more than US$1.5 billion over the past decade and are wheat with iron in western China and in Morocco and fish beginning to see results, such as the development of new drugs sauce with vitamin A in Vietnam. for malaria and TB, promising vaccines for malaria and AIDS, Another recent example of DAH going directly to the pri- and microbicides to protect against HIV infection (IPPPH vate sector is Avahan, the innovative AIDS prevention program 2004; Rockefeller Foundation 2004). The largest partnership, that the Bill & Melinda Gates Foundation is financing in six the IAVI, illustrates the innovative nature of these partnerships Indian states. The program uses external financing to leverage and the effective use of DAH (box 13.4). 254 | Disease Control Priorities in Developing Countries | Robert Hecht and Raj Shah Box 13.4 The International AIDS Vaccine Initiative The International AIDS Vaccine Initiative was established velopers, while ensuring that developing countries will in 1996 with support from the Rockefeller Foundation as have access to the vaccine at an affordable price if it turns an innovative way to give a boost to AIDS vaccine R&D. out to be effective. Optimism about AIDS vaccines in the late 1980s had given Since IAVI embarked on these vaccine development way to a series of failures and to discouragement by the partnerships in 1999, it has spent a total of about end of the decade. R&D efforts were spending less than US$200 million in this area. Five vaccine candidates are US$100 million a year. Neither governments nor private undergoing clinical trials in eight countries in Africa, Asia, vaccine companies were investing much in research into Europe, and North America. IAVI is poised to spend AIDS vaccines. another US$300 million in R&D during 2005­7 in an IAVI's mission was defined as ensuring the develop- effort to accelerate the discovery, development, and licens- ment of a safe, effective, and accessible vaccine for use ing of a vaccine to prevent HIV infection. IAVI is also try- throughout the world. IAVI's activities were to include a ing to stimulate expanded use of donor funding for R&D combination of global advocacy, policy analysis and in the field of AIDS vaccines and is calling for govern- reform, and investments in carefully chosen R&D projects ments to increase public financing from the current focusing on the most promising vaccine candidates. amount of around US$600 million a year to US$1.2 bil- IAVI's collaboration with the private sector has lion annually. At the same time, IAVI has proposed that occurred at several levels. Funding for IAVI has come from donors create a purchase fund of several billion dollars to six governments (Canada, Denmark, the Netherlands, serve as a promise to buy large numbers of doses of an Norway, the United Kingdom, and the United States); the efficacious vaccine from qualified manufacturers. The European Union; and the World Bank, as well as from pri- U.K. government has committed itself to joining such an vate foundations and companies. IAVI's vaccine develop- advance purchase fund and is urging others to join it ment partnerships take many forms. They typically ("Gordon Brown to Earmark" 2004). include an academic developer and a biotechnology com- The health and economic stakes are enormous. pany, plus researchers, laboratories, and clinical trial sites Without improved HIV prevention tools, an additional in developing countries such as India, Kenya, and Uganda. 100 million HIV infections are likely over the next two Private companies generally manufacture test lots of the decades, resulting in huge economic losses. IAVI estimates AIDS vaccines and undertake bioengineering studies and that an efficacious vaccine could prevent 2 million AIDS enhancements to the vaccine. IAVI generally shares the deaths a year and generate billons of dollars in lives saved risks and costs of the partnerships with the private code- and antiretroviral treatment costs averted. Source: IAVI 2004. Getting Funds to the Front Line · a strong financial capacity in NGOs and private providers in Central government funds can easily leak as they move through cases in which the government's strategy for local develop- the pipeline from the center to local levels. In addition, in the ment is to rely on private institutions absence of local initiative and the right incentives, service pro- · a government body that is appropriately equipped and vision can fail to reflect the views of local people. Effective DAH responsible for regulating the quality of public and private needs to address those impediments. It needs to channel tech- providers nologies, ideas, finance, and technical assistance closer to · a balanced approach to community-driven development in households, health providers, and supervisory officials in ways health to ensure that financing for community health initia- that are consistent with national policies and are amenable to tives of the social fund type is sustainable. monitoring and reporting. Examples of DAH reaching frontline workers in an expedi- Development assistance for health is more likely to reach tious and sustainable way include block grants for districts in communities if they have the following: Uganda; social development funds in Central America; con- · a decentralized system of fiduciary and technical manage- tracts with urban and rural NGOs under India's Reproductive ment in the public sector and Child Health Program; and support to community-led Recent Trends and Innovations in Development Assistance for Health | 255 initiatives under the Multi-Country AIDS Program, which adequate funds flow to the upstream stages of R&D on new financed an average of 10,000 local initiatives in each of its first health technologies, where the private sector lacks the market four years in several African countries (World Bank 2004c). incentives to invest and where national research bodies have so far not been up to the task. The multilateral banks do not have CONCLUSIONS the instruments to channel major funding to global, as opposed to national, health technology programs, and the bilateral agen- Despite the promising trends in DAH over the past five to seven cies alone are not equal to this task. Even with the modest exter- years, the outlook for the next few years is uncertain. What nal funds allocated to the recently established public-private happens will depend on overall trends and innovations in partnerships (such as the IAVI, the Medicines for Malaria development assistance, which in turn are driven by such fac- Venture, the Global TB Drug Alliance, and the International tors as political changes and the rate of economic growth in Partnership for Microbicides), new technologies are emerging. OECD countries, and the willingness of high-income countries However, they need to be reinforced with additional funding. to honor their pledges to increase the share of GDP they devote One option would be to design a new funding facility within to development. the multilateral banks that would allow them to allocate signif- Under most scenarios, the share of overall assistance going to icant resources to global health research and product develop- health will likely continue its recent rise, given the current polit- ment. Another would be to use the nascent international ical focus on the global AIDS pandemic and the growing aware- finance facility to provide funds for global research. ness of the challenges and opportunities associated with the To address those challenges, strong political will is the essen- MDGs for maternal and child health and communicable tial baseline ingredient, as recent experience with HIV/AIDS diseases. has clearly demonstrated. The United Nations and the multi- Under a more conservative scenario, a number of factors lateral agencies must remain firmly behind more robust DAH, could have a negative impact on DAH. Those factors could as they are currently doing by means of the Millennium Project include an overall slowing in the rate of growth of development and the High-Level Forum on the Health-Related MDGs. financing and the donor fatigue that could set in if the larger allo- Individual bilateral donors and foundations must continue to cations for HIV/AIDS, TB, and malaria are not fully disbursed, demonstrate their leadership. Finally, and perhaps most impor- are misused, or yield disappointing results on the ground. tant, leaders and civil society organizations in poor countries Under a more optimistic scenario, DAH will continue to need to continue to speak out for more and more effective grow as developing countries and donors find new ways to dis- DAH, indicating that health is their priority and that they are burse a higher volume of funds and use them effectively--for prepared to commit domestic resources to match the larger example, through subcontracts with NGOs and private health external flows provided through DAH. service delivery organizations. Sectorwide approaches and budget support through national poverty reduction programs may also result in expanded flows of DAH. New generations of REFERENCES technologies adapted to the developing world, such as more Bhushan, I., S. Keller, and B. Schwartz. 2002. Achieving the Twin Objectives effective antimalarials, better TB diagnostics, and a vaccine to of Efficiency and Equity: Contracting for Health Services in Cambodia. prevent HIV infection, would almost certainly attract increas- Policy Brief Series 6. Manila: Asian Development Bank. ing amounts of DAH. Burnside, C., and D. Dollar. 2000. "Aid, Growth, the Incentive Regime, and Even under the more optimistic scenario, DAH will still face Poverty Reduction." In The World Bank: Structure and Policies, ed. C. L. Gilbert and D. Vines, 210­27. Cambridge, U.K.: Cambridge University major challenges. The expected volume of financial assistance Press. is unlikely to match the large needs of the developing world Central TB Division, Directorate General of Health Services, Ministry of and the requirements to attain the MDGs. Countries and Health and Family Welfare. 2002. TB India 2002: RNTCP Status Report. donors will therefore face difficult decisions in relation to pri- New Delhi: Directorate General of Health Services. ority setting and require better tools to make such allocation CMH (Commission on Macroeconomics and Health). 2001. Macroeconomics and Health: Investing in Health for Economic decisions. Cost-effectiveness analysis offers one such tool. Development--Report of the Commission on Macroeconomics and Effective absorption of DAH will also continue to pose diffi- Health. Geneva: World Health Organization. culties for countries with weak capacity. In such cases, Devarajan, S., M. J. Miller, and E. V. Swanson. 2002. "Goals for increased use of NGOs and the private sector in general to Development: History, Prospects and Costs." Discussion Paper 2819, complement public sector action may help make the informa- World Bank, Washington, DC. tion and services that poor households need to improve their Eichler, R., P. Auxilia, and J. Pollock. 2001. Output-Based Health Care: Paying for Performance in Haiti. Washington, DC: Abt Associates. health status more accessible. Foster, M., A. Brown, and T. Conway. 2000. Sector-Wide Approaches for Another issue is that the current architecture of develop- Health Development: A Review of Experience. Geneva: World Health ment assistance does not contain a mechanism to ensure that Organization. 256 | Disease Control Priorities in Developing Countries | Robert Hecht and Raj Shah GAVI (Global Alliance for Vaccines and Immunization). 2004a. GAVI Stop TB Partnership. 2003. Report of the DOTS Working Group for the 22 Awards for Top Performing Countries. Geneva: GAVI. High-Burden Countries. Geneva: World Health Organization. ------. 2004b. Programmatic Aspects of the IFF Immunization Pilot. Swedish International Development Cooperation Agency. 2003. Mapping Geneva: GAVI. of Sector-Wide Approaches in Health. London: Institute for Health "Gordon Brown to Earmark 200 Million Pounds a Year to Fund AIDS Sector Development. Vaccine." 2004. Independent (London), December 1, p. 18. UN (United Nations) Millennium Project. 2004. "Millennium Heaver, R. 2002. Thailand's National Nutrition Program: Lessons in Development Goals Needs Assessment: Case Studies of Bangladesh, Management and Capacity Development. Washington, DC: World Cambodia, Ghana, Tanzania, and Uganda." Working Paper, United Bank. Nations, New York. Hecht, R. 2004. "Making Health Care Accountable: The New Focus on UNAIDS (Joint United Nations Programme on HIV/AIDS). 2001. Report Performance-Based Funding of Health Services." Finance and on the Global HIV/AIDS Epidemic. Geneva: UNAIDS. Development 41 (March): 16­19. ------. 2004. Global Expenditures and Requirements to Address the IAVI (International AIDS Vaccine Initiative). 2004. The IAVI Strategic Plan HIV/AIDS Pandemic. Geneva: UNAIDS. 2005­07. New York: IAVI. World Bank. 1993. World Development Report 1993: Investing in Health. IPPPH (Initiative for Public-Private Partnerships for Health). 2004. New York: Oxford University Press. Combating Diseases Associated with Poverty: Financing Strategies for ------. 1998. Assessing Aid: What Works, What Doesn't, and Why. Oxford, Product Development and the Potential Role of Public-Private U.K.: Oxford University Press. Partnerships. Geneva: IPPPH. ------. 2000. "Large-Scale Government Contracting of NGOs to Extend Jamison, D. T. 2004. "External Finance of Immunization Programs: Time Basic Health Services to Poor Populations in Guatemala." Paper pre- for a Change of Paradigm?" In Vaccines: Preventing Disease and pared for the Challenge of Health Reform: Reaching the Poor, Europe Protecting Health, ed. C. de Quadros, 325­32. Washington, DC: Pan- and Americas Forum, San José, Costa Rica, May 24­28. American Health Organization. ------. 2002. "India--Cataract Blindness Control Project." Imple- Jorgensen, S., and J. V. Domelen. 2001. Helping the Poor Manage Risks mentation Completion Report 25232, World Bank, Washington, DC. Better: The Role of Social Funds. Washington, DC: Brookings ------. 2003a. Attaining the Millennium Development Goals in India: How Institution. Likely and What Will It Take? Washington, DC: World Bank. Levine, R., and What Works Working Group. 2004. Millions Saved. ------. 2003b. "Private Health: Policy and Regulatory Options for Private Washington, DC: Center for Global Development. Participation." Private Sector and Infrastructure Network Note 264, Michaud, C. 2003. Development Assistance for Health: Recent Trends and World Bank, Washington, DC. Resource Allocation. Boston: Harvard Center for Population ------. 2003c. Progress Report and Critical Next Steps in Scaling up Development. Education for All, Health, HIV/AIDS, and Water and Sanitation. OECD (Organisation for Economic Co-Operation and Development). Washington, DC: World Bank and International Monetary Fund 2004a. CRS Online Database on Aid Activities. http://www.oecd.org/ Development Committee. dac/stats/idsonline. ------. 2004a. "The Mauritania Health System and Implementation of ------. 2004b. Survey on Harmonisation and Alignment: Preliminary the Poverty Reduction Strategy." Africa Human Development Working Edition. Paris: OECD. Paper 39, World Bank, Washington, DC. Rajkumar, A., and V. Swaroop. 2002. "Public Spending and Outcomes: ------. 2004b. The Millennium Development Goals for Health: Rising to Does Governance Matter?" Policy Research Working Paper 2840, the Challenges. Washington, DC: World Bank. World Bank, Washington, DC. ------. 2004c. Experience in Scaling Up Support to Local Response in Rockefeller Foundation. 2004. Partnering to Develop New Products for Multi-Country AIDS Programs (MAP) in Africa. Washington, DC: Diseases of Poverty. New York: Rockefeller Foundation. World Bank. Sengupta, J., and J. Sinha. 2004. "Battling AIDS in India." McKinsey ------. 2004d. "Financing Modalities toward the Millennium Quarterly 3. New Delhi: McKinsey. Development Goals: Progress Note." World Bank, Washington, DC. Soucat, A., W. Van Lerberghe, F. Diop, S. Nguyen, and R. Knipperberg. World Bank and Ministry of Health, Ethiopia. 2005. Ethiopia: A Country 2003. Marginal Budgeting for Bottlenecks: A New Costing and Status Report on Health and Poverty. Volumes I and II. Washington, Resource Allocation Practice to Buy Health Results. Washington, DC: DC: World Bank. World Bank. Recent Trends and Innovations in Development Assistance for Health | 257 Chapter 14 Ethical Issues in Resource Allocation, Research, and New Product Development Dan W. Brock and Daniel Wikler The ethical justification for developing and providing the the production of health. Costs are measured in monetary means to reduce the burden of disease in developing countries terms; benefits are measured in health improvements. By divid- is self-evident. Nevertheless, those who pursue these laudable ing costs by benefits, one can obtain a cost-to-effectiveness ratio ends encounter ethical dilemmas at every turn. The develop- for each health intervention, and interventions can be ranked ment of new interventions requires testing with human by these ratios. Although a CEA is typically an economic analy- subjects, an activity fraught with controversy since the dawn of sis performed by health economists, it is also a measure of scientific medicine and especially problematic with poor and one ethical criterion for the evaluation of health programs. vulnerable participants in developing countries. Ethical dilem- Cost-effectiveness is not merely an economic concern, because mas arising in setting priorities among interventions and improving people's health and well-being is a moral concern, among individuals in need of care are most acute when needs and an allocation of resources that is not cost-effective pro- are great and resources few. duces fewer benefits than would have been possible with a dif- We address some of these concerns in this chapter, identify- ferent allocation. Producing more rather than fewer benefits for ing some of the principal ethical issues that arise in the devel- people is one important ethical consideration in evaluating opment and allocation of effective interventions for developing actions and social policies. countries and discussing some alternative resolutions. We omit Second, the allocation should be equitable or just; equity is discussion of two other aspects of these ethical decisions: concerned with the distribution of benefits and costs to distinct ensuring that the process of decision making is fair and individuals or groups. The maximization of benefits, which is involves the subject population (Daniels 2000; Holm 1998), associated with the general philosophical moral theory of util- and respecting legal obligations under international human itarianism or consequentialism, however, is routinely criticized rights treaties (Gruskin and Tarantola 2001). for ignoring those considerations (Rawls 1971). Equity in health care distribution is complex and embodies several dis- tinct moral concerns or issues that this chapter delineates HEALTH RESOURCE ALLOCATION (Brock 2003a). There is no generally accepted methodology comparable to CEA for determining how equitable a distribu- Resource allocation in health and elsewhere should satisfy two tion is; nevertheless, allocations are unsatisfactory if equity main ethical criteria. First, it should be cost-effective--limited considerations are ignored. resources for health should be allocated to maximize the health Efficiency and equity can sometimes coincide. In some of benefits for the population served. A cost-effectiveness analysis the world's poorest countries, for example, health budgets (CEA) of alternative health interventions measures their respec- support tertiary care and travel to clinics abroad for the elite tive costs and benefits to determine their relative efficiency in and the well connected, even as the poor are denied effective, 259 low-cost prevention or treatment for life-threatening diseases undiminished function (or health) is generally determined by (Birdsall and Hecht 1995). Moreover, because equity concerns soliciting a group of individuals' preferences for life in that state the relative treatment of different individuals, CEA is largely using standard gambles, time tradeoffs, visual analog scales, or unobjectionable when it is used only for evaluating alternative person tradeoffs. In all these methods, a common issue is health interventions that would serve the same patients. whose preferences to use for valuing health states. The main However, considerations of equity may conflict with cost- debate has been whether to use a randomly selected group of effectiveness and so may provide moral reasons for an alloca- citizens or to use people who have the particular disability or tion that is not cost-effective. The discussion in this chapter limitation in function being evaluated. accepts that CEA identifies one important ethical criterion in This issue matters because a number of studies have shown evaluating health care interventions--producing the most ben- that persons without disabilities generally evaluate the quality efits possible for individuals served by those interventions-- of life with a particular disability as significantly worse than do and then focuses on the other ethical criterion of ensuring persons who have that same disability (Menzel and others equitable distribution of those benefits. 2002). If the preferences of persons without disabilities are This chapter considers two types of equity issues: first, those used, their lower evaluation of quality of life with various dis- that arise in the general construction of a CEA--that is, in abilities will mean that fewer QALYs will be produced by life- determining the form of a CEA; second, those that arise in the saving interventions for persons with disabilities than if the use of the results of a CEA for resource allocation in the health preferences of persons with disabilities had been used. sector. It is worth noting that, when applied appropriately and However, if we use the preferences of persons with disabilities, broadly to all social conditions and programs that significantly then both prevention and rehabilitation will receive less value influence health, CEA may often support using resources to than if the preferences of persons without disabilities had been affect the so-called social determinants of health--which used. largely affect the incidence of disease, disability, and premature This difference in evaluations in part results from igno- mortality--rather than using those resources on health care to rance, prejudice, and stereotypes on the part of persons treat disease. However, we shall focus largely on CEA in the without disabilities about what it is like to live with various evaluation of health care and public health programs. disabilities. The difference results as well from the process of adaptation to disability in which disabled persons adjust by learning new skills, cope by adjusting their expectations to their Issues in the Construction of a Cost-Effectiveness Analysis new circumstances, and accommodate by substituting new Cost-effectiveness analyses require decisions about which costs aims and activities for ones made difficult or impossible by to include, which if any financial gains should be counted as off- their disabilities (Solomon and Murray 2002). They thus adopt setting costs, whether to include benefits beyond the effects of a new valuational perspective for making health and quality- the intervention on health, and whether all health gains should of-life evaluations. Because the adoption of this new perspec- be valued alike. None of those decisions, in our view, is exclu- tive resulted from a disability, it will represent a set of values for sively a technical issue, and CEA results reflect the analysts' making choices that reflects a restricted set of abilities. ethical judgments on those issues. Nevertheless, neither the nondisabled perspective nor the adapted disabled perspective is mistaken; they are only different Evaluation of Benefits. Evaluating health benefits within a (Brock 1995). These differences create controversy in the litera- CEA involves several issues. This chapter assumes that some ture over which perspective is correct for cost-effectiveness version of a quality-adjusted life year (QALY) is used to com- evaluations in health care. bine the two main benefits of health care--(a) protecting or A second issue is whether, in evaluating interventions that improving health or health-related quality of life and (b) pre- preserve or extend life, we should use life years saved (as QALYs serving life. Disability-adjusted life years (DALYs) are a variant do) or lives saved. Certainly individuals offered two interven- of QALYs in that they measure the losses from disability or pre- tions that would preserve their lives for different lengths of time mature death; a CEA will determine which interventions will would prefer, all other things being equal, the alternative with maximize QALYs or minimize DALYs. Calculating QALYs the longer period of survival. Moreover, when the differences requires a metric evaluating the effect of different states of lim- are extreme--for example, extending group A's lives by a week itations in function on health-related quality of life, such as the or extending an equally numerous group B's lives by 10 years-- Health Utilities Index (Horsman and others 2003). The Disease virtually everyone would judge this difference to support giving Control Priorities Project uses the health state valuations or priority to group B. This fact suggests that even the proponent disability weights of the World Health Organization (WHO). of counting lives saved should require that the lives saved for a The relative value of any particular health state, typically on a shorter period of time must still be saved for a significant scale in which "0" represents death and "1" represents full, period of time; what is significant will depend in part on the 260 | Disease Control Priorities in Developing Countries | Dan W. Brock and Daniel Wikler duration of lives saved by the alternative with which it is being the moral urgency, grounded in fairness, of preventing the compared. Some empirical studies indicate that ordinary peo- death. The younger a person is, the greater is the moral value of ple tend not to give much weight to differences in the duration providing a QALY to him or her. This view leaves open to what of health benefits to different groups of persons when priori- extent the moral value of QALYs should decline with the age of tizing between them, as long as the lesser duration benefits are the recipient. This age weighting to favor the young has been viewed as significant; this attitude suggests that they favor lives attacked by some as unjust age discrimination, but because an saved over life years saved (Nord and others 1996). The life explicit moral justification in terms of fairness is offered for it, years saved versus lives saved controversy remains unsettled. critics must show why that justification is unsound. Should Life Years Be Age Weighted? The standard assumption What Costs Should Count in Health Cost-Effectiveness in most CEAs using QALYs is that one QALY has the same Analyses? No controversy surrounds the inclusion in a CEA of social value, regardless of the age of the recipient (Gold and direct costs of a health intervention program or direct health others 1996). Thus, equality is adopted as the weighting for benefits to the intervention's recipients. Ethical issues do, how- QALYs achieved by recipients at different ages, and that is the ever, arise in other aspects of the cost calculation (Brock approach adopted in this volume. The use of any age weighting 2003b). A full CEA of alternative health programs should take that gives less value to benefits for the elderly than for younger account of all the economic effects on public or private expen- persons is often charged as unjust age discrimination. Even the ditures of the alternative health interventions or programs use of equally weighted QALYs is often charged as unjust age under analysis. An example is provided in the consideration of discrimination because, other things being equal, saving the treatment for two alternative health conditions judged to have lives of younger persons will produce more QALYs than saving equally detrimental effects on patients' health: the first condi- the lives of older persons. The goal of lives saved, as opposed to tion permits patients to continue working, and the second life years saved, removes this disadvantage to the elderly from interferes with regular work and so has large economic costs to CEAs that use QALYs. However, if the relevant benefit is adding the patients' employers. Should the costs of treating the second years to life, then standard CEA is neutral or impartial regard- be reduced by the cost savings to the employers from returning ing age, in the sense that it gives the same value to a year of life the patients to work on a regular basis? If so, the second treat- extension whatever the age of its recipient. ment program will have a more favorable cost-effectiveness WHO, in its burden-of-disease and resource prioritization ratio than the first, even if it may be no better or worse without studies that use DALYs, rejected the equal age weighting that is consideration of those economic effects. The same issue arises standard with QALYs. Instead, it gave less value to DALYs in many other contexts. prevented for infants, young children, and the elderly, in com- From the moral perspectives of both a consequentialist and parison with persons in their productive adult years. WHO jus- a standard CEA, these indirect economic effects for others are tified this weighting by the fact that the very young and the eld- real benefits or cost reductions and should be part of the CEA. erly both tend to be economically, socially, and psychologically The fundamental moral objection to giving higher priority to dependent on adults during those adults' productive working treating those who can be treated at lower net cost because of and child-rearing years (Murray 1994). This justification is eth- the economic savings to their employers is the same as that ically problematic, however, because it assigns different value to with WHO's instrumental rationale for its age weighting. One meeting people's health needs on the basis of differences in the condition or group of patients gets higher priority solely instrumental value to others of meeting their needs. This because treating it or them is a means to producing economic approach differentiates people solely on whether they are a benefits to others, thereby reducing the net social costs of their means to benefiting others. The same reasoning would justify treatment. This approach violates the Kantian injunction giving priority to rich over poor patients with the same med- against treating people solely as means--the first group has ical needs because the rich are more socially productive than lower priority for treatment solely because treating that group the poor, a practice that would be widely regarded as unjust. is not a means to the economic savings to employers. It fails to Writers in this field have provided different reasons for giv- give equal moral concern to the health needs of each group of ing greater value to QALYs for younger patients, however, that patients because it discriminates against the less socially valu- are not subject to this moral objection and that are specifically able patients. Conversely, at the macro level of the allocation of grounded in fairness. For example, Alan Williams has devel- resources to health instead of other social goods, the WHO oped an argument to the effect that fairness requires that indi- Commission on Macroeconomics and Health has supported viduals should each receive "fair innings" of QALYs in their increasing health investments in developing countries because lives (Williams 1997). In this view, the earlier a preventable such investments often more than pay for themselves in their death could occur and the worse a person's past health is, the economic and development benefits (CMH 2001). Using a greater is the unfairness the person suffers--so the greater is "separate spheres" view, only the health benefits and health Ethical Issues in Resource Allocation, Research, and New Product Development | 261 costs of alternative health interventions should determine The ethical issue about discounting is whether, after taking their priority for obtaining resources, but this view remains account of such considerations, a health benefit of the same controversial. size has progressively less social value the farther into the future Another aspect of cost calculation concerns whether future that it occurs. To make the issue more concrete, suppose we health care and other costs, such as old-age payments, that will must decide between two programs: one will save 100 lives be incurred as a result of a person's life being saved should be now, and the other, say a hepatitis vaccination program, will added to the costs of treating that person now. Persons who do save 200 lives in 30 years. The vaccination program will save not die now because of a life-saving intervention will typically twice as many lives, but if we apply even a 3 percent discount go on to incur future health costs that would not have been rate to the future lives saved, they are equivalent to only 78 lives incurred had they died now. The U.S. Public Health Service saved now, and we should prefer the first program. This exam- Panel on cost-effectiveness recommended that inclusion of ple illustrates not only the theoretical issue, but its practical these costs be optional in CEAs (Gold and others 1996). Others import, too, because discounting future health benefits will sys- have argued that, if CEA is designed to maximize lifetime util- tematically tend to disadvantage prevention programs that ity, the future costs should be included (Meltzer 1997). These must be undertaken now but whose benefits occur only at are costs that would not be incurred if the patient was not some point in the future. This reasoning applies not only to saved, but virtually no one would argue that, because of those many vaccination programs, but also to most programs to costs, we should judge a life-saving intervention as not cost- change unhealthy behaviors in which the benefits generally effective and thus deserving of lower priority than interven- occur at some later time. tions that do not have those effects. What does this thinking Arguments for discounting health benefits at the same rate show? That we are not prepared to allocate health resources on as costs have included consistency arguments (Weinstein and the basis of a full CEA that accounts for all the costs incurred Stason 1997), avoidance of paradoxes in allocation concerning and saved by those interventions--that is, that some should be research and deferral of spending (Keeler and Cretin 1983), disregarded on ethical grounds. individual or social rates of time preference, and so forth. Those arguments cannot be reviewed here, but whether to dis- Should Health Benefits and Costs Be Discounted in count health benefits is squarely an ethical question about the Cost-Effectiveness Analyses? As standard practice in CEAs, valuing of health benefits over time and should be explicitly both health care costs and benefits are discounted at the same addressed as such in allocating resources. rate, for example, 3 percent or 5 percent, and the Disease Control Priorities Project applies a 3 percent discount rate to costs and benefits (Gold and others 1996). Little controversy Issues in the Use of Cost-Effectiveness Analysis surrounds the idea that future monetary costs and benefits for Resource Allocation should be discounted to their present value in a CEA. The same It is now widely recognized that CEA alone is not a satisfactory amount of money is worth more if received today than in guide to resource allocation in all cases. CEA, as customarily 10 years because it can be invested at the market rate of inter- formulated, measures the sum of costs and benefits and largely est if received today. For the same reason, costs that can be ignores the pattern of their distribution across the affected deferred require fewer present dollars to meet them. population. In some cases, the resulting allocation will strike The controversial issue is whether health benefits should be most observers as unfair. Health resource allocators need to discounted--that is, whether the same magnitude of health take distributional issues into account along with cost- benefit has progressively less social value the farther into the effectiveness. future it occurs. This issue is complex and has engendered an extensive literature that cannot be reviewed here, but we can at Priority to the Worst Off. Justice requires a special concern for least try to focus the issue. It is appropriate to discount for the the worst off, as is reflected in aphorisms such as "you can tell uncertainty about whether potential beneficiaries will survive the justice of a society by how it treats its least well-off mem- to receive a future health benefit and to discount for any bers," in the well-known Difference Principle in John Rawls's increased uncertainty about whether a benefit will occur theory of justice, and by the special concern for the poor with- because it is more distant. However, these uncertainties are in many religious traditions (Brock 2002; Rawls 1971). This reflected in the calculation of expected future benefits and do concern is often understood to reflect a concern for equality-- not require that future benefits be discounted. Likewise, if indi- in particular, equality in outcomes or welfare between people. viduals receive a health benefit (such as regaining mobility) In the health context, it takes the form of a concern for reduc- sooner rather than later, their total lifetime benefit may be ing inequalities in health between persons or groups. A variety greater, but this fact, too, is reflected in the estimation of the of ethical bases underpin a concern for equality in general and total benefit without discounting. for equality in health in particular, and they cannot be explored 262 | Disease Control Priorities in Developing Countries | Dan W. Brock and Daniel Wikler here. It is important, however, to understand that concern Finally, how much priority should the worst off receive? for the worst off is different from a concern for equality, Giving absolute priority to the worst off is implausible because because the two can be and often are confused. Raising the it faces the bottomless pit problem--using very great amounts position of the worst off will typically reduce inequality, but it of resources to produce very limited or marginal gains in the need not always do so. Sometimes improving the position of health-related quality of life of the severely ill or disabled. the worst off may unavoidably improve the position of those However, there is no apparent principled basis for determining who are better off even more and thereby increase inequality. how much priority the worst off should receive. Moreover, the concern for equality in outcomes is subject to the "leveling down" objection, in which equality is achieved by Aggregation and Cost Differences. The aggregation problem making the better-off members worse off, even when doing so occurs when determining at what point small benefits to a large in no way benefits those who are worst off. In the face of that number of persons should take priority over very large benefits objection, many have rejected equality in outcomes in favor of to a few, because the former result in greater aggregate or total a prioritarian view, according to which benefiting people has benefits (Daniels 1993; Kamm 1993). The issue can be illus- greater moral value the worse off those people are (Parfit 1991). trated by the initial effort to prioritize different treatment- A number of possible lines of reasoning support prioritari- condition pairs in the Medicaid program in the U.S. state of anism. For example, the worse off that people are, the greater is Oregon by what was essentially a cost-effectiveness standard. the relative improvement that a given size of benefit will pro- As was widely reported, capping teeth for exposed pulp was vide them, so the more the benefit may matter to them. ranked just above performing appendectomies for acute Alternatively, the greater the undeserved health deprivation or appendicitis, even though appendicitis is a life-threatening need that an individual suffers, the greater is the moral claim to condition. A variety of methodological problems affected have it alleviated or met. Oregon's analysis, but this kind of result is to be expected from However priority to the worst off is justified, an important CEA. The Oregon Health Services Commission estimated that issue is who the worst off are. In the context of resource alloca- it was possible to provide a tooth capping for more than tion in health care, the worst off might be those who are glob- 100 patients for the cost of one appendectomy, so the aggregate ally worst off, those with the worst overall well-being (such as benefits of the many tooth cappings were estimated to exceed the poor), or those with the worst health (that is, the sickest). the benefit of one appendectomy. As a consequence of results General theories of justice usually focus on people's overall of this sort, the commission fundamentally changed its priori- well-being, often allowing a lower level in one domain of well- tization methodology to largely ignore cost differences, except being to be compensated for by a higher level in another in the case of roughly equally beneficial interventions. The domain. However, there are both moral and pragmatic reasons commission essentially adopted what might be called a relative for what has been called a separate spheres view, according to effectiveness or benefit standard (Hadorn 1991). which the worst off for the purpose of health resource alloca- What Oregon's experience shows is that most people's sense tion should be considered to be those with worse health. of priorities is determined by a one-to-one comparison of the Morally, for example, Scanlon has argued that "for differences benefits of different interventions, in which case appendec- in level to affect the relative strength of people's claims to help, tomies are clearly a higher priority than tooth capping. That these differences have to be in an aspect of welfare that the ignores the great differences in costs between different health help in question will contribute to" (Scanlon 1997, 227). interventions that a CEA will reflect. Is it then simply a mistake Pragmatically, it may generally be too difficult, costly, intrusive, to ignore those cost differences in allocating health resources? and controversial, as well as too subject to mistake and abuse, At least two moral considerations suggest not. First, empirical to have to inquire into all aspects of people's overall levels of studies have shown that many people ignore the cost differ- well-being. ences because they believe that patients should not be at a Even if health allocation to the worst off should be based on disadvantage in priority for treatment simply because their levels of health, other issues remain. For example, are those condition happens to be more expensive to treat than are other with worse health those who are sickest now, at the time a patients' conditions (Nord and others 1995). Second, according health intervention would be provided for them, or those with to many moral theories, individuals should confront other worse health over time, taking into account past and perhaps competitors for scarce resources as individuals, and their prior- expected future health? The latter would give special weight to ity for treatment should be determined by the urgency of their meeting the health needs of those with long-term chronic dis- individual claims to treatment (Scanlon 1997). eases and disabilities. Separate spheres would still include past Then again, most people and most moral theories do not and future health. Should special priority also be given to those reject all aggregation of different sizes and costs of health ben- whose health is not worse now but is especially vulnerable to efits in setting priorities and allocation, although there is no becoming worse? consensus either on when aggregation should be permitted or Ethical Issues in Resource Allocation, Research, and New Product Development | 263 for what reasons. However, at a minimum, we suggest that indi- groups are treated. In the extreme case, some live and others viduals should not be denied very great health benefits--in the die. The better outcome is produced by funding program A extreme case, life-saving interventions--merely to provide rather than program B, but that additional good is insufficient small health benefits to a large number of other persons. to justify morally the huge difference in the way the two groups of patients are treated. The conflict between fair chances and Fair Chances and Best Outcomes. The thesis that resources best outcomes can arise in a variety of contexts (Kamm 1993). should be targeted to interventions in which they will do the Preferring the most cost-effective program can also seem most good ascribes a higher priority to those who can be helped unfair because it compounds existing unfair inequalities. For more easily or cheaply. This thinking, in turn, implies that some example, screening slum-dwelling black men for hypertension patients will lose out simply because their needs are more diffi- targets the group with the highest incidence and greatest risk of cult or expensive to meet. Consider, for example, a ward with premature death. However, it is more cost-effective to 100 patients,50of whomrequireonepilland50of whomrequire target well-to-do suburban white men, because they have more two pills to recover. The patients are otherwise similar. The clinic ordered lives, comply better, have personal doctors and has 50 pills and must decide how to distribute them. To achieve the means to obtain medical services, are more educated, the best outcome, all 50 pills should be given to the patients who and are more likely to modify their lifestyles wisely. However, if need only one to recover. However, to give each patient an equal the poor black men are not screened for this reason, it only chance to recover, entitlement to treatment should be awarded compounds their existing unjust deprivation and, of course, is randomly. Seventeen fewer cures would result. also in conflict with giving priority to the worst off. Limited surveys indicate a sharp difference between health If those who need a less cost-effective program deserve a fair professionals and the general public in their responses to this chance to have their needs met, what would be a fair chance? conflict. Most health professionals favor distribution to one- Some argue that a fair chance is an equal chance, so some ran- pill patients only, and most members of the general public dom method of selecting which program to fund should be insist that people should not be penalized for needing two pills used (Broome 1991). Others suggest proportional chances or a (Nord 1999). This division of opinion goes to the heart of CEA, weighted lottery, in which the chance of each program being which is precisely a guide to identifying the route to the best selected is proportional to the amount of health benefit each outcomes that can be hoped for with existing resources. It also would produce, as a way of balancing fair chances against best creates a dilemma for those health professionals who maintain outcomes (Brock 1988). Alternatively, some resources might go that health policy should be based on values most frequently to each program (which is usually possible at the macro level), endorsed by the population affected. thereby benefiting some patients in each group--at least if The conflict between fair chances and best outcomes arises their relative benefits are not strikingly dissimilar--instead of not only from differences in the costs of treating otherwise sim- all going to the most cost-effective programs. ilar groups of patients, but also when one group of patients will Another consideration supports spreading some resources receive somewhat greater benefits than another at the same to less cost-effective programs instead of devoting them all to cost. The appeal of a fair-chances solution is greater when the the most cost-effective: to give all--or at least more--patients difference in cost-effectiveness between the two programs is a reason to hope that their health needs will be met. This con- relatively small compared with the potential gain or loss to sideration may be especially important in developing countries individual patients. Suppose that health program A will pro- where resource scarcity is more severe and adhering strictly to duce 5,000 QALYs while program B will produce 4,500 QALYs cost-effectiveness criteria could result in large numbers of and that the effect on the health or life of each patient served is patients with serious--or even life-threatening--health needs large--in the extreme, life saving. Patients who would be served having no hope that their needs will be met. by program B could complain that it is not fair that all the resources go to program A and none to B when they have Discrimination against Persons with Disabilities. The use of nearly as pressing health needs and would be benefited by CEA in resource allocation to maximize the QALYs produced treatment nearly as much as the patients served by program A. by available resources will often discriminate against persons If all cannot be treated, they might go on to argue, they deserve with disabilities. Many persons with disabilities such as cystic a fair chance to have their needs met rather than having no fibrosis, HIV/AIDS, and chronic pulmonary or heart disease chance for treatment only because treating them would pro- have reduced life expectancies or health-related quality of life duce slightly less benefit than treating the patients served by as a result of their disabilities. Life-extending health care for program A. The small difference in benefits produced for the those people will produce fewer QALYs than for people with- two groups--for example, a slightly greater life expectancy or out them, all else being equal. more serious disability averted in program A--they argue, is When health interventions are aimed at improving quality too small to justify the tremendous difference in how the two of life rather than extending life, similar discrimination can 264 | Disease Control Priorities in Developing Countries | Dan W. Brock and Daniel Wikler arise. The presence of disabilities can act as comorbidities, the practice or norms of medicine, which have the goal of making treatment less effective or more expensive (or both) meeting patients' medical needs. than it would otherwise be, thereby worsening its cost- There are strong moral reasons for considerable caution in effectiveness ratio relative to comparable treatment for persons letting health resource allocation depend on individuals' without disabilities. These effects of treatment can result from responsibility for their health needs (Wikler 2002). For that a disability that exists before treatment and is unrelated to the practice to be fair to those whose needs receive lower priority treatment provided. So it seems that a cost-effectiveness stan- because of behavior, (a) the needs must have been caused by the dard for resource allocation discriminates against such persons behavior, (b) the behavior must have been voluntary, and (c) specifically because of their disabilities. Moreover, this effect the persons must have known that the behavior would cause the will arise not only in the case of preexisting disabilities, but also health needs and that if they engaged in it their health needs in the case of patients who become disabled as a result of treat- resulting from it would receive lower priority. Smoking shows ment that is only partially effective. that these conditions are not easily satisfied. Smoking is one Several strategies to avoid this discrimination in resource causal factor in much cancer and heart disease, but many smok- allocation have been suggested. Perhaps the most plausible, at ers do not get those diseases, indicating that other factors, no least for the case of life-sustaining treatment, is to ignore dif- doubt in part genetic differences for which individuals are not ferences in patients' posttreatment quality of life as long as each responsible, also play an important causal role. Smoking is typ- patient accepts and values that quality of life and to ignore dif- ically begun when individuals are young adolescents, and as dis- ferences in life expectancy after treatment as long as each will cussed in chapter 46, it is highly addictive, which undermines receive a significant gain in life extension; obviously, what the voluntariness of continuing to smoke. Individuals in indus- counts as significant is vague and needs finer definition. trial countries are now generally familiar with the health risks of Ignoring differences in life expectancy posttreatment fits with smoking, but this is less true among less educated populations empirical evidence that individuals give little weight to dura- in developing countries, where smoking is an increasing prob- tion of benefits in prioritizing between health interventions lem. No one anywhere has been informed before they smoke that serve different individuals. that, if they do, their health needs from smoking will receive lower priority for treatment than will other health needs. Thus, it would generally be unfair to give smokers lower pri- Cutoffs for Cost per Quality-Adjusted Life Year. It is not ority for treatment of smoking-related diseases on the grounds uncommon in health care allocation to suggest the use of cut- that they were morally responsible for those health needs, offs tied to cost per QALY, although the cutoffs suggested vary although there may be other behaviors for which individuals substantially depending on the overall wealth of the country could more justifiably be held responsible. Moreover, attempt- and on the amount that it spends on health care. The cutoffs ing to make those judgments in individual cases would be can be of some value in identifying health interventions that extremely difficult and controversial. Given the difficulty of are either good or poor buys, given the society's overall wealth instituting a fair practice that allocates health resources accord- and overall level of health spending. However, it is important to ing to people's moral responsibility for their health needs, we be clear that such cutoffs should never function as anything generally have good moral reason to preserve the egalitarian more than a rough initial guide in health resource allocation. feature of the practice of medicine that looks to patients' needs The various equity considerations discussed briefly above can for care rather than to whether they deserve care. serve as justification for departing from or violating any cutoffs related to cost per QALY. ETHICS IN RESEARCH AND NEW Responsibility for Health Needs. Some have suggested that PRODUCT DEVELOPMENT health needs for which individuals are morally responsible should have lower priority than health needs for which indi- All new drugs and other medical products must be tested on viduals are not responsible (Moss and Siegler 1991). If individ- human subjects before they are sold. Although participation uals are responsible for their health needs and could have taken in health research is often a valuable opportunity for partici- steps to avoid them, they have weaker claims on social pants, what happens to them is determined not only by their resources to meet those needs than do individuals whose health clinicians' therapeutic intent (if any) but also by the need to needs are no fault of their own and could not have been pre- ensure that the research yields useful information. Managing vented. Smoking and substance abuse are two of the most the potential conflict between those motivations is often an prominent examples of behaviors often cited. However, differ- ethically challenging task, and the issues become particularly entiating patients by whether they deserve care on the basis of contentious when research is conducted in developing whether they are responsible for their health needs does not fit countries. Ethical Issues in Resource Allocation, Research, and New Product Development | 265 Developing Consensus on Ethics and Human scientists must explore the range of reasonable therapeutic Subjects Research alternatives with the patient. Potential subjects must under- The central ethical question in health research that involves stand that their participation is completely voluntary and that human subjects is what may be asked of some individuals so they may withdraw at any time and for any reason. Because that others may benefit. The question arises in any research in they cannot voluntarily shoulder risks, further protection must which human subjects are asked to participate, but is most be provided to those who cannot give consent. Such people pressing if the care that is offered to subjects provides no ther- include, among others, mentally incompetent or immature apeutic benefit or if that care is compromised by the require- participants and those involved in research (chiefly in social ments of the study design. Informed consent, while in most psychology) that requires initial deception. Consent, however, cases a requirement for ethical justification of research involv- is not sufficient to ensure fairness; there should be additional ing risk, does not relieve the scientist of responsibility. The eth- safeguards against unfair distribution of the burdens and ben- ical question is what potential subjects may be recruited for, efits of research. Finally, all research that involves potential risks even if they do consent. should be reviewed by an IRB acting on the basis of interna- A rough consensus exists worldwide on the elements of tionally recognized ethical principles. research ethics and, increasingly, on the central role of the eth- The global acceptance of these principles and the rapid ical review committee, or institutional review board (IRB). development of capacity for ethical review attest to the per- This consensus can be traced back to the post­World War II ceived validity of this system of rules and procedures of ethi- international determination to ensure that the kind of barbaric cal review. However, there has been relatively little research on research practiced by Nazi scientists would not again stain the how IRBs actually perform. Many IRBs in smaller institutions good name of medical science.1 Three advisory documents lack the necessary expertise to review novel or complex pro- have been particularly influential. The Nuremberg Tribunal posals, and their institutional setting creates a potential con- that conducted the postwar Doctors' Trial promulgated a code flict of interest. Government investigations of the adequacy of of conduct for medical research that stressed the requirement IRBs for the tasks that are now assigned to them have often of informed consent. The World Medical Association issued the been critical (for example, Office of the Inspector General first version of its Declaration of Helsinki in 1964 and has 1998, 2000). IRBs are often overworked and understaffed, revised it several times. A further set of guidelines, issued in resulting in ever-lengthening delays between initial submis- 1993 and revised a decade later, was published by the Geneva- sion of protocols and final approval. Regardless of the value of based Council for International Organizations of Medical IRBs, predicting what will pass through them and what might Sciences. Although they lack the force of law, these documents provoke delay or rejection has become a significant concern are widely acknowledged as international standards. Indeed, for medical researchers. The system thus has costs as well as the World Medical Association's periodic revisions of the benefits, a fact that lends additional gravity to the controversies Declaration of Helsinki have become focal points for interna- that it must resolve. tional debates over outstanding issues in research ethics. The most elaborate codification of research ethics is the so- called Common Rule of Conduct of the U.S. Code of Federal Goals of Ethical Review of Research Regulations (title 45, section 46), which derived from the work Although the overall purpose of ethical review is to ensure that of the National Commission for the Protection of Human research with human subjects is ethically defensible, the inter- Subjects of Biomedical and Behavior Research of the mid national consensus specifies several distinct goals that are 1970s. In addition to proposing rules governing many aspects sometimes in tension with each other: of research with human subjects, the commission proposed that the IRB be given the central role in research ethics and be · Protection. Ethical review committees can protect subjects responsible for prior review of research proposals.2 The IRB by alerting investigators to unforeseen hazards and by sug- was a compromise, granting a measure of self-regulation to sci- gesting research designs that can avoid unnecessary risk or entists and an assurance of ethical conduct to the government reduce the number of subjects exposed to risk. By insisting and the public for publicly funded investigations. that a clear explanation of risks and benefits be provided to The basic elements of research ethics engender little dis- potential participants, ethical review committees also help agreement. The research must never be brutal or inhumane, potential participants to protect themselves. Ethical review and all unnecessary risks should be eliminated. Any net risks to committees often take the name "Committee for the subjects must be justified by the prospect of potential benefits Protection of Human Subjects," reflecting a central preoccu- to others. Prospective participants must be told that they are in pation of research ethics today. a study and must be informed of its nature and its risks and · Assurance that participation is voluntary. Some research can- benefits. In the case of research that offers therapeutic benefit, not be conducted without asking some participants to 266 | Disease Control Priorities in Developing Countries | Dan W. Brock and Daniel Wikler endure discomfort or pain, to delay relief from symptoms of Current Controversies in Research Ethics their disease, or to risk other harm so that future patients Some of the most sharply disputed issues have arisen in inter- may benefit. Permitting investigators to approach potential national collaborative research involving scientists and spon- subjects in these cases requires an ethical judgment. In sors from wealthy countries conducting experiments in devel- approving such a proposal, the function of the ethical oping countries. Some of the problems are procedural. For review committee is not, strictly speaking, only to protect example, U.S. agencies have insisted on the same kind of the subjects (the goal of protection would often be served recordkeeping for IRBs in developing countries that is required more effectively by declining to do the research), but also to of IRBs in U.S. research institutions. IRBs in developing coun- permit them to be enlisted in the effort to improve health tries may accept the same principles of accountability, but they care for others. Thus, a second function of ethical review is do not have the elaborate staffs and budgets that leading IRBs to ensure that those who agree to participate do so volun- rely on. tarily and freely and that they understand what is being The most difficult disputes involving the ethics of research asked of them. in developing countries are, however, substantive rather than · Equality and fairness. Although research ethics committees procedural. have little authority to address persistent social injustices, a third concern of research ethics is that the benefits and bur- Standard of Care. The international guidelines used in navi- dens of health research be distributed fairly. This function gating the ethical dilemmas of research in developing countries receives relatively little attention in the literature of research were created for the very different purpose of ensuring that ethics, despite its prominence in such well-known docu- what happened at Dachau and Auschwitz would not recur. It is ments as the Belmont Report of the National Commission not clear whether those rules usefully resolve the kinds of for the Protection of Human Subjects (1979). Many of the dilemmas that arise in, say, Uganda or Peru. most notorious abuses of research subjects, including the The Declaration of Helsinki, following the Nuremberg Nazi investigations in the concentration camps, the Japanese Tribunal, requires informed consent of all competent research biowarfare experiments on Chinese and other civilians, subjects, and in section 29 states that "the benefits, risks, bur- and the Tuskegee research on African Americans suffer- dens, and effectiveness of a new method should be tested ing from syphilis, were committed on subjects chosen against those of the best current prophylactic, diagnostic, and exclusively from disadvantaged groups. therapeutic methods." To its supporters, any departure from the letter of the Those three goals of ethical review--protecting subjects; Declaration of Helsinki that would permit an experiment in a ensuring voluntary, informed participation; and reviewing the poor country that would be forbidden in a rich one would fairness of recruitment--are promulgated in the international constitute a double ethical standard. In their view, this clause of guidelines and in the Common Rule (and in the regulations of the Declaration of Helsinki affirms the equal importance of other countries), but they do not always point in the same human lives, regardless of wealth or nationality, and stands as a direction. For example, a research project that asks participants safeguard against exploitation of those made vulnerable by to endure a burden or risk--thus failing to offer full poverty, sickness, and absence of governmental protection. protection--can still meet the requirement of equality if the Opponents, however, argue that this position seems to rule burden is equally shared. out the possibility of testing cheap new products that may be Ethical review, thus, is not a matter of applying a checklist, effective, although perhaps not as effective as other products but it imposes an obligation of substantial ethical judgment. A that the population could not afford. If so, it would be difficult key challenge for IRBs is to earn and retain the trust of partic- to understand whom the single-standard-of-care position ipants and of the public, a task made more difficult by the would be protecting, for surely it is better for a seriously ill per- unavoidable absence of explicit criteria for approval. This prob- son to receive a good drug, even if it is not the best, than to lem is exacerbated by the institutional conflict of interest receive no drug at all.3 inherent in the placement of the IRB within the research Both points of view deserve respect. The single-standard institution, which prompts concern that the committees will approach is consistent with the postwar consensus on princi- downplay risks to subjects for projects that profit or benefit the ples of research ethics, and it offers a bright line between institution or its influential staff members. Conversely, IRBs research that amply respects human subjects and that which that are fearful of institutional embarrassment or legal sanction might result if sponsors and scientists were tempted to roam the in the event of any harm befalling research participants might globe in search of human subjects who could be used as exper- lean too far in the direction of overprotection of subjects, at the imental material with a minimum of expense or trouble. expense of important scientific research initiatives. Both con- Opponents of the universal-standard view, however, chal- cerns have been raised about the IRB system. lenge its premise. It made sense to insist on a single, universal Ethical Issues in Resource Allocation, Research, and New Product Development | 267 standard when the problem was Nazi barbarity, because the beneficiaries of the research and possibly against the immedi- prevailing standard was high and the medical criminals in the ate interests of the women in the control group. death camps denied it to the imprisoned minority--people unjustly stripped of their entitlements. In Uganda or Nepal, Rights of Host Communities. Ideally, research involving however, care at the highest world standard is available, if at all, human subjects would be a cooperative endeavor for mutual to only a small elite. advantage among free citizens who understand and endorse A full reconciliation of those points of view may not be pos- the need for research and who expect to share both in the bur- sible. The authors suggest that a relativized standard should be den of serving as research subjects and in the eventual benefit considered only when the beneficiaries will include the impov- of improved health care. Societies that recruit subjects prima- erished, sick population. Even in those cases, however, the local rily from lower socioeconomic strata fall short of this ideal; standard of care could be adopted in the experiment only if it those that do not offer new advances in care to all of their citi- met or exceeded the standard provided by other countries at zens fall even further short, raising serious questions about similar levels of development. fairness. Furthest of all from this ideal are some instances of the increasingly common practice of recruiting research subjects Placebo Controls and Other Issues Involving Research among the poorest people in the poorest countries. The means Design. For certain purposes, scientists use a placebo control for protecting human subjects in these countries are often non- even though a proven treatment exists. Patients in these control existent. Most of their citizens will be unable to afford new groups thus receive care that is inferior to what they would expe- drugs developed by firms in industrial countries. It is not clear rience in good clinical care. Until very recently, the Declaration that these subjects participate voluntarily. Their lack of scien- of Helsinki flatly condemned this practice (its current language tific education or even literacy limits their ability to understand is somewhat less restrictive), but the U.S. Food and Drug the terms of the proposed agreement with the scientists and Administration (FDA) accepted results of these trials in applica- sponsors (particularly when consent forms, on legal advice, run tions for approval of new drugs. The FDA's justification for this to 20 dense pages), and poverty often deprives them of any acceptance rests on two claims, one scientific and one ethical. alternative means of recovering their health. The first is that in certain contexts (for example, for conditions Despite these potential ethical shortcomings, international such as depression, in which eligibility criteria and outcomes are collaborative research is assured of continued growth. Some of subjective, to an appreciable extent, and in which symptoms this research targets diseases affecting mainly poor people, who fluctuate in both treated and untreated patients), active controls as a group suffer more from too little research on their popula- may produce misleading indications of equivalency, yielding tions than from too much. Even research intended to develop seemingly positive results that may be spurious. The second is therapies that will be affordable only to much wealthier that when only placebo controls can be informative, it is some- patients can be defended. Individual participants may receive times justifiable to ask participants to be randomized with better care than they would otherwise, and visiting scientists placebo and thereby to risk discomfort and distress (but not any offer employment and technical training. appreciable risk of death or long-term impairment). To right the perceived imbalance in what is asked of research Debates over placebo controls are often joined in the con- subjects in poor countries and the value that is obtained by sci- text of disputes over the appropriate standard of care that arise entists who experiment on them, some have proposed that in the case of research in developing countries, but placebo sponsors of research in the poorest countries compensate their controls are controversial in trials in high-income countries, hosts by offering a supplementary benefit (Glantz and others too. Placebo controls are one instance of a large category of 1998). One much discussed option is access following the end ethical issues in research that require weighing the importance of the study to any drugs or other therapies whose effectiveness of a scientifically ideal research design against the well-being of is confirmed in the research. The most limited proposals would participants. For example, a study of long-term chemotherapy restrict this entitlement to individuals who were enrolled in the to prevent the recurrence of breast cancer was halted before the study (those who received placebo as members of a control designated endpoints had been reached after the study's Data group, for instance), and time limits (such as three years) have Safety and Monitoring Board decided that continuing the been proposed in the case of chronic diseases such as study after a strong trend had been established favoring the HIV/AIDS. More expansive community benefit proposals have chemotherapy would be unfair to the control group. It is called for lifetime access to the treatments by all participants, notable that in this instance severe criticism of this decision their families, other members of the local community, or even was voiced by an organization representing women at risk for all citizens of the country. Other proposed benefits include a breast cancer, as well as by the editorial board of the New York specified amount of technology transfer, including scientific Times. Critics of the early termination of the trial were, training and the construction of clinics and laboratories, and in effect, aligning themselves with the interests of future cash payments earmarked for health care. A moderate proposal 268 | Disease Control Priorities in Developing Countries | Dan W. Brock and Daniel Wikler is to encourage these benefits but to require only that they be be viewed as evidence of coercion by some observers but noth- discussed and agreed on before investigations are initiated ing more than common sense by others. (National Bioethics Advisory Commission 2001). Most of these controversies can be traced back to underde- These proposals are intended to restore fairness to the rela- velopment and the inequalities of wealth and education that tionship between participants and those who benefit from prevail among and within nations today, but progress in resolv- research, including scientists and their sponsors and also future ing the ethical controversies that have become obstacles to beneficiaries of advances in medical science. Among the poten- badly needed health research must be made even as these dis- tial drawbacks are the inability to specify, even roughly, how parities persist. Viewing health research in the context of devel- much is owed to host communities; the inability to determine opment and emphasizing research that is targeted to the needs whether community benefit should be required even of of the poor majorities in poor countries can provide a context research funded by governments or philanthropists for the in which trust rather than fear or suspicion is the default benefit of people living in the host communities; and the risk response in host countries. Efforts to build capacity for ethical that placing these demands on proposed research projects will review within the host countries, such as financial support for drive them away from these very needy sites. Some of these ethical review committees, can place the locus of decision mak- uncertainties may be resolved over time as a variety of ing closer to the people who serve as subjects. Research on the approaches are attempted, particularly if they are studied and effectiveness of current ethical and regulatory requirements reported to officials in potential research sites. and mechanisms might enhance the process of ethical review These international collaborations would draw less scrutiny while reducing its bureaucratic burden. Meanwhile, the quality if it were clear that all subjects knew what they were getting into and appropriateness of ethical review of this research that takes and participated of their own free will. Although evidence on place in the sponsors' countries would be enhanced by eliciting this point is mixed, special circumstances in some countries the views of officials in developing countries, clinicians, scien- introduce problems that will have to be addressed over the long tists, and community leaders. run. Cultural differences between host populations and scien- tists may lead to conflicts over who has the authority to speak for the individuals invited to participate in a given study. NOTES Regulatory authorities in high-income countries have been 1. Because our current system of ethical review and regulation of reluctant to accept permission by a woman's husband or by a research with human subjects derives from our resolve to prevent the village chief on behalf of his people in lieu of individual con- recurrence of earlier abuses, it deserves mention that the standard histor- sent. It is often unclear--particularly from the vantage point of ical account of research ethics has been seriously incomplete. While the an IRB in Europe or the United States--whether the cultural Allies sat in judgment of the Nazi scientists at Nuremberg, abuses of sim- ilar scope and savagery practiced by Japanese biowarfare researchers on norms of the host population designate the husband or village Chinese and other civilians and prisoners of war were kept secret (and chief as decision makers in these transactions and whether their perpetrators were unpunished) following a pact with the criminal insistence on concurrent individual consent would be viewed scientists to exchange data for war crimes immunity. Moreover, the Allied governments did not always honor the Nuremberg principles. In the as intrusive or insulting. Soviet Union, scientists attempting to develop for clandestine operations Another recurring issue is whether people enrolled in a trial poisons that would not be identified on autopsy practiced their craft, with of a promising therapy who are ill and very poor can rightly be predictably lethal results, on hapless prisoners (Birstein 2001). Abuses in the United States, such as the Tuskegee syphilis study (Brandt 2000), have viewed as volunteers. The prospect of a cure for a person who been more widely publicized, but ethical lapses in large-scale Cold would otherwise die would seem to be irresistible, even if the War­related studies, ranging from radiation studies on urban populations treatment is not up to the standards that even less well-to-do (Advisory Committee on Human Radiation Experiments 1995) to surrep- citizens of richer countries would expect. Financial incentives, titious administration of mind-altering drugs such as LSD (Rockefeller Commission 1975), were state secrets. too, would predictably have a powerful effect on an individual 2. In the United States, the Office of Human Research Protections, an who may always be looking for a day's wage to feed hungry agency of the Department of Health and Human Services, has overall children. Some IRBs limit payments to compensation for lost responsibility for oversight of IRBs administering research using U.S. government funds. Its Web site is http://ohrp.osophs.dhhs.gov/. wages and travel expenses, but even at this level researchers are 3. Supporters of the single-standard view might point out that, in its asked to change the amounts offered to avoid forcing a choice current version, the Declaration of Helsinki does not require that everyone on the potential participant. As with alleged cultural differences in an experiment receive the best available care, but rather that new treat- ments be tested against the best available care. But this defense faces further regarding individual informed consent, IRBs operate with objections. In some cases, testing against the best available care (rather than scant evidence on this point. It is difficult from a long distance against the care currently provided to the population or against placebo) to decide what amount of compensation undermines freedom will fail to provide the evidence needed to convince the ministry of health of choice. It is also unclear whether the moral categories used or potential donors that funds should be provided. There is a potential con- tradiction in any view that claims both that all patients in experiments in these disputes have been adequately thought through. The deserve the best care and that it is ethically acceptable to test a new treat- fact that a poor person finds an attractive offer irresistible will ment that is not expected to be quite as good as the best currently available. Ethical Issues in Resource Allocation, Research, and New Product Development | 269 REFERENCES Meltzer, D. 1997. "Accounting for Future Costs in Medical Cost- Effectiveness Analysis." Journal of Health Economics 16 (1): 33­64. Advisory Committee on Human Radiation Experiments. 1995. Final Menzel, P., P. Dolan, J. Richardson, and J. A. Olsen. 2002. "The Role of Report. Washington, DC: U.S. Government Printing Office. Adaptation to Disability and Disease in Health State Valuation: A Birdsall, N., and R. Hecht. 1995."Swimming against the Tide: Strategies for Preliminary Normative Analysis." Social Science and Medicine 55 (12): Improving Equity in Health." Human Capital Development and 2149­58. Operations Policy Working Paper 55, World Bank, Washington, DC. Moss, A. H., and M. Siegler. 1991. "Should Alcoholics Compete Equally for Birstein, V. 2001. The Perversion of Knowledge: The True Story of Soviet Liver Transplantation?" Journal of the American Medical Association Science. Boulder, CO: Westview. 265 (10): 1295­98. Brandt, A. 2000. "Racism and Research: The Case of the Tuskegee Syphilis Murray, C. J. L. 1994. "Quantifying the Burden of Disease: The Technical Experiment." In Tuskegee's Truths: Rethinking the Tuskegee Syphilis Basis for Disability-Adjusted Life Years." In Global Comparative Study, ed. S. M. Reverby. Chapel Hill, NC: University of North Carolina Assessments in the Health Sector: Disease Burden, Expenditures, and Press. Intervention Packages, ed. C. J. L. Murray and A. D. Lopez. Geneva: Brock, D. 1988. "Ethical Issues in Recipient Selection for Organ World Health Organization. Transplantation." In Organ Substitution Technology: Ethical, Legal, and National Bioethics Advisory Commission. 2001. "When Research Is Public Policy Issues, ed. D. Mathieu. Boulder, CO, and London: Concluded--Access to the Benefits of Research by Participants, Westview. Communities, and Countries." In Ethical and Policy Issues in ------. 1995. "Justice and the ADA: Does Prioritizing and Rationing International Research: Clinical Trials in Developing Countries. Health Care Discriminate against the Disabled?" Social Philosophy and Washington, DC: U.S. Government Printing Office. Policy 12 (2): 159­84. National Commission for the Protection of Human Subjects of ------. 2002. "Priority to the Worst Off in Health Care Resource Biomedical and Behavioral Research. Office of the Secretary, U.S. Prioritization." In Medicine and Social Justice, ed. M. Battin, R. Rhodes, Department of Health, Education, and Welfare. 1979. "The Belmont and A. Silvers. New York: Oxford University Press. Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research." U.S. Department of Health, Education, and ------. 2003a. "Ethical Issues in the Use of Cost Effectiveness Analysis for Welfare, Washington, DC. the Prioritization of Health Care Resources." In Making Choices in Nord, E. 1999. Cost-Value Analysis in Health Care: Making Sense out of Health: WHO Guide to Cost-Effectiveness Analysis, ed. T. Tan-Torres QALYs. Cambridge, U.K.: Cambridge University Press. Edejer, R. Baltussen, T. Adam, R. Hutubessy, A. Acharya, D. B. Evans, and C. J. L. Murray. Geneva: World Health Organization. Nord, E., J. Richardson, H. Kuhse, and P. Singer. 1996. "The Significance of Age and Duration of Effect in Social Evaluation of Health Care." ------. 2003b. "Separate Spheres and Indirect Benefits." Cost-Effectiveness Health Care Analysis 4 (2): 103­11. and Resource Allocation 1 (1): 4. Nord, E., J. Richardson, A. Street, H. Kuhse, and P. Singer. 1995. "Who Broome, J. 1991. "Fairness." Proceedings of the Aristotelian Society 91 (1): Cares about Cost? Does Economic Analysis Impose or Reflect Social 87­102. Values?" Health Policy 34 (2): 79­94. CMH (Commission on Macroeconomics and Health). 2001. Office of the Inspector General, U.S. Department of Health and Human Macroeconomics and Health: Investing in Health for Economic Services. 1998. Institutional Review Boards: A Time for Reform. Development. Geneva: World Health Organization. Washington, DC: U.S. Government Printing Office. Daniels, N. 1993. "Rationing Fairly: Programmatic Considerations." ------. 2000. Protecting Human Subjects: Status of Recommendations. Bioethics 7 (2­3): 224­33. Washington. DC: U.S. Government Printing Office. ------. 2000. "Accountability for Reasonableness." British Medical Journal Parfit, D. 1991. "Equality or Priority." Lindley Lecture. Department of 321: 1300­1. Philosophy, University of Kansas, Lawrence, KS. Glantz, L. H., G. J. Annas, M. A. Grodin, and W. K. Mariner. 1998. Rawls, J. 1971. A Theory of Justice. Cambridge, MA: Harvard University "Research in Developing Countries: Taking `Benefit' Seriously." Press. Hastings Center Report 28 (6): 38­42. Rockefeller Commission (Commission on CIA Activities within the Gold, M. R., J. E. Siegel, L. B. Russell, and M. C. Weinstein, eds. 1996. Cost- United States). 1975. Report to the President. Washington, DC: U.S. Effectiveness in Health and Medicine. New York: Oxford University Press. Government Printing Office. Gruskin, S., and D. Tarantola. 2001."Health and Human Rights." In Oxford Scanlon, T. M. 1997. What We Owe to Each Other. Cambridge, MA: Textbook on Public Health, ed. R. Detels and R. Beaglehole. Oxford, Harvard University Press. U.K.: Oxford University Press. Solomon, J., and C. Murray. 2002. "A Conceptual Framework for Hadorn, D. 1991. "Setting Health Care Priorities in Oregon." Journal of the Understanding Adaptation, Coping, and Adjustment in Health State American Medical Association 265 (17): 2218­25. Valuations." In Summary Measures of Population Health, ed. C. Murray, Holm, S. 1998. "Goodbye to the Simple Solutions: the Second Phase of J. Salomon, C. Mathers, A. Lopez, and J. Lozano. Geneva: World Health Priority Setting in Health Care." British Medical Journal 317: 1000­7. Organization. Horsman, J., W. Furlong, D. Feeny, and G. Torrance. 2003. "The Health Weinstein, M., and W. B. Stason. 1997. "Foundations of Cost-Effectiveness Utilities Index (HUI®): Concepts, Measurement Properties, and Analysis for Health and Medical Practices." New England Journal of Applications." Health and Quality of Life Outcomes 1 (1): 54. Medicine 296 (13): 716­21. Kamm, F. M. 1993. Morality/Mortality. Volume One. Death and Whom to Wikler, D. 2002. "Personal and Social Responsibility for Health." Ethics and Save from It. Oxford, U.K.: Oxford University Press. International Affairs 16 (2): 47­55. Keeler, E. B., and S. Cretin. 1983. "Discounting of Life-Saving and Other Williams, A. 1997. "Intergenerational Equity: An Exploration of the Non-monetary Effects." Management Science 29 (3): 300­6. `Fair Innings' Argument." Health Economics 6 (2): 117­32. 270 | Disease Control Priorities in Developing Countries | Dan W. Brock and Daniel Wikler a N(a)YLL(a) a YLL(a = Chapter 15 Cost-Effectiveness Analysis YLL(pop) for Priority Setting Philip Musgrove and Julia Fox-Rushby The economic analyses in this volume focus on activities whose The subsequent section deals with the costs of interventions: main objective is to improve health. Although the chapters vary first with the question of which costs to include in the analysis, considerably, all possess, nonetheless, a common core of defi- and then with the conversion of costs in national currencies to nitions, assumptions, and methods of analysis. These are equivalents in U.S. dollars for international comparisons. drawn primarily from concepts and applications in the Oxford Despite the common assumptions and parameter values, Textbook of Public Health (Jamison 2002), drawing partly on the economic analyses differ from chapter to chapter in how the 1993 World Development Report (World Bank 1993). In this comprehensive and how exact they are, including how they chapter, we summarize and explain the common features and deal with a variety of approximations and how the results vary some of the variations of economic analysis and point the from place to place or according to differences in the assump- reader to examples throughout the book. tions. This section also contains a brief description of the dif- First is a general discussion of cost-effectiveness analysis ferences in the quality of the basic evidence and in how widely (CEA), which is the principal analytic tool used here. Here we conclusions are applicable. explain what such analysis does and does not provide, how it is Estimates of the cost-effectiveness of interventions often related to the concept of burden of disease, and how it can be describe what happens at the level of the individual patient or used, along with other criteria, in setting priorities. beneficiary. In the next section, we suggest two ways to consider Because CEA is applied to specific interventions, the next costs and outcomes at the population level, allowing for large section describes the several meanings of that term and the way differences among countries in the size of population; the inci- that interventions are classified and evaluated. It is essential to dence or prevalence of a disease, condition, or risk factor; and understand what is being analyzed before considering in detail the amount spent or available to spend on an intervention. how the analysis is conducted. In the final section, we indicate how the type of analysis pre- Estimating the effectiveness of an intervention requires sented in this volume might be improved and how it can be specifying the units in which that concept is measured. This applied to help set priorities among the large number of inter- action in turn requires choices of several parameter values, ventions to which limited resources can be applied. including, in the analyses reported here, the discount rate applied to future years; the disability weights that describe the COST-EFFECTIVENESS AND PRIORITY SETTING severity of diseases and conditions, corresponding to the health losses that they cause; and the life expectancy at different ages, The principal analytic tool throughout this volume is which determines how many years of healthy life can be saved CEA, which compares the cost of an activity, called an inter- by averting a death or preventing or treating a long-term health vention, with the known or expected health gain. The result is problem. We also consider briefly the nonhealth benefits that summarized in a cost-effectiveness ratio (CER), as explained may result from a health intervention. more fully below. This ratio corresponds to the concept of 271 (health) value for money. Favoring activities that are more circumstances); vertical equity (priority for people with worse cost-effective over those that are less so is consistent with the problems); adequacy of demand; and public attitudes and ethical view that "limited resources for health should be allo- wants. Two criteria--whether an intervention is a public good cated to maximize the health benefits for the population and whether it yields substantial externalities--are classic justi- served" (chapter 14). fications for public intervention, because private markets could Cost-effectiveness provides the clearest simple way to pro- not supply them efficiently, just as in other sectors. As noted in mote value for money in health: hence, the emphasis on it here. chapter 1, the interventions analyzed in this volume are not CEA allows comparisons throughout the health sector and not limited to public or semipublic goods. The emphasis is on value only for the same health outcome. It does not allow comparison for money--that is, whether an intervention is worth buying, to nonhealth outcomes unless these outcomes can be incorpo- not who pays for it. Nonetheless, when one is choosing which rated into costs, and the calculation of the CER by itself makes public goods to buy, several criteria become irrelevant, and no pretense of monetizing the intrinsic value of health. To use cost-effectiveness can be used as the chief or even the only con- CERs for choosing what to buy and what not to, decision mak- sideration. Cost-effectiveness can similarly determine what to ers must determine a maximum willingness to pay for units of include in a mandatory universal public package of health care health gain, unless other criteria are considered to justify buy- alongside competitive voluntary insurance (Smith 2005). ing something with relatively poor cost-effectiveness. Cost-effectiveness can conflict with both kinds of equity-- For risk factors, CEA requires estimating the gain in health that is, the more cost-effective of two interventions may also consequent on introducing an intervention to reduce the risk of lead to a less equitable distribution of health benefits. Equity acquiring or transmitting a condition. For packages of inter- and cost-effectiveness are compatible when a cost-effective ventions or elements of the health system, such as hospitals, intervention is provided to only part of the population that effectiveness is estimated by judging how much mortality and would benefit from it because everyone in the group suffers morbidity would be reduced by providing the whole package or from the same problem. Then expanding coverage will gener- set of services or by operating the facility. With some excep- ally also promote horizontal equity. These equity effects are tions, the analyses may describe but do not quantify the non- reinforced when those who are better off already benefit while health benefits of an intervention. Apart from the difficulty of the poorer and sicker population does not. Choices about ver- obtaining enough data, such quantification requires attaching tical equity--doing more for those in greatest need--are more values to nonhealth outcomes, which is problematic when com- complicated. Doing very little for people with severe health parisons are made over large cultural and income differences. problems--because the available interventions for those prob- All comparisons are relative, with no absolute distinction lems are not very effective at reducing their suffering--is not between being and not being cost-effective. In assigning priori- necessarily preferable to doing more for people with less severe ties among interventions for public funding or for other policy problems that are more amenable to intervention. When an actions, one must also consider the magnitude of health prob- intervention is reaching only part of a potential beneficiary lems to which interventions apply because that affects what is population and those not benefiting tend to have more severe affordable. Calculations of the effect of spending US$1 million illness, then expanding coverage can improve both horizontal or the total cost and health gain in a population of 1 million and vertical equity. Where possible, chapters consider the equity people offer ways of looking at such choices. Equity, poverty, effects of expanding or changing interventions. and risk of impoverishment from ill health may also influence priorities; so do the budgets available--and the decisions of Cost-Effectiveness and Disease Burden how much to make available--for buying interventions. Finally, Cost-effectiveness and disease burden are related because the effectiveness of an intervention and, therefore, the degree to effectiveness is the reduction in burden caused by an interven- which it deserves priority depend on how far it is culturally tion. This relationship holds true at the individual level. The appropriate or acceptable for the population it is intended to magnitude of a health problem--the total burden in the benefit. The identical intervention, technically speaking, may population--is irrelevant for marginal changes in resource lead to different degrees of use or compliance in different pop- allocation. However, it matters for large changes from the sta- ulation groups, and information and incentives may be needed tus quo. Health interventions demand managerial capacity as to achieve the full potential outcomes. well as financial and physical resources, and managerial ability Cost-effectiveness is only one of at least nine criteria rele- may be stretched thin if it has to deal with a large number of vant for priority setting in health if the object is to decide how interventions. In consequence, it may be efficient to concen- to spend public funds (Musgrove 1999). Cost matters by itself, trate on relatively few and somewhat less cost-effective inter- as do the capacities of potential beneficiaries to pay for an ventions, provided they attack substantial burdens, rather than intervention. The other criteria that may affect priorities many other interventions that are more cost-effective but include horizontal equity (equal treatment for people in equal affect only small burdens. Moreover, even for a cost-effective 272 | Disease Control Priorities in Developing Countries | Philip Musgrove and Julia Fox-Rushby intervention, high prevalence or incidence may make the cost cial resources in a deliberate attempt to improve health by of covering the whole potential beneficiary population prohib- reducing the risk, duration, or severity of a health problem itive. The authors of chapter 21 indicate how expensive it (Jamison 2002, table 2). The term usually refers to an activity would be to protect all at-risk African children from malaria undertaken by a health system rather than by an individual. with bednets, even though bednets are highly cost-effective. The emphasis on a deliberate, systemic effort means that an Conversely, an intervention that costs more per health gain intervention is not simply anything that improves health; for may be affordable and given priority if it treats a manageable example, if more rainfall leads to higher crop yields and better burden of disease and corresponds to a small beneficiary nutritional status, the rain does not count as an intervention. group. Priority turns on the available budget relative to the cost Similarly, although breastfeeding protects infants' health, it is of a program; on how divisible a program is (that is, how easily not itself an intervention as the word is used in this volume. In it can be operated at different scales, as a technical or political contrast, a program to encourage new mothers to breastfeed is matter); and on whether interventions are mutually exclusive an intervention (as described in chapter 27). How effective (Karlsson and Johannesson 1996). such a program is, of course, depends on how many mothers it Because of the interaction between cost-effectiveness, persuades to adopt the practice when they are neither currently disease burden, and available funds, no single threshold of breastfeeding nor planning to do so. maximum cost per health gain exists below which an interven- Interventions can be directed against an injury or disease tion is "cost-effective." A rule of thumb, such as that any inter- (such as trachoma), a condition associated with or deriving vention is worthwhile if it costs less than two or three times from a disease (such as blindness), or a risk factor that makes income per capita, ignores this interaction and is an inadequate the disease or condition more likely (such as the lack of hygiene guide to priority setting. However, even an intervention that is that leads to trachoma). An intervention may pursue primary considered justified by cost-effectiveness may be infeasible to prevention at the population level--promoting personal deliver, for example, if the costs are monetary and come from behavior change, controlling environmental hazards, or deliv- the public budget but the benefits are nonmonetary and dif- ering a medical intervention such as immunization to a large fused over the population. Economic theory would suggest population--or individual action for primary prevention, removing the current budget constraint by raising more rev- cure, acute management, chronic management, secondary pre- enue until the marginal social cost of the interventions plus the vention, rehabilitation, or palliation. Box 15.1 defines these cost of obtaining the revenue equals the marginal social bene- terms, and the figure in the box illustrates how interventions fit. Although theoretically attractive, this escape from resource may prevent ill health events or deal with their consequences. limitation may not be possible because of political reasons or Characterizing an intervention fully also means distinguishing because the economic cost of raising extra taxes is prohibitive. the level at which it is delivered (home, primary care facility, Because so many criteria can affect priority setting and district hospital, or referral hospital); indicating whether it because evidence on cost-effectiveness in low- and middle- involves drugs, immune enhancement, surgery, or physical or income countries is so scarce, health system policies and psychological therapy; and determining whether it requires a budgets seldom derive purely from considerations of cost ver- physician or uses diagnostic, laboratory, or imaging proce- sus outcomes. Even in high-income countries, where more such dures. Such procedures are most often evaluated relative to the analyses are available, their effect has been limited, although it is interventions they screen for or lead to, because they produce growing (Gabbay and le May 2004; Glick, Polsky, and Schulman no health gain by themselves (although the information they 2001; Hoffmann and others 2002; McDaid, Cookson, and provide can be valuable for reassurance or for promoting ASTEC Group 2003; Sheldon and others 2004; Taylor and NICE behavioral changes). 2002). Cost-effectiveness studies are now required by, for An intervention in the everyday sense includes such activi- example, the U.S. Food and Drug Administration for labeling ties as immunizing a child, performing a surgical procedure, or claims, the National Institute for Clinical Excellence before treating an infection with antibiotics. The authors of some advising national policy on treatments and care in England, chapters use the term only in this sense--for example, in dis- and the Ministry of Health in the Netherlands for new drugs cussing interventions that contribute to meeting the (iMTA 2005). Millennium Development Goals (chapter 9). Authors of other chapters use the term in several other senses as well. It can DEFINITION AND CHARACTERISTICS mean modifying an existing intervention--for example, OF INTERVENTIONS adding Haemophilus influenzae type B (Hib) antigen to the Expanded Program on Immunizations (EPI). Immunization The object of a CEA--the thing to which it is applied, the costs against Hib is itself an intervention, but instead of analyzing it and outcomes of which are to be analyzed--is an intervention. separately, one can use CEA to evaluate the additional cost of An intervention is an activity using human, physical, and finan- incorporating that antigen and the additional health gain that Cost-Effectiveness Analysis for Priority Setting | 273 Box 15.1 Intervention Categories, with Examples The following figure illustrates how interventions are · Primary prevention--to reduce the level of one or more related to a health event; the definitions of these categories risk factors, to reduce the probability of initial occur- are given below. rence of disease (medication for hypertension to pre- vent stroke or heart attack), or to reduce the likelihood Before Event After of disease when the risk factor is already present (pro- Primary prevention Secondary prevention phylaxis for sickle cell anemia). of the occurrence of a risk of another event of the · Secondary prevention following the occurrence of factor or same kind (for example, disease--either to prevent another event of the same of an existing risk a second stroke) or factor developing into an of a related event of a kind or to reduce the risk of a different but related event adverse health event different kind (for (medication to reduce the likelihood of a second coro- example, heart attack after stroke) nary event or a first heart attack after stroke). · Cure--to remove the cause of a condition and restore Case management function to the status quo ante (surgery for appendicitis) Cure Acute care · Acute management--short-term activity to decrease Chronic care the severity of acute events or the level of established Rehabilitation Palliation risk factors, to minimize their long-term impacts (thrombolytic medication following heart attack, Interventions Related to the Occurrence of an Adverse Health angioplasty to reduce stenosis in coronary arteries). Event · Chronic management--continued activity to decrease Population-based interventions all aim at primary preven- the severity of chronic conditions or prevent deteriora- tion (as defined later), are directed to entire populations tion (medication for unipolar depression, insulin for or large subgroups, and fall into three categories: diabetes). Chronic management can include some sec- ondary prevention. · Promoting personal behavior change (diet, exercise, · Rehabilitation--full or partial restoration of physical, smoking, sexual activity) psychological, or social function that has been damaged · Control of environmental hazards (air and water pollu- by a previous disease or condition (therapy following tion, disease vectors) musculoskeletal injury, counseling for psychological · Medical interventions (immunization, mass chemopro- problems). phylaxis, large-scale screening, referral). · Palliation--to reduce pain and suffering from a condi- tion for which no cure or rehabilitation is currently Personal interventions are directed to individuals and can available (analgesics for headache, opiates for terminal be intended for the following: cancer). Source: Authors. is expected to result (see chapter 20). The intervention studied (chapter 20) and malaria (chapter 21) provide explicit esti- is then not Hib immunization as such but the change in the full mates of the differential costs of expanding coverage. vaccination procedure. A change in the scale of an existing Adding one intervention to another to deal with the same activity can also be considered an intervention, even if the disease or condition is also an intervention, and combinations activity itself is unchanged: that is, one can analyze the change of interventions can be analyzed to determine which is most in costs and in outcomes associated with expanding or cost-effective or how the cost-effectiveness of one intervention contracting the coverage of the activity--for example, extend- depends on the other activities with which it is combined. ing antiretroviral treatment for HIV and AIDS to a larger pop- Examples include successively adding drugs for treatment of ulation (chapter 18) or screening more newborns for sickle cell epilepsy (chapter 32) or secondary prevention of cardiovascular anemia (chapter 34). In most chapters, the authors assume that disease (chapter 33) or combining several quite different inter- expansion affects costs and outcomes linearly, so that the CER ventions to control tobacco addiction or alcohol (chapters 46 does not change. The chapters on vaccine-preventable diseases and 47, respectively). The analyses of community health and 274 | Disease Control Priorities in Developing Countries | Philip Musgrove and Julia Fox-Rushby nutrition programs (chapter 56) and integrated management of interventions in themselves, and as such are crucial for infant and childhood illness (IMCI; chapter 63) define "the controlling HIV and AIDS (chapter 18), promoting better intervention" as a whole program incorporating several differ- infant and child care (chapters 20 and 27), preventing inherit- ent activities. Generally, even less empirical evidence exists con- ed disorders (chapter 34), encouraging healthful diets and cerning combinations of interventions than for individual exercise (chapters 44 and 45), and avoiding addiction (chap- activities, but IMCI is an exception; it has been evaluated more ters 46­48). But they can also be used to improve the effec- thoroughly than most single interventions. tiveness of other interventions by increasing awareness and Box 15.2 includes a more detailed discussion, using a hypo- demand, combating mistaken beliefs about diseases and risks, thetical example of three different ways to deliver immuniza- or reducing anxiety and stigma. In that sense, information, tion, of how CEA can be applied to four of the meanings of education, and communication delivered to consumers or intervention used here: an existing intervention at its current providers or both are examples of policy instruments. They coverage, changes in the scale of that intervention, the addition can facilitate or promote the use of such interventions as con- of one intervention to another when expanding coverage, and dom distribution, screening for diseases or congenital disor- the complete shift from one intervention to a different (and ders, prenatal care, or immunization. more cost-effective) one. Other activities that can be classified either as interventions Depending on the comparison undertaken, the result may or as policy instruments include the following: be an average cost-effectiveness ratio (ACER) or an incremen- tal cost-effectiveness ratio (ICER). The former compares total · Measures to increase the quality of care, such as some kinds of costs and total results, starting from zero, whereas the latter staff training or the introduction of better recordkeeping. compares additional costs and additional results, starting from These activities may simultaneously affect a large number of the current or some other level of coverage of an intervention. specific interventions in a health facility (chapter 70). Either shifting completely from one intervention to another · Legislation and regulation to impose an intervention (for or partially replacing one with another may reduce costs while example, limiting the salt content of foods, chapter 45, producing more health gain. For example, if spending is high or requiring that salt be iodized, chapter 28); to limit or on hospitalization for acute myocardial infarction, a program prohibit an intervention that is ineffective or dangerous or using a "polypill" (several medications in a single pill) would to reduce unhealthful behavior such as smoking and exces- reduce expenditures by lowering incidence (chapter 33) and sive drinking (chapters 46­47); or to codify how an inter- would be cost saving, because less hospitalization would be vention should be delivered and determine who may pro- needed. If the status quo is no hospitalization (as is typical at vide it, as by licensing doctors, nurses, and health facilities low incomes), a polypill program increases costs but may more (chapter 71). than correspondingly increase health gains and therefore be · Economic incentives, which can take the form of subsidies or more cost-effective. If the polypill both reduces costs and taxes (chapter 11) for particular items of consumption other improves outcomes compared with hospitalization, it is said than health goods or services, such as tobacco and alcohol to dominate a hospital-only strategy. The second figure in (chapters 46­47) or condoms to reduce HIV transmission box 15.2 illustrates the concept of dominance; table 45.4 and (chapter 18), or can be provided through protection of box 45.1 of chapter 45 provide examples of interventions that property rights, as for patented drugs (chapter 72). are dominated by others. Unfortunately, reliable information on current intervention coverage, costs, and results is not always available even in high- These activities of informing, mandating, legislating, regu- income countries (iMTA 2005) and is extremely scarce in low- lating, and taxing or subsidizing, which are at one remove or and middle-income countries. Studies showing whether an more from medical interventions, are also often called func- intervention is effective or cost-effective seldom cover the tions of the health system (WHO 2000, chapter 2; see also entire potential beneficiary population, and service provision chapter 9 in this volume). Several of these instruments may be in the private sector is often not recorded. Many chapter used together, such as increased taxes on tobacco or alcohol authors describe only the ACER of an existing or potential along with measures to educate consumers and to restrict the intervention, whereas others explicitly compare alternatives to times, places, or quantities of consumption. Sometimes the current practice (for an example, see chapter 16). instrument is needed before introducing or expanding an Many of the activities analyzed here aim at promoting intervention to overcome barriers to its use or to make it cost- changes in personal behavior, by informing and persuading effective enough to be worth pursuing. Educating the affected individuals to eat differently, to avoid smoking and excessive population, for example, is crucial to screening and treatment alcohol, to reduce the risks of sexually transmitted infections, of cancers and hemoglobin disorders. The need for a particular or to practice better hygiene. Such efforts can be considered instrument may vary from country to country even if the Cost-Effectiveness Analysis for Priority Setting | 275 Box 15.2 Average and Incremental Cost-Effectiveness and Intervention Choices In the figure below, which compares three ways of deliver- Raising immunization coverage at an affordable cost ing immunization, point X describes the status quo of a may require adopting the alternative of mobile vaccina- current intervention, delivering immunization by means tion teams, intervention Y. The hypothetical combination of fixed facilities. At point X, the intervention achieves a of fixed facilities and such teams allows increasing the total effect E2 (measured as coverage or as disease reduc- effect to E4 (complete or nearly complete immunization) tion) at a total cost C2. The ratio C2 to E2 is the average at a total cost of C4. The ICER of the mobile teams is cost-effectiveness ratio (ACER), shown by the slope of the shown by the slope of the line X­Y and the resulting over- line O­X. Beyond point X, expanding coverage becomes all or combined ACER by the slope O­Y. Adopting inter- very costly, perhaps because the population not yet immu- vention Y would be clearly preferable to trying to expand nized is dispersed and hard to reach. (Chapter 20 includes coverage through intervention X by building and staffing estimates of how costs increase as immunization coverage more fixed facilities. expands but without introducing a sharp increase in An alternative even better than Y might subsequently costs.) Expansion to point X1, which increases the cost be developed, represented by point Z--for example, com- from C2 to C3, yields only a small increment E3­E2 in munity-based immunization teams that could operate effect. The slope of the line X­X1 represents the incre- either near or far from fixed facilities because they use heat- mental cost-effectiveness ratio (ICER) of that expansion, stable vaccines that do not require a cold chain. The ICER which would raise the ACER to line O­X1. The line X­X2 of turning to that choice, represented by the line X­Z, is not shows the alternative of reducing coverage, which would only more favorable than intervention Y, but it is even bet- improve the average cost-effectiveness (to C1/E1) because ter than the current ACER, and preferable to intervention X marginal costs are rising steeply near point X. The ICER at any coverage level beyond X2. The cost-effective choice, of the reduction in coverage is the ratio of C2­C1 to therefore, is not to retain intervention X at its current level E2­E1. and add Z beyond that point but to switch entirely from X (or from X plus Y, if Y has already been adopted) to Z. Because it costs less but provides a better outcome, Z is said to dominate both X and Y. The following figure illustrates Total cost dominance of one intervention by another, as well as cases in which neither of two interventions is dominant. Through mobile teams C4 Y Maximum acceptable New incremental cost per intervention disability life year more costly Through community averted C3 teams using heat-stable X1 vaccines (no cold chain) Old intervention X New intervention W Z dominates more effective but more costly Through fixed facilities C2 New New X Z 2 intervention X intervention less effective more effective C1 X2 New intervention New intervention Z V less effective dominates but less costly New O E1 E2 E3 E2* E4 intervention Total effect (coverage or health gain) less costly Average and Incremental Cost-Effectiveness and Intervention Comparison of Cost and of Effectiveness between Interventions: Choices: Comparison of Three Ways to Deliver Immunization Conditions for Dominance 276 | Disease Control Priorities in Developing Countries | Philip Musgrove and Julia Fox-Rushby If intervention Z is divisible (meaning that it can be the upper left quadrant. An intervention such as V or W operated at any desired scale, such as Z2), then it is prefer- may or may not be considered preferable to X (V is cheap- able to X at a cost of C2 because of the additional effect er but also less effective, and W is more effective but also E2*­E2. It can be extended all the way to E4, just as with more costly). Whether either such intervention would be intervention Y, provided only that the ICER represented by selected over X depends on the relation of the increased the slope X­Z is still acceptable to decision makers choosing (or decreased) cost to the increased (or decreased) effec- how far to expand the intervention. That is, the cost must tiveness. That ratio corresponds to an ICER. If a maxi- still appear to be justified by the increased coverage. Under mum acceptable, or threshold, value for the ICER is deter- either of these conditions, an obstacle to switching, or to mined, as shown by the dashed diagonal line, then any doing so quickly, would exist only if substantial fixed costs intervention that falls below the dashed line would be accompanied the transition from one intervention to the acceptable (preferable to X), and those that fall above the other, such as recruiting or retraining staff, building health dashed line would not be. Uncertainty about the estimates posts in communities, or setting up the system for dis- of cost and effectiveness means that, instead of a sharp line tributing the new heat-stable vaccines. as in the figure, the division of preferable from nonprefer- Compared with intervention X, intervention Z is better able interventions corresponds to a zone of some width in both dimensions (lower cost and greater effectiveness), that depends on the confidence intervals around the esti- so it is to be preferred, and is said to dominate X. However, mates. This kind of comparison can start from an existing intervention X would dominate any other treatment that intervention such as X in the first figure or, when there is is both more costly and less effective and, therefore, falls in currently no intervention, from point O in the first figure. Source: Authors. intervention that it facilitates is identical, because the legal, uncertainty about the future and preferences for timing of con- regulatory, or financial environment differs. sumption, and it avoids two problems. First, outcomes that potentially generate benefits forever, such as smallpox eradica- tion, appear to have infinite benefits if the future is not ESTIMATING EFFECTIVENESS IN HEALTH discounted and therefore seem to justify any finite cost at all. Second, it makes little sense to postpone interventions forever Using cost-effectiveness for resource allocation requires health simply because funds to finance them could be invested today effects to be represented in common units in order to facilitate and be worth more tomorrow. Even discounting the future at comparison across interventions, diseases, or conditions. All the low rate of 3 percent annually has a substantial effect--that analyses start with some natural unit: cases of disease or injury, is, dividing the values for future years by successive powers of deaths, or numbers of people who quit smoking or adopt some 1.03. That means dividing values for year 1 by 1.03; those for other health-improving behavior. All interventions that avert year 2 by 1.03 squared, or 1.0609; and so on. At that rate, avert- death are alike in that regard. Preventing a child's death at a par- ing an infant death saves not all the 80 calendar years of life ticular age, independent of the cause, means that the averted expectancy at birth (or fewer in low-income countries) but at death alone is an adequate measure of outcome. However, when most 30 discounted years. lives are saved at different ages--averting death from malaria at For interventions that avert mortality, analysis starts by age 2 versus death from beta thalassemia at age 10­15--the out- estimating the deaths prevented, uses age at death to yield come is no longer identical, and some measure must account numbers of life years saved, and then discounts those years as for the difference in years of life saved. These cases provide described above. When interventions improve health by avert- another natural unit, subject to estimating how much longer a ing or reducing nonfatal disability, different disabilities must be person spared death might live. The choice of life expectancy to compared in severity. As with mortality, age at the time of assume for such calculations is discussed later. intervention matters for long-lasting conditions, and so does The unit of time becomes a less natural and more synthetic discounting. In contrast, age is irrelevant for episodes of illness measure if the future is discounted, as in all these analyses. or injury that are self-limited or quickly resolved by interven- Discounting means reducing the value of each variable in each tion, because the duration of ill health does not depend on age, future year by an amount that increases the further in the and all ages are treated alike in this analysis. Discounting also future that year is. The discounting procedure reflects inherent makes little difference over short intervals. Cost-Effectiveness Analysis for Priority Setting | 277 Unit of Measurement of Health costs and effects, permitting all such comparisons. Knowing The common unit of health loss or gain used here takes into that one intervention achieves the same results as another at account duration and severity, as well as discounting the future. lower cost, which would be relevant if they were the only two The disability-adjusted life year (DALY) is a unit introduced by possible interventions against a common problem, is not the World Health Organization (WHO) and the World Bank enough. Comparing both with another intervention with dif- (Jamison and others 1993; Murray 1996; World Bank 1993). As ferent effects may also be necessary. For example, a coronary previously discussed, the DALY incorporates assumptions and artery bypass graft for myocardial infarction costs, on average measurements about severity of nonfatal conditions, age at across regions, US$37,000 per DALY gained, compared with an incidence or intervention, duration with and without interven- average of only US$409 for the polypill discussed earlier. tion, and remaining life expectancy at that age. For interven- However, both are much more expensive than saving life years tions directed to risk factors rather than diseases, the analysis for a middle-aged person by treating active tuberculosis (and incorporates estimates of reductions in diseases that result thereby preventing transmission), an intervention that costs from changes in the level of risks. Smoking cessation, for only US$15 per DALY in the absence of HIV infection, or example, reduces deaths from both cardiovascular disease and US$102 on average where coinfection makes treatment more cancer (chapter 46). difficult. (In only a few cases do chapters deal explicitly with Published analyses, particularly in high-income countries, comorbidity, in part because the DALY approach considers often use not DALYs but quality-adjusted life years (QALYs), an conditions only individually.) alternative measure of how much a year of life is worth if a per- Parameter values for effectiveness are required in order to son suffers one or more limitations of various kinds and conduct CEA: how to value disability, compared with mortality; degrees. QALYs can be estimated directly using a valuation how to treat the future; and whether to distinguish people method such as the time tradeoff (comparing and varying the according to age, sex, or other characteristics. Because effec- time spent in one health state with time spent in another state tiveness is related to reduction in disease burden, nearly all until the quality of life is judged the same in both). these parameter choices coincide with those adopted to esti- Alternatively, a prescored questionnaire such as the EQ5D (a mate the existing burden (see Mathers and others 2005 for a European quality of life measure) could be used. The EQ5D full explanation). distinguishes three grades--no problem, an extreme problem, Because disease burden estimates discount the future at and total disability--on each of five dimensions of life 3 percent annually, CEA in this volume does the same, for both quality--mobility, self-care, performance of usual activities, effects and costs. This method follows the recommendations of pain or discomfort, and anxiety or depression (Brooks, Rabin, the U.S. Public Health Service Panel on Cost-Effectiveness in and de Charro 2003). Discounting of QALYs occurs as an addi- Health and Medicine (Gold and others 1996) and appears tional step, although some concern exists that discounting appropriate whenever the benefits of an intervention begin values derived from the time tradeoff approach is double immediately. Constant discounting (using the same percentage discounting (Dolan and Jones-Lee 1997). rate each year) undervalues interventions for which the bene- QALYs allow comparison among interventions and can eas- fits appear long after the costs have been paid. Immunization ily account for comorbidity. Although the concept of DALYs against hepatitis B can prevent liver cancer decades later (chap- averted by an intervention is similar to that of QALYs gained, ter 20) but, compared with the costs incurred at the moment of no systematic formula exists for converting between DALYs vaccination, appears less cost-effective if the health gain is and QALYs except in broad approximations (Fox-Rushby heavily discounted during that interval. Slow discounting, with 2002). This gap is partly because DALY disability weights are the rate falling close to zero for the more distant future, would specific to diseases whereas the QALY system of evaluation is yield a higher present value of benefits (Jamison and Jamison not (it is based on overall health status). Authors sometimes 2003), but given the absence of consensus on the correct form, report effectiveness results in QALYs, because they cite studies the analyses here use constant discounting. in high-income countries that often use QALYs. When some The limitations from a disease or condition in the absence of interventions are evaluated in DALYs and others in QALYs, intervention are measured by disability weights (Mathers and ranking interventions according to cost-effectiveness may still others 2005), despite some controversy as to whether they ade- be possible (see chapter 29 for examples of the use of both quately capture all the disability (see chapter 24 on helminthic units). infections). These weights range from zero for perfect health to Priority setters sometimes stop exercises in priority setting 1.0 for death. Authors have made their own estimates whenever after concluding that something is or is not cost saving, without WHO did not provide any disability weight because the asking whether an intervention yielding a different outcome corresponding condition was not explicitly included in the (against a different disease, for example) would be still more burden of disease. For example, weights for anemia caused by cost-effective. In this volume, the intent is to estimate both hemoglobin disorders (chapter 34) were taken from other 278 | Disease Control Priorities in Developing Countries | Philip Musgrove and Julia Fox-Rushby causes of anemia. Note that years lost to early death also are transmission of HIV and AIDS or prevented all deaths from DALYs, since they include the disability weight of 1.0. malaria would do that. Given the absence of evidence that any When an intervention prevents or completely cures a condi- intervention actually has such a substantial effect, it is assumed tion, the postintervention disability is zero. For partially suc- in this volume that individuals face the same probability of cessful interventions leaving residual disability, the disability death at each subsequent age as the existing population does. weight is reduced but not eliminated. WHO has sometimes This assumption makes interventions appear less effective estimated weights for "treated" as opposed to "untreated" con- when overall mortality is high than when mortality is low. For ditions (Murray and Lopez 1996, annex table 3) without spec- example, averting an infant death in Sub-Saharan Africa will ifying the intervention. This distinction is introduced for some save, on average, only 44 to 49 undiscounted life years consequences of chronic conditions: cancers not yet in the ter- and should not be credited with saving 80 or more. Cost- minal stage, diabetic conditions, major psychological disorders, effectiveness calculations and estimates of burden of disease are cataracts, various cardiovascular conditions, chronic respira- inconsistent in that fully effective interventions appear able to tory conditions, ulcers, arthritis, cleft lip and palate, edentulism deal with only part of the burden they aim to control. Regional (total loss of teeth), and some burns. Chapters 31 and 33 use rather than standard life expectancy also makes interventions these values to describe intervention outcomes. in a high-mortality region appear more effective relative to WHO burden-of-disease estimates used in the first edition outcomes where mortality is lower, when they avert deaths later of Disease Control Priorities in Developing Countries (Jamison in life. and others 1993) incorporated age weights--that is, numbers Nonhealth benefits of health interventions sometimes attempting to describe the relative value of life at different ages. should be taken into account, because many health interven- These numbers were calculated to keep the discounted integral tions also yield other kinds of benefits. They often make bene- from age 0 to 80 the same, as if no age distinction were made. ficiaries mentally or physically more productive, better able to The weights are zero at birth, ignoring health losses from still- continue in and learn from school or to work and earn more. birth prior to live birth; reach a maximum at age 25; and This benefit occurs particularly with interventions against non- decline almost to zero at advanced age. They are a particularly fatal consequences, as indicated in the chapters on malnutri- controversial element in the burden estimates (Musgrove 2000) tion (chapter 28), malaria (chapter 21), helminthic infections because they value some years of life more than others, and lit- (chapter 24), tropical diseases (chapters 22­23), psychiatric dis- tle evidence suggests what an appropriate weight should be. In orders (chapter 31), and learning and developmental disorders consequence, only constant age weighting (treating all years (chapter 49). Interventions that prevent injury or restore work alike) is used in these analyses. Removing age weights makes no capacity also have such effects (chapters 39­40 and 60), as difference to an intervention that averts an infant death, but it do interventions against diseases that kill in the prime of changes the relative importance of interventions at later ages. life, notably tuberculosis and AIDS (chapters 16 and 18). Because life is more highly valued at advanced ages, death and Nonhealth benefits also occur as time is saved when piped disability after age 38 become more important compared with water is made available, as less cleaning is needed when air events before that age, and interventions later in life become pollution is reduced (chapter 42), or as property damage is more cost-effective. Some estimates used here therefore differ reduced by improved traffic safety (chapter 39). from those published previously by WHO even when all the Several chapters include discussions of the nature and, other parameters are unchanged. where possible, the magnitude of nonhealth benefits from The only parameters for CEA that differ from those in the health interventions. This factor is important when the health burden of disease concern life expectancy. In estimating bur- benefits, although substantial, are so costly that interventions den, people at any age and in all regions are assumed, on the do not appear cost-effective on health grounds alone but may ethical criterion of valuing all lives equally, to have the same life be justified by large nonhealth returns. Safe water and sanita- expectancy. The only exception is that at birth males appear to tion services are the classic example (chapter 41). Because dif- have a biologically determined (not behavior-related) life ferent types of benefits--health gains, increased income, time expectancy of 80 years, which is shorter than the life expectancy saved--cannot be compared directly, the only way of combin- of females by 2.5 years. However, applying these expectancies to ing them into a single expression is usually to evaluate all out- CEA will overstate the effects of interventions when life comes in monetary terms. (An exception occurs when some expectancy in a population is low. Averting a death at age 5 in monetary gains can be measured directly--for example, Sub-Saharan Africa or South Asia does not confer a high prob- increased worker productivity from better health. Those gains ability of living to age 80 or longer. Competing causes of death can be subtracted from costs and incorporated into CEA with- reduce the effectiveness of any single intervention, unless it out attributing a monetary value to the health gains.) Most affects so large a population that it actually increases life chapters that account for any nonhealth benefits simply offer expectancy. An intervention that completely interrupted the descriptions of them rather than incorporating them into Cost-Effectiveness Analysis for Priority Setting | 279 monetary outcome indicators. Chapter 7 is an exception; it producing an intervention, the U.S. Public Health Service compares gains in welfare from living longer or in better health guidelines (Gold and others 1996) recommend including the with those from higher income. indirect costs to patients and their families of consuming it. Estimating the monetary value of all benefits and adding This recommendation means, in particular, the value of time them together for comparison with cost is what cost-benefit needed for travel, waiting, and undergoing medical tests and analysis does. Interventions are considered justified in absolute procedures, or the value of time used in caregiving, as well as terms if the benefits exceed the costs. However, when faced with any income forgone during treatment. Externalities, or costs constrained budgets that cannot finance all interventions imposed on third parties, such as on the school system or the whose benefits are greater than costs, policy makers need to environment, should also be included. The analyses in this vol- establish some minimum acceptable rate of return. This choice ume generally exclude such costs and report only the direct is parallel to the need to set a maximum on cost per unit of costs of delivering interventions, partly because published health gain when choosing according to cost-effectiveness. analyses seldom include the various indirect costs, and they are Published analyses of health interventions sometimes use harder to estimate. Walker and Fox-Rushby (2000) found that cost-benefit analysis, so results following that method are only 20 of 101 studies included some element of indirect cost- incorporated in some chapters here. The decision to emphasize ing. Valuing time according to local wages or income, for exam- CEA instead derives from two chief considerations. One is that, ple, may underestimate how valuable time actually is to poor for most interventions, the health consequences seem more rel- people. Estimating such costs, even if time is not valued in evant or more important than any nonhealth outcomes. The money, may show whether time or monetary costs or both other is that, conceptually, it is unclear what dollar value to account for a relatively low level of use and therefore impede assign to improved health, as would have to be done in most expanding coverage. Applying one or more of the policy cases. Two approaches to valuing health, particularly for judging instruments discussed earlier, along with the intervention, may how much a life is worth, are known respectively as (a) the value then be important in order for it to be cost-effective. of a statistical life, or the human capital approach, and (b) the Including such costs also raises a question of interpretation. willingness-to-pay valuation, or contingent valuation. The for- If an intervention appears low in cost-effectiveness because it mer depends on estimating earnings lost from premature death requires much travel or waiting time, the fault may lie not with or retirement, and the latter on what people pay or indicate they the intervention itself but with health facilities that are located would pay for care to protect or restore their health. too far from the beneficiary population, are understaffed, or Both approaches reflect a society's level of income. Although are inefficiently managed. For this reason, cost-effectiveness is they may be appropriate within a homogeneous society, if estimated assuming a functional health system that does not applied globally they imply that better health is worth less impose prohibitive time costs on users. among poor populations than among those who are better off. Not only the characteristics of the interventions themselves, Both methods are arguably more appropriate for marginal but also the capacity to deliver interventions greatly affect cost- improvements (like saving travel time for commuters) than for effectiveness across many activities. In a complete analysis, each valuing life-or-death differences, although willingness to pay is intervention is characterized by how demanding it is of mana- sometimes used in analyses of policies to reduce mortality. gerial or institutional capacity. This element is difficult to Avoiding monetary evaluation of health benefits sidesteps most measure directly, but authors often provide at least an intuitive of the ethical problems of valuing individual lives and requires description of how easy or hard delivery of an intervention is fewer assumptions about what benefits are worth. The cost of or what factors facilitate or impede its implementation. Where this simplification is that occasionally substantial nonhealth capacity to deliver several interventions together is important, benefits are not explicitly valued, so interventions may look less authors deal explicitly with the issue, as in the chapters on justified than they would be if all benefits were analyzed. health facilities (chapters 64­66), resources (chapters 71­72), service management (chapter 73), and whole packages of inter- ventions (chapters 56 and 63). DETERMINING COSTS FOR INTERVENTIONS Dollar values of unit costs need to be calculated for interna- tional comparisons. The inputs used to produce an Whatever outcome measures are used to evaluate an interven- intervention--the time (and training) of human resources; tion, its costs must be estimated. This need raises several ques- drugs and supplies; and depreciation or rental value of equip- tions about which costs to attribute to the intervention and ment, vehicles, and buildings--are either produced in the how some of them should be valued. country or imported. If the latter, they already have prices in U.S. Direct and indirect costs should be distinguished, and dollars; if the former, prices in local currency must be converted choices should be made about which, if any, of the latter to to U.S. dollars for comparison with other interventions include. In addition to the direct costs to the health system of and other countries. The usual distinction between tradable 280 | Disease Control Priorities in Developing Countries | Philip Musgrove and Julia Fox-Rushby and nontradable goods is that tradables move from producing cost, not according to what they would cost if prices were more to importing countries at relatively constant "world" prices. In nearly uniform among countries. If, in local currency, physi- fact, the same good may be imported at different prices to dif- cians are paid little more than nurses are, it may make sense to ferent countries or may be imported to one country but locally employ more doctors per nurse--even if at international prices produced in another, so that it has both an international and a doctors would cost much more and should be replaced by local price. This situation is increasingly true of drugs and sup- nurses when possible. Of course, the staffing decision turns plies, which middle-income countries (Brazil, for example) and on the competencies of the two groups as well as on their costs; some low-income countries (India) now produce and some- for certain health problems, more nurses might be the better times export. choice even if they cost more. Prices in local currency can be converted to U.S. dollars by Two other reasons besides that of efficiency in buying inter- exchange rates or by purchasing-power parity rates (as esti- ventions support basing cost-effectiveness on exchange rate mated in World Bank 2003). The former may reflect under- or prices. First, authors who have used published costs (which overvaluation of the local currency, making goods systemati- usually involve exchange rates) rather than building up esti- cally cheaper or more expensive than at world prices, and they mates from individual prices and quantities seldom break may change quickly and substantially in response to changes down costs into imported and domestic components. Local in a country's trade balance, indebtedness, or capital flows. inputs cannot be repriced at purchasing-power parity rates or Nonetheless, they represent what is actually paid for locally can only be repriced very approximately. Second, for readers produced inputs at any given moment. Purchasing-power par- accustomed to dealing with prices converted using exchange ity rates, in contrast, attempt to say what local currency is rates, real resource estimates may simply appear to penalize the worth in purchasing power, correcting for systematic price dif- use of local inputs by valuing them at unrealistic prices. The ferences. Such rates can be calculated for the country as a problem with exchange rate prices, in contrast, is that when whole, for the health sector, or for specific inputs or combina- rates change, so may the relative cost-effectiveness of interven- tions thereof (Wordsworth and Ludbrook 2004). This calcula- tions, as imported inputs become relatively more or less expen- tion means valuing local inputs at external prices, assuming sive. Cost-effectiveness is not static or intrinsic but depends they are equally productive or of equal quality in the particular on prices as well as on quantities and on the results of an country as in the countries from which purchasing-power par- intervention--and prices can change individually or generally, ity rates are derived. A doctor in South Asia or Sub-Saharan through exchange movements. Priorities sometimes need to Africa is treated as costing just as much as a physician in high- shift because of such price changes, as well as because of tech- income countries. This approach approximates measuring the nological changes that make interventions more effective, and real resource cost of intervention by comparing quantities so analyses should be kept up to date. of inputs among countries, eliminating price differences as a source of cost variation. Estimates of real national income are derived this way, making poor countries usually look less poor MORE AND LESS COMPREHENSIVE DATA in dollar terms than if income in local currency were valued at AND ANALYSIS exchange rates. Granted that purchasing-power parity rates are reasonable Several authors (Drummond and others 1997; Gold and others for comparing large aggregates such as income across coun- 1996; Sloan 1995) provide similar guidance and recommenda- tries, but they bear little relation to the allocation of resources tions for relatively comprehensive economic evaluation in and budgetary choices within a country. The cost calculations general or for specific medical procedures. This volume aims at in this volume are, therefore, all based on exchange rates. estimating cost-effectiveness for interventions against many Exchange rates more accurately reflect what a domestic different problems in all low- and middle-income regions, for buyer--or a foreign donor or investor--has to pay for which varying amounts and quality of information are avail- imported versus domestic inputs and, therefore, are more rele- able. It has therefore not always been possible to conduct as vant for choices between interventions with high or low complete an analysis as would be desirable. Some degree of imported content. (If exchange rates are artificially fixed, the modeling is usually inescapable (Buxton and others 1997). country pays a cost for that distortion that affects all interven- More complete analysis starts by characterizing, in each tions to the extent that they require foreign exchange.) In gen- regional setting where an intervention is relevant (where the eral, the more an intervention is produced with local inputs, the health problem causes some measurable burden and the inter- more cost-effective it will appear when priced using exchange vention appears feasible), estimates of the quantities of inputs rates, compared with its cost at purchasing-power parity rates. required (Q), the unit costs of those inputs (P), and the effec- For decision makers and purchasers in the country, efficiency tiveness or health gain (E). Authors were provided regional esti- means choosing interventions according to what they actually mates of unit costs for the major inputs--salaries, facility costs, Cost-Effectiveness Analysis for Priority Setting | 281 fuel and vehicle operation, drugs, representative equipment, Local cost and outcome estimates that have not been con- diagnostic tests, and buildings (Mulligan and others 2003). The structed transparently from inputs and prices provide a less total cost of delivery is the sum of the input costs PQ, which is complete basis for secondary analysis. Where such estimates are compared with effectiveness E and the CER calculated from the used, information about how costs are constructed or how total costs and total effects of the proposed intervention or results vary with the scale of the intervention is usually not from the changes in those costs and outcomes compared with available, but the data may explicitly show regional differences current practice. in one or both elements, thereby permitting regionally differ- The data on unit costs, quantities, and outcomes may all entiated recommendations (or may show differences so small derive from published literature; what is original is how that that recommendations need not differ regionally). If costs and information is combined to calculate cost-effectiveness rather results refer to only one moment or are specified year by year, than taking the ratios from existing studies. Estimates are built they can be discounted at 3 percent. Published analyses often up using prices and physical inputs in the chapters on tubercu- use higher constant rates of 5, 6, or even 10 percent and may losis (chapter 16), vaccine-preventable diseases (chapter 20), specify only total costs and outcomes rather than the respective malaria (chapter 21), cancers (chapter 29), psychiatric streams through time. In that case, both costs and health gains disorders (chapter 31), neurological disorders (chapter 32), occurring in the future are valued less, but conversion to a CER cardiovascular disease (chapter 33), hemoglobin disorders based on 3 percent discounting may be impossible and is at (chapter 34), water and sanitation (chapter 41), indoor air pol- best only approximate. Some published analyses discount costs lution (chapter 42), tobacco (chapter 46), alcohol (chapter 47), but not health outcomes, which makes interventions look more community programs (chapter 56), family planning (chapter cost-effective when costs are spread over long intervals (for 57), surgery (chapter 67), emergency care (chapter 68), and examples, see chapter 29). Imported estimates of cost and complementary medicine (chapter 69). Several chapters ana- effect--that is, estimates from other regions, commonly from lyze some interventions more fully and others less fully, high-income countries--are often all that is available. depending on the available information. Sometimes data on costs and outcomes derive from the same As indicated in box 15.2, expanding or contracting the scale source; in other cases, they come from different sources and of an intervention may change the CER because of difficulty even different regions and are difficult to compare directly. in reaching more of the population. The ratio may also vary More appropriate adjustments to total costs are possible with because of the cost of identifying who would benefit most from information on quantities of inputs. In the absence of data on the intervention--for example, whether to screen all newborns quantities of resources used, differences in average cost can for sickle cell disease or only those of African origin (chap- sometimes be calculated using estimates of input proportions. ter 34). And expansion may change the cost-effectiveness Such an approximation characterizes the analysis for diabetes because it would require considerable fixed investment. The (chapter 30), in which proportions were known in one region costs of expanding capacity to deliver an intervention, includ- and assumed to be the same elsewhere and costs were estimated ing physical capital and training of human resources, should be from regional cost ratios. amortized over a reasonable interval (10 years is the standard in this volume) and included in the total costs. Ideally, one would know the complete production function of the inter- Variation of Results and Uncertainty of Estimates vention, including the possibilities of substituting one input for Variation and uncertainty are two different aspects of cost- another to minimize costs in response to differences in prices. effectiveness estimates that also need to be accounted for in set- However, analysis of this level of complexity is difficult to ting priorities. Because costs of inputs differ among regions, achieve, so most chapters assume fixed input proportions. Q, intervention costs vary even if effectiveness does not--and there then, does not depend on P, and the CER varies (at most) only are often reasons why the same intervention is more effective in with coverage, prices, and outcomes. This result could be an one place than another. Such variation means that a single underestimate of the true cost-effectiveness if much substitu- estimate of incremental or average cost-effectiveness of an inter- tion is possible (see chapter 16 on tuberculosis). vention is not universally applicable.All estimates should ideally Approximations are required when the average and incre- be local, and regional values capture only part of the real varia- mental CERs have to be taken directly from the literature and tion. For example, the average cost per DALY of chemotherapy when key parameter values are not easily available. Existing for active or contagious tuberculosis, in the absence of HIV and estimates of total cost or effectiveness may or may not AIDS, is US$15, but that figure varies from US$6 to US$31 incorporate the standard assumptions about discounting, dis- across regions, and such wide variation is common in many ability weights, and life expectancy. Authors then need to chapters. Whenever the estimates of cost-effectiveness in differ- judge how to adjust the available estimates for a more consis- ent chapters use the same input prices, their results are compa- tent analysis. rable within a given region. Analyses that draw on published 282 | Disease Control Priorities in Developing Countries | Philip Musgrove and Julia Fox-Rushby estimates for price or unit cost information are necessarily less The quality and relevance of evidence can vary considerably, comparable across interventions and introduce another ele- depending on whether information comes from randomized ment of variation, even in the same locale. Still more variation controlled trials or systematic overviews, nonrandomized stud- arises when costs or outcomes are extrapolated from one coun- ies with multivariate analyses and well-defined endpoints, or try or region to another. case studies or expert opinion. For these analyses, the quality of Because the CER depends on many parameters and vari- evidence also depends on geographic coverage, as distinguished ables, of which only the discount rate and the disability weights in chapter 2: are uniform, good analytic practice calls for sensitivity analysis to see how the ACERs and ICERs change with plausible varia- · literature review of one cost-effectiveness study, in one tion in one or more parameters. Many chapters (such as chapter country 26) provide such analyses, varying one value at a time, to sketch · literature review of several studies in different countries in the likely range of estimates. This method is one way of dealing different regions with uncertainty (which differs from real, known variation) · literature review of several studies in different countries in about the true values of the data and seeing whether the ranking the same region of interventions changes when those values change. Such analy- · original analyses starting with price and quantity data in ses do not indicate the probability that the true CER falls in a one country particular interval, only under what input values it would do · original analyses starting with price and quantity in one or so. Estimating such probabilities requires knowing or assuming more regions. the statistical distributions of the parameters in question and using that information to derive confidence intervals The first three categories differ in how representative published around the point estimates. Guides to CEA recommend these findings are; the latter two categories differ according to the approaches (Gold and others 1996), and the National Institute data used in constructing total effects and total costs. for Clinical Excellence requires probabilistic sensitivity analysis Besides the quality of the evidence at its source, how the before approving medical treatments in the United Kingdom results will apply to other settings matters, particularly when (NICE 2004). the data are limited to high-income countries. The more that Data for estimating probability distributions around mean outcomes depend on underlying biology, the more the find- parameter estimates are seldom available in low- and middle- ings will apply to low- and middle-income countries. Outcomes income countries. Simply having available several different depending more on cultural or environmental factors are less estimates of a parameter is inadequate for deriving a distribu- readily transferred and require judgment and evidence as to tion, because the differences may be caused by variation in their applicability elsewhere. Sometimes the only detailed stud- regional costs or expected life years rather than uncertainty. ies refer to high-income countries, as for abuse of substances However, assumptions about the shape of distributions can other than alcohol and tobacco (chapter 48). At the other be applied within modeling exercises to give an indication of extreme, in a few cases all or nearly all the information comes the likely distribution of ICERs. Only a few chapters, therefore, from low- and middle-income countries, and there is no need include confidence intervals. The analyses for tuberculosis to extrapolate, as for nutritional interventions (chapter 28) and (chapter 16) and malaria (chapter 21) do, but ranges associated community health and nutrition programs (chapter 56). with most cost-effectiveness estimates (see chapter 2) reflect other causes of variation, not statistical accuracy. Although calculations are often reported to several signifi- COST-EFFECTIVENESS AND POPULATION IMPACT cant digits, such precision is not really feasible given the uncertainties in the original data: "economics is a one- or at An intervention CER, whether average or incremental, is based most a two-digit science" (Morgenstern 1963). However, even on assumptions about introduction, expansion, contraction, or crude findings can be valuable, either as guides to value for modification of the activity compared with current (or some- money if inaccuracies do not affect the relative order of times "best") practice. Comparison of ratios indicates whether magnitude of the results or for understanding and exploring one intervention offers better or worse value for money than the sources of variation and their effect on priorities as well as another at the individual level but says nothing about how indicating future research needs (Claxton, Sculpher, and either one affects the whole population. The analysis, therefore, Drummond 2002). These issues arise, for example, when con- includes, wherever possible, two ways of describing the latter sidering whether to expand the EPI or to add new antigens effect. One is to consider a population of 1 million, with a typ- (chapter 20), how far to extend screening procedures (chapters ical regional age and sex structure, and to suppose that the 29 and 34), and when to change drugs in response to vector or intervention were delivered to all the potential beneficiaries. parasite resistance (see chapters 21 and 23). That number of people is just the prevalence or incidence of Cost-Effectiveness Analysis for Priority Setting | 283 the condition times 1 million. The total cost would then be the and to develop priorities on the basis of analyses appropriate to unit cost times that number (or the cost of reaching that many local circumstances. The methods used here are intended to help people if the unit cost varies with coverage). The total health guide such efforts, and they can and should be refined through gain would be the individual effectiveness times that same research to provide more robust help to policy. number (or the overall outcome if that depends on externali- Finally, a more concerted approach is needed for clarifying ties, such as the transmission of communicable disease, that are the options facing different decision makers and incorporating sensitive to coverage). Standardizing on a population of 1 mil- the results from systematic literature reviews into analytic lion allows comparisons among regions and interventions in models that compare the costs and effects of alternative inter- which the incidence or prevalence may vary greatly. ventions (Buxton and others 1997; Kuntz and Weinstein 2001). A second approach standardizes not on population but on Modeling encourages explicit decision making and can deal expenditure: if an additional US$1 million were devoted to the comprehensively with the inputs and outcomes of decision resources needed for an intervention, how many people could options, which allows a range of uncertainties to be reflected. benefit from it and how large would the health gain be? The Thus, hypotheses about interventions can be formulated and coverage of the intervention would be US$1 million divided by tested statistically. Specifying models explicitly (as in chapter the average cost, and the total gain in DALYs would be that 16, for example) can also help identify gaps in current evidence number of people times the average effectiveness. This and can capture details specific to particular populations and approach is applied in relatively few chapters because of the settings. information requirements; its advantage is to facilitate judg- ments as to where increased spending would be most ACKNOWLEDGMENTS justified--where it would yield the largest improvement in health, reach the most people, or account for the largest share The authors are grateful to Jo Mulligan for the estimation and of burden from a condition. Table 1.3 in chapter 1 provides explanation of input costs used in this volume; to Sonbol A. examples for some interventions to reduce child mortality, pre- Shahid-Salles for help in drawing methodological examples vent or treat HIV and AIDS, reduce smoking prevalence, treat from numerous chapters; to George Alleyne, Joel G. Breman, heart attack and stroke, detect and treat cervical cancer, and Mariam Claeson, Anthony R. Measham, and Elinor Schwartz operate a basic surgical ward. The estimates of DALYs gained for helpful comments; and to Dean T. Jamison and Anne Mills per US$1 million vary from less than 100 to more than for overall guidance. 100,000--a thousandfold difference in value for money. Annex 26.A of chapter 26 provides both kinds of calculations, per mil- lion population, to compare the cost-effectiveness of interven- REFERENCES tions for improved maternal health in South Asia and Sub- Brooks, R., R. Rabin, and F. de Charro, eds. 2003. The Measurement and Saharan Africa. Valuation of Health Status Using the EQ-5D: European Perspective. (Evidence from the EuroQol BIOMED research program.) Dordrecht, Netherlands: Kluwer Academic Publishers. Buxton, M. J., M. F. Drummond, B. A.Van Hout, R. L. Prince, T. A. Sheldon, IMPROVEMENTS AND FURTHER APPLICATIONS T. Szucs, and M. Vray. 1997. "Modeling in Economic Evaluation: An Unavoidable Fact of Life." Health Economics 6 (3): 217­27. What would improve the kind of estimates and conclusions Claxton, K., M. Sculpher, and M. F. Drummond. 2002. "A Rational reported in this volume? Most crucially, more and better data are Framework for Decision Making by the National Institute for Clinical Excellence (NICE)." Lancet 360 (9334): 711­15. needed in low- and middle-income countries to reduce reliance Dolan, P., and M. Jones-Lee. 1997. "The Time Trade-Off: A Note on the on extrapolation from high-income countries and on expert Effect of Lifetime Reallocation of Consumption and Discounting." judgments. The need for information starts, in some cases, with Journal of Health Economics 16: 731­39. better estimates of incidence and prevalence, but even where the Drummond, M. F., B. O'Brien, G. L. Stoddart, and G. W. Torrance. 1997. epidemiology is well known, data on coverage and outcomes of Methods for the Economic Evaluation of Health Care Programmes. Oxford, U.K.: Oxford Medical Publications. existing interventions are scarce. Evidence of what it would cost Fox-Rushby, J. A. 2002. Disability-Adjusted Life Years (DALYs) for Decision- to change coverage of existing interventions or add new inter- Making? London: Office of Health Economics. ventions, and with what results, is particularly scarce and Gabbay, J., and A. le May. 2004. "Evidence Based Guidelines or Collectively depends heavily on assumptions. This situation is sometimes Constructed `Mindlines?' Ethnographic Study of Knowledge true even for activities that have been conducted widely for Management in Primary Care." British Medical Journal 329 (7473): many years and have been extensively analyzed, notably the EPI 1013. (chapter 20). Analyses should when possible be conducted at Glick, H., D. Polsky, and K. Schulman. 2001. "Trial-Based Economic Evaluations: An Overview of Design and Analysis." In Economic Eval- the level of a country or even smaller units, to take full account uation in Health Care: Merging Theory with Practice, ed. M. Drummond of all the reasons cost-effectiveness varies from place to place and A. McGuire, 113­40. Oxford, U.K.: Oxford University Press. 284 | Disease Control Priorities in Developing Countries | Philip Musgrove and Julia Fox-Rushby Gold, M. R., J. E. Siegel, L. B. Russell, and M. C. Weinstein, eds. 1996. Murray, C. J. L. 1996."Rethinking DALYs." In The Global Burden of Disease, Cost-Effectiveness in Health and Medicine. New York: Oxford University ed. C. J. L. Murray and A. D. Lopez, 1­98. Cambridge, MA: Harvard Press. University Press. Hoffmann, C., B. A. Stoykova, J. Nixon, J. M. Glanville, K. Misso, and M. F. Murray, C. J. L., and A. D. Lopez, eds. 1996. The Global Burden of Disease. Drummond. 2002. "Do Health-Care Decision Makers Find Economic Cambridge, MA: Harvard University Press. Evaluations Useful? The Findings of Focus Group Research in U.K. Musgrove, P. 1999. "Public Spending on Health Care: How Are Different Health Authorities." Value Health 5 (2): 71­78. Criteria Related?" Health Policy 47: 207­23. iMTA (Institute for Medical Technology Assessment). 2005. Newsletter 3 ------. 2000. "A Critical Review of `A Critical Review': The Methodology (1): 1­3. of the 1993 World Development Report, Investing in Health." Health Jamison, D. T. 2002. "Cost-Effectiveness Analysis: Concepts and Policy and Planning 15 (1): 110­15. Applications." In Oxford Textbook of Public Health, 4th ed., ed. R. G. NICE (National Institute for Clinical Excellence). 2004. Guide to the Detels, J. McEwen, R. Beaglehole, and H. Tanaka, 903­19. Oxford, Methods of Technology Appraisal. London: NICE. http://www.nice. U.K.: Oxford University Press. Also published as Disease Control org.uk. Priorities Project Reprint 3. Sheldon, T. A., N. Cullum, D. Dawson, A. Lankshear, K. Lowson, I. Watt, Jamison, D. T., and J. S. Jamison. 2003. "Discounting." Disease Control and others. 2004. "What's the Evidence That NICE Guidance Has Been Priorities Project Working Paper 4, World Bank, Washington, DC. Implemented? Results from a National Evaluation Using Time Series Jamison, D. T., W. H. Mosley, A. R. Measham, and J. L. Bobadilla, eds. 1993. Analysis, Audit of Patients' Notes, and Interviews." British Medical Disease Control Priorities in Developing Countries. New York: Oxford Journal 329 (7473): 999. University Press. Sloan, F. A. 1995. "Introduction." In Valuing Health Care: Costs, Benefits, Karlsson, G., and M. Johannesson. 1996. "The Decision Rules of Cost- and Effectiveness of Pharmaceuticals and Other Technologies, ed. F. A. Effectiveness Analysis." Pharmacoeconomics 9 (2): 113­20. Sloan. New York: Cambridge University Press. Kuntz, K. M., and M. C. Weinstein. 2001. "Modelling in Economic Smith, P. C. 2005. "Statutory Packages of Health Care Alongside Voluntary Evaluation." In Economic Evaluation in Health Care: Merging Theory Insurance: What Treatments Should Be Covered?" York, U.K.: Centre with Practice, ed. M. Drummond and A. McGuire, 141­71. Oxford, for Health Economics, University of York. U.K.: Oxford University Press. Taylor, R., and NICE (National Institute for Clinical Excellence). 2002. Mathers, C. D., A. Lopez, C. Stein, D. Ma Fat, C. Rao, M. Inoue, and oth- "HTA Rhyme and Reason?" International Journal of Technology ers. 2005. "Deaths and Disease Burden by Cause: Global Burden of Assessment in Health Care 18 (2): 166­70. Disease Estimates for 2001 by World Bank Country Groups." Disease Walker, D., and J. A. Fox-Rushby. 2000. "Critical Review of Economic Control Priorities Project Working Paper 18, World Bank, Washington, Evaluations of Communicable Disease Interventions in Developing DC. Countries." Health Economics 9 (8): 681­98. McDaid, D., R. Cookson, and ASTEC Group. 2003. "Evaluating Health WHO (World Health Organization). 2000. World Health Report 2000. Care Interventions in the European Union." Health Policy 63 (2): "Health Systems: Improving Performance." Geneva: WHO. 133­39. Wordsworth, S., and A. Ludbrook. 2004. "Comparing Costing Results in Morgenstern, O. 1963. On the Accuracy of Economic Observations. Across Country Economic Evaluations: The Use of Technology Princeton, N.J.: Princeton University Press. Specific Purchasing Power Parities." Health Economics 14 (1): 93­96. Mulligan, J., J. A. Fox-Rushby, T. Adam, B. Johns, and A. Mills. 2003. World Bank. 1993. World Development Report: Investing in Health. New "Unit Costs of Health Care Inputs in Low and Middle Income York: Oxford University Press. Regions." Disease Control Priorities Project Working Paper 9, World ------. 2003. World Development Indicators. Washington, DC: World Bank, Washington, DC. Bank. Cost-Effectiveness Analysis for Priority Setting | 285 Part Two Selecting Interventions · Infectious Disease, Reproductive Health, and Undernutrition · Noncommunicable Disease and Injury · Risk Factors · Consequences of Disease and Injury Chapter 16 Tuberculosis Christopher Dye and Katherine Floyd Despite the availability of drugs to cure tuberculosis (TB) since tries. The results have not been fully synthesized but may the 1940s, TB remains an important cause of death from an suggest ways to enhance DOTS. infectious agent, second only to the human immunodeficiency · Second, striking increases in TB have been associated with virus, or HIV (WHO 2004f). TB control is high on the inter- the spread of HIV infection and drug resistance, suggesting national public health agenda, not only because of the enor- that DOTS alone may not be enough to bring TB under mous burden of disease, but also because short-course control, especially in Africa and in the countries of the chemotherapy (SCC) is recognized as one of the most cost- former Soviet Union. effective of all health interventions (Jamison and others 1993). · Third, there is now substantially more investment in new That recognition is partly attributable to an influential series of tools for TB control, including multimillion-dollar initia- studies done in three of the poorest countries of southeastern tives to develop better diagnostics, drugs, and vaccines, Africa (Malawi, Mozambique, and Tanzania), which suggested many of which operate under the umbrella of the Stop TB that a year of healthy life could be gained for less than about Partnership (see http://www.stoptb.org). Some of the possi- US$5 (de Jonghe and others 1994; Murray and others 1991). ble products of this new research would stimulate reevalua- This evidence has been central to the global promotion of the tions of the current reliance on chemotherapy, especially the DOTS strategy, the package of measures combining best development of a new high-efficacy vaccine. practices in the diagnosis and treatment of patients with active · Fourth, interest in TB is renascent, not simply as the out- TB, in which direct observation of treatment during SCC is a come of mycobacterial infection, but also as the conse- key element (WHO 2002a, 2004c). quence of exposure to exacerbating risks, such as tobacco Although the World Health Organization (WHO) has fos- smoke, air pollution, malnutrition, overcrowding, and poor tered the implementation of DOTS over the past decade, four access to health services. Research directed at quantifying recent developments have drawn attention to a wider range of these risks will also suggest ways to minimize them. options for TB control: These developments set a big agenda for analysis. To make · First, many more studies have investigated the costs, efficacy, some inroads, this chapter presents an overview of the value and cost-effectiveness of different approaches to TB control. for money and potential effect of the principal modes of They are mostly studies of ways to improve the delivery TB control around the world. The starting point is a review of first-line drug treatment for active disease, but they of the natural history and clinical characteristics of TB and include some investigations of preventive therapy (treat- the geographical distribution of and trends in TB cases and ment of latent infection), treatment of multidrug-resistant deaths. This introduction sets the context for a discussion of TB (MDR-TB) using both first- and second-line drugs, and the interventions that are now available to control TB and of different approaches to diagnosis. They have been carried how they have been used. We use a new method for evaluating out in a variety of settings, in richer as well as poorer coun- the cost-effectiveness of infectious disease control and apply 289 this method systematically to four groups of TB interventions The most common clinical manifestation is pulmonary as they could be implemented in six regions of the world. disease, typically in the parenchyma of the middle and lower The internationally agreed-on targets for TB control, lung. In the most infectious patients, bacilli can be seen micro- embraced by the United Nations Millennium Development scopically on stained sputum smears (60 to 70 percent of Goals (MDGs), are to detect 70 percent of sputum-smear- pulmonary cases; Marais and others 2004; Styblo 1991). positive cases and successfully treat 85 percent of such cases by Smear-negative patients may also be infectious but, per patient, the end of 2005. The expectation is that, if these targets can be contribute relatively little to transmission (Behr and others reached and maintained, incidence rates will be falling by 2015, 1999; Hernandez-Garduno and others 2004). Extrapulmonary and the TB prevalence and death rates of 1990 will be halved by tuberculosis accounts for 10 to 30 percent of the disease but is 2015. Meeting these targets requires a set of interventions that more common among women and children (particularly lym- are not only cost-effective but also affordable and capable of phatic TB) and in people infected with HIV (Aaron and others having an effect on a large scale. The final sections of the chap- 2004; Rieder 1999; Rieder, Snider, and Cauthen 1990; Shafer ter discuss the absolute costs and benefits of global TB control and Edlin 1996). and the potential for achieving the effect defined within the In the absence of other predisposing conditions, only MDG framework. The main themes of the text that follows about 5 percent of infected people develop progressive pri- are elaborated in a series of annexes available online at http:// mary disease within five years of infection (Comstock, www.fic.nih.gov/dcpp as well as at http://www.who.int/tb/ Livesay, and Woolpert 1974; Sutherland 1968, 1976). After publications/en/. five years, the annual risk of developing TB by the reactivation of latent infection is much lower( 10 4per capita per year). The risk of progressing to active disease is relatively high in TUBERCULOSIS INFECTION, DISEASE, AND DEATH infancy and lower in older children; it increases quickly dur- ing adolescence (earlier in girls) and then more slowly Human TB is caused by infection with mycobacteria, princi- throughout adulthood (Comstock, Livesay, and Woolpert pally Mycobacterium tuberculosis. Individuals with pulmonary 1974; Nelson and Wells 2004; Sutherland, Svandova, and or laryngeal TB produce airborne droplets while coughing, Radhakrishna 1982; Vynnycky and Fine 1997). Whether latent sneezing, or simply talking. Inhaled infectious droplets lodge in bacilli remain viable for the full life span of all infected peo- the alveoli, and bacilli are taken up there by macrophages, ple is unknown, but the risk of reactivation certainly persists beginning a series of events that results in either the contain- into old age. The lifetime risk of developing TB following ment of infection or the progression to active disease (Frieden infection clearly depends on the prevailing transmission rate; and others 2003). Following uptake by macrophages, M. tuber- the rule of thumb is 10 percent, but it has been calculated at culosis replicates slowly but continuously and spreads through 12 percent for all forms of pulmonary disease in England and the lymphatic system to hilar lymph nodes. In most infected Wales during the second half of the 20th century (Vynnycky people, cell-mediated immunity, associated with a positive and Fine 2000). tuberculin test, develops two to eight weeks after infection. Besides the strong innate resistance to developing disease, Activated T lymphocytes and macrophages form granulomas, infection is associated with an acquired immune response. This which limit the further replication and spread of bacilli. Unless response is only partially protective (Dye and others 1998; a later defect occurs in cell-mediated immunity, the infection Sutherland, Svandova, and Radhakrishna 1982; Vynnycky and remains contained within the granulomas. Fine 1997), which helps explain why developing an effective The immune mechanisms are, in their details, far more vaccine has been difficult (few manufactured vaccines are complex. For example, following antigenic challenge, a suite of more protective than natural immunity; Andersen 2001; different T cells is responsible for the induction and suppres- Fordham von Reyn and Vuola 2002; Young and Stewart 2002). sion of protective immunity, delayed hypersensitivity, cytolysis, Consequently, individuals who carry a latent infection and who and the production of antibodies and memory cells. Helper T continue to be exposed are at risk of TB following reinfection. cells mature into two functionally different populations: in M. The importance of reinfection remains controversial, but tuberculosis infection, the TH1 response is associated with gran- mathematical modeling shows that the decline of TB in Europe uloma formation and protection, whereas the TH2 response cannot easily be explained without reinfection (Dye and others results in tissue-necrotizing hypersensitivity and the progres- 1998; Vynnycky and Fine 1997). In addition, molecular finger- sion of disease. The processes that determine the balance of the printing has produced direct evidence that TB commonly two responses affect, for example, the interaction between arises from infection and reinfection in endemic areas (de M. tuberculosis and other infectious agents (Grange 2003). Viedma and others 2002; Richardson and others 2002; van Rie When the immune response cannot suppress replication, and others 1999; Verver and others 2004), especially where sub- primary infection leads to active TB (progressive primary TB). jects are infected with HIV (Glynn and others 2004). 290 | Disease Control Priorities in Developing Countries | Christopher Dye and Katherine Floyd The low incidence of infection and the low probability of Globally, the TB incidence rate per capita appears to be breakdown to disease explain why TB is relatively rare. Its growing slowly (online annex 2). Case numbers have been importance among infectious diseases is attributable not so declining more or less steadily for at least two decades in much to the number of cases as to the high case-fatality rate Western and Central Europe, the Americas, and the Middle among untreated or improperly treated patients. About two- East. Striking increases have occurred in countries of Eastern thirds of untreated smear-positive patients will die within five Europe (mainly the former Soviet republics) since 1990 and to eight years, the majority within the first 18 months (Styblo in Sub-Saharan Africa since the mid 1980s, although trends 1991). Most of those who are still alive after eight years will in case notifications suggest that the rate of increase in both have quiescent TB (self-cures, susceptible to relapse), and a few regions has slowed significantly since the mid 1990s (WHO will become chronic excretors of bacilli. The case-fatality rate 2005). for untreated smear-negative cases is lower, but still of the TB has increased in Eastern European countries because of order of 10 to 15 percent (Krebs 1930; Rieder 1999). Even economic decline and the general failure of TB control and among smear-positive patients receiving antituberculosis other health services since 1991 (Shilova and Dye 2001). drugs, the case-fatality rate can exceed 10 percent if adherence Periodic surveys indicate that more than 10 percent of new TB to treatment is low or if rates of HIV infection and drug resist- cases in Estonia, Latvia, and some parts of the Russian ance are high (WHO 2004c). Federation are multidrug-resistant--that is, resistant to at least Online annex 1 contains more information about factors isoniazid and rifampicin, the two most effective anti-TB drugs that affect the risk to individuals of contracting infection and (Espinal and others 2001; WHO 2004a). Drug resistance is developing disease and the distribution of TB in populations. likely to be a by-product of the events that led to TB resurgence in these countries, not the primary cause of it, for three rea- sons. First, resistance is generated initially by inadequate treat- EPIDEMIOLOGICAL BURDEN AND TRENDS ment caused, for example, by interruption of the treatment schedule or use of low-quality drugs. Second, resistance tends Surveys of the prevalence of infection and disease, assessments to build up over many years, and yet TB incidence increased of the performance of surveillance systems, and death registra- suddenly in Eastern European countries after 1991. Third, tions yield an estimated 8.8 million new cases of TB in 2003, although formal calculations have not been done, resistance fewer than half of which were reported to public health author- rates are probably too low to attribute all of the increase in ities and WHO (online annex 2). Approximately 3.9 million caseload to excess transmission from treatment failures. cases were sputum-smear positive, the most infectious form of Globally, 12 percent of new adult TB cases were infected the disease (Corbett and others 2003; Dye and others 1999; with HIV in 2003, but there was marked variation among WHO 2005). The African region has the highest estimated inci- regions--from an estimated 33 percent in Sub-Saharan Africa dence rate (345 per 100,000 population annually), but the most to 2 percent in East Asia and the Pacific (online annex 2). HIV populous countries of Asia harbor the largest number of cases: infection rates in TB patients have so far remained below 1 per- Bangladesh, China, India, Indonesia, and Pakistan together cent in Bangladesh, China, and Indonesia. The increase in TB account for half the new cases arising each year. In terms of incidence in Africa is strongly associated with the prevalence of the total estimated number of new TB cases arising annually, HIV infection (Corbett and others 2002), and in populations about 80 percent of new cases occur in the top-ranking with higher rates of HIV infection, women 15­24 years old 22 countries. constitute a higher proportion of TB patients (Corbett and In most countries (but not all), more cases of TB are others 2002). The rise in the number of TB cases in Africa is reported among men than women. This differential is partly slowing, almost certainly because HIV infection rates are also because women have less access to diagnostic facilities in some beginning to stabilize or fall (Asamoah-Odei, Garcia Calleja, settings (Hudelson 1996), but the broader pattern also reflects and Boerma 2004). HIV has probably had a smaller effect on real epidemiological differences between men and women, TB prevalence than on incidence because HIV significantly both in exposure to infection and in susceptibility to disease reduces the life expectancy of TB patients (Corbett and others (Borgdorff and others 2000; Hamid Salim and others 2004; 2004). Where HIV infection rates are high in the general pop- Radhakrishna, Frieden, and Subramani 2003). Where the ulation, they are also high among TB patients; estimates transmission of M. tuberculosis has been stable or increasing for 2003 suggested that more than 50 percent of TB patients for many years, the incidence rate is highest among young infected with HIV in Botswana, South Africa, Zambia, and adults, and most cases are caused by recent infection or rein- Zimbabwe, among other countries. fection. As transmission falls, the caseload shifts to older age Approximately 1.7 million people died of TB in 2003 groups, and a higher proportion of cases comes from the reac- (Corbett and others 2003), including 229,000 patients who tivation of latent infection. were also infected with HIV (online annex 2). Although these Tuberculosis | 291 are usually reported as AIDS deaths under the International Reported BCG vaccination coverage has increased through- Statistical Classification of Diseases and Related Health out the world during the past 25 years, reaching about 100 mil- Problems, 10th revision (ICD-10), and by WHO, TB control lion infants, or 86 percent of all infants, in 2002. An estimated programs need to know the total number of TB deaths, what- 92 percent of children were vaccinated in Europe and 62 per- ever the underlying cause. cent in Africa in 2002 (WHO 2001). During the past 15 years, coverage has generally been most variable among African countries and least variable in Europe and the Americas. The INTERVENTIONS AGAINST TUBERCULOSIS most complete analysis of the effect of BCG vaccination sug- gests that BCG given to children born in 2002 prevents about TB can be controlled by preventing infection, by stopping pro- 29,700 cases of childhood meningitis and 11,500 cases of mil- gression from infection to active disease, and by treating active iary TB during the first five years of life, or one case for every disease. The principal intervention is the DOTS strategy and its 3,400 and 9,300 vaccinations, respectively (Bourdin Trunz, variations, centered on the diagnosis and treatment of the most Fine, and Dye, forthcoming). severe and most infectious (smear-positive) forms of TB but including treatment for smear-negative and extrapulmonary cases as well. Anti-TB drugs can also be used to treat latent Treatment of Latent Infection M. tuberculosis infection and active TB in patients with HIV Individuals at high risk of TB who have a positive tuberculin coinfection, and the widely used bacillus Calmette-Guérin skin test but not active disease (for example, associates of active (BCG) vaccine prevents (mainly) severe forms of TB in child- cases, especially children and immigrants to low-incidence hood. These biomedical interventions directed specifically countries) can be offered treatment for latent TB infection against TB can be implemented in a variety of ways through (TLTI), most commonly with the relatively safe and inexpensive medical services and public action and can be supported by other drug isoniazid. Studies among those who have contacts with efforts to reduce environmental risk factors (online annex 1). active cases have demonstrated that 12 months of daily isoni- azid gives 30 to 100 percent protection against the development of active TB (Cohn and El-Sadr 2000; Comstock 2000). For Vaccination patients who may be carrying a strain resistant to isoniazid, Currently, the only means of immunizing against TB is with the rifampicin daily for 4 months is an acceptable alternative (or live attenuated vaccine BCG, although other vaccines are under rifabutin, if used with protease inhibitors for HIV-infected peo- development (Fruth and Young 2004; Goonetilleke and others ple; Cohn 2003; Menzies and others 2004). Nevertheless, TLTI is 2003; Horwitz and others 2000; Letvin, Bloom, and Hoffman not widely used. The main reason is that compliance with long- 2001; Reed and others 2003; Young and Stewart 2002). term daily treatment tends to be poor among healthy people-- Randomized controlled trials and case-control studies have a relatively high risk of TB among those who are latently shown consistently high protective efficacy of BCG against seri- infected is usually still a low risk in absolute terms. An addi- ous forms of disease in children (73 percent [95 percent tional reason is that the tuberculin skin test tends to be less spe- confidence limits 67­79 percent] for meningitis and 77 percent cific when applied to individuals who have been vaccinated with [95 percent confidence limits 58­87 percent] for miliary TB) BCG. Although it is sometimes possible to make separate esti- but highly variable--and often very low--efficacy against mates of the number of individuals in a population who have pulmonary TB in adults (Bourdin Trunz, Fine, and Dye, forth- been infected and who have received BCG (Neuenschwander coming; Fine 2001; Rieder 2003). Thus, even with the high cov- and others 2002), distinguishing the responses to BCG and erage now achieved, BCG is unlikely to have any substantial infection is harder in any given individual. effect on transmission. In parts of Europe and North America The exceptionally high risk of TB among persons coinfected that did and did not use BCG, TB declined at rates that were with M. tuberculosis and HIV is a reason for encouraging wider not measurably different (Styblo 1991). In areas of high inci- use of TLTI, especially in Africa. However, there are significant dence, BCG vaccination is recommended for children at birth barriers to making TLTI effective for coinfected individuals liv- or at first contact with health services. Vaccination is being dis- ing in areas of high transmission (in addition to those listed ear- continued in many low-incidence countries because the risk of lier). Although trials of TLTI with individuals infected with HIV infection is low and because the response to BCG confounds whose tuberculin skin test was positive have averaged about the interpretation of tuberculin skin tests used to track persons 60 percent protection for up to three years (with a good deal of infected during occasional outbreaks. BCG may have substan- variability), the effects have been lost soon afterward, and little tial nonspecific effects on child mortality--that is, in reducing or no effect has been seen on mortality (Bucher and others deaths from causes other than TB--but this possibility is still 1999; Johnson and others 2001; Mwinga and others 1998; controversial (Kristensen, Aaby, and Jensen 2000). Quigley and others 2001; Whalen and others 1997; Wilkinson, 292 | Disease Control Priorities in Developing Countries | Christopher Dye and Katherine Floyd Squire,and Garner 1998).In addition,identifying M. tuberculosis found soon enough to eliminate transmission. In general, the infection is more difficult in HIV-positive individuals than in decline will be faster where a larger fraction of cases arises from those who are HIV-negative because the former are often aner- recent infection (that is, in areas where transmission rates have gic and are, therefore, unresponsive to tuberculin. Early studies recently been high) and slower where there is a large backlog of have also experienced problems with uptake and compliance. In asymptomatic infection. As TB transmission and incidence go a pilot project in Zambia, for example, only 35 percent of HIV- down, a higher proportion of cases comes from the reactivation infected individuals identified through HIV testing and coun- of latent infection and the rate of decline in incidence slows. seling services actually started TLTI, and, of those who started, These facts explain why it should be easier to control epidemic only 23 percent completed at least six months of treatment than endemic disease: during an outbreak in an area that pre- (Terris-Prestholt and Kumaranayake 2003). viously had little TB, the reservoir of latent infection will be TLTI has been used as a component of intensive, local con- small, and most new cases will come from recent infection. trol campaigns, such as those carried out for North American In the control of endemic TB, largely by chemotherapy, the and Greenland Eskimos, but probably had effects secondary to best results have been achieved in communities of Alaskan, the prompt treatment of active disease (Comstock, Baum, and Canadian, and Greenland Eskimos, where incidence was Snider 1979; Styblo 1991). At present, TLTI plays no more than reduced at 13 to 18 percent per year from the early 1950s an accessory role in TB control in any setting, although the (Styblo 1991). Over a much wider area in Western Europe, TB number of recipients around the world has been neither declined at 7 to 10 percent per year after drugs became widely directly quantified nor indirectly estimated. available during the 1950s, although incidence was already falling at 4 to 5 percent per year before chemotherapy (Styblo 1991). More recently, between 1994 and 2000, the incidence of Treatment of Active Disease: The DOTS Strategy pulmonary TB among Moroccan children 0 to 4 years of age fell at more than 10 percent per year, suggesting that the risk of The cornerstone of TB control is the prompt treatment of active infection was falling at least as quickly (S. Ottmani, personal cases with SCC using first-line drugs, administered through communication 2005). The overall reduction in pulmonary TB the DOTS strategy (WHO 2002a) within targets framed by the was only 4 percent per year, in part because of the large reser- MDGs. The DOTS strategy has five elements: voir of infection in adults. DOTS was launched in Peru in 1991, and high rates of case detection and cure appear to have pushed · political commitment down the incidence rate of pulmonary TB by 6 percent per · diagnosis primarily by sputum-smear microscopy among year (Suarez and others 2001). For epidemic TB, as a result of patients attending health facilities aggressive intervention following an outbreak in New York · SCC with effective case management (including direct City, the number of MDR-TB cases fell at a rate of more than observation of treatment) 40 percent per year (Frieden and others 1995). · a regular drug supply Although the long-term aim of TB control is to eliminate · systematic monitoring to evaluate the outcomes of every all new cases, cutting prevalence and death rates is arguably patient started on treatment. more important. About 86 percent of the burden of TB, as measured in terms of disability-adjusted life years (DALYs) lost, Standard SCC can cure more than 90 percent of new, drug- is attributable to premature death rather than illness, and preva- susceptible TB cases, and high cure rates are a prerequisite for lence and mortality can be reduced faster than incidence in expanding case finding. Although the DOTS strategy aims pri- chemotherapy programs. Thus, the TB death rate among marily to provide free treatment for smear-positive patients, Alaskan Eskimos dropped at an average of 30 percent per year most DOTS programs also treat smear-negative patients, usu- in the period 1950­70 and at an average of 12 percent per year ally without a fee. DOTS can be used as the basis for more com- throughout the Netherlands from 1950 to 1990. Indirect assess- plex TB control strategies where rates of drug resistance or HIV ments of the effect of DOTS suggest that 70 percent of the TB infection are high. deaths expected in the absence of DOTS were averted in Peru Mathematical modeling and practical experience suggest between 1991 and 2000, and more than half the TB deaths that the incidence of TB will decline at 5 to 10 percent per year expected in the absence of DOTS are prevented each year in when 70 percent of infectious cases are detected through pas- DOTS provinces of China (Dye and others 2000; Suarez and sive case finding and 85 percent of these cases are cured, even others 2001). There have been few direct measures of the reduc- though that level represents a treatment success rate among all tion in TB prevalence over time, but surveys done in China in infectious cases of only 60 percent (Dye 2000; Dye and others 1990 and 2000 showed a 32 percent (95 percent confidence 1998). In principle, TB incidence could be forced down more limits 9­51 percent) reduction in the prevalence rate of all quickly, by as much as 30 percent per year, if new cases could be forms of TB in DOTS areas, as compared with the change in the Tuberculosis | 293 prevalence rate in other parts of the country (China three drugs to which bacilli are fully susceptible. Of greatest Tuberculosis Control Collaboration 2004; PRC Ministry of importance is resistance to the two principal first-line drugs, Health 2000). These findings imply that the targets of halving isoniazid and rifampicin (that is, MDR-TB). The introduction of prevalence and death rates between 1990 and 2015 are techni- resistance testing, second-line drugs, longer treatment regimens cally feasible, at least in countries that are not burdened by high (12 to 18 months), and rigorous bacteriological and clinical rates of HIV infection or drug resistance. monitoring all increase program costs without necessarily ensur- Many of the 182 national DOTS programs in existence by ing high cure rates (equal to or greater than 85 percent). Indeed, the end of 2003 have shown that they can achieve high cure achieving the same cure rates for MDR-TB patients as for rates: the average treatment success rate was 82 percent (that is, patients carrying fully susceptible strains may not be possible. the percentage that were sputum-smear negative at the end of The cost-effectiveness of this component of a TB control pro- treatment plus the percentage that had completed treatment gram is therefore lower by an amount that depends on the nature but for whom cure was not confirmed by sputum smear), not of the resistance, the methods of testing and monitoring, and the far below the 85 percent international target (WHO 2005). choice of regimen. The higher costs and lower cure rates associ- The outstanding deviations below that average were in Africa ated with treating drug-resistant TB are part of the argument for (73 percent) and some former Soviet republics (for example, preventing the spread of resistance in the first place, as can be 67 percent in Russia). Although the completion of treatment investigated with models of selection and transmission (Dye and was almost a guarantee of cure before the spread of HIV and Espinal 2001; Dye and others 2002; Dye and Williams 2000). drug resistance, "completed" is an unsatisfactory way to report Suarez and others (2002) have investigated the cost-effectiveness the outcome of treatment if cure is in doubt. of managing drug-resistant TB in Peru, but because studies in Although most TB patients probably receive some form of other settings have yet to be published, an empirical overview is treatment, only 45 percent of all estimated new smear-positive not yet possible. Further data will be available from studies in cases were reported by DOTS programs to WHO in 2003. The Estonia, the Philippines, and Russia in 2005. case-detection rate in DOTS programs has been accelerating globally since 2000, but the annual increment must be still greater if the 70 percent target is to be reached by the end of Treatment of HIV Coinfection 2005. Observations on the way DOTS is presently implemented Antiretroviral therapy for HIV-positive individuals is unlikely to suggest that a ceiling on case detection might be reached prevent a large fraction of TB cases unless treatment can be at about 50 to 60 percent (Dye and others 2003; WHO 2005). given shortly after HIV infection is acquired (Sonnenberg and This fraction is about the same as the percentage of all cases others 2005; Williams and Dye 2003). In general, antiretroviral reported annually to WHO from all sources (that is, from therapy is likely to be most effective, not in reducing TB inci- DOTS and non-DOTS programs). The problem is that, as dence, but in extending the life expectancy of HIV-positive DOTS programs have expanded geographically, they have not patients successfully treated for TB (Friedland and others 2004). yet reached far beyond existing public health reporting systems. Antiretroviral therapy and DOTS are formally synergistic, because without undergoing both together, HIV-infected TB patients have a short life expectancy, typically less than five years. ALTERNATIVE AND COMPLEMENTARY Where the prevalence of HIV infection has been rising APPROACHES TO THE DIAGNOSIS quickly, as in eastern and southern Africa, even the most ener- AND TREATMENT OF ACTIVE DISEASE getic programs of TB chemotherapy may not be able to reverse the rise in TB incidence. However, mathematical modeling indi- The limitations of the DOTS strategy have stimulated numerous cates that, even in the midst of a major HIV epidemic, early initiatives to improve program performance (including treat- detection and cure are the most cost-effective ways of minimiz- ment protocols for patients carrying drug-resistant bacilli or ing TB cases and deaths (Currie and others, 2005). One reason who are infected with HIV), active case finding, collaborations is that DOTS programs treat all TB cases, not just those linked within and between public and private sector health services, with HIV. The alternatives--the prevention of HIV infection, schemes for outpatient and community-based treatment, and TLTI, and antiretroviral therapy--are less promising strategies integration of the management of TB and other illnesses. to control TB, at least for the coming decade, although they could be used in combination with DOTS. Management of Drug-Resistant Disease The higher the proportion of patients carrying drug-resistant Active Case Finding bacilli is, the greater the need for accurate resistance testing and The DOTS strategy is based on passive case detection for three for the provision of alternative regimens that include at least reasons: (a) the majority of incipient TB cases develop active 294 | Disease Control Priorities in Developing Countries | Christopher Dye and Katherine Floyd smear-positive, infectious disease more quickly than any rea- and others 2003; Floyd and others 2003; Floyd, Wilkinson, and sonable interval between successive rounds of mass screening Gilks 1997; Moalosi and others 2003; Okello and others 2003; for TB symptoms or x-ray abnormalities; (b) the majority of Sinanovic and others 2003; Vassall and others 2002; Wilkinson, patients severely ill with a life-threatening disease are likely to Floyd, and Gilks 1997). Various schemes have been used to seek help quickly (Toman 1979); and (c) countries that have provide TB care in the community, in which nongovernmental not yet implemented effective systems for passive case detec- organizations, volunteers (Okello and others 2003), or tion are not in a position to pursue cases more actively. The appointed "guardians" (Floyd and others 2003) supervise drawback of passive case finding is that the delays to diagnosis treatment, sometimes with financial incentives (Sinanovic and among symptomatic patients are often long, and health ser- others 2003). Consequently, community-based care is being vices never see some patients. To shorten delays and increase adopted in some countries (for example, Uganda) as standard the proportion of cases detected, studies of risk can identify procedure. subpopulations in which TB tends to be relatively common. Systematic surveys of these subpopulations for active TB may Integrated Management of Tuberculosis and Other be logistically feasible and affordable. The target populations Respiratory Illnesses include individuals infected with HIV, refugees (Marks and others 2001), contacts of active cases (Claessens and oth- Surveys in nine countries found that up to one-third of patients ers 2002; Noertjojo and others 2002), health workers over five years of age attending primary health centers had res- (Cuhadaroglu and others 2002), and drug users and prisoners piratory symptoms, of whom 5 to 10 percent were TB suspects, (Nyangulu and others 1997). Despite the practical possibili- but only 1 to 2 percent had TB (WHO 2004e). Because TB ties and the potential effect on transmission (Murray and is rare among respiratory diseases, comanaging TB with other Salomon 1998), active case finding is rarely done in high- conditions has clear advantages. The purpose of the WHO's burden countries, where the emphasis is still on implementing Practical Approach to Lung Health (PAL) project is to encour- the basic DOTS strategy. age a syndromic approach to management of patients, to stan- dardize health service delivery through the development and implementation of clinical guidelines, and to promote the nec- Case Finding and Treatment in the Private Sector essary coordination within national health services. Preliminary It is well known that many TB patients first seek treatment from investigations in the Kyrgyz Republic and Morocco suggest that private practitioners and that diagnosis and treatment in the PAL projects can improve the accuracy of diagnosis, encourage private sector often do not meet internationally accepted stan- better practice in prescribing drugs, and strengthen primary dards (Uplekar, Pathania, and Raviglione 2001). A new scheme care. However, a full analysis of costs and effects in the nine- to deliver DOTS through the private sector (Public-Private Mix country study remains to be done. DOTS) operates through the provision of free drugs, by infor- mation exchange and patient referral, and with some financial COST-EFFECTIVENESS OF INTERVENTIONS support from participating governments. Two pilot projects in AGAINST TUBERCULOSIS Hyderabad and Delhi, India, improved case-detection rates by 26 percent and 47 percent, respectively, and maintained treat- Some questions about investing in TB control are broad and ment success close to the target of 85 percent (WHO 2004b). strategic (for example, should money be spent on the control Other such projects are under way elsewhere in India as well as of TB rather than on the control of some other condition?); in Bangladesh, Indonesia, Nepal, the Philippines, and Vietnam others are specific and technical (for example, which laboratory (WHO 2004d). diagnostic procedures should be used?). On whatever level the question is posed, cost-effectiveness analysis (CEA) has become a prominent method for evaluating and choosing among dif- Outpatient and Community-Based Treatment ferent health interventions. Early studies of the cost-effectiveness of TB control found that full ambulatory treatment, eliminating hospitalization during the first two months (intensive phase), was cheaper and did not Background compromise cure rates (de Jonghe and others 1994; Murray Between 1980 and 2004, 32 studies of the cost-effectiveness of and others 1991). Partly as a result, ambulatory treatment has TB control were published from the low- and middle-income become the standard of care in many high-burden countries. countries considered by the Disease Control Priorities Project The natural extension, to home- and community-based treat- (table 16.1; online annex 3 summarizes the 32 studies that have ment, has proved to be just as effective in several African set- been published according to the country and year of publica- tings, and even lower in cost (Adatu and others 2003; Dudley tion, the question being addressed, the strategies compared, the Tuberculosis | 295 Table 16.1 Number of Studies on the Cost-Effectiveness of TB Control by Topic and Region, 1980­2004 East Asia Europe and Latin America Middle East Number that and the Central and the and Sub-Saharan consider Intervention Pacific Asia Caribbean North Africa South Asia Africa World Total transmission BCG vaccination 1 0 0 0 0 0 0 1 0 TLTI 0 0 0 0 0 3 0 3 3 Treatment of 4 2 0 1 0 2 0 9 4 active disease: the DOTS strategy Variations on DOTS: Management of drug- 0 0 1 0 0 1 0 2 1 resistant disease Treatment of HIV 0 0 0 0 0 1 0 1 0 coinfection Active case finding 0 1 1 0 0 4 1 7 1 and diagnosis Outpatient and 0 0 0 0 2 7 0 9 0 community-based treatment All interventions 5 3 2 1 2 18 1 32 9 Source: Authors. subjects and costs considered, the effectiveness of measures cases through reduced transmission.) Little work has been done used, whether or not transmission is considered, and the main in China, India, and other large countries in Asia, even though results and conclusions). Almost all of these studies (28, or Asia carries the largest burden of TB, and only limited informa- 88 percent) have concerned ways of finding, diagnosing, and tion is available for Europe and Central Asia, Latin America and treating patients with active TB, and most (18, or 56 percent) the Caribbean, and the Middle East and North Africa. Of the have been done in eight countries in Sub-Saharan Africa (Floyd 32 studies, only 10 used a measure of effectiveness that allows 2003). Three studies (all in Sub-Saharan Africa) have investi- comparison with other diseases (table 16.2), and only 9 gated TLTI, and one study in Indonesia has examined BCG attempted to include an estimate of the benefits gained from vaccination. The principal findings are that short-course reduced transmission (table 16.1). The benefits from reduced chemotherapy for active TB is a comparatively cost-effective transmission are usually assessed through mathematical model- intervention and one of the most cost-effective of all health ing (using computer simulations) for a given epidemiological interventions. TB patients can be treated more cheaply and situation, an approach that produces specific solutions for each conveniently outside hospitals on an ambulatory basis, by setting rather than results that are generally applicable. In addi- health staff or with the help of family and community tion, although the benefits from prevented transmission are members, without compromising the success of treatment. lower when TB is endemic, existing studies do not make a clear Supplementary methods, such as standardized second-line distinction between the cost-effectiveness of interventions in drug treatment for MDR-TB, appear to be affordable and cost- epidemic (outbreak) and endemic situations. effective in some settings. What does not emerge from this compilation of data is a comprehensive overview of the value for money provided by Methods current and potential interventions against TB in all major In this study, a general analytical framework was used to eval- regions of the world, expressed using a common measure of uate the total costs and total effects (defined as cases prevent- effectiveness and based on a consistent approach to the evalua- ed, deaths averted, and DALYs gained) of the principal tion of transmission. (The returns on investment in infectious interventions against TB across six regions of the world (see disease control include the immediate benefits to individuals online annexes 4­7 for further details). A dynamic infectious treated--for example, those vaccinated or given drug therapy-- disease model (online annex 4) was used to derive general plus the longer-term benefits gained by preventing secondary formulas for calculating the cost-effectiveness of interventions 296 | Disease Control Priorities in Developing Countries | Christopher Dye and Katherine Floyd Table 16.2 Number of Studies on the Cost-Effectiveness of TB Control by Effectiveness Measure and Intervention, 1980­2004 Cases detected Cure or or cases Cases successful Deaths Years of QALYs DALYs Intervention diagnosed prevented treatment rate prevented life saved gained gained BCG vaccination 0 1 0 1 0 0 0 TLTI 0 3 0 0 0 1 0 Treatment of active disease: 0 1 6 1 3 0 1 the DOTS strategy Variations on DOTS: Management of drug- 0 0 2 2 0 0 1 resistant disease Treatment of HIV 0 0 0 1 0 0 0 coinfection Active case finding and 5 0 0 0 1 0 2 diagnosis Outpatient and community- 0 0 10 0 0 1 0 based treatment All interventions 5 5 18 5 4 2 4 Source: Authors. QALY quality-adjusted life year. Note: The total for all interventions is greater than the number of studies because some studies use more than one measure of effectiveness. to control endemic (online annex 5) and epidemic (online Costs were considered from a health system or provider per- annex 6) TB in a wide variety of settings. The formulas are spective. They were calculated by combining estimates of the approximate, but they are simple and able to provide insights quantities of resources required for each intervention (per into the strategies that give value for money under a wide patient or per person treated) with the unit prices of those variety of epidemiological circumstances. The model was then resources (in 2001 U.S. dollars) using the cost categories and supplied with cost and efficacy data (online annex 7) for unit prices defined in the Disease Control Priorities costing each of the six World Bank regions for four main groups of guidelines. interventions: · immunization with BCG (proportion of infants, m, COST-EFFECTIVENESS OF MANAGING assumed to be protected against severe, noninfectious child- ENDEMIC TUBERCULOSIS hood TB only), or a new vaccine that prevents infection and The primary problem in global TB control is the management progression to pulmonary and extrapulmonary TB in chil- of disease in countries where incidence has been roughly stable dren and adults for many years (that is, where TB is endemic). · isoniazid treatment of latent TB infection (TLTI, given at per capita rate ), for people infected with M. tuberculosis, with or without HIV coinfection and with or without the Cost per Case Prevented use of radiography to exclude patients with active disease In monetary terms, the cost-effectiveness (C/E) of a new pro- · short-course chemotherapy, delivered as a component of gram of treatment for active infectious disease (here defined as the DOTS strategy, for smear-positive or smear-negative sputum-smear positive), per case prevented, can be calculated pulmonary disease and extrapulmonary disease (with a from C/E P/ kT, where P is the cost of treatment, is the effi- combination of drugs given at per capita rate ), and for cacy of treatment, k is a constant determined by the mode of patients infected with HIV, with or without supporting anti- action of the intervention, and T is the duration of the inter- retroviral therapy vention in years (online annex 5). The cost per case prevented · treatment for MDR-TB using a standardized regimen is mostly in the range of US$1,000 to US$10,000, depending on including first- and second-line drugs or using individual- the region of the world (figure 16.1). The exception is Europe ized regimens of first- and second-line drugs that are tai- and Central Asia, where costs are high because patients are cur- lored to each patient's drug susceptibility pattern. rently treated for long periods in hospitals rather than on an Tuberculosis | 297 Cost per case prevented (US$) 1,000,000 100,000 10,000 1,000 100 10 BCG New vaccine TLTI TLTI Treat infectious Treat infectious Treat Treat MDR Treat MDR ( x-ray) ( x-ray) ( trans) ( trans) noninfectious (standard) (individual) Cost per death prevented (US$) 1,000,000 100,000 10,000 1,000 100 BCG New vaccine TLTI TLTI Treat infectious Treat infectious Treat Treat MDR Treat MDR ( x-ray) ( x-ray) ( trans) ( trans) noninfectious (standard) (individual) Cost per DALY gained (US$) 100,000 Cost-effectiveness of 10,000 US$1 per day of healthy life gained. 1,000 365 100 10 0 BCG New vaccine TLTI TLTI Treat infectious Treat infectious Treat Treat MDR Treat MDR ( x-ray) ( x-ray) ( trans) ( trans) noninfectious (standard) (individual) East Asia and the Pacific Europe and Central Asia Latin America and the Caribbean Middle East and North Africa South Asia Sub-Saharan Africa Sub-Saharan Africa­outpatient Source: Authors. Note: Where shown, bar 7 is for ambulatory (outpatient) treatment in Sub-Saharan Africa. The treatment of active disease saves no additional cases of TB when the effects of reducing transmission are excluded, so the cost per case prevented cannot be calculated. Cost-effectiveness of vaccination and TLTI is calculated for an initial incidence rate of 100 per 100,000 population per year. Cost-effectiveness ratios are plotted on a logarithmic scale. Error bars are 90 percent confidence limits. The horizontal gray line in the third chart marks a cost- effectiveness of US$1 per day of healthy life gained. Figure 16.1 Cost-Effectiveness of Different Interventions against Endemic TB 298 | Disease Control Priorities in Developing Countries | Christopher Dye and Katherine Floyd ambulatory basis. These cost-effectiveness ratios (CERs) are (greater than US$90), assuming resistant bacilli are as trans- computed from the total costs and total effects of treatment. missible and pathogenic as susceptible bacilli. Costs are therefore the same as the incremental costs for new BCG vaccination is not much less cost-effective than the programs. If costs and effects are compared with those of a pre- treatment of active disease (US$40 to US$170 per DALY vious treatment program, CERs for the treatment of active dis- gained; higher where the risk of infection is lower). If a new ease are often negative; that is, the program sooner or later vaccine with 75 percent efficacy against pulmonary disease and saves money, as well as preventing TB cases. The positive CERs other forms of TB costs the same as BCG, it would be almost reported here for new treatment programs are, in this sense, as cost-effective (US$20 to US$100 per DALY gained) as the upper estimates. ambulatory treatment of active TB. As expected from the pre- The cost of TLTI per active case prevented also depends on ceding analysis, TLTI is much more expensive than all other the initial incidence rate (I) and is calculated from C/E P/ kIT options (US$5,500 to US$26,000 per DALY gained) and most (online annex 5). The cost is substantially higher than that for costly where the death rate from TB among adults is already the treatment of active TB: US$20,000 to US$40,000 when relatively low--for example, because an effective DOTS radiography is used to exclude patients with active disease, program already exists. Although the cost-effectiveness of but it is less (US$13,000 to US$20,000) if active TB can be each intervention varies among regions, the variation among ruled out on the basis of symptoms and clinical examination strategies is much greater, whatever the outcome measure (figure 16.1). TLTI is less cost effective than the treatment of (figure 16.1). active TB because preventive treatment would be given to latently infected individuals, most of whom were not recently COST-EFFECTIVENESS OF MANAGING infected and who are at small risk of developing active disease. TUBERCULOSIS OUTBREAKS In an endemic setting, there is no feasible method of identify- ing individuals who have recently acquired infection and who The basic case reproduction number, R0, is a ready-made epi- will proceed rapidly to active TB. demiological tool for relating effort and reward in the manage- A new vaccine that prevents infection and, hence, the ment of outbreaks. R0 is the average number of secondary cases progression to pulmonary TB among people who were previ- generated by a primary case introduced into a previously unin- ously uninfected would be extremely competitive (US$90 to fected population (Anderson and May 1991). No country is US$200) per case prevented if the costs were the same as presently free of TB, but some countries have recently suffered those for BCG. BCG is cheap to manufacture and administer "epidemic" increases in incidence from previously low levels. (US$1 to US$3 per dose) but less cost-effective (US$2,000 to The algebraic expression of R0 for TB reveals how the various US$8,500 per case prevented) than the treatment of active dis- components of a disease's natural history and the different ease because it is assumed to protect against severe forms of kinds of intervention interact with each other to influence childhood TB only and because it does not affect transmission transmission and the generation of new cases (online annex 4). (figure 16.1). For example, the cost-effectiveness of chemotherapy per M. tuberculosis generation is C/E P / R0, where is the number of TB patients treated per prevalent case per unit time, and is Cost per Death Prevented and DALY Gained the proportion of new cases that is infectious. The wider benefits of treating active TB are revealed when The biggest resurgences of TB in recent history have been allowing for the additional reduction in case fatality. For a driven by the spread of HIV in Africa and are linked to the rise 10-year program of treatment for infectious TB, the cost per of drug resistance in former Soviet republics; this analysis is death prevented is typically US$150 to US$750, and the cost confined to interventions associated with these two phenomena per DALY gained is US$5 to US$50 for all regions except (figure 16.2; online annex 6). Indeed, in this study, interventions Europe and Central Asia (figure 16.1). When TB is close to the related to TB with HIV are considered only in the epidemic endemic equilibrium, the extra benefits gained from reducing context. transmission under DOTS are small: the cost per DALY gained If multidrug-resistant strains of M. tuberculosis are assumed is only 60 percent higher when transmission benefits are to have the same intrinsic transmissibility and pathogenicity as excluded. The treatment of noninfectious TB is less cost- drug-susceptible strains, and given the spread of MDR-TB as effective (US$60 to US$200 per DALY gained), not primarily an independent epidemic (Dye and Williams 2000), then treat- because transmission is unaffected, but because the case fatality ment of MDR-TB with a standard regimen including second- of untreated smear-negative and extrapulmonary disease is line drugs is more costly per DALY gained than treatment relatively low. Treating infectious MDR-TB is between two and of fully susceptible disease in Sub-Saharan Africa, but it is ten times more costly than treating drug-susceptible TB per marginally less costly than TLTI (with an x-ray screen) over death prevented (greater than US$2,000), or per DALY gained most rates of case detection and treatment (online annex 6). Tuberculosis | 299 Cost per case prevented (US$) 100,000 10,000 1,000 100 10 0 TLTI TLTI HIV Treat Treat MDR Treat MDR Treat active Treat active ( x-ray) ( x-ray) infectious standard individual HIV HIV antiretroviral ( trans) therapy Cost per death prevented (US$) 100,000 10,000 1,000 100 10 0 TLTI TLTI HIV Treat Treat MDR Treat MDR Treat active Treat active ( x-ray) ( x-ray) infectious standard individual HIV HIV antiretroviral ( trans) therapy Cost per DALY gained (US$) 10,000 Cost-effectiveness of US$1 per day of healthy life gained. 1,000 365 100 10 0 TLTI TLTI HIV Treat Treat MDR Treat MDR Treat active Treat active ( x-ray) ( x-ray) infectious standard individual HIV HIV antiretroviral ( trans) therapy East Asia and the Pacific Europe and Central Asia Latin America and the Caribbean Middle East and North Africa South Asia Sub-Saharan Africa­outpatient Source: Authors. Note: Five interventions used in the management of TB epidemics that are linked with HIV and MDR-TB (TLTI for people coinfected with TB and HIV, treatment of infectious MDR-TB with a standard or individual regimen, treatment of HIV-infected TB patients with TB drugs, treatment of HIV-infected TB patients with TB and antiretroviral drugs) are compared with two standard methods (TLTI, with active disease excluded by x-ray screen, and treatment of active infectious disease, allowing for transmission). Cost-effectiveness ratios (plotted on a logarithmic scale) vary with the treatment rate (online annex 6); for illustration here, 20 percent of eligible people are treated annually with each intervention. The horizontal gray line in the third figure marks a cost-effectiveness of US$1 per day of healthy life gained. Error bars are 90 percent confidence limits. Figure 16.2 Cost-Effectiveness of Managing Epidemic TB 300 | Disease Control Priorities in Developing Countries | Christopher Dye and Katherine Floyd For example, at the fixed rate of treatment used to generate with HIV. Neither does this analysis address all the important figure 16.2, treatment of MDR-TB with a standard regimen questions about managing outbreaks of drug-resistant or HIV- costs US$91 to US$846 per DALY gained, depending on the related TB. Fuller investigations should assess, for example, the region, as compared with US$6 to US$31 for the treatment of benefits to whole populations of giving antiretroviral therapy drug-susceptible TB. The treatment of MDR-TB with regimens to HIV-infected individuals before they develop TB and of tailored to the resistance patterns of individual patients is more investing in DOTS to prevent multidrug-resistant epidemics costly but also more efficacious than standardized treatment from arising in the first place. for MDR-TB and, therefore, almost equally cost-effective under this set of assumptions. SUMMARY OF COST-EFFECTIVENESS ANALYSES TB patients infected with HIV are more costly to treat per DALY gained than HIV-negative patients, either without anti- Box 16.1 summarizes the results of these calculations of the retroviral therapy (low cost, short life expectancy) or with such cost-effectiveness of managing epidemic and endemic TB. The therapy (high cost, long life expectancy). TLTI is a more attrac- findings are one justification for maintaining and expanding tive option for the management of epidemic TB than for DOTS programs, on the basis of SCC for patients with active endemic TB (compare figures 16.1 and 16.2), because during disease, as the dominant mode of TB control around the world. an outbreak, TLTI is directed at recent rather than remote BCG vaccination and the treatment of MDR-TB (standard or infection. TLTI is even more cost-effective in the control of TB individualized regimens) or HIV-infected TB patients (with or and HIV coinfection, because it prevents the rapid breakdown without supporting antiretroviral therapy) are more costly in to active disease caused by immunodeficiency. absolute terms, but they typically cost less than US$1 per day of These results are indicative rather than definitive, because healthy life gained, which is less than the average economic the calculations assume, among other things, that HIV-infected productivity of workers in the least developed countries. TLTI populations exist in isolation; in reality, HIV-infected people appears to be relatively poor value for money, even though this also acquire TB infection from TB patients who are not infected analysis assumes that one course of treatment prevents active Box 16.1 Cost-Effectiveness of TB Interventions: Main Findings · The cost effectiveness of TB control depends not only · Even with relatively favorable assumptions, the treat- on local costs but also on the local characteristics of TB ment of latent TB infection where TB is endemic and epidemiology (for example, epidemic or endemic, low populations are unaffected by HIV is the least cost- or high rates of HIV infection and drug resistance) and effective of the interventions examined here (US$5,500 on the rate of application of any chosen intervention. to US$26,000 per DALY gained). TLTI is more cost- · Short-course chemotherapy for the treatment of infec- effective during outbreaks (US$150 to US$500 per tious and noninfectious TB patients through the DOTS DALY gained) and for people who are coinfected with strategy is highly cost-effective for the control of either TB and HIV (US$15 to US$300 per DALY gained). epidemic or endemic TB (US$5 to US$50 per DALY · BCG vaccination to prevent severe forms of childhood gained, for regions excluding Eastern and Central TB is much less effective than SCC but nearly as cost- Europe). When a new treatment program is compared effective (US$40 to US$170 per DALY gained). with a previous program, DOTS often saves money as · A new vaccine that prevents pulmonary TB with high well as preventing cases and deaths. efficacy (equal to or greater than 75 percent) would be · Some variations on DOTS are less cost-effective but more cost-effective than BCG if the cost of immuniza- still good value for money, including the treatment of tion were the same as BCG (US$20 to US$100 per patients with MDR-TB (standard or individualized DALY gained). drug regimens) and with HIV infection (with or with- · For any intervention with the potential to cut trans- out supporting antiretroviral therapy). For these addi- mission (that is, excluding BCG vaccination), control tional interventions, the cost per DALY gained is less of epidemic disease produces more favorable cost- than the annual average economic productivity per effectiveness ratios than control of endemic disease, capita in the least developed countries. because the benefits gained from reduced transmission are greater during outbreaks. Source: Authors. Tuberculosis | 301 TB for life. TLTI is more cost-effective in epidemic than in children between 1990 and 2003. If chemotherapy is assumed to endemic settings, and it is more cost-effective when it is used reduce only the case-fatality rate and to have no effect on trans- to treat individuals coinfected with TB and HIV. A new, high- mission and incidence, 23 million deaths (44 percent) would efficacy vaccine that prevents infection and the progression have been saved in non-DOTS treatment programs. The expan- to pulmonary TB in adults, to be directed at the control of sion to 45 percent case-detection rate under DOTS during the endemic TB, would be more cost-effective than BCG at the same period saved an estimated 2.3 million ( 5 percent) addi- same price and almost as cost-effective as SCC. tional deaths, the largest numbers in Sub-Saharan Africa (1.1 million), East Asia and the Pacific (558,000), and South Asia Averted and Avertable Burden of Tuberculosis (408,000). Further analysis shows that, if 70 percent of TB cases (smear positive and smear negative) can be treated under DOTS Trends in case notifications can be used, judiciously, to assess before MDG target year 2015, an estimated 1.9 million TB regional and global trends in TB incidence, but no satisfactory deaths (26 per 100,000) will occur in that year, a greater num- large-scale analysis has been done of the number of cases ber than in 1990, but a 7 percent lower death rate per capita prevented by chemotherapy (as distinct from the reductions (Dye and others 2005). in transmission and susceptibility associated with improved The calculations for Africa assume that treatment cures TB living standards). One approach to evaluating the averted and in the majority of HIV-infected patients even though, without avertable burden of TB begins with the observation that antiretroviral therapy, many of these patients will die anyway. 86 percent of the years of healthy life lost that are attributable Despite these favorable assumptions, the number of TB deaths to TB are from premature death, and only 14 percent are from was evidently still rising in Africa in 2003, whereas it was falling illness. Because DALYs lost are dominated by premature death, in Asia, aided by the large programs of DOTS expansion in a conservative estimate of the burden of TB alleviated can be China (1991­97) and India (from 1998). obtained in terms of the number of deaths and associated Reducing the TB death rate sufficiently to meet the MDG DALYs gained, regionally and globally, since the introduction target requires a significant cut in incidence, as well as in case of the DOTS strategy in 1991. fatality. An extension of this assessment suggests that case detec- Figure 16.3 is derived from recent estimates of cases and tion must reach at least 70 percent and the TB incidence rate deaths and their trends by region, including those attributable must fall by 5 to 6 percent annually between 2003 and 2015 to HIV coinfection (Corbett and others 2003; WHO 2004c). (Dye and others 2005). For the world, excluding Sub-Saharan In the MDG baseline year, 1990, approximately 1.5 million TB Africa and former Soviet republics, the incidence rate would deaths (28 per 100,000) occurred. BCG vaccination saved have to fall at a more modest 2 percent per year. roughly 650,000 deaths from extrapulmonary TB among New diagnostics, drugs, and vaccines would also help reduce the global TB burden more quickly. The most desirable of these is a vaccine that prevents pulmonary disease, whether or not Number of deaths per year (millions) vaccination subjects are already infected (a pre- or postexposure No BCG 4 vaccine), and that confers lifetime immunity (Andersen 2001; No treatment Fordham von Reyn and Vuola 2002; McMurray 2003; Young and Stewart 2002). A new vaccine with high efficacy against 3 pulmonary TB would almost certainly change immunization No DOTS practice: mass vaccination campaigns among adults (rather 2 DOTS than infants) would have dramatic effects, going far beyond the expectations of DOTS programs (figure 16.4; Dye 2000). A 1 DOTS 70 percent case-detection rate postexposure vaccine that stops progression to disease among those already infected, as well as preventing infection in others, 0 1990 1992 1994 1996 1998 2000 2002 would have greater effect than a preexposure vaccine that only Source: Authors. prevents infection (Lietman and Blower 2000). However, such Note: Broken and gray lines represent various hypothetical scenarios; the solid black line calculations are at present highly speculative, because the mode represents DOTS programs. The interventions are, from top to bottom: no BCG vaccination and no anti-TB treatment, no treatment, no DOTS programs, DOTS expansion from zero of action and efficacy of any new vaccine is unknown. to 45 percent case detection over the period 1990­2003, and DOTS with 70 percent case- detection rate throughout the period 1990­2003. To make a conservative assessment of effect, the treatment of active TB is assumed to change the case-fatality rate without affecting the TB incidence rate. Economic Benefits of Tuberculosis Control Preventing TB deaths brings no savings in the costs of TB con- Figure 16.3 Estimated Number of TB Deaths Worldwide under Various Hypothetical Scenarios and the Estimated Effect of DOTS trol unless it is accompanied by a reduction in incidence so that Programs, 1990­2003 fewer patients require treatment. The prompt and effective 302 | Disease Control Priorities in Developing Countries | Christopher Dye and Katherine Floyd Incidence rate per 100,000 population per year for TB are likely to give net returns on investment or at least to 140 appear to be good value for money in ways that go beyond the 120 arguments from cost-effectiveness (Jack 2001). The analysis earlier in this chapter showed that SCC typi- 100 1 cally costs up to US$30 per DALY gained for the treatment of 80 2 infectious TB and up to US$200 per DALY gained for the treat- 60 ment of noninfectious TB (excluding Europe and Central 3 40 Asia). These figures can be compared with a recent estimate of 4 US$1.5 billion as the annual global cost of treating 70 percent 20 5 6 of cases with 85 percent cure (WHO 2004c). Reaching these 0 10 0 10 20 targets would prevent approximately 2.1 million of all the TB Years from start of intervention deaths expected if no treatment were available in 2003, includ- ing 391,000 deaths prevented by DOTS (figure 16.3). Because Source: Dye 2000. Note: Lines and points show: (1) no intervention in a population where TB is already each TB death prevented gains approximately 20 DALYs in slow decline, as in many countries in Asia and Latin America; (2) a postexposure (WHO 2002b), the total cost per DALY gained would be about vaccine given annually to infected infants so that 20 percent are immunized; (3) a postexposure vaccine given annually to infected infants so that 70 percent are US$36. This rough calculation excludes any benefits in reduced immunized; (4) DOTS reaching 70 percent case detection and 85 percent cure by year 5 and maintained at these levels thereafter; (5) one-time mass immunization with a transmission but includes the costs of treating smear-negative preexposure vaccine giving 70 percent protection to uninfected people, followed by and extrapulmonary TB and is of the same order of magnitude annual vaccination of infants with the same fraction protected; and (6) one-time mass immunization with a postexposure vaccine giving 70 percent protection to uninfected as the results from CEA. people, followed by annual vaccination of infants with the same fraction protected. However the calculation is done, the cost of gaining a year Figure 16.4 Hypothetical Effect of New Vaccines on TB Incidence of healthy life under DOTS is substantially less than the annual Rate average productivity per capita in the low-income (gross national income [GNI] less than or equal to US$735) or least treatment of active disease is almost certainly reducing trans- developed (GNI average US$290, http://www.worldbank.org) mission around the world, but because the effect on incidence countries, and it is probably less than the marginal productiv- is necessarily slow, it has been hard to quantify in all but a few ity of labor in the poorest communities. It is also less than countries, notably Peru (Suarez and others 2001). twice the average annual income per capita, which has also The monetary savings implied by a reduction in incidence been proposed as a benchmark for assessing whether an inter- of one-quarter (26 percent) between 2000 and 2015--which vention is cost-effective (Garber and Phelps 1997). Moreover, may be enough to achieve the MDG targets--could be magni- it is less than the World Bank's definition of absolute poverty fied or diminished by adjustments to the DOTS strategy. On (living on US$1 per day or less, close to average GNI per capita the one hand, without compromising cure rates, chemotherapy for the least developed countries) and is certainly less than the can be delivered more cheaply to outpatients than inpatients monetary values that are typically placed on the value of a and with less reliance on x-ray diagnosis and surgical proce- human life year (for example, a life was valued at US$100,000 dures. On the other hand, various additions to DOTS--contact by the 2004 Copenhagen Consensus panel, http://www. tracing, active case finding, antiretroviral therapy for HIV- copenhagenconsensus.com). All these comparisons suggest infected patients, second-line drugs for patients carrying resist- not only that the basic DOTS strategy, and perhaps even an ant bacilli, or joint public-private schemes for the management enhanced DOTS strategy, are cost-effective but that they also of TB--might be desirable but more costly per year of healthy have very favorable cost-benefit ratios. life gained. Whether the savings made by reducing incidence and improving efficiency offset the costs of DOTS add-ons will, therefore, depend on the setting. RESEARCH AND DEVELOPMENT Besides the possibility of reducing diagnostic and treatment costs, improved health and longevity yield other economic The preceding review and analysis suggest at least six areas for benefits, but the quantification of those benefits is always con- economic and epidemiological research and development: troversial. This difficulty is reflected in the limited number of cost-benefit analyses of TB control; among the few examples, 1. DOTS expansion. Refinement of existing cost estimates of one detailed study in India estimated the potential societal ben- scaling up DOTS programs to reach and move beyond efits of DOTS to be worth US$8.3 billion in 1993­94, or 4 per- targets for case detection (70 percent) and cure (85 percent) cent of the gross domestic product (Dholakia 1996). Without in the poorest countries--notably in Africa--through more attempting to extend such analyses here, we note that the comprehensive planning and budgeting exercises. The preceding results also imply that large-scale treatment programs analyses should include the costs of developing fully staffed Tuberculosis | 303 health services, with expanded and renovated infrastructure Among a growing list of new vaccine antigens (Fruth and and improved management capacity where necessary, and Young 2004), three of the most promising are now undergo- the costs of the new initiatives that will be required to ing phase 1 safety trials in humans. One trial has evaluated improve case detection and cure rates. mycobacterial antigen 85, delivered as a recombinant smallpox 2. Service delivery. Assessment of the potential for health vaccine (Goonetilleke and others 2003). Another is testing a service restructuring to detect, diagnose, and treat TB live attenuated BCG bacterium (rBCG30) that overexpresses patients more efficiently through syndromic management of antigen 85B protein and that provides guinea pigs with greater respiratory diseases at primary health centers and through protection than BCG alone (Horwitz and others 2000). A third collaborations between public and private health services, trial is assessing a fusion protein of two different antigens in between different parts of the public sector health service, adjuvant, referred to as Mtb72f, that is likely to be used as a and between TB and HIV/AIDS control programs. booster to either BCG or rBCG30 (Reed and others 2003). 3. Complementary strategies. Further investigation of the costs Compounds that could form the basis of new drugs and and effectiveness of strategies that are potentially comple- new drug regimens include the nitroimidazopyran PA-824. mentary to DOTS, including active case finding and TLTI in Experiments with a mouse model of TB have shown that high-risk populations, and the management of drug resist- PA-824 has bactericidal activity similar to that of isoniazid and ance and of patients infected with HIV. sterilizing activity that may rival that of rifampicin and that it 4. Impact and targets. Evaluation of the actual and potential is particularly active against dormant bacilli. effects of the tools (mostly drugs) now being used for TB Among the most important recent discoveries is a control. This research requires a better understanding of the diarylquinoline with a novel mode of action on the ATP syn- ways human population density, age structure, migration, thase of M. tuberculosis that powerfully inhibits both drug- HIV coinfection, and drug resistance affect TB epidemi- sensitive and drug-resistant strains of bacilli (Andries and ology. The analyses should check the internal consistency others 2004). Alongside these laboratory studies, analytical of international targets for the implementation and effect of and operational research are needed to find out what kinds of chemotherapy programs, as defined by the MDGs. The new tools will give the best returns on investment. Investigations analyses should also make better use of the rich body of of this kind will contribute to the introduction of new vac- routine surveillance data collected by all national TB control cines, drugs, and diagnostics and will inform the work of the programs around the world. Foundation for Innovative New Diagnostics (http://www. 5. Risk factors. Assessment of the reductions in TB cases and finddiagnostics.org), the Global Alliance for TB Drug deaths that could be made by reducing exposure to environ- Development (http://www.tballiance.org), and the AerasGlobal mental risk factors, notably indoor and outdoor air pollu- TB Vaccine Foundation (http://www.aeras.org). tion, tobacco smoking, and malnutrition. These risk factors affect the establishment of infection, the progression to active disease, and the outcome of treatment. CONCLUSIONS 6. New diagnostics, drugs, and vaccines. A sensitive and specific test for active TB that is cheap and simple to use at the first After more than a decade of climbing incidence rates in Africa point of contact between patients and health services would and former Soviet republics, the global TB epidemic appears be a major advance in diagnosis. Mycobacterial culture, once again to be on the threshold of decline. The spread of HIV which detects a higher proportion of active TB patients than and drug resistance, respectively, in those two regions has exac- sputum-smear microscopy, is a prerequisite for screening erbated the problems of TB control, but at the same time it for drug resistance. However, present culture methods are has helped keep TB on the international public health agenda. slow, taking four to six weeks to obtain a result. Technology The global incidence rate was still rising in 2003, but more based on phage amplification and nucleic acid amplification slowly each year. This slowdown is not only (or even mainly) can establish whether cultures are positive in days or hours, because of direct intervention through DOTS programs but but this technology needs to be packaged for use in devel- because HIV epidemics are approaching peak levels in Africa oping countries (Albert and others 2002, 2004; Johansen and because incidence is now starting to fall again in some and others 2003; Woods 2001). The tuberculin skin test former Soviet republics, including Russia. Where TB incidence is being superseded in many developed countries by more is already falling, prevalence and death rates should be drop- specific methods for detecting infection (Doherty and ping more quickly, although little evidence demonstrates this others 2002; Pai, Riley, and Colford 2004). A test that can decrease yet. predict who will progress from latent to active disease, as yet The prompt diagnosis and treatment of active TB has been hypothetical, would greatly increase the feasibility of treat- the mainstay of TB control and will continue to be so for the ing latent infection. foreseeable future. Short-course chemotherapy, delivered 304 | Disease Control Priorities in Developing Countries | Christopher Dye and Katherine Floyd through the DOTS strategy, is, at typically US$5 to US$350 per such a complete analysis. The results of CEA are therefore DALY gained, the most cost-effective among current methods typically used more informally, along with other evidence and for the management of TB, and in most high-burden countries, constraints, when a mix of health interventions is chosen. the cost is toward the lower end of this range. A comparison of Although this problem will recur in discussions about allo- the costs of treating active TB with the costs of running a pre- cating health budgets, the case for large-scale programs of TB vious program suggests that DOTS could actually save money treatment has now been accepted in many parts of the world. in the long run. In addition, DOTS provides an operational That is the fruit of more than 10 years' work on burden, cost, framework for the introduction of more specialized methods efficacy, effectiveness, and cost-effectiveness. The governments in certain risk groups. The extensions to DOTS investigated of the less poor members of the group of 22 high-burden coun- here include the treatment of MDR-TB with second-line drugs, tries have demonstrated that they can budget for, and provide, preventive therapy (TLTI) during outbreaks and for people most of the funds needed to reach target levels of case detection coinfected with M. tuberculosis and HIV, and antiretroviral and cure (WHO 2004c, 2005). Some of the poorer countries therapy for HIV-infected TB patients. Those interventions cost among the 22 are now receiving sufficient external assistance more than the basic DOTS strategy but are still less than a to fill the gaps in their budgets for TB control, principally dollar for each day of healthy life gained, which provides an from the Global Fund to Fight AIDS, Tuberculosis, and economic argument for their integration into enhanced DOTS Malaria. Consequently, the total reported budget deficit for the programs. high-burden countries in 2005 was remarkably small--just Although the analyses in this chapter show that DOTS and US$119 million--and concentrated in the poorest countries its extensions are good value for money, they conceal various (WHO 2005). features of health systems, as yet poorly defined, that may facil- From those findings and observations arise two key ques- itate the implementation of treatment programs. For example, tions for global TB control: If the estimated budget gap is filled, if broader investment in the health sector is needed before TB would the money be enough to ensure that enhanced DOTS control programs can work in some parts of some countries, programs reach 70 percent case detection and 85 percent then the full cost of DOTS could be greater. By contrast, a more cure--and by when? And if those targets are reached, will the integrated approach to the management of TB and other respi- effort be sufficient to achieve the MDG objectives of halving ratory diseases in primary health facilities could lead to cost prevalence and death rates by 2015? savings. Those possibilities have not yet been investigated. As yet, there are only partial answers. On the costs, it is clear The only development that could radically alter the current that, by moving treatment out of hospitals and into the com- approach to TB control--shifting the emphasis from cure to munity, DOTS can often be made cheaper and more conven- prevention--is the discovery of a new vaccine that protects ient for patients and health services without compromising adults against infectious pulmonary disease. Whether such a treatment outcome. However, planning for TB control in the vaccine would be more or less cost-effective than BCG (US$40 poorest countries is still inadequate, and budgets commonly to US$1,600 per DALY gained) depends on price and efficacy, understate the real costs of scaling up national TB control pro- but the potential epidemiological effect would be far greater grams (WHO 2004c, 2005). Despite those weaknesses in the than that of BCG, perhaps justifying mass adult vaccination. If budgeting and funding process, the overall expenditure on TB research and development proceed according to plan, a new control in high-burden countries has increased since 2000, and vaccine of some kind could be licensed between 2010 and 2015. the injection of extra effort and money has led to a small accel- New drugs and diagnostics should be available earlier, shorten- eration in case finding globally. As a result, case detection under ing the delay to, and duration of, treatment. DOTS could reach 50 to 60 percent by 2005, and treatment Although cost-effectiveness studies show that DOTS is a success should be close to the target level of 85 percent. good investment, they do not formally show that the strategy is A case-detection rate of 50 to 60 percent may not be affordable. The analytical difficulty is that CEA does not solve enough. The analysis in this chapter suggests that the MDG the practical problem of how to allocate money to TB control objective of halving the death rate can be reached with 70 per- in combination with other interventions, or even how to com- cent case detection globally, provided this case detection also bine different approaches to TB control (Tan-Torres Edejer and generates a 5 to 6 percent annual reduction in the incidence others 2004). Interpreted literally, CEA says that the best return rate between 2003 and 2015. The DOTS program in Peru gen- on total investment is obtained by ranking interventions erated a 6 to 7 percent annual reduction in the incidence rate of according to CER and then fully implementing each interven- pulmonary TB, but that result has not yet been repeated in tion, from smallest to largest CER, allowing for diminishing other high-burden countries with good control programs (for returns, until the total budget is spent. This method is unlikely example, India, Morocco, and Vietnam). It is unlikely to be to lead to a balanced health care portfolio in the poorest achieved in African countries that currently have high rates of countries. Besides, the evidence is rarely available to carry out HIV infection. Tuberculosis | 305 Although others have emphasized that the costs of infec- Asamoah-Odei, E., J. M. Garcia Calleja, and J. T. Boerma. 2004. "HIV tious disease control can be related to the benefits in complex Prevalence and Trends in Sub-Saharan Africa: No Decline and Large Subregional Differences." Lancet 364: 35­40. ways (Brandeau, Zaric, and Richter 2003), we advocate the use Behr, M. A, S. A. Warren, H. Salamon, P. C. Hopewell, A. Ponce de Leon, of a powerful new method of carrying out CEA, which is based C. L. Daley, and P. M. Small. 1999. "Transmission of Mycobacterium on the observation that mathematical models can be used to tuberculosis from Patients Smear-Negative for Acid-Fast Bacilli." Lancet generate simple (albeit approximate) and general formulas that 353: 444­49. relate reward to effort in the management of both epidemic Borgdorff, M. W., N. J. Nagelkerke, C. Dye, and P. Nunn. 2000. "Gender and Tuberculosis: A Comparison of Prevalence Surveys with (based on R0) and endemic (based on dynamics in the vicinity Notification Data to Explore Sex Differences in Case Detection." of equilibrium) TB. The results are similar to those obtained by International Journal of Tuberculosis and Lung Disease 4: 123­32. using more complex simulations in specific settings, and they Bourdin Trunz, B., P. E. M. Fine, and C. Dye. Forthcoming. Global Impact are accurate enough to offer a choice between interventions of BCG Vaccination on Childhood Tuberculous Meningitis and Miliary Tuberculosis. (Currie and others, 2005). The generality of the method expos- Brandeau, M. L., G. S. Zaric, and A. Richter. 2003. "Resource Allocation for es more clearly the reasons some interventions are compara- Control of Infectious Diseases in Multiple Independent Populations: tively cost-effective and indicates the range of conditions under Beyond Cost-Effectiveness Analysis." Journal of Health Economics 22: which specific cost-effectiveness results apply. The scope for 575­98. using this approach for other infectious diseases remains to be Bucher, H. C., L. E. Griffith, G. H. Guyatt, P. Sudre, M. Naef, P. Sendi, explored, but it should be readily applicable in the evaluation and M. Battegay. 1999. "Isoniazid Prophylaxis for Tuberculosis in HIV Infection: A Meta-Analysis of Randomized Controlled Trials." AIDS of new approaches to TB control, whether through vaccina- 13: 501­7. tion, drug treatment, the reduction of environmental risks, or China Tuberculosis Control Collaboration. 2004. "The Effect of improved service delivery. Tuberculosis Control in China." Lancet 364: 417­22. Claessens, N. J. M., F. F. Gausi, S. Meijnen, M. M. Weismuller, F. M. Salaniponi, and A. D. Harries. 2002. "High Frequency of Tuberculosis in Households of Index TB Patients." International Journal of ACKNOWLEDGMENTS Tuberculosis and Lung Disease 6: 266­69. Cohn, D. L. 2003. "Treatment of Latent Tuberculosis Infection." Seminars The authors wish to thank Uli Fruth, Kreena Govender, in Respiratory Infections 18: 249­62. Ulla Griffiths, Mehran Hosseini, Anne Mills, Mark Perkins, Cohn, D. L., and W. M. El-Sadr. 2000. "Treatment of Latent Tuberculosis Catherine Watt, Diana Weil, and Brian Williams for help of Infection." In Tuberculosis: A Comprehensive International Approach, various kinds during the preparation of this chapter. ed. L. B. Reichman and E. S. Hershfield, 471­502. New York: Marcel Dekker. Comstock, G. W. 2000. "How Much Isoniazid Is Needed for Prevention of Tuberculosis among Immunocompetent Adults? In Reply." REFERENCES International Journal of Tuberculosis and Lung Disease 4: 485­86. Aaron, L., D. Saadoun, I. Calatroni, O. Launay, N. Memain, V. Vincent, and Comstock, G. W., C. Baum, and D. E. Snider. 1979. "Isoniazid Prophylaxis others. 2004. "Tuberculosis in HIV-Infected Patients: A Comprehensive among Alaskan Eskimos: A Final Report of the Bethel Isoniazid Review." Clinical Microbiology and Infection 10: 388­98. Studies." American Review of Respiratory Disease 119: 827­30. Adatu, F., R. Odeke, M. Mugenyi, G. Gargioni, E. McCray, E. Schneider, Comstock, G. W., V. T. Livesay, and S. F. Woolpert. 1974. "The Prognosis and D. Maher. 2003. "Implementation of the DOTS Strategy for of a Positive Tuberculin Reaction in Childhood and Adolescence." Tuberculosis Control in Rural Kiboga District, Uganda, Offering American Journal of Epidemiology 99: 131­38. Patients the Option of Treatment Supervision in the Community, Corbett, E. L., S. Charalambous, V. M. Moloi, K. Fielding, A. D. Grant, 1998­1999." International Journal of Tuberculosis and Lung Disease 7: C. Dye, and others. 2004. "Human Immunodeficiency Virus and the S63­71. Prevalence of Undiagnosed Tuberculosis in African Gold Miners." Albert, H., A. Heydenrych, R. Brookes, R. J. Mole, B. Harley, E. Subotsky, American Journal of Respiratory Critical Care Medicine 170: 673­79. and others. 2002. "Performance of a Rapid Phage-Based Test, Corbett, E. L., R. W. Steketee, F. O. ter Kuile, A. S. Latif, A. Kamali, and FASTPlaqueTBTM, to Diagnose Pulmonary Tuberculosis from Sputum R. J. Hayes. 2002. "HIV-1/AIDS and the Control of Other Infectious Specimens in South Africa." International Journal of Tuberculosis and Diseases in Africa." Lancet 359: 2177­87. Lung Disease 6 (6): 529­37. Corbett E. L., C. J. Watt, N. Walker, D. Maher, B. G. Williams, M. C. Albert, H., A. Trollip, T. Seaman, and R. J. Mole. 2004. "Simple, Phage- Raviglione, and C. Dye. 2003. "The Growing Burden of Tuberculosis: Based (FASTPlaque) Technology to Determine Rifampicin Resistance Global Trends and Interactions with the HIV Epidemic." Archives of of Mycobacterium tuberculosis Directly from Sputum." International Internal Medicine 163: 1009­21. Journal of Tuberculosis and Lung Disease 8: 1114­19. Cuhadaroglu, C., M. Erelel, L. Tabak, and Z. Kilicaslan. 2002. "Increased Andersen, P. 2001. "TB Vaccines: Progress and Problems." Trends in Risk of Tuberculosis in Health Care Workers: A Retrospective Survey at Immunology 22: 160­68. a Teaching Hospital in Istanbul, Turkey." BioMed Central Infectious Anderson, R. M., and R. M. May. 1991. Infectious Diseases of Humans: Diseases 2: 14. Dynamics and Control. Oxford, U.K.: Oxford University Press. Currie, C. S. M., K. Floyd, B. G. Williams, and C. Dye. 2005. "Cost Andries, K., P. Verhasselt, J. Guillemont, H. W. Gohlmann, J. M. Neefs, Affordability and Cost-Effectiveness of Strategies to Control H. Winkler, and others. 2004. "A Diarylquinoline Drug Active on the Tuberculosis in Countries with High HIV, Prevalence." BMC Public ATP Synthase of Mycobacterium Tuberculosis." Science 307: 223­27. Health 5 (1): 130. 306 | Disease Control Priorities in Developing Countries | Christopher Dye and Katherine Floyd de Jonghe, E., C. J. Murray, H. J. Chum, D. S. Nyangulu, A. Salomao, and International Journal of Tuberculosis and Lung Disease 7 (Suppl. 1): K. Styblo. 1994. "Cost-Effectiveness of Chemotherapy for Sputum S29­37. Smear-Positive Pulmonary Tuberculosis in Malawi, Mozambique and Floyd, K., D. Wilkinson, and C. Gilks. 1997. "Comparison of Cost Effec- Tanzania." International Journal of Health Planning and Management 9: tiveness of Directly Observed Treatment (DOT) and Conventionally 151­81. Delivered Treatment for Tuberculosis: Experience from Rural South de Viedma, D. G., M. Marin, S. Hernangomez, M. Diaz, M. J. R. Serrano, Africa." British Medical Journal 315 (7): 1407­11. L. Alcala, and E. Bouza. 2002. "Reinfection Plays a Role in a Population Fordham von Reyn, C., and J. M. Vuola. 2002. "New Vaccines for the Whose Clinical/Epidemiological Characteristics Do Not Favor Prevention of Tuberculosis. Clinical Infectious Diseases 35: 465­74. Reinfection." Archives of Internal Medicine 162: 1873­79. Frieden, T. R., P. I. Fujiwara, R. M. Washko, and M. A. Hamburg. 1995. Dholakia, R. 1996. The Potential Economic Benefits of the DOTS Strategy "Tuberculosis in New York City--Turning the Tide." New England against TB in India. Geneva: World Health Organization. Journal of Medicine 333: 229­33. Doherty, T. M., A. Demissie, J. Olobo, D. Wolday, S. Britton, T. Eguale, and Frieden, T., T. R. Sterling, S. S. Munsiff, C. J. Watt, and C. Dye. 2003. others. 2002. "Immune Responses to the Mycobacterium tuberculosis­ "Tuberculosis." Lancet 362: 887­99. Specific Antigen ESAT-6 Signal Subclinical Infection among Contacts of Tuberculosis Patients." Journal of Clinical Microbiology 40 (2): Friedland, G., S. Abdool Karim, Q. Abdool Karim, U. Lalloo, C. Jack, 704­6. N. Gandhi, and W. El Sadr. 2004. "Utility of Tuberculosis Directly Observed Therapy Programs as Sites for Access to and Provision of Dudley, L., V. Azevedo, R. Grant, J. H. Schoeman, L. Dikweni, and D. Maher. Antiretroviral Therapy in Resource-Limited Countries." Clinical 2003."Evaluation of Community Contribution to Tuberculosis Control Infectious Diseases 38 (Suppl. 5): S421­28. in Cape Town, South Africa." International Journal of Tuberculosis and Lung Disease 7 (Suppl. 1): S48­55. Fruth, U., and D. Young. 2004. "Prospects for New TB Vaccines: Stop TB Working Group on TB Vaccine Development." International Journal of Dye, C. 2000. "Tuberculosis 2000­2010: Control, but Not Elimination." Tuberculosis and Lung Disease 8: 151­55. International Journal of Tuberculosis and Lung Disease 4 (Suppl. 2): S146­52. Garber, A. M., and C. E. Phelps. 1997. "Economic Foundations of Cost- Effectiveness Analysis." Journal of Health Economics 16: 1­31. Dye, C., and M. A. Espinal. 2001. "Will Tuberculosis Become Resistant to All Antibiotics?" Proceedings of the Royal Society of London, Series B, Glynn, J. R., M. D. Yates, A. C. Crampin, B. M. Ngwira, F. D. Mwaungulu, Biological Sciences 268: 45­52. G. F. Black, and others. 2004. "DNA Fingerprint Changes in Tuberculosis: Reinfection, Evolution, or Laboratory Error?" Journal of Dye, C., G. P. Garnett, K. Sleeman, and B. G. Williams. 1998. "Prospects for Infectious Diseases 190: 1158­66. Worldwide Tuberculosis Control under the WHO DOTS Strategy." Lancet 352: 1886­91. Goonetilleke, N. P., H. McShane, C. M. Hannan, R. J. Anderson, R. H. Brookes, and A. V. Hill. 2003. "Enhanced Immunogenicity and Dye, C., S. Scheele, P. Dolin, V. Pathania, and M. C. Raviglione. 1999. Protective Efficacy against Mycobacterium tuberculosis of Bacilli "Global Burden of Tuberculosis: Estimated Incidence, Prevalence, and Calmette-Guérin Vaccine Using Mucosal Administration and Boosting Mortality by Country." Journal of the American Medical Association with a Recombinant Modified Vaccinia Virus Ankara." Journal of 282: 677­86. Immunology 171: 1602­9. Dye, C., C. J. Watt, D. M. Bleed, S. M. Hosseini, and M. C. Raviglione. 2005. Grange, J. 2003. "Immunophysiology and Immunopathology." In Clinical "The Evolution of Tuberculosis Control, and Prospects for Reducing Tuberculosis, 3rd ed., ed. P. D. O. Davies, 88­104. London: Arnold. Incidence, Prevalence and Deaths Globally." Journal of the American Medical Association 293: 2767­75. Hamid Salim, M. A., E. Declercq, A. Van Deun, and K. A. R. Saki. 2004. "Gender Differences in Tuberculosis: A Prevalence Survey Done in Dye, C., C. J. Watt, D. M. Bleed, and B. G. Williams. 2003. "What Is the Bangladesh." International Journal of Tuberculosis and Lung Disease 8: Limit to Case Detection under the DOTS Strategy for Tuberculosis 952­57. Control?" Tuberculosis 83: 35­43. Hernandez-Garduno, E., V. Cook, D. Kunimoto, R. K. Elwood, W. A. Black, Dye, C., and B. G. Williams. 2000. "Criteria for the Control of Drug- and J. M. FitzGerald. 2004. "Transmission of Tuberculosis from Smear Resistant Tuberculosis." Proceedings of the National Academy of Sciences Negative Patients: A Molecular Epidemiology Study." Thorax 59: USA 97: 8180­85. 286­90. Dye, C., B. G. Williams, M. A. Espinal, and M. C. Raviglione. 2002."Erasing Horwitz, M. A., G. Harth, B. J. Dillon, and S. Maslesa-Galic. 2000. the World's Slow Stain: Strategies to Beat Multidrug-Resistant "Recombinant Bacillus Calmette-Guérin (BCG) Vaccines Expressing Tuberculosis." Science 295: 2042­46. the Mycobacterium tuberculosis 30-kDA Major Secretory Protein Dye, C., F. Zhao, S. Scheele, and B. G. Williams. 2000."Evaluating the Impact Induce Greater Protective Immunity against Tuberculosis Than of Tuberculosis Control: Number of Deaths Prevented by Short-Course Conventional BCG Vaccines in a Highly Susceptible Animal Model." Chemotherapy in China." International Journal of Epidemiology 29: Proceedings of the National Academy of Sciences USA 97: 13853­58. 558­64. Hudelson, P. 1996. "Gender Differentials in Tuberculosis: The Role of Espinal, M. A., A. Laszlo, L. Simonsen, F. Boulahbal, S. J. Kim, A. Reniero, Socio-Economic and Cultural Factors." Tubercle and Lung Disease 77: and others. 2001. "Global Trends in Resistance to Antituberculosis 391­400. Drugs." New England Journal of Medicine 344: 1294­1303. Jack, W. 2001. "The Public Economics of Tuberculosis Control." Health Fine, P. E. M. 2001. "BCG Vaccines and Vaccination." In Tuberculosis: A Policy 57: 79­96. Comprehensive International Approach, ed. L. B. Reichman, and E. S. Jamison, D. T., W. H. Mosley, A. R. Meashem, and J. L. Bobadilla. 1993. Hershfield, 503­24. New York: Marcel Dekker. Disease Control Priorities in Developing Countries. New York: Oxford Floyd, K. 2003. "Costs and Effectiveness: The Impact of Economic Studies University Press. on TB Control." Tuberculosis (Edinburgh) 83: 187­200. Johansen, I. S., B. Lundgren, A. Sosnovskaja, and V. Ø. Thomsen. 2003. Floyd, K., J. Skeva, T. Nyirenda, F. Gausi, and F. Salaniponi. 2003."Cost and "Direct Detection of Multidrug-Resistant Mycobacterium tuberculosis Cost-Effectiveness of Increased Community and Primary Care Facility in Clinical Specimens in Low- and High-Incidence Countries by Line Involvement in Tuberculosis Care in Lilongwe District, Malawi." Probe Assay." Journal of Clinical Microbiology 41 (9): 4454­56. Tuberculosis | 307 Johnson, J. L., A. Okwera, D. L. Hom, H. Mayanja, C. Mutuluuza Kityo, Okello, D., K. Floyd, F. Adatu, R. Odeke, and G. Gargloni. 2003. "Cost and P. Nsubuga, and others. 2001. "Duration of Efficacy of Treatment of Cost-Effectiveness of Community-Based Care in Rural Uganda." Latent Tuberculosis Infection in HIV-Infected Adults." AIDS 15: International Journal of Tuberculosis and Lung Disease 7 (Suppl. 1): 2137­47. S72­79. Krebs, W. 1930. "Die Fälle von Lungentuberkulose in der aargauischen Pai, M., L. W. Riley, and J. M. Colford Jr. 2004. "Interferon-gamma Assays Heilstätte Barmelweid aus den Jahren 1912­1927." Beiträge zur Klinik in the Immunodiagnosis of Tuberculosis: A Systematic Review." Lancet der Tuberkulose 74: 345­79. Infectious Diseases 4 (12): 761­76. Kristensen, I., P. Aaby, and H. Jensen. 2000. "Routine Vaccinations and PRC (People's Republic of China) Ministry of Health. 2000. Report on Child Survival: Follow Up Study in Guinea-Bissau, West Africa." British Nationwide Random Survey for the Epidemiology of Tuberculosis in Medical Journal 321: 1435­38. 2000. Beijing: PRC Ministry of Health. Letvin, N. L., B. R. Bloom, and S. L. Hoffman. 2001."Prospects for Vaccines Quigley, M. A., A. Mwinga, M. Hosp, I. Lisse, D. Fuchs, J. D. H. Porter, and to Protect against AIDS, Tuberculosis, and Malaria." Journal of the P. Godfrey-Faussett. 2001. "Long-Term Effect of Preventive Therapy American Medical Association 285: 606­11. for Tuberculosis in a Cohort of HIV-Infected Zambian Adults." AIDS 15: 215­22. Lietman, T., and S. M. Blower. 2000. "Potential Impact of Tuberculosis Vaccines as Epidemic Control Agents." Clinical Infectious Diseases 30 Radhakrishna, S., T. R. Frieden, and R. Subramani. 2003. "Association of (Suppl. 3): S316­22. Initial Tuberculin Sensitivity, Age, and Sex with the Incidence of Tuberculosis in South India: A 15-Year Follow-Up." International Marais B. J., R. P. Gie, H. S. Schaaf, A. C. Hesseling, C. C. Obihara, L. J. Journal of Tuberculosis and Lung Disease 7: 1083­91. Nelson, and others. 2004. "The Clinical Epidemiology of Childhood Pulmonary Tuberculosis: A Critical Review of Literature from the Pre- Reed, S. G., M. R. Alderson, W. Dalemans, Y. Lobet, and Y. A. W. Skeiky. Chemotherapy Era." International Journal of Tuberculosis and Lung 2003."Prospects for a Better Vaccine against Tuberculosis." Tuberculosis Disease 8: 278­85. 83: 213­19. Richardson, M., N. M. Carroll, E. Engelke, G. D. Van Der Spuy, F. Salker, Marks, G. B., J. Bai, G. J. Stewart, S. E. Simpson, and E. A. Sullivan. 2001. Z. Munch, and others. 2002. "Multiple Mycobacterium Tuberculosis "Effectiveness of Postmigration Screening in Controlling Tuberculosis Strains in Early Cultures from Patients in a High-Incidence among Refugees: A Historical Cohort Study, 1984­1998." American Community Setting." Journal of Clinical Microbiology 40: 2750­54. Journal of Public Health 91: 1797­99. Rieder, H. L. 1999. "Epidemiologic Basis of Tuberculosis Control." Paris: McMurray, D. N. 2003. "Recent Progress in the Development and Testing International Union against Tuberculosis and Lung Disease. of Vaccines against Human Tuberculosis." International Journal of Parasitology 33: 547­54. ------. 2003. "BCG Vaccines." In Clinical Tuberculosis, 3rd ed., ed. P. D. O. Davies, 337­53. London: Arnold. Menzies, D., M. J. Dion, B. Rabinovitch, S. Mannix, P. Brassard, and K. Schwartzman. 2004. "Treatment Completion and Costs of a Rieder, H. L., D. E. Snider Jr., and G. M. Cauthen. 1990. "Extrapulmonary Randomized Trial of Rifampin for 4 Months versus Isoniazid for Tuberculosis in the United States." American Review of Respiratory 9 Months." American Journal of Respiratory and Critical Care Medicine Disease 141: 347­51. 170: 445­49. Shafer, R. W., and B. R. Edlin. 1996. "Tuberculosis in Patients Infected with Human Immunodeficiency Virus: Perspective on the Past Decade." Moalosi, G., K. Floyd, J. Phatshwane, T. Moeti, N. Binkin, and T. Kenyon. Clinical Infectious Diseases 22: 683­704. 2003. "Cost-Effectiveness of Home-Based Care versus Hospital Care for Chronically Ill Tuberculosis Patients, Francistown, Botswana." Shilova, M. V., and C. Dye. 2001. "The Resurgence of Tuberculosis in International Journal of Tuberculosis and Lung Disease 7 (Suppl. 1): Russia." Philosophical Transactions of the Royal Society of London, S80­85. Series B, Biological Sciences 356: 1069­75. Murray, C. J. L., E. de Jonghe, H. J. Chum, D. S. Nyangulu, A. Salomao, and Sinanovic, E., K. Floyd, L. Dudley, V. Azevedo, R. Grant, and D. Maher. K. Styblo. 1991. "Cost Effectiveness of Chemotherapy for Pulmonary 2003. "Cost and Cost-Effectiveness of Community-Based Care for Tuberculosis in Three Sub-Saharan African Countries." Lancet 338: Tuberculosis in Cape Town, South Africa." International Journal of 1305­8. Tuberculosis and Lung Disease 7 (Suppl. 1): S56­62. Sonnenberg, P., J. R. Glynn, K. Fielding, J. Murray, P. Godfrey-Faussett, and Murray, C. J. L., and J. A. Salomon. 1998. "Modeling the Impact of Global S. Shearer. 2005. "How Soon after Infection with HIV Does the Risk of Tuberculosis Control Strategies." Proceedings of the National Academy Tuberculosis Start to Increase? A Retrospective Cohort Study in South of Sciences USA 95: 13881­86. African Gold Miners." Journal of Infectious Diseases 191: 150­58. Mwinga, A., M. Hosp, P. Godfrey-Faussett, M. Quigley, P. Mwaba, B. N. Styblo, K. 1991. Epidemiology of Tuberculosis. 2nd ed. The Hague: Royal Mugala, and others. 1998. "Twice Weekly Tuberculosis Preventive Netherlands Tuberculosis Association. Therapy in HIV Infection in Zambia." AIDS 12: 2447­57. Suarez, P. G., K. Floyd, J. Portocarrero, E. Alarcon, E. Rapiti, G. Ramos, Nelson, L. J., and C. D. Wells. 2004. "Global Epidemiology of Childhood and others. 2002. "Feasibility and Cost-Effectiveness of Standardised Tuberculosis." International Journal of Tuberculosis and Lung Disease 8: Second-Line Drug Treatment for Chronic Tuberculosis Patients: A 636­47. National Cohort Study in Peru." Lancet 359: 1980­89. Neuenschwander, B. E., M. Zwahlen, S. J. Kim, E. G. Lee, and H. L. Rieder. Suarez, P. G., C. J. Watt, E. Alarcon, J. Portocarrero, D. Zavala, R. Canales, 2002. "Determination of the Prevalence of Infection with and others. 2001. "The Dynamics of Tuberculosis in Response to Mycobacterium tuberculosis among Persons Vaccinated against Bacillus 10 Years of Intensive Control Effort in Peru." Journal of Infectious Calmette-Guérin in South Korea." American Journal of Epidemiology Diseases 184: 473­78. 155: 654­63. Sutherland, I. 1968. "The Ten-Year Incidence of Clinical Tuberculosis Noertjojo, K., C. M. Tam, S. L. Chan, J. Tan, and M. Chan-Yeung. 2002. Following `Conversion' in 2,550 Individuals Aged 14 to 19 Years." "Contact Examination for Tuberculosis in Hong Kong Is Useful." Unpublished progress report of the Tuberculosis Surveillance and International Journal of Tuberculosis and Lung Disease 6: 19­24. Research Unit, KNCV, The Hague, Netherlands. Nyangulu, D. S., A. D. Harries, C. Kang'ombe, A. E. Yadidi, K. Chokani, ------. 1976. "Recent Studies in the Epidemiology of Tuberculosis, Based T. Cullinan, and others. 1997. "Tuberculosis in a Prison Population in on the Risk of Being Infected with Tubercle Bacilli." Advances in Malawi." Lancet 350: 1284­87. Tuberculosis Research 19: 1­63. 308 | Disease Control Priorities in Developing Countries | Christopher Dye and Katherine Floyd Sutherland, I., E. Svandova, and S. Radhakrishna. 1982."The Development Tuberculosis in Ugandan Adults Infected with the Human Immunode- of Clinical Tuberculosis Following Infection with Tubercle Bacilli: 1. ficiency Virus." New England Journal of Medicine 337: 801­8. A Theoretical Model for the Development of Clinical Tuberculosis WHO (World Health Organization). 2001. Vaccine Preventable Diseases: Following Infection, Linking from Data on the Risk of Tuberculosis Monitoring System--2001 Global Summary. Geneva: WHO, Depart- Infection and the Incidence of Clinical Tuberculosis in the ment of Vaccines and Biologicals. Netherlands." Tubercle 63: 255­68. ------. 2002a. An Expanded DOTS Framework for Effective Tuberculosis Tan-Torres Edejer, T., R. Baltussen, T. Adam, R. Hutubessy, A. Acharya, Control. Geneva: WHO. D. B. Evans, and C. J. L. Murray, eds. 2004. WHO Guide to Cost- Effectiveness Analysis. Geneva: World Health Organization. ------. 2002b. The World Health Report: Reducing Risks, Promoting Healthy Life. Geneva: WHO. Terris-Prestholt, F., and L. Kumaranayake. 2003. "Cost Analysis of the Zambian ProTEST Project: A Package to Reduce the Impact of ------. 2004a. Anti-Tuberculosis Drug Resistance in the World. Report 3. Tuberculosis and Other HIV-Related Diseases." Unpublished report, Geneva: WHO. London School of Hygiene and Tropical Medicine. ------. 2004b. Cost and Cost-Effectiveness of Public-Private Mix DOTS: Toman, K. 1979. Tuberculosis Case-Finding and Chemotherapy. Questions Evidence from Two Pilot Projects in India. Geneva: WHO. and Answers. Geneva: World Health Organization. ------. 2004c. Global Tuberculosis Control: Surveillance, Planning, Uplekar, M., V. Pathania, and M. Raviglione. 2001. "Private Practitioners Financing. Geneva: WHO. and Public Health: Weak Links in Tuberculosis Control." Lancet 358: ------. 2004d. Public-Private Mix for DOTS: Global Progress. Geneva: 912­16. WHO. van Rie, A., R. Warren, M. Richardson, T. C. Victor, R. P. Gie, D. A. Enarson, ------. 2004e. Respiratory Care in Primary Care Services--A Survey in 9 and others. 1999. "Exogenous Reinfection as a Cause of Recurrent Countries. Geneva: WHO. Tuberculosis after Curative Treatment." New England Journal of ------. 2004f. World Health Report 2004: Changing History. Geneva: WHO. Medicine 341: 1174­79. ------. 2005. Global Tuberculosis Control: Surveillance, Planning, Vassall, A., S. Bagdadi, H. Bashour, H. Zaher, and P. V. Maaren. 2002."Cost- Financing. Geneva: WHO. Effectiveness of Different Treatment Strategies for Tuberculosis in Egypt and Syria." International Journal of Tuberculosis and Lung Wilkinson, D., K. Floyd, and C. F. Gilks. 1997."Costs and Cost-Effectiveness Disease 6: 1083­90. of Alternative Tuberculosis Management Strategies in South Africa-- Implications for Policy." South African Medical Journal 87 (4): 451­55. Verver, S., R. M. Warren, Z. Munch, E. Vynnycky, P. D. van Helden, M. Richardson, and others. 2004. "Transmission of Tuberculosis in a Wilkinson, D., S. B. Squire, and P. Garner. 1998. "Effect of Preventive High Incidence Urban Community in South Africa." International Treatment for Tuberculosis in Adults Infected with HIV: Systematic Journal of Epidemiology 33: 351­57. Review of Randomised Placebo Controlled Trials." British Medical Vynnycky, E., and P. E. M. Fine. 1997. "The Natural History of Journal 317: 625­29. Tuberculosis: The Implications of Age-Dependent Risks of Disease and Williams, B. G., and C. Dye. 2003. "Antiretroviral Drugs for Tuberculosis the Role of Reinfection." Epidemiology and Infection 119: 183­201. Control in the Era of HIV/AIDS." Science 301: 1535­37. ------. 2000. "Life Time Risks, Incubation Period, and Serial Interval of Woods, G. L. 2001. "Molecular Techniques in Mycobacterial Detection." Tuberculosis." American Journal of Epidemiology 152: 247­63. Archives of Pathology and Laboratory Medicine 125 (1): 122­26. Whalen, C. C., J. L. Johnson, A. Okwera, D. L. Hom, R. Huebner, Young, D. B., and G. R. Stewart. 2002. "Tuberculosis Vaccines." British P. Mugyenyi, and others. 1997. "A Trial of Three Regimens to Prevent Medical Bulletin 62: 73­86. Tuberculosis | 309 Chapter 17 Sexually Transmitted Infections Sevgi O. Aral and Mead Over, with Lisa Manhart and King K. Holmes Sexually transmitted infections (STIs) are responsible for an neonatal and infant infections and blindness in infants; infer- enormous burden of morbidity and mortality in many devel- tility in both men and women; urethral strictures in men; gen- oping countries because of their effects on reproductive and ital malignancies, such as cancer of the cervix uteri, vulva, child health (Wasserheit 1989) and their role in facilitating the vagina, penis, and anus; arthritis secondary to gonorrhea and transmission of HIV infection (Laga, Diallo, and Buvé 1994). chlamydia; liver failure and liver cancer secondary to hepatitis B or human T cell lymphotropic virus type I; and central nerv- ous system disease secondary to syphilis (Holmes and Aral INTRODUCTION 1991; Meheus, Schulz, and Cates 1990; van Dam, Dallabetta, and Piot 1999). Thus, STI sequelae affect mostly women and Largely because of the HIV epidemic, interest in STIs has children. increased over the past two decades. During that time, the epi- In developing countries, high levels of STIs and high rates of demiology of STIs has changed in developing countries, partly complications and sequelae result largely from inadequacies in as a result of modifications in STI case management approaches health service provision and health care seeking (Aral and and partly because of behavioral changes in response to the Wasserheit 1999). STI care is provided by a large variety of HIV epidemic. At the same time, advances in STI prevention health care providers, many of whom are poorly trained in STI have enhanced understanding of the intricacies of STI trans- case management, and the quality of care they provide is often mission dynamics and the role of interventions in the control less than desirable (Moses and others 1994; WHO 1991). of STIs. However, what has not changed is as significant as what Health care seeking for STIs is frequently inadequate, particu- has changed: the epidemiology of STIs still differs substantially larly among women (van Dam 1995), because of the low levels in the industrial countries and the developing world. The of awareness regarding sexual health, the stigmatization associ- sociocultural and economic contexts in developing countries ated with genital symptoms, and the asymptomatic nature of influence the epidemiology of STIs and help make them an many STIs. A study in Nairobi, Kenya, found that 42 percent of important public health priority. patients had been symptomatic for more than a week before Incidence and prevalence rates of STIs are generally high in coming to a clinic and that 23 percent had been symptomatic both urban and rural populations and vary considerably across for more than two weeks (Moses and others 1994). areas. Because diagnosis and treatment of STIs are often Setting up good-quality STI services is considerably more delayed, inadequate, or both, rates of STI complications are difficult in resource-poor settings than elsewhere. Variables also high in developing countries. Those complications include that affect the duration of infectiousness include adequacy of pelvic inflammatory disease, ectopic pregnancy, and chronic health workers' training, attitudes of health workers toward abdominal pain in women; adverse pregnancy outcomes, such marginalized groups as sex workers, patient loads at including abortion, intrauterine death, and premature delivery; health centers, availability of drugs and clinic supplies, and 311 costs of care (Moses and others 2002). Thus, improvements vaccines may be able to help prevent genital and anal cancers in pertaining to all these factors would greatly improve STI- the foreseeable future. Researchers are evaluating multivalent related services, help reduce the duration of infectiousness, and vaccines for preventing moderate to severe cervical dysplasia as decrease the incidence of STIs (Aral 2002a). However, in many well. Other advances include easier episodic treatment of geni- countries in the developing world, worsened economic condi- tal herpes (Strand and others 2002) and the use of suppressive tions and the increasing burden of HIV/AIDS have negatively therapy to reduce the transmission of genital herpes to regular affected these variables. For example, in South Africa, the ratio partners (Corey and others 2004). In a related development, of hospital beds to population declined from 6.5 per 1,000 in a prophylactic vaccine against herpes simplex virus type 2 1976 to 2.3 in 1996; during 1999, approximately 300 profes- (HSV-2) has shown limited efficacy in that it has proved partly sionally trained nurses left the country each month; and stu- effective for HSV-seronegative women, but not for men or dent enrollments in nursing school declined from 12,282 in herpes simplex virus type 1 (HSV-1) seropositive women 1996 to 10,398 in 1999 (Aral 2002a). (Stanberry and others 2002). Prevention successes of the recent Sexual behaviors also contribute to the STI burden in devel- past include STI sequelae, such as pelvic inflammatory disease oping countries. These behaviors are heavily influenced by the and cervical cancer. A randomized controlled trial showed that sociocultural, economic, and political contexts, which in the selective screening of women for Chlamydia trachomatis signif- past two decades have deteriorated at an accelerated rate in icantly reduced the incidence of pelvic inflammatory disease many areas. Societal change has included rising levels of (Scholes and others 1996). inequality within countries, growing inequality between coun- Widespread implementation of syndromic management as tries, increased levels of globalization, increased proportions of an approach to STI case management has apparently had a people who live in cultures they were not born in, and a larger considerable effect on the epidemiology of STIs, particularly in proportion of the world's population living in postconflict resource-poor settings (King Holmes and Michael Alary, per- societies (Aral 2002a). One effect of these changes is an increase sonal communication, May 15, 2003). in multipartner sexual activity, which in turn increases the rate In some developing countries, including Cambodia, the at which infected and susceptible individuals are sexually Dominican Republic, and Thailand, sexual risk behaviors have exposed to each other and consequently the rate at which STIs been changing over the past decade. In Uganda, for example, spread. the age of sexual debut has increased, the frequency of sex with casual partners has decreased, and the use of condoms has increased (Stoneburner and Low-Beer 2004). During the Changes in STI Epidemiology, Management, and Prevention 1990s, demographic and health surveys in 29 developing coun- since 1993 tries asked individuals if they had done anything to avoid AIDS Since 1993, STI epidemiology and management have evolved (Low-Beer and Stoneburner 2003): almost 80 percent of men interactively, particularly in developing countries. Techno- and 50 percent of women surveyed reported that they had. logical advances in diagnosis, screening, and treatment; evalua- Specific behavior changes reported included increased mono- tion and widespread implementation of new case-management gamy, reduced number of partners, avoidance of sex workers, algorithms; and changes in risk behaviors in response to the and increased condom use. AIDS epidemic have all influenced the dynamic typology of By contrast, in developed countries, recent years have seen STIs (Wasserheit and Aral 1996). behavior changes in the opposite direction; for example, in The introduction of nucleic acid amplification tests, which many European countries and in the United States, risk behav- have improved the sensitivity and expanded the repertoire of iors among men who have sex with men have increased signif- usable specimens, has heralded a new era in STI diagnosis. icantly (CDC 2004; L. Doherty and others 2002). In addition, The use of urine and vaginal swabs in diagnosis has enabled Grémy and Beltzer (2004) report declines in condom use providers to supply diagnostic and screening services outside among heterosexual adult populations in Europe. Investigators traditional clinical facilities and has greatly enhanced the cov- attribute increases in risk behaviors to the introduction and erage of outreach activities (Schachter 2001). Unfortunately, availability of antiretroviral therapy for HIV infection and the many of these tests are currently too expensive for routine use difficulties in sustaining preventive behaviors in the long term, in developing countries. Single-dose oral azithromycin has referred to as prevention fatigue. Some researchers speculate improved the treatment of several bacterial STIs (Lau and that the widespread introduction of antiretroviral therapy in Qureshi 2002), but quinolones are apparently becoming inef- developing countries may have a similar disinhibitory effect on fective for gonorrhea in some locations (Donovan 2004). sexual behaviors and that changes in sexual behavior may off- A major recent advance in STI prevention is the early suc- set the beneficial effect of antiretroviral therapy (Blower and cess of a prophylactic, monovalent human papillomavirus others 2001; Blower and Farmer 2003; Blower and Volberding (HPV) type 16 vaccine (Koutsky and others 2002); HPV 2002; Over and others 2004). 312 | Disease Control Priorities in Developing Countries | Sevgi O. Aral and Mead Over with others Advances in STI prevention in recent decades have vary considerably within and across developing countries. At enhanced understanding of transmission dynamics and the the same time, health care delivery for STIs varies by type of role of interventions. Investigators have articulated the follow- institution and location, although inadequate resources are ing five emergent insights about STI epidemiology and preven- universal in the developing world, as are recordkeeping, data tion over the past two decades: management, and data analysis. The limited data that are avail- able suggest that STIs are a major public health burden in the · Populations consist of many diverse subpopulations, and developing world. Although the prevalence and incidence of each population-level epidemic trajectory consists of many bacterial STIs have apparently declined because of expanded distinct subpopulation epidemic trajectories (Pisani and syndromic management, changes in sexual behavior, and death others 2003). The epidemic trajectories of specific STIs of high-risk populations, the prevalence and incidence of viral differ depending on when and where the infection was STIs seem to have increased over the past decade. introduced; the natural history and transmissibility of the infection; the structure of sexual networks; the demographic, Syndromic Management. Health systems can use three differ- economic, social, and epidemiological context; and the state ent approaches to manage patients presenting with symptoms of the health system (Aral and others 2005). suggestive of an STI. First, etiology-based management relies · Temporal dimensions are important in relation to STI epi- on identifying causative micro-organisms or detecting specific demiology (Aral and Blanchard 2002). At the individual antibodies. It requires costly and often technically complex lab- level, concurrency of partnerships and gaps between part- oratory diagnosis, trained personnel, quality assurance pro- nerships are risk factors for the acquisition and transmission grams, and infrastructure. Second, clinical diagnosis­based of STIs (Adimora and others 2002; Agrawal, Gillespie, and management is rapid, inexpensive, and requires less infrastruc- Foxman 2001; Kraut and Aral 2001). At the population level, ture than etiology-based management; however, clinical diag- investigators have described the evolution of STI epidemics nosis is often inaccurate, may miss multiple infections, and through sometimes predictable phases, characterized by may result in undertreatment or overtreatment. Third, syn- changing patterns in the distribution and transmission dromic management, which is based on the recognition of a of STI pathogens within and between subpopulations constellation of clinical signs and symptoms, is inexpensive, (UNAIDS and WHO 2000; Wasserheit and Aral 1996). can be standardized, and can be used by both physicians and · Sexual networks are important in the transmission dynam- paramedical personnel, though it often results in some ics of STIs at the population level, and position in a sexual overtreatment. Nevertheless, syndromic management has network is important in the transmission and acquisition of been recommended as a realistic approach for managing STIs at the individual level (Morris 2004). symptomatic patients in developing countries (Over and Piot · Trajectories whereby STI epidemics evolve differ for 1993). Implementation issues associated with the syndromic different types of population-pathogen interactions (Aral management approach involve inadequate local evaluation of 2002a; Blanchard 2002; Garnett 2002). Whereas highly treatment algorithms because of a lack of local data, inconsis- infectious, short-duration bacterial STIs--for instance, tencies in implementation, and inadequate monitoring gonorrhea--depend on the presence of core groups marked (Dallabetta, Gerbase, and Holmes 1998; Hawkes and Santhya by multiple sex partnerships of short duration for their 2002; WHO 2001b). spread, less infectious, long-duration viral STIs--for exam- Limitations of the syndromic management approach ple, HSV--depend on the presence of multiple partnerships include the inability to directly target the subclinical STI pool, of longer duration. the variability of STI symptoms and signs, the potential for · Interactions among sexually transmitted pathogens affect wasting antibiotics, the risk of promoting drug resistance, and STI epidemic trajectories at the population level (Wasserheit the unintended consequence of decreasing the skill levels of 1991). The inconsistent findings of three landmark random- health care providers (Dallabetta Gerbase, and Holmes 1998; ized community trials evaluating the effect of STI treatment Donovan 2004). Moreover, syndromic management tends to on HIV transmission (Grosskurth and others 1995; Kamali undermine STI surveillance efforts because cases are managed and others 2003; Wawer and others 1999) can be accounted and treated in the absence of a specific clinical or laboratory for by the complex, multifactorial, multilevel, and phase- diagnosis (O'Farrell 2002). specific nature of STI epidemics (Orroth 2003). Role of Core Groups and Bridge Populations. Core groups-- that is, groups of individuals who have large numbers of sex Epidemiology and Control partners who themselves have large numbers of sex partners-- The epidemiology of STI pathogens, the local prevention and play an important role in the spread and persistence of STIs care infrastructure, and the cultural and sociopolitical context and are characterized by a high prevalence of STIs. Examples of Sexually Transmitted Infections | 313 core groups include sex workers, drug users, truck drivers, and behaviors, STIs and HIV may facilitate each other's bar girls. Because a case treated or prevented in a core group transmission. member tends to prevent that person from infecting several others, interventions that target core groups tend to be more effective and more cost-effective than interventions that target BACTERIAL AND VIRAL STIs AND THEIR the general population (Ainsworth and Over 1997; Over 1999; SEQUELAE Over and Piot 1993). In situations in which a high prevalence of STIs is concentrated in core groups, so-called bridge popu- Both bacterial and viral STIs are widespread in developing lations (individuals who have sexual links with members of countries; recently, incidence of bacterial STIs has declined both high- and low-prevalence subpopulations) may play an while that of viral STIs has been increasing. important role in disseminating infection from core groups to the general population (Aral 2000; Aral and others 1999; Gorbach and others 2000; Morris and others 1996). Natural History of Bacterial STIs and Their Sequelae Several variables influence the relative importance of core Chancroid is a genital ulcer disease caused by Haemophilus groups in the spread of STIs, including the characteristics of ducreyi. Its incidence has declined greatly in both developed the specific pathogen, such as its transmissibility and duration and developing countries. This decline has been associated of infectiousness; the phase of a particular epidemic; and the with the provision of STI diagnostic and therapeutic services to duration of sexual partnerships among those involved in mul- sex workers (Steen 2001) and with improved syndromic man- tipartner sexual activity (Aral 2002a, 2002b; Blanchard 2002; agement of genital ulcers. Like other genital ulcer diseases, Garnett 2002; Wasserheit and Aral 1996). The role of core chancroid is associated with increased acquisition and trans- groups in STI dissemination tends to be greater during the ini- mission of HIV (Donovan 2004). tial and later phases of epidemics, when infection is highly con- Syphilis is a genital ulcer disease caused by Treponema pal- centrated in small, high-risk subpopulations, than during the lidum. In 1999, the World Health Organization (WHO) esti- middle phases of epidemics, when infection tends to be widely mated the global prevalence of syphilis at 12 million (WHO spread across subpopulations. The importance of core groups 2001a), with high prevalence rates in South and Southeast Asia appears to be greater in populations in which most people are and Sub-Saharan Africa. Those most likely to be affected are involved in sexual activity with a single partner and only a populations in developing countries and disadvantaged sub- small minority of people engage in short-term sexual partner- populations in developed countries. Since 1999, syphilis out- ships with a large number of sex partners (Laumann and Youm breaks have reemerged in many developed countries among 1999). men who have sex with men (CDC 2004; L. Doherty and oth- ers 2002). Among heterosexuals, sexual contact with sex Antibiotic Use and Drug Resistance. Antibiotic use is unreg- workers is an important risk factor. If untreated, syphilis dur- ulated in many developing countries, and antibiotics are fre- ing pregnancy may lead to stillbirth and congenital syphilis quently misused and overused, which results in drug resistance. (Genc and Ledger 2000). Resistance to antimicrobial drugs is increasing mortality and Gonorrhea is a discharge disease caused by N. gonorrhoeae. morbidity from infectious diseases (Hart and Kariuki 1998). In 1999, WHO estimated its global prevalence at 62.4 million STIs are among the most frequently occurring infections (WHO 2001a). Like syphilis, its prevalence is high in South worldwide, with more than 76 percent estimated to occur in and Southeast Asia and Sub-Saharan Africa, in many develop- the developing world (WHO 2001a). Neisseria gonorrhoeae has ing countries elsewhere, and among high-risk groups and shown great versatility in developing resistance to antimicro- disadvantaged subpopulations in developed countries. bial drugs, including sulfonamides, penicillins, and tetracy- Community surveys reveal a substantial pool of asymptomatic cline. Fluoroquinolones such as ciprofloxacin and ofloxacin gonococcal infections (Chandeying and others 2000; Turner have proved highly effective in treating gonorrhea, but after and others 2002). Following the emergence of AIDS, gonor- widespread and often inappropriate use of fluoroquinolones, rhea cases declined among men having sex with men, sex resistant N. gonorrhoeae has emerged. In some areas, such workers, and the general population in the developed world resistance leaves third-generation cephalosporins as the only and among sex workers in many developing countries predictably effective antibiotic treatment for gonorrhea. (Donovan 2004). In most populations tested, infection with Chlamydia tra- STIs and HIV/AIDS. Because HIV is a sexually transmitted chomatis is the most common bacterial STI. In 1999, WHO infection, people who are infected with another STI also tend estimated the global prevalence of chlamydial infection to be to be at increased risk of HIV infection and vice versa. 92 million (WHO 2001a). Chlamydial infection is common However, beyond this correlation resulting from common risk in most countries, especially among young people. Key risk 314 | Disease Control Priorities in Developing Countries | Sevgi O. Aral and Mead Over with others factors are being younger than 25 and having a new sex part- HSV-2 include severe primary disease, meningitis, hepatitis, ner. Many women with uncomplicated infection are asympto- erythema multiforme, and neonatal herpes (Donovan 2004). matic or have mild symptoms. Like untreated gonococcal Infected neonates may die or develop severe neurological infection, untreated chlamydial infection can cause pelvic sequelae despite antiviral therapy. In contrast to bacterial STIs, inflammatory disease, chronic pelvic pain, and ectopic preg- HSV-2 may be transmitted to sex partners many years after ini- nancy. Chlamydial infection is an important acquired cause of tial infection and during periods when the infected individual infertility in women (Simms and Stephenson 2000). Roughly may be asymptomatic. Infection with HSV-2 is now one of the half of men with urethral chlamydial infection develop symp- most common STIs worldwide and is the most frequent cause tomatic urethritis, chlamydial infection is the most common of genital ulcers in almost all areas; however, this observation cause of epididymitis in young men, and both men and women may be related to better diagnostic technologies rather than a may develop chlamydial conjunctivitis or reactive arthritis genuine alteration in the spectrum of genital ulcer disease (Stamm 1999). Research also suggests that chlamydial infection (Corey and Handsfield 2000). Improved control of chancroid in men may be associated with reduced fecundity among cou- and syphilis as well as actual increases in the sexual transmis- ples (Idahl and others 2004). In addition, chlamydial infection sion of HSV-2 in areas with advanced HIV epidemics, where can affect neonates: many delivered vaginally become infected, HIV-related immunosuppression causes more frequent and developing conjunctivitis or, less often, chlamydia pneumonia more severe HSV-2 disease, may also play a role. Estimates indi- (Donovan 2004). The role of C. trachomatis in preterm births cate that 10 to 30 percent of adults worldwide are infected with and in cervical cancer awaits further clarification through HSV-2 (Brugha and others 1997). Prevalence increases with age research (Samoff and others 2004; Wallin and others 2002). and is higher in women and high-risk populations. In the absence of control programs, the prevalence of HPV types are grouped into low-risk (nononcogenic) and Trichomonas vaginalis varies greatly across countries, ranging high-risk (oncogenic) types. Low-risk types, including types 6 from less than 1 percent among urban women to more than and 11, cause benign anogenital warts, whereas high-risk 20 percent in underserved populations in the same country types, including HPV 16, 18, 31, and 45, occasionally lead to (Brown and Brown 2000), and may increase with age. WHO genital and anal squamous cell cancers. The introduction of estimated the global prevalence of T. vaginalis at 174 million in nucleic acid amplification tests revealed that genital and anal 1999. The introduction of nucleic acid amplification tests high- HPV infection is common even among relatively sexually lighted the poor sensitivity of microscopy in the detection of inexperienced individuals (Giuliano and others 2002; Stone T. vaginalis. Even though most infected people are asympto- and others 2002). Investigators believe that most adults matic, T. vaginalis can cause vaginitis with vaginal discharge in become infected with HPV but that only a few develop warts women and urethritis in men. T. vaginalis has been associated or genital or anal cancer. Infection with a high-risk HPV type with preterm birth and may promote the sexual transmission is implicated in nearly all cases of invasive cervical cancer of HIV (Laga and others 1993). However, a randomized con- (Walboomers and others 1999) and with vaginal, vulvar, and trolled trial did not show that screening and treatment for anal cancers. T. vaginalis to prevent preterm birth were effective (Klebanoff In developed countries, hepatitis B virus is spread predomi- and others 2001). nantly by sexual and injecting drug-use transmission. Indeed, Like T. vaginalis, bacterial vaginosis and vulvovaginal can- the first three trials of hepatitis B vaccine successfully demon- didiasis cause vaginal symptoms in women, are extremely strated prevention of sexual transmission of hepatitis B virus in prevalent in developing countries, and in one or more studies men who have sex with men (Manhart and Holmes 2005). In have been associated with HIV acquisition or HIV genital shed- developing countries, hepatitis B is more often acquired peri- ding by women (Donovan 2004). Although often referred to as natally or during childhood, but a rise in seroincidence in ado- reproductive tract infections rather than STIs, they are man- lescence and young adulthood in some countries probably aged in conjunction with STIs, and bacterial vaginosis is asso- reflects sexual or injecting drug-use transmission. Hepatitis B ciated with some of the same risk factors as other STIs. virus causes acute hepatitis and in some people causes chronic hepatitis that can lead to cirrhosis and liver cancer. Human T cell lymphotropic virus type I (and perhaps in Viral STIs and Their Sequelae more cases type II) is, like hepatitis B virus, transmitted Both HSV-1 and HSV-2 infect the genital and anal areas, but perinatally and sexually. In some high-risk populations, for HSV-2 causes the most clinical recurrences in the genital tract. example, female sex workers in Latin America, human T cell Symptoms are mild in most of those infected and tend to go lymphotropic virus type I infection is substantially more com- unrecognized and undiagnosed (Corey 2000; Scoular 2002). mon than HIV infection. This infection causes a serious form Genital herpes establishes a lifelong infection that in some peo- of spastic paralysis or human T cell lymphotropic-associated ple is associated with significant morbidity. Complications of myelopathy, as well as T cell lymphoma or leukemia. Sexually Transmitted Infections | 315 Coinfection with Sexually Transmitted Pathogens status (Fenton, Johnson, and Nicoll 1997). In all societies, ado- The epidemiology and natural history of coinfection with lescents and young people are at greater risk for acquiring most more than one sexually transmitted pathogen may have impor- STIs. Women tend to have a higher prevalence and incidence of tant intervention and economic implications. Coverage by all STIs (except for men who have sex with men) and suffer clinical services, outreach, access, partner management, and more of the serious complications, such as pelvic inflammatory treatment may be different with coinfection than with inde- disease, ectopic pregnancy, infertility, and chronic abdominal pendent infections. Although coinfections with HIV and other pain. For many STIs, the probability of transmission from an STIs have received a great deal of attention in recent years, infected man to a susceptible woman is higher than from an researchers have not focused on overlaps among non-HIV STIs infected woman to a susceptible man. Social and behavioral in a similar systematic manner. A number of biological mech- patterns also increase women's vulnerability to STIs; for anisms may lead to coinfection with STIs: infection with one instance, many men have concurrent sex partnerships, which pathogen may increase the probability of acquiring or trans- increase their risk for transmitting infection to their female sex mitting another pathogen; infection with one pathogen may partners. In addition, many young women have sex with older increase or decrease the frequency, the severity, or both of male partners, who expose them to the higher STI prevalence symptoms associated with another sexually transmitted rates in older age groups. pathogen; and presence of one STI may affect the natural his- In most societies, minority racial ethnic groups have higher tory of another STI. High-risk behaviors and networks often STI rates than other groups. Both in the United Kingdom lead to coinfection. (Fenton, Johnson, and Nicoll 1997) and in the United States Empirical data on coinfection are limited. Most studies have (Laumann and Youm 1999), assortative sexual mixing and been conducted in developed countries and have focused on higher rates of sexual mixing with members of core groups co-occurrences of chlamydial and gonococcal infection. Earlier emerge as determinants of ethnicity differentials in STI rates. studies of coinfection assessed the proportion of gonorrhea The prevalence of concurrent partnerships is also higher among cases with concurrent chlamydial infection in a variety of clin- racial ethnic minorities (Kraut-Becher and Aral 2003). The ical settings. Reported levels of coinfection were 4 to 64 percent relative inadequacy of STI health services and of health among attendees at STI clinics, 46 percent among prenatal care­seeking behaviors among minority racial ethnic groups clinic attendees, and 4 to 25 percent at primary health care may also contribute to their higher prevalence of STIs (Aral and facilities (Creighton and others 2003). The proportion of those Wasserheit 1999). with chlamydia who also have gonorrhea has been assessed less Socioeconomic status differentials in STI prevalence and well, and estimates have ranged between 3 and 4 percent incidence are similar to ethnicity differentials. However, the (Creighton and others 2003). multicollinearity between the two factors makes delineating the independent contributions of either variable to differentials in STI prevalence and incidence difficult. Sexual Behavior and Sexual Health Care Unprotected sex with an infected partner is the most important Behavioral Risk Factors for Exposure to Infected Sex risk factor for acquiring an STI. This risk is influenced by the Partners. Most sexual behaviors of individuals are associated behaviors of the individual and the probability that the partner with exposure to sex partners infected with sexually trans- is infected, which is determined by the prevalence and distri- mitted pathogens and, consequently, with acquisition of STI. bution of infection in the population as well as the partner's These behavioral factors include number of sex partners behaviors. Current approaches to STI epidemiology recognize over the individual's lifetime, over the past year, and over a at least three distinct components of transmission dynamics at short term (Fenton and others 2001; Laumann and Youm the population level: likelihood of sexual exposure between 1999); frequency or number of sexual encounters (Garnett and infected and uninfected individuals, transmissibility of infec- Rottingen 2001); having sex with members of groups with tion upon exposure between an infected and an uninfected high STI prevalence, such as core groups and sex workers person, and duration of infection among those infected (Aral (Fenton and others 2001; Laumann and Youm 1999) or older and Holmes 1999; Over and Piot 1993). The first of these com- age groups (Service and Blower 1996); and position in a sexual ponents is entirely behavioral, and behavior plays an important network (I. A. Doherty and others 2005). Some sexual behav- role in the last two--for example, condom use, sexual practices, iors of individuals are associated with transmission of STIs, and health care­seeking behaviors. and for those infected the behaviors increase the probability that people will transmit their infections to susceptible sex Demographic and Social Risk Markers. The prevalence and partners. These behaviors include having concurrent incidence of STIs vary across societies and subpopulations partnerships (Koumans and others 2001; Kraut-Becher and defined by age, gender, race and ethnicity, and socioeconomic Aral 2003; Morris and Kretzchmar 1995) and having short 316 | Disease Control Priorities in Developing Countries | Sevgi O. Aral and Mead Over with others gaps between sex partners in serial monogamous partnerships (including opportunity costs) of treatment, perceived and (Kraut-Becher and Aral 2003). actual quality of care, and beliefs about the appropriate provider Sex partners' behaviors are also critical determinants of to consult. The proportion of those infected seeking care is exposure to infection. Investigators use many behavioral and highly variable, and delays in seeking treatment can be substan- epidemiological indicators to assess partners' risk of having tial. In many places, the proportion of people seeking timely infection, including existence and number of new sex partners; care from appropriately trained providers is limited (Hawkes presence of concurrent partnerships; gap between sex part- and Santhya 2002; Moses and others 2002; Rekart 2002). ners; partners' number of partners; and risk status of partners' Behaviors on the part of health care providers that ensure partners--for example, if they have sex with sex workers or timely and accurate diagnosis, appropriate treatment, and non- men who have sex with men (Aral 2002b). judgmental attitudes toward those infected would also help reduce the duration of infectiousness of STIs. However, estab- Behavioral Risk Factors Associated with STI Acquisition and lishing effective, accessible, affordable, and decentralized Transmission on Exposure to Infected Partners. Certain services is difficult (Over 2004; World Bank 2003). The major behaviors influence the likelihood of an infected person's trans- barriers Moses and others (2002) identified in Nairobi reflect mitting infection to a susceptible partner, including condom the situation in many developing countries. Those barriers use, sexual practices such as anal intercourse, vaginal douching, include inadequate basic training and inefficient deployment and use of drying agents in the vagina (Bailey, Plummer, and of health workers; attitudes of health workers toward margin- Moses 2001; Donovan 2000a, 2000b). The probability of trans- alized groups (for instance, female sex workers); high patient mission varies depending on the pathogen and is much higher loads at health centers; lack of supportive supervision; inade- for bacterial STIs, such as gonorrhea, syphilis, and chlamydia, quate referral systems; chronic shortages of supplies and drugs; than for other STIs, such as HIV infection. Thus, preventive and inadequate recording of health information. User fees can behaviors such as condom use may be more effective in pre- be a substantial additional barrier, though they may contribute venting the latter than the former (National Institute for to the sustainability of the treatment program and improve the Allergy and Infectious Diseases 2001). In addition, the proba- provider's incentives. bility of both acquisition and transmission is significantly affected by such nonbehavioral cofactors as circumcision status Behavioral Interactions. Both at the individual and the pop- (Aral and Holmes 1999). ulation levels, people's risk behaviors respond to changing Overall, oral sex and anal sex tend to be practiced less often circumstances. In many developing countries where HIV in the developing world than in the developed world (Vos incidence has been high, people have adopted compensatory 1994). Insertion of herbs to tighten or dry the vagina and behavior changes, such as delayed age of sexual debut, reduced other practices of vaginal clearing and wiping are widespread number of sex partners, and increased use of condoms (Brown and Brown 2000). Condom use is increasing in some (Shelton and others 2004; Stoneburner and Low-Beer 2004), countries--for example, India, Thailand, and Uganda-- especially with high-risk partners (Peterman and others 2000). especially during high-risk encounters. Some people now seek health care when they suspect they have been exposed to an STI. Behaviors Associated with the Duration of Infectiousness. At the same time, risk behaviors can overlap: people who The duration of infectiousness is an important component of initiate sexual activity early in life tend to have many partners, transmission dynamics. Because effective treatment curtails the and people who engage in risky sex tend to also use drugs and duration of curable STIs, the speed with which infected alcohol. A history of sexual abuse or of being an abuser is also individuals seek treatment and the speed and effectiveness with positively associated with high-risk sexual behaviors and drug which health care providers supply effective treatment together use (Aral 2004). The adoption of preventive behaviors raises determine duration. To the extent that suppressive therapies the possibility that people will compensate by changing other truncate the period of infectiousness of viral STIs, as they do behaviors in response; for example, many believe that the for HSV-2 and HIV infection, duration is also important in the widespread adoption of antiviral therapy or condom use may transmission dynamics of incurable viral STIs. lead to increases in the numbers of sex partners (Blower and Behaviors that can reduce the average duration of infec- others 2001; Blower and Farmer 2003; Over and others 2004). tiousness include timely and appropriate health care seeking, Although constructing mathematical models to explore the effective participation in risk assessment, and compliance with effects of such changes in behavior is helpful, empirical therapy and prevention recommendations on the part of those research in varied contexts is urgently needed to identify the infected and at risk (Aral and Wasserheit 1999). Health care variables that determine patterns of interaction among risky seeking depends on perceived seriousness and causality of and preventive behaviors. Two such variables may be individ- symptoms, availability and accessibility of health care, costs ual autonomy and awareness of the epidemiological context. Sexually Transmitted Infections | 317 Societal Determinants of STIs. Sexual networks and patterns Income and Inequality. A cross-country database (George of sexual partnership formation and dissolution constitute a Schmid, personal communication, September 15, 2004) major mechanism through which the political economy and enables us to analyze the association between national STI the sociolegal system influence the rate of spread of STIs in a prevalence rates and two important economic variables: gross population. Sexual networks that are highly critical to the rate national income per capita and the degree of income inequal- of spread of STIs include those involving sex work; exchange of ity as measured by the Gini coefficient.1 As table 17.1 shows, sex for drugs, gifts, or material needs; and anonymous sex. The these two variables explain 45 percent of the variation in STI frequency of sex in exchange for money or other goods appears prevalence in low-risk groups and 16 percent of STI prevalence to be highly sensitive to changes in the political economy and in high-risk groups. Figure 17.1 illustrates the relationship the sociolegal system. Internal conflicts, war, economic crises, between each of these two economic variables and STI and social collapse are accompanied by the establishment of prevalence. major sex markets or the expansion of existing ones. For exam- Poor countries' higher prevalence rates of STIs are unsur- ple, following the collapse of the former Soviet Union, the prising and could be explained by the fact that people in richer number and size of commercial sex and sex-drug networks countries are likely to seek and find care for STIs more quickly. expanded significantly (Aral and St. Lawrence 2002; Aral and More notable is that income inequality is such a strong predic- others 2005). The availability and use of condoms also influ- tor of STI prevalence even after controlling for gross national ence the rate of spread of STIs. income per capita. Furthermore, income inequality is a strong Many developing countries continuously face political con- predictor of STI prevalence among high-risk groups, where flict, war, economic deterioration, mass migration, and increas- income per capita performs less well. A possible explanation for ing inequality plus the effect of globalization. In addition, in this finding is that greater inequality creates more active mar- most developing societies, gender power relationships are kets for commercial and casual sex as higher-income men marked by great inequality. Those contextual factors lead to negotiate for the sexual services of lower-income sex workers sexual networks and sexual mixing patterns that are highly (Aral 2002b; Over 1998). conducive to the spread of STIs. Sexual partnerships are often not stable, and in the long-term absence of a spouse, both men and women (but especially men) have other partners. Table 17.1 Cross-Country Regressions of Average STI In addition, as economic needs rise, the number of women Prevalence on Per Capita Income and the Gini Coefficient who exchange sex for material needs increases. Wilson and of Inequality for Low- and High-Risk Groups, Selected others (1989) estimate that approximately 10 percent of the Countries female population in Bulawayo, Zimbabwe, had engaged in full- or part-time sex work at some time in their lives, whereas Average prevalence Average prevalence of STIs in of STIs in Aral and St. Lawrence's (2002) estimates for Saratov, Russian Category low-risk groups high-risk groups Federation, are closer to 25 percent. As the supply of sex work- ers increases, the demand for their services often increases in Gross national income 0.862** 0.194 purchasing power parity parallel. Economic need also affects sexual mixing patterns. In per capita, log, 19~2001 most developing countries, young girls commonly have "sugar Gini index, 1990s 7.731** 2.73* daddies"--that is, older, often married men who provide them Dummy for syphilis 0.093 0.051 with material goods in return for sex while also exposing them to chronic STIs typical of relatively older cohorts (Gregson and Dummy for chlamydia 1.992** 0.308 others 2002). Dummy for herpes 3.611** 1.507** Gender inequalities put women in a highly vulnerable posi- Constant 0.515 1.751 tion in many ways. For example, Decosas and Padian (2002) R 2 0.45 0.16 find that, among women attending family-planning and pri- Number of countries 204 147 mary health care clinics in Zimbabwe, 17 percent had at some Source: Authors' calculations from Schmid and others 2004. time received a gift in exchange for sex, 22 percent had been Note: * probability of less than 0.05 percent; ** probability of less than 0.01 percent. forced to have sex with a steady partner, 5 percent had been Positive numbers indicate a probability of 0.1 to 0.2 percent. Both regressions pool data across these STIs: syphilis, chlamydia, and herpes. The coefficients of the dummies show that the esti- forced to have sex with a nonsteady partner, 35 percent were mated prevalence rates are significantly higher for herpes than for the other diseases. Among certain their steady partner had other partners, 27 percent said low-risk groups, the prevalence of chlamydia is higher than that of syphilis. Standard errors (not shown) are estimated using the White correction for heteroskedasticity under the assumption their partners had STI symptoms, 24 percent said their partner that the observations for a single country are comparable to a cluster of data. Other explanatory was intoxicated during sexual intercourse more than half the variables that were unsuccessful in explaining a significant proportion of the variance included percentage of the population foreign born, percentage of the population that is Muslim, male-to- time, and only 10 percent had used a condom in the previous female literacy gap, urban male-to-female population ratio, and military personnel per 1,000 three months. urban population. 318 | Disease Control Priorities in Developing Countries | Sevgi O. Aral and Mead Over with others a. Gross national income per capita BURDEN OF STIs AND BENEFITS OF CONTROL Average national prevalence of STIs among low-risk groups On the basis of an independent analysis of cross-country data (deviations from mean) on STI prevalence, we believe that WHO may have underesti- 5 mated the burden of STIs relative to that of HIV and other dis- eases (www.fic.nih.gov/dcpp/gbd.html). Adjusting the WHO estimates on the basis of our calculations increases the estimate 0 of years of life lost burden by about 18.1 percent and the over- all estimate of disability-adjusted life years (DALYs) lost by 5 about 8.2 percent. Coefficient 0.86, t 5.14 10 DALYs Gained from Effectively Preventing or Treating STIs 2 1 0 1 2 Gross national income per capita In the first edition of this volume, chapter 20 presented esti- deviations from mean adjusted for Gini coefficient mates of the so-called static and dynamic burdens of prevent- ing or curing a single case of an STI and of HIV (Over and Piot b. Income inequality 1993). We reproduce those estimates in table 17.2 for four of Average national prevalence of STIs among low-risk groups the STIs.2 The static benefit column estimates the average (deviations from mean) number of DALYs saved for a single person by curing or pre- 5 venting his or her own case of each disease. These estimates are based on specific assumptions regarding the distribution of incidence, case-fatality rates, and severity across age ranges. 0 Although updating these estimates to 2004 and varying them by region would ideally be possible, we have not found any 5 more recent data. If the person who is cured of an STI ceases to be sexually active, the static benefit would be the only benefit of Coefficient 7.73, t 4.38 curing or preventing his or her case; however, most individuals 10 .2 .1 0 .1 .2 .3 who have contracted an STI remain sexually active and are Income inequality measured by Gini coefficient therefore likely to communicate that STI to others. An STI deviations from mean adjusted for gross national income per capita prevented or cured in a sexually active person will prevent Source: Authors' calculations based on personal communication with George Schmid additional cases in that person's sex partners, in the sex part- (April 5, 2004). ners' partners, and so on. Thus, the dynamic benefit columns Figure 17.1 Association between Per Capita Income and Income indicate the magnitude of those additional benefits. Inequality with STI Prevalence for Low-Risk Groups, Selected The key finding is that preventing or curing a case of any Countries of the STIs in a core group member generates approximately Table 17.2 Discounted DALYs Saved Per Case Prevented or Cured Dynamic benefits Total benefits Static Member of Member of Member of Member of STI benefits noncore group core group noncore group core group Chancroid 0.2 0.2 1.6 0.4 1.8 Chlamydia 1.1 4.4 43.0 5.5 44.1 Gonorrhea 0.9 3.6 36.4 4.5 37.3 Syphilis 3.8 16.0 157.0 19.8 160.8 For comparison HIV without ulcers 19.5 35.1 340.1 54.6 359.6 HIV with ulcers 19.5 39.2 410.7 58.7 430.2 Source: Over and Piot 1993, table 20-16, appendix 20B. Sexually Transmitted Infections | 319 10 times the dynamic benefits of the same intervention in a per- country regressions are presented in table 17.3.3 Columns (1) son in a noncore group. This result is driven by the assumption and (2) of table 17.3 present the results of regressions estimated that a member of a noncore group has a new sexual contact on the subsets of countries for which data are available on all every 50 days, or about seven new contacts per year, whereas a eight explanatory variables and on the dependent variable (2002 member of a core group has 10 times as many contacts. Within urban HIV prevalence). Their specifications differ only by the this model, the results are proportional to the frequency of replacement of the prevalence of syphilis as an explanatory vari- partner change, so that the dynamic benefits of curing or pre- able in column (1) with the prevalence of gonorrhea in column venting a case in a sex worker who has two partners a day would (2). Columns (3) and (4) repeat the same two regressions by be approximately 10 times as great as for a member of the core replacing missing values of the two prevalence rates with esti- group in table 17.2. The implication is that preventing or cur- mates that are based on a regression of these rates on the other ing a case of syphilis in a sex worker can result in up to 1,600 variables in the regression. This procedure expands the samples DALYs of benefit, a health effect that is likely to be competitive dramatically from 56 and 38 to 181 and 180, respectively.4 with any discussed in the other chapters in this volume. In interpreting these regressions, note first that all four spec- ifications explain more than half of the variance in 2002 urban HIV prevalence, a remarkably good fit for cross-sectional Impact of STIs on HIV regressions. In all these specifications, the lagged value of an The preceding discussion does not address the possibility that STI prevalence is a statistically significant predictor of HIV STI infections increase HIV transmission. On this point the prevalence approximately seven years later. The coefficient for evidence is mixed, with a study in Mwanza, Tanzania gonorrhea is larger than the coefficient for syphilis and is more (Grosskurth and others 1995), demonstrating a statistically sig- statistically significant in the augmented sample, though less so nificant 40 percent reduction in HIV incidence attributable to in the basic sample. an STI intervention, while two studies (Kamali and others After the age of the epidemic is controlled for, several other 2003; Wawer and others 1999) in Uganda failed to show any variables contribute to explaining the variation in HIV preva- such effect. Recent reanalyses (Orroth 2003) of the data from lence. These variables include national income per capita these studies suggest that the effect of an STI intervention on (richer countries have lower infection rates), the percentage of an HIV epidemic will vary depending on the sexual activity and the population that is Muslim (a higher percentage is associ- resulting prevalence of STIs among those being treated. ated with lower infection rates), and the ratio of males to None of the randomized controlled trials of the effect of STI females in the sexually active age range (higher ratios are asso- treatment on HIV prevention exclusively targeted the most sex- ciated with higher infection rates). ually active people in the community. In Mwanza, Tanzania, The major difference between the regressions using the aug- and Masaka, Uganda, treatment was provided to those who mented sample and those using the basic sample is in the sta- sought it at health care clinics. In Rakai, Uganda, treatment was tistical significance of the estimated coefficient of income given to all adults in all households, regardless of whether the inequality as measured by the Gini coefficient. When the sam- individual complained of STI symptoms. Data on the preva- ple is expanded to take advantage of the available data, the coef- lence of HIV infection in the three communities suggest that ficient stabilizes at about 5.3 and is statistically significant at the the HIV and STI epidemics were both at an earlier stage in 0.01 probability value, suggesting that an increased degree of Mwanza and were therefore more concentrated among those income inequality is associated with increased HIV infection more sexually active. Thus, the people who became sympto- even after controlling for STI prevalence. This result lends sup- matic and sought treatment were among the most sexually port to the idea that income inequality is just as important as active people in Mwanza, and treating them would therefore poverty in setting the stage for HIV transmission. have had a greater effect on HIV incidence than would treating As with the results of any ecological or cross-sectional analy- an average person in the two Ugandan sites. Conversely, in the sis, questions of attribution and interpretation arise. Is the sta- more generalized epidemics in Uganda, a larger proportion of tistically significant coefficient of syphilis or gonorrhea captur- new infections occurred within stable HIV-1 serodiscordant ing a biological effect of an STI on increasing the transmission couples. probability during sexual intercourse? Or is the coefficient An alternative, less rigorous way to test for the effect of STI instead simply reflecting the fact that greater sexual activity prevalence on HIV infection is to study the cross-sectional cor- spreads all STIs, including gonorrhea, syphilis, and HIV? Is the relation in ecological data. In a replication of an earlier study coefficient of the percentage of the population that is Muslim (Over 1998), we have attempted to explain urban HIV preva- capturing differential sexual activity or the prevalence of male lence in a cross-country sample by the prevalence of syphilis and circumcision, which is increasingly recognized as biologically gonorrhea seven years earlier after controlling for six other protective? A biological interpretation of both the STI and the potentially confounding variables. The results of these cross- Muslim coefficients is suggested by the fact that the variable 320 | Disease Control Priorities in Developing Countries | Sevgi O. Aral and Mead Over with others Table 17.3 Multiple Regression of circa 2002 Urban HIV Prevalence on circa 1995 STI Prevalence and Other Socioeconomic Variables (1) (2) (3) (4) Augmented Augmented Basic sample Basic sample sample with sample with Category with syphilis with gonorrhea syphilis gonorrhea Age of epidemic (urban low) in 2002a 0.177 0.131 0.12 0.08 (4.07)*** (2.00)* (2.90)*** (1.70)* Per capita national income, median 1996­2001 1.344 1.177 0.929 0.949 (2.82)*** (2.30)** (3.63)*** (4.45)*** Income equality, Gini index, 1990s 0.125 3.186 5.349 5.258 (0.04) (0.83) (2.77)*** (2.79)*** Percentage of the population that is Muslim, 1999 0.026 0.031 0.015 0.02 (4.05)*** (3.50)*** (3.45)*** (4.10)*** Urban male-to-female ratio for those age 20­39, 1999 6.178 6.199 3.27 3.038 (2.03)** (1.40) (1.29) (1.19) Urban high-risk population dummy 1.17 0.564 0.894 0.37 (1 if yes, 0 is no) (2.83)*** (0.85) (2.64)*** (0.71) Logit syphilis in low-risk group, 1995b 0.31 0.233 (2.46)** (1.78)* Logit gonorrhea in low-risk group, 1995b 0.502 0.479 (1.92)* (1.86)* Constant 0.297 0.791 2.117 0.054 (0.05) (0.11) (0.63) (0.02) Number of observations 56 38 181 180 Number of countries 40 29 101 100 R 2 0.57 0.56 0.5 0.51 Source: Authors' calculations. Syphilis and gonorrhea prevalence in 1995 (George Schmid, personal communication, April 5, 2004). HIV prevalence circa 2002 is from the urban low-risk tables in the U.S. Bureau of Census database on HIV prevalence. Other variables are from World Bank data. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. Note: The figures in parentheses are robust t-statistics. a. Age of epidemic is defined as the number of years since the first case of HIV/AIDS was reported. b. The logarithm of the ratio of the prevalence to 1 minus the prevalence of the given STI for the low risk population in 1995. urban male-to-female ratio is probably already capturing much the Mwanza trial found that a reduction in the prevalence of of the variation in the most risky sexual behavior: the practice male urethritis of 0.6 percent was associated with a decrease of of prostitution. 0.7 percent in the incidence of HIV (Grosskurth and others Increasing the availability of treatment for STIs and for HIV 1995). Thus, the present study suggests an effect about one- infection reduces the prevalence of the former and increases fourth as strong as that of the Mwanza study. the prevalence of the latter. Thus, this statistical relationship between STI prevalence and HIV prevalence, even if once valid, will no longer obtain. Under current conditions, estimating the EFFECTIVENESS OF THE PRINCIPAL effect of a change in the prevalence rate of an STI on the inci- INTERVENTIONS dence rate of HIV would be more relevant. If we assume that in 2002 the HIV epidemic was approach- Unlike HIV interventions, STI interventions benefit from a ing equilibrium in many urban settings and that prior to anti- large body of rigorous evaluations. STI interventions that have retroviral treatment the median duration of the illness was been rigorously evaluated for effectiveness can be organized about 10 years, the prevalence of HIV infection is approxi- by intervention level (that is, individual, group, or commu- mately equal to 10 times the incidence rate. Thus, a 10 percent- nity); by the outcomes measured; and by the intervention age point increase in the prevalence of syphilis or gonorrhea is modality used (for example, behavior change, vaccination, estimated to increase the incidence of HIV by 0.27 percentage topical microbicide use, screening, or treatment). Prevention points for syphilis and 0.57 for gonorrhea. For comparisons, outcomes may measure the prevention of acquisition, of Sexually Transmitted Infections | 321 transmission, and of complications of STIs (Manhart and or recurrent gonococcal or chlamydial infection in the index Holmes 2005). This section reviews the interventions for which patient (Golden and others 2005). the strongest evidence exists. Preventing Complications. Risk-based screening for C. tra- chomatis infection resulted in a 56 percent reduction in the Individual-Level Interventions subsequent risk of incident pelvic inflammatory disease A large number of STI interventions that have been rigorously (Scholes and others 1996). Several trials have shown that evaluated are individual-level interventions. antiviral suppression decreases clinical and virological recur- rences of genital herpes (Corey and Handsfield 2000). Preventing Acquisition. The following have been the main means of preventing STI acquisition: Group-Level Interventions · Behavior change. Counseling on risk reduction was the most Studies of behavior-change methods in small-group settings to frequently used behavior-change approach. Most studies reduce the acquisition of STIs had mixed outcomes. Behavior- showed a reduction in risk behaviors as a result of counsel- change approaches resulted in significant reduction in incident ing, and some showed decreases in STI outcomes (Kamb STIs; antimicrobial prophylaxis and provision of female con- and others 1998). doms did not. · Antimicrobial prophylaxis. Two studies (Harrison and others 1979; Kaul and others 2004) showed reductions in the inci- dence of gonococcal, chlamydial, or trichomonal infections Community-Level Interventions following antimicrobial prophylaxis. Four community-level randomized trials have sought to reduce · Vaccines and passive immunization. A yeast-derived HPV the prevalence and transmission of STIs by shortening the type 16 vaccine was 100 percent efficacious in preventing duration of infectiousness within the general population persistent HPV-16 infection in young college women (Manhart and Holmes 2005). (Koutsky and others 2002), and a bivalent HPV type 16 and The "Mema Kwa Vigara" study in Mwanza, Tanzania, ran- type 18 vaccine was also highly efficacious in preventing domized 20 communities to intervention and control commu- those infections. An HSV-2 glycoprotein D-adjuvant vac- nities. The intervention consisted of school-based sexual and cine among those with no serological evidence of prior reproductive health education, enhanced reproductive health HSV-1 infection partially protected women, but not men, services for youths, condom distribution, and community from experiencing genital herpes disease, with 73 percent activities. Knowledge and reported behaviors improved; how- efficacy for such women in one trial and 74 percent in ever, no differences were apparent between the intervention another (Stanberry and others 2002). and control communities in relation to HIV or HSV-2 seroin- · Microbicides. To date, studies have not identified any effica- cidence, incidence of other STIs, or pregnancy outcomes cious topical microbicides. (Hayes and others 2003). · Male circumcision. Even though cross-sectional evidence A second study in Mwanza, Tanzania, randomized commu- suggesting that male circumcision decreases the risk of nities to intervention and control conditions. The intervention acquiring chancroid and HIV is strong, outcome data are consisted of syndromic treatment of STIs. The results showed not yet available from ongoing randomized trials in Kenya, a 40 percent reduction in HIV incidence and reductions in South Africa, and Uganda. symptomatic urethritis in men and prevalence of syphilis seroreactivity; the prevalence of gonorrheal or chlamydial Preventing Transmission. All individual-level interventions infection in prenatal women did not change (Grosskurth and aimed at preventing transmission have involved curative or others 1995; Mayaud and others 1997). suppressive therapy. Giving tinidazole to male partners of In a community randomized trial in Masaka, Uganda, one females treated for vaginal trichomoniasis infections signifi- community received information, education, and communica- cantly reduced recurrences in the females; administering vala- tion; a second community received information, education, and cyclovir to positive members of HSV-2 serodiscordant couples communication plus syndromic management of STIs; and the reduced the incidence of symptomatic genital herpes and control received community development assistance. The HSV-2 seroconversion in the uninfected partners; patient- results showed no differences in HIV-1 incidence. The incidence delivered therapy to partners of women with chlamydial infec- of HSV-2 seroconversion declined in the community receiving tion demonstrated a nonsignificant trend toward reduced risk information, education, and communication only; the inci- of reinfection with C. trachomatis; and expedited partner ther- dence of syphilis and of gonorrhea decreased in the community apy (usually patient delivered) significantly reduced persistent receiving information, education, and communication plus 322 | Disease Control Priorities in Developing Countries | Sevgi O. Aral and Mead Over with others STI syndromic management; and condom use increased in all Organization of STI Control Activities in Poor Countries three communities (Kamali and others 2003). In poor countries, patients can typically obtain treatment for an In Rakai, Uganda, a community randomized trial evaluated STI in a public sector health care facility. Many countries have the efficacy of repeated mass treatment of STIs. Relative to con- publicly funded, stand-alone STI clinics, but the typical pattern trol communities, in intervention communities the prevalence is for health care personnel to provide care for STIs as part of of T. vaginalis in women was reduced significantly, but no sig- their regular practice in general outpatient clinics. Despite the nificant reduction was apparent in prevalence of gonorrhea, availability of such publicly funded care, or perhaps because of chlamydial infection, new syphilis seroreactivity, and bacterial concerns about anonymity, many STI patients in poor coun- vaginosis; in HIV incidence; or in history of urethral or vaginal tries avoid public facilities in favor of traditional healers and discharge or genital ulcer disease (Wawer and others 1999). A private pharmacies. A recent study of the cost-effectiveness of subanalysis of pregnant participants showed a reduction in the delivering STI treatment through trained pharmacists in prevalence of several STIs in women tested near delivery and Peruvian cities included reports of the popularity of self- in potentially STI-related pregnancy, puerperal, and neonatal treatment in Brazil, Cameroon, Ghana, Nepal, South Africa, morbidity (Gray and others 2001). Thailand, Vietnam, and Zambia (Adams and others 2003). After reviewing this literature, Adams and others selected the point estimate of 0.4 as their best guess for the proportion of Conclusions on Interventions STI patients seeking treatment from a pharmacy in Lima. The review of STI intervention research suggests several, per- haps counterintuitive, insights: Determinants of the Costs of Interventions We adopt a government perspective in analyzing the costs and · First, most evidence is on individual-level interventions cost-effectiveness of interventions. Thus, we define the costs of aimed at reducing STI acquisition, even though individual- an activity as the total budgetary expenditure attributable to level interventions may be costly and difficult to sustain. that activity--that is, the total budget for buildings, equipment, · Second, behavior change is the most commonly evaluated personnel, and supplies, with adjustments made when build- modality, followed by treatment. ings are used for multiple purposes. · Third, theory-based behavioral interventions failed to show We define the unit costs of an activity as the total budgetary an effect as often as behavioral interventions not based on expenditure during a stated time period divided by the number theory. of units of output during that same period. Because the same · Fourth, behavioral interventions delivered in small group activity can have several outputs, this definition necessarily settings were as effective as those delivered to individuals entails some ambiguity. For example, an intermediate output of (Manhart and Holmes 2005). the delivery of STI treatment services is the patient treated, · Fifth, the effect of a particular behavior change on STI risk whereas a more final output is the patient cured. An even more depended on the type of STI; however, the number of part- complete measure of output would include the secondary ners may be more predictive of risk for highly infectious infections averted as a result of the cure. STIs than for HIV, and unprotected sex acts may be more One of the reasons that many economists prefer cost-benefit predictive of risk for HIV than for highly infectious STIs analysis to cost-effectiveness analysis is that the former attaches (Semaan and others 2002). Thus, behavioral interven- a dollar value to each of the outputs of an activity and then tions may have different effects on STIs of differing aggregates across the outputs to construct a summary measure infectiousness. of the total benefit of the activity. However, this simple result · Finally, the number of intervention trials that demonstrate hides many arbitrary assumptions that are required to value declines in risk behaviors combined with either no effect on the separate outputs. One of the most arbitrary of these STIs or increases in STIs is increasing. This observation calls assumptions is the assignment of a dollar value to a healthy life into question the use of behavioral outcome measures as year. So instead we present costs and cost-effectiveness denom- indicators of biomedical outcomes (Aral and Peterman inated in the outputs for which we have data. We then go as 2002). far as we can toward aggregation by adopting the conventions of the healthy life year and the disability-adjusted healthy life year. INTERVENTION COSTS AND Kumaranayake and others' (2004) background study for this COST-EFFECTIVENESS chapter reviews the literature on the unit costs of STI treat- ment. They identify 35 studies on this topic that provide a total Widespread implementation of effective interventions depends of 77 unit cost estimates. These are grouped in table 17.4 by the on cost and cost-effectiveness considerations. disease or syndrome being treated and by the output that was Sexually Transmitted Infections | 323 Table 17.4 Average Estimated Costs per Unit of Output, by · population composition and concentration Disease or Syndrome and Type of Output · resource combinations and input prices (2001 US$) · incentives to providers for high quality and quantity of serv- ice delivery Disease or syndrome Treatment Cure Average · willingness to pay for treatment as a function of price, Syphilis 36.04 n.a. 36.04 income, and distance (5.91) n.a. (5.91) · stigmatization Urethral discharge 14.29 89.07 29.25 · disutility of condom use. (20.68) (0) (37.94) Genital ulcer 23.16 100.60 48.97 (21.73) (83.74) (59.56) SCALING UP CONTROL STRATEGIES Venereal disease 25.47 82.65 31.83 (18.56) (111.55) (37.12) Throughout the history of STI control, tension has been appar- ent between those who support prioritizing resources for the Pelvic inflammatory disease 7.12 n.a. 7.12 (3.09) n.a. (3.09) small proportion of people with the most sexual contacts and Vaginal discharge 48.23 102.92 81.04 those who advocate spreading prevention, screening, and treat- (0) (89.63) (70.10) ment resources more thinly over the entire population. Total 24.05 96.10 39.49 Opponents of prioritization argue that most of the people who (19.04) (73.44) (47.23) practice the riskiest behavior are hard to find and that attempts to find them would expose those individuals to stigmatization Source: Authors' calculations based on a literature review done as a background study for this chapter by Kumaranayake and others 2004. and repressive measures. n.a. not applicable. A group of researchers at the University of North Carolina Note: The figures in parentheses are standard deviations. The 42 observations in the table are individual unit cost estimates distributed across separate studies. has developed and applied a novel approach to STI interven- tions that has demonstrated in several countries that finding costed. Of the 46 estimates of the unit cost of treatment, only the people who practice the riskiest sexual behavior without 33 could be interpreted as, or converted to, 2001 U.S. dollars, targeting them as individuals is possible (Weir and others and the same applied to 9 of the 10 estimates of the unit cost of 2003). As an example of this approach, consider its application a cure. Table 17.4 summarizes the results of these studies by the to Madagascar, a country where risky sexual behavior had pre- disease or syndrome that occasioned the treatment. viously been thought to be too common to be identified or to The most notable thing about the summary statistics is their be distinguished from less risky behavior. variability: the cost per unit of output can vary by a factor of In May 2003, the Malagasy Steering Committee, which con- 100 or more. Even though the mean dollar cost per cure could sisted of representatives of the Ministry of Health, local gov- plausibly be almost four times that of treatment alone, the ernment officials, and other knowledgeable experts, selected standard errors are so large as to make the difference statisti- five towns judged to be at high risk for STIs for a pilot study. cally insignificant. The Priorities for Local AIDS Control Effort (PLACE) method The same point--that is, that unit costs vary enormously applied in these towns was, first, to interview adults at random from one site to another--is made by the preliminary results of on the streets of the city to find out where people go to meet a study by Dandona and others (2005). The cost per case and socialize and, second, to visit and collect data on these loca- treated in the study varies by a factor of 10 across the 14 sites. tions and the people who frequent them. Furthermore, the two sites that treat the fewest and the most In each of the five cities, the informants tended to agree on cases per year also display the highest costs, a finding that sug- the most frequented sites. They identified between 70 and gests the existence of both economies and diseconomies of 267 unique socialization sites of various types, ranging from scale. bars and restaurants to beaches and brothels. Interviews with The variables that determine the costs--and therefore the people frequenting the sites revealed them to be much more cost-effectiveness--of STI treatment include the following: sexually active than the average Malagasy adult. According to the 1997 demographic and health survey, 13 percent of women · delivery by the public or private sector outside the capital city, Antananarivo, had two or more sexual · economies of scale partners in the previous year. In contrast, the percentage of · economies of scope women at the socialization sites who had had more than two · prevalence and incidence partners ranged from 46 to 68 percent. According to the · epidemic phase demographic and health survey, only about 3 percent of · transmission efficiency women outside Antananarivo had four or more partners in the · health system characteristics previous year, but the percentage of women at the study sites 324 | Disease Control Priorities in Developing Countries | Sevgi O. Aral and Mead Over with others having this many partners was 10 times larger. The men inter- ative spillover effects (externalities). The right combination viewed at these sites were even more sexually active than the of patient and provider incentives needs to be found that women. will maximize the beneficial spillovers while minimizing the The pilot study also investigated whether information on or harmful ones. products for prevention of STIs, HIV, or both were available at · The implementation of studies in support of global elimi- the socialization sites. In the five towns, the proportion of sites nation programs. where condoms were available on the day of the visit varied from 27 to 54 percent. These percentages are not negligible Because of the clandestine nature of most sexual behavior, and are undoubtedly much higher than they were 10 years ago, STIs are probably massively underreported, which in turn leads and the availability of condoms at so many of these sites is a to an underestimation of their importance. New survey and tribute to the success of the condom social-marketing cam- measurement tools have been developed. They now need to be paign. However, the statistics also indicate substantial room for applied to populations in poor countries to improve these improvement, for example, by distributing condoms in 100 per- estimates. cent of identified socialization sites. The feasibility of such a In addition, randomized controlled trials need to be con- program is enhanced by information from the PLACE study ducted in different settings to test the hypothesis that treating that more than 80 percent of the owners or managers of these or preventing STIs in high-risk individuals has beneficial sites expressed their willingness to host STI and HIV preven- spillover effects by preventing infections among low-risk indi- tion programs, and more than half were willing to sell viduals. An improved understanding of the determinants of condoms. high-risk sexual behavior and the role that such behavior can Researchers have carried out similar PLACE studies in sometimes play in helping women to escape from poverty and Burkina Faso, Ghana, South Africa, and elsewhere. Un- helping men to cope with it is also needed, as is a better under- fortunately, in none of these countries has this extensive risk standing of the full range of benefits of effective STI interven- mapping been followed by the implementation of prevention tions for high-risk individuals and their dependents. programs at all the identified locations. Until such programs As concerns disease modeling and surveillance, further are implemented and evaluated, no African country will be able improvements are needed in understanding the implications to claim that it has scaled up the most effective type of STI pre- for interventions of different kinds of local, regional, and inter- vention to population levels. national sex networks. Tools RESEARCH AND DEVELOPMENT AGENDA Because of the difficulty of persuading patients to adhere to a Priorities for global STI research include the following: course of medication for the prescribed period, single-dose therapy would be valuable. Rapid, point-of-care diagnostics · The development and evaluation of therapeutic (drug treat- are also a high priority, so that drugs can be targeted at ment or vaccines), behavioral, and structural interventions pathogens more accurately. New approaches for treating to prevent or reduce STIs and their sequelae. Given the chronic STIs should be incorporated into prevention strategies, spread of drug-resistant strains of gonorrhea and other and blister packs of antibiotics that can be sold over the counter STIs, new pharmaceutical products and new combination for syndromic management of STIs should be available. therapies are needed to prevent and treat STIs. Vaccines would be particularly valuable, both for preventing · The development and evaluation of mechanisms to accu- and for potentially treating chronic viral STIs and chlamydial rately quantify the disease burden in order to prioritize infection, which is often asymptomatic but is responsible for activities. considerable morbidity. · The development and evaluation of inexpensive and practi- Diagnostics tests that could be used at home or at social cal rapid diagnostic tests to permit early detection and treat- meeting spots may help people decide whether to engage in ment of STIs. risky sex. Packages of diagnostic tests that change color when · The conducting of studies to evaluate effective prevention the contents expire would also be useful. modalities for persons at highest risk for STIs. · The undertaking of health services research to gain an understanding of practical and cost-effective STI prevention Intervention Methods strategies or systems that ideally can be integrated into exist- Syndromic management algorithms have now existed for more ing public health infrastructure. When an individual is than a decade. However, treatment algorithms are sensitive to treated for an STI, this treatment has both positive and neg- changes in the relative prices of pharmaceuticals and diagnostic Sexually Transmitted Infections | 325 reagents, in the prevalence of the various STIs, in pathogens' As we learn more about the complexities of delivering STI resistance, and so on. Thus, every country needs some ability to treatment services and take into account the diversity of risk respond to local changes by developing or modifying algo- behavior, the ease with which STI interventions can be ascribed rithms as needed. a simple cost-effectiveness ratio has declined. If no easy way to As concerns intervention packaging, many policy makers summarize experience to date with a simple cost-effectiveness continue to believe that the most sexually active people are ratio is available, how should we analyze economic investments hard to find. This belief hampers efforts to target these people in STI treatment? We believe that the way forward is a better with STI prevention programs. The PLACE methods devel- understanding of why STI treatment and other health services oped at the University of North Carolina offer an opportunity vary so much in terms of their efficiency and effectiveness from to correct that impression and should be packaged with other one setting to another. By studying the determinants of this urban public health functions. Packaging sex education into variation, we should gain an improved understanding of the school curricula is a challenge in most of the world. As enroll- full costs of high-quality STI service delivery and its place in the ment rates for poor children, especially girls, rise, the presence health sector investment picture. of a strong, culturally appropriate, sex education curriculum will lay the foundation for strong STI prevention and treatment campaigns. ACKNOWLEDGMENTS As this chapter has argued, the determinants of the unit costs of STI treatment and prevention are largely unknown. We The authors would like to acknowledge Becca Feldman, recommend health services and operations research to study Patricia Jackson, Dilip Parajuli, and Melanie Ross for their out- the determinants of the unit costs of STI prevention and treat- standing support in the preparation of this chapter. ment services. The purpose of this research would be to learn not only how to deliver care in the most cost-effective ways, but also how to build systems that achieve that technological fron- NOTES tier in a high percentage of public and private facilities and pass 1. The Gini coefficient is a measure of inequality that here we apply to those cost savings on to the government and to patients. Given income. If income is distributed equally in the population, the coefficient the beneficial spillover effects from effective STI prevention is equal to 0, and if a few individuals hold almost all the wealth, the coef- and treatment among those who are most sexually active, ficient is close to 1. research is needed to learn how the PLACE approach to target- 2. Given that WHO has expended enormous efforts to estimate dis- counted years of life lost, disability years lost, and DALYs lost as a result of ing can be implemented most cost-effectively in different cul- STIs, the simplest and most direct approach for computing the DALY ben- tural contexts. efits of preventing or curing an STI in a single patient would be to use Finally, improved understanding of the best way to design WHO's years of life lost, disability years lost, or DALY per case assump- tions. Alternatively, one could simply divide WHO's aggregate values of an STI treatment system, including the rewards and penalties these indicators by the incidence rate of each disease in each region to that best motivate providers (to be polite, discrete, prompt, effi- obtain the estimated burden per incident case. Unfortunately, neither the cient, and accurate in following best practice, evidence-based case-specific burden numbers nor the incidence rates that correspond to the DALY aggregates are available from WHO. treatment protocols) and patients (to seek and then to adhere 3. In contemporaneous data, STIs can either affect or be affected by to treatment) is a priority. Improved data on the costs of each HIV prevalence. To focus on the effects of an STI on HIV infection, we lag STI intervention at the pilot stage and after scaling up to the infection by an STI by seven years. Though partially correcting for simul- national level are also necessary. taneity bias, this strategy does not allow us to identify whether lagged STI prevalence is directly affecting HIV infection or only serving as a proxy for the risky sexual behavior that drives both epidemics. 4. The samples include two measures of HIV prevalence (low- and high-risk groups) for some countries. These measures enable us to expand CONCLUSIONS the sample used in the column (1) regression from 40 countries to 56 sep- arate observations. Equations are estimated with Stata's cluster option to Regarding the cost-effectiveness of STI control, the position correct the standard errors of the coefficients for the correlation between this chapter takes is "it depends." The health benefit in terms of the errors on separate observations from the same country. The variable urban high-risk dummy is used to shift the intercept coefficient for the numbers of disability-adjusted, discounted, healthy life years high-risk sample in comparison with the low-risk one. saved by curing or preventing a case of syphilis varies from 3 years in a person who has ceased all sexual activity to as many as 161 years in a sex worker with two partners a day. The cost REFERENCES of treating that prostitute for syphilis varies from US$5 to US$100. Thus, the cost per DALY of syphilis treatment can Adams, E. J., P. J. Garcia, G. P. Garnett, W. J. Edmunds, and K. K. Holmes. 2003. "The Cost-Effectiveness of Syndromic Management in range from 100/3 or US$33 per DALY to 5/161 or less than Pharmacies in Lima, Peru." Sexually Transmitted Diseases 30 (5): US$0.05 per DALY. 379­87. 326 | Disease Control Priorities in Developing Countries | Sevgi O. Aral and Mead Over with others Adimora, A. A., V. J. Schoenbach, D. M. Bonas, F. E. Martinson, K. H. Brown, J. E., and R. C. Brown. 2000. "Traditional Intravaginal Practices Donaldson, and T. R. Stancil. 2002. "Concurrent Sexual and the Heterosexual Transmission of Disease: A Review." Sexually Partnerships among Women in the United States." Epidemiology 13 (3): Transmitted Diseases 27 (4): 183­87. 320­27. Brugha, R., K. Keersmaekers, A. Renton, and A. Meheus. 1997. "Genital Agrawal, D., B. Gillespie, and B. Foxman. 2001."Sexual Behavior across the Herpes Infection: A Review." International Journal of Epidemiology Lifespan: Results from a Random-Digit Dialing Survey of Women 26 (4): 698­709. Aged 60­94." International Journal of STD and AIDS 12 (Suppl. 2): 186. CDC (U.S. Centers for Disease Control and Prevention). 2004. "Increases Ainsworth, M., and M. Over. 1997. Confronting AIDS: Public Priorities in a in Fluoroquinolone-Resistant Neisseria gonorrhoeae among Men Who Global Epidemic. New York: Oxford University Press. Have Sex with Men: United States 2003 and Revised Re- Aral, S. O. 2000. "Behavioral Aspects of Sexually Transmitted Diseases: commendations for Gonorrhea Treatment 2004." Morbidity and Core Groups and Bridge Populations." Sexually Transmitted Diseases Mortality Weekly Review 53 (16): 335­38. 27 (6): 327­28. Chandeying, V., S. Skov, P. Duramad, B. Makepeace, M. Ward, and ------. 2002a. "Determinants of STD Epidemics: Implications for Phase P. Khunigij. 2000. "The Prevalence of Urethral Infections amongst Appropriate Intervention Strategies." Sexually Transmitted Infections 78 Asymptomatic Young Men in Hat Yai, Southern Thailand." (Suppl. 1): i3­13. International Journal of STD and AIDS 11 (6): 402­5. ------. 2002b. "Understanding Racial-Ethnic and Societal Differentials in Corey, L. 2000."Herpes Simplex Type 2 Infection in the Developing World: STI: Do We Need to Move beyond Behavioral Epidemiology?" Sexually Is It Time to Address this Disease?" Sexually Transmitted Diseases Transmitted Infections 78 (1): 2­3. 27 (1): 30­31. ------. 2004. "Editorial Response: Mental Health--A Powerful Predictor Corey, L., and H. H. Handsfield. 2000. "Genital Herpes and Public Health: of Sexual Health?" Sexually Transmitted Diseases 31 (1): 13­14. Addressing a Global Problem." Journal of the American Medical Aral, S. O., and J. F. Blanchard. 2002. "Phase Specific Approaches to the Association 283 (6): 791­94. Epidemiology and Prevention of Sexually Transmitted Diseases." Corey, L., A. Wald, R. Patel, S. L. Sacks, S. K. Tyring, T. Warren, and others. Sexually Transmitted Infections 78 (Suppl. 1): i1­2. 2004. "Once-Daily Valacyclovir to Reduce the Risk of Transmission of Aral, S. O., and K. K. Holmes. 1999. "Social and Behavioral Determinants Genital Herpes." New England Journal of Medicine 350 (1): 11­20. of the Epidemiology of STDs: Industrialized and Developing Creighton, S., M. Tenant-Flowers, C. B. Taylor, R. Miller, and N. Low. 2003. Countries." In Sexually Transmitted Diseases, 3rd ed., ed. K. K. Holmes, "Co-infection with Gonorrhoea and Chlamydia: How Much Is There P. F. Sparling, P.-A. Mardh, S. M. Lemon, W. E. Stamm, P. Piot, and and What Does It Mean?" International Journal of STD and AIDS J. N. Wasserheit, 139­76. New York: McGraw Hill. 14 (2): 109­13. Aral, S. O., J. P. Hughes, B. Stoner, W. Whittington, H. H. Handsfield, R. M. Dallabetta, G. A., A. C. Gerbase, and K. K. Holmes. 1998. "Problems, Anderson, and others. 1999. "Sexual Mixing Patterns in the Spread of Solutions, and Challenges in Syndromic Management of Sexually Gonococcal and Chlamydial Infections." American Journal of Public Transmitted Diseases." Sexually Transmitted Infections 74 (Suppl. 1): Health 89 (6): 825­33. S1­11. Aral, S. O., and T. A. Peterman. 2002. "A Stratified Approach to Untangling Dandona, L., P. Sisodia, Y. K. Ramesh, S. G. Kumar, A. A. Kumar, M. C. Rao, the Behavioral/Biomedical Outcomes Conundrum." Sexually and others. 2005. "Cost and Efficiency of HIV Voluntary Counselling Transmitted Diseases 29 (9): 530­32. and Testing Centres in Andhra Pradesh, India." National Medical Aral, S. O., and J. S. St. Lawrence. 2002. "The Ecology of Sex Work and Journal of India 18 (1): 26­31. Drug Use in Saratov Oblast, Russia." Sexually Transmitted Diseases 29 Decosas, J., and N. Padian. 2002. "The Profile and Context of the (12): 798­805. Epidemics of Sexually Transmitted Infections Including HIV in Aral, S. O., J. S. St. Lawrence, R. Dyatlov, and A. Kozlov. 2005."Commercial Zimbabwe." Sexually Transmitted Infections 78 (Suppl. 1): i40­46. Sex Work, Drug Use, and Sexually Transmitted Infections in St. Doherty, I. A., N. S. Padian, C. Marlow, and S. O. Aral. 2005."Determinants Petersburg, Russia." Social Science and Medicine 60 (10): 2181­90. and Consequences of Sexual Networks as They Affect the Spread of Aral, S. O., and J. N. Wasserheit. 1999. "STD-Related Health Care Seeking Sexually Transmitted Infections." Journal of Infectious Diseases 191 and Health Service Delivery." In Sexually Transmitted Diseases, 3rd ed., (Suppl. 1): S42­54. ed. K. K. Holmes, P. F. Sparling, P.-A. Mardh, S. M. Lemon, W. E. Doherty, L., K. A. Fenton, J. Jones, T. C. Paine, S. P. Higgins, D. Williams, Stamm, P. Piot, and J. N. Wasserheit, 1295­306. New York: McGraw and A. Palfreeman. 2002."Syphilis: Old Problem, New Strategy." British Hill. Medical Journal 325 (7356): 153­56. Bailey, R. C., F. A. Plummer, and S. Moses. 2001. "Male Circumcision and Donovan, B. 2000a. "The Repertoire of Human Efforts to Avoid Sexually HIV Prevention: Current Knowledge and Future Research Directions." Transmissible Diseases: Past and Present. Part 1--Strategies Lancet Infectious Diseases 1 (4): 223­31. Used before or Instead of Sex." Sexually Transmitted Infections 76 (1): Blanchard, J. F. 2002. "Populations, Pathogens, and Epidemic Phases: 7­12. Closing the Gap between Theory and Practice in the Prevention of Sexually Transmitted Diseases." Sexually Transmitted Infections 78 ------. 2000b. "The Repertoire of Human Efforts to Avoid Sexually (Suppl. 1): i183­88. Transmissible Diseases: Past and Present. Part 2--Strategies Used dur- ing or after Sex." Sexually Transmitted Infections 76 (2): 88­93. Blower, S. M., A. N. Aschenbach, H. B. Gershengorn, and J. O. Kahn. 2001. "Predicting the Unpredictable: Transmission of Drug-Resistant HIV." ------. 2004. "Sexually Transmissible Infections Other Than HIV." Lancet Nature Medicine 7 (9): 1016­20. 363 (9408): 545­56. Blower, S. M., and P. Farmer. 2003."Predicting the Public Health Impact of Fenton, K. A., A. M. Johnson, and A. Nicoll. 1997. "Race, Ethnicity, and Antiretrovirals: Preventing HIV in Developing Countries." AIDScience. Sexual Health." British Medical Journal 314 (7096): 1703­4. http://www.aidscience.org/Articles/AIDScience033.asp. Fenton, K. A., C. Korovessis, A. M. Johnson, A. McCadden, S. McManus, K. Blower, S. M., and P. Volberding. 2002. "What Can Modeling Tell Us about Wellings, and others. 2001. "Sexual Behaviour in Britain: Reported the Threat of Antiviral Drug Resistance?" Current Opinion in Infectious Sexually Transmitted Infections and Prevalent Genital Chlamydia Diseases 15 (6): 609­14. Trachomatis Infection." Lancet 358 (9296): 1851­54. Sexually Transmitted Infections | 327 Garnett, G. P. 2002. "The Geographical and Temporal Evolution of Transmission of HIV-1 in Rural Uganda: A Community Randomized Sexually Transmitted Disease Epidemics." Sexually Transmitted Trial." Lancet 361 (9358): 645­52. Infections 78 (Suppl. 1): i14­19. Kamb, M. L., M. Fishbein, J. M. Douglas Jr., F. Rhodes, J. Rogers, G. Bolan, Garnett, G. P., and J. A. Rottingen. 2001. "Measuring the Risk of HIV and others. 1998. "Efficacy of Risk-Reduction Counseling to Prevent Transmission." AIDS 15 (5): 641­43. Human Immunodeficiency Virus and Sexually Transmitted Diseases: Genc, M., and W. J. Ledger. 2000. "Syphilis in Pregnancy." Sexually A Randomized Controlled Trial." Journal of the American Medical Transmitted Infections 76 (2): 73­79. Association 280 (13): 1161­67. Kaul, R., J. Kimani, N. J. Nagelkerke, K. Fonck, E. N. Ngugi, F. Keli, and oth- Giuliano, A. R., R. Harris, R. L. Sedjo, S. Baldwin, D. Roe, M. R. Papenfuss, ers. 2004 "Monthly Antibiotic Chemoprophylaxis and Incidence of and others. 2002. "Incidence, Prevalence, and Clearance of Type- Sexually Transmitted Infections and HIV-1 Infection in Kenyan Sex Specific Human Papillomavirus Infections: The Young Women's Workers: A Randomized Controlled Trial." Journal of the American Health Study." Journal of Infectious Diseases 186 (4): 462­69. Medical Association 291 (21): 2555­62. Golden, M. R., W. L. Whittington, H. H. Handsfield, J. P. Hughes, W. E. Klebanoff, M. A., J. C. Carey, J. C. Hauth, S. L. Hillier, R. P. Nugent, E. A. Stamm, M. Hogben, and others. 2005. "Effect of Expedited Treatment Thom, and others. 2001. "Failure of Metronidazole to Prevent of Sex Partners on Recurrent or Persistent Gonorrhea or Chlamydial Preterm Delivery among Pregnant Women with Asymptomatic Infection." New England Journal of Medicine 352 (7): 676­85. Trichomonas vaginalis Infection." New England Journal of Medicine Gorbach, P. M., H. Sopheab, T. Phalla, H. B. Leng, S. Mills, A. Bennett, and 345 (7): 487­93. others. 2000. "Sexual Bridging by Cambodian Men: Potential Koumans, E. H., T. A. Farley, J. J. Gibson, C. Langley, M. W. Ross, M. Importance for General Population Spread of STD/HIV Epidemics." McFarlane, and others. 2001. "Characteristics of Persons with Syphilis Sexually Transmitted Diseases 27 (6): 320­26. in Areas of Persisting Syphilis in the United States: Sustained Gray, R. H., F. Wabwire-Mangen, G. Kigozi, N. K. Sewankambo, D. Transmission Associated with Concurrent Partnerships." Sexually Serwadda, L. H. Moulton, and others. 2001. "Randomized Trial of Transmitted Diseases 28 (9): 497­503. Presumptive Sexually Transmitted Disease Therapy during Pregnancy Koutsky, L. A., K. A. Ault, C. M. Wheeler, D. R. Brown, E. Barr, F. B. Alvarez, in Rakai, Uganda." American Journal of Obstetrics and Gynecology and others. 2002. "A Controlled Trial of a Human Papillomavirus Type 185 (5): 1209­17. 16 Vaccine." New England Journal of Medicine 347 (21): 1645­51. Gregson, S., C. A. Nyamukapa, G. P. Garnett, P. R. Mason, T. Zhuwau, M. Kraut, J., and S. O. Aral. 2001. "Patterns of Age Mixing Are Associated with Carael, and others. 2002."Sexual Mixing Patterns and Sex-Differentials STDs in the USA." International Journal of STD and AIDS 12 (Suppl. in Teenage Exposure to HIV Infection in Rural Zimbabwe." Lancet 359 2): 188. (9321): 1896­903. Kraut-Becher, J. R., and S. O. Aral. 2003. "Gap Length: An Important Grémy, I., and N. Beltzer. 2004. "HIV Risk and Condom Use in the Adult Factor in Sexually Transmitted Disease Transmission." Sexually Heterosexual Population in France between 1992 and 2001: Return to Transmitted Diseases 30 (3): 221­25. the Starting Point?" AIDS 18 (5): 805­9. Kumaranayake, L., P. Vickerman, D. Walker, S. Samoshkin, V. Romantzov, Grosskurth, H., F. Mosha, J. Todd, E. Mwijarubi, A. Klokke, K. Senkoro, Z. Emelyanova, and others. 2004. "The Cost-Effectiveness of HIV and others. 1995. "Impact of Improved Treatment of Sexually Preventive Measures among Injecting Drug Users in Svetlogorsk, Transmitted Diseases on HIV Infection in Rural Tanzania: Belarus." Addiction 99 (12): 1565­76. Randomized Controlled Trial." Lancet 346 (8974): 530­36. Laga, M., M. O. Diallo, and A. Buvé. 1994. "Interrelationship of STD and Harrison W. O., R. R. Hooper, P. J. Wiesner, A. F. Campbell, W. W. Karney, HIV: Where Are We Now?" AIDS 8 (Suppl.): S119­24. G. H. Reynolds, and others. 1979 "A Trial of Minocycline Given after Laga, M., A. Manoka, M. Kivuvu, B. Malele, M. Tuliza, N. Nzila, and oth- Exposure to Prevent Gonorrhea." New England Journal of Medicine 300 ers. 1993. "Non-ulcerative Sexually Transmitted Diseases as Risk (19): 1074­78. Factors for HIV-1 Transmission in Women: Results from a Cohort Hart, C. A., and S. Kariuki. 1998. "Antimicrobial Resistance in Developing Study." AIDS 7 (1): 95­102. Countries." British Medical Journal 317 (7159): 647­50. Lau, C. Y., and A. K. Qureshi. 2002. "Azithromycin versus Doxycycline for Hawkes, S., and K. G. Santhya. 2002. "Diverse Realities: Sexually Genital Chlamydial Infections: A Meta-Analysis of Randomized Transmitted Infections and HIV in India." Sexually Transmitted Clinical Trials." Sexually Transmitted Diseases 29 (9): 497­502. Diseases 78 (Suppl. 1): i31­39. Laumann, E. O., and Y. Youm. 1999. "Racial-Ethnic Group Differences in Hayes, R., J. Chagalucha, H. Grosskurth, A. Obasi, J. Todd, B. Cleophas- the Prevalence of Sexually Transmitted Diseases in the United States: Mazige, and others. 2003. "Mema Kwa Vijana, a Randomized A Network Explanation." Sexually Transmitted Diseases 26 (5): 250­61. Controlled Trial of an Adolescent Sexual and Reproductive Health Low-Beer, D., and R. L. Stoneburner. 2003. "Behaviour and Com- Intervention Programme in Rural Mwanza, Tanzania: 1. Rationale and munication Change in Reducing HIV: Is Uganda Unique?" African Trial Design" (Abstract 0695). Paper presented at the 15th Biennial Journal of AIDS Research 2: 9­21. Congress of the International Society of Sexually Transmitted Diseases Manhart, L., and K. K. Holmes. 2005. "Randomized Controlled Trials of Research, Ottawa, July 27­30. Individual-Level, Population-Level, and Multilevel Interventions for Holmes,K.K.,and S.O.Aral.1991."Behavioral Interventions in Developing Preventing Sexually Transmitted Infections: What Has Worked?" Countries." In Research Issues in Human Behavior and STD in the AIDS Journal of Infectious Diseases 191 (Suppl. 1): S7­24. Era, ed. J. N. Wasserheit, S. O. Aral, and K. K. Holmes, 318­44. Mayaud, P., F. Mosha, J. Todd, R. Balira, J. Mgara, B. West, and others. 1997. Washington, DC: American Society of Microbiology Publications. "Improved Treatment Services Significantly Reduce the Prevalence of Idahl, A., J. Boman, U. Kumlin, and J. I. Olofsson. 2004. "Demonstration of Sexually Transmitted Diseases in Rural Tanzania: Results of a Chlamydia trachomatis IgG Antibodies in the Male Partner of the Randomized Controlled Trial." AIDS 11 (15): 1873­80. Infertile Couple Is Correlated with a Reduced Likelihood of Achieving Meheus, A., K. F. Schulz, and W. Cates Jr. 1990. "Development of Pregnancy." Human Reproduction 19 (5): 1121­26. Prevention and Control Programs for Sexually Transmitted Diseases Kamali, A., M. Quigley, J. Nakiyingi, J. Kinsman, J. Kengeya-Kayondo, R. in Developing Countries." In Sexually Transmitted Diseases, 2nd ed., Gopal, and others. 2003. "Syndromic Management of Sexually- ed. K. K. Holmes, P.-A. Mardh, and P. F. Sparling, 1041­46. New York: Transmitted Infections and Behaviour Change Interventions on McGraw-Hill. 328 | Disease Control Priorities in Developing Countries | Sevgi O. Aral and Mead Over with others Morris, M., ed. 2004. Network Epidemiology: A Handbook for Survey Design Schachter, J. 2001. "NAATs to Diagnose Chlamydia trachomatis Genital and Data Collection. Oxford, U.K.: Oxford University Press. Infection: A Promise Still Unfulfilled." Expert Review of Molecular Morris, M., and M. Kretzchmar. 1995. "Concurrent Partnerships and Diagnostics 1 (2): 137­44. Transmission of Dynamics in Networks." Social Networks 17: 299­318. Schmid, G. P., A. Buve, P. Mugyenyi, G. P. Garnett, R. J. Hayes, B. G. Morris, M., C. Podhisita, M. J. Wawer, and M. S. Handcock. 1996. "Bridge Williams, and others. 2004. "Transmission of HIV-1 Infection in Sub- Populations in the Spread of HIV/AIDS in Thailand." AIDS 10 (11): Saharan Africa and Effect of Elimination of Unsafe Injections." Lancet 1265­71. 363 (9407): 482­88. Moses, S., E. N. Ngugi, J. E. Bradley, E. K. Njeru, G. Eldridge, E. Muia, and Scholes D., A. Stergachis, F. E. Heidrich, H. Andrilla, K. K. Holmes, and others. 1994. "Health Care­Seeking Behavior Related to the W. E. Stamm. 1996. "Prevention of Pelvic Inflammatory Disease by Transmission of Sexually Transmitted Diseases in Kenya." American Screening for Cervical Chlamydial Infection." New England Journal of Journal of Public Health 84 (12): 1947­51. Medicine 334 (21): 1362­66. Moses, S., E. N. Ngugi, A. Costigan, C. Kariuki, I. Maclean, R. C. Brunham, Scoular, A. 2002. "Using the Evidence Base on Genital Herpes: Optimizing and others. 2002. "Response of a Sexually Transmitted Infection the Use of Diagnostic Tests and Information Provision." Sexually Epidemic to a Treatment and Prevention Program in Nairobi, Kenya." Transmitted Infections 78 (3): 160­65. Sexually Transmitted Infections 78 (Suppl. 1): i114­20. Semaan, S., L. Kay, D. Strouse, E. Sogolow, P. D. Mullen, M. S. Neumann, National Institute for Allergy and Infectious Diseases. 2001. Scientific and others. 2002. "A Profile of U.S.-Based Trials of Behavioral and Evidence on Condom Effectiveness for Sexually Transmitted Disease Social Interventions of HIV Risk Reduction." Journal of Acquired Prevention. Workshop report. Bethesda, MD: National Institute for Immune Deficiency Syndrome 30 (Suppl. 1): S30­50. Allergy and Infectious Diseases. Service, S. K., and S. M. Blower. 1996. "Linked HIV Epidemics in San O'Farrell, N. 2002. "Donovanosis." Sexually Transmitted Infections 78 (6): Francisco" (letter). Journal of Acquired Immune Deficiency Syndromes 452­57. and Human Retrovirology 11 (3): 311­13. Orroth, A. K. 2003. "Investigations of the Proportion of HIV Infections Shelton, J. D., D. T. Halperin, V. Nantulya, M. Potts, H. D. Gayle, and K. K. Attributable to Sexually Transmitted Diseases in Sub-Saharan Africa Holmes. 2004. "Partner Reduction Is Crucial for Balanced `ABC' Based on Data from the Mwanza and Rakai Trials." Ph.D. thesis, Approach to HIV Prevention." British Medical Journal 328 (7444): Faculty of Medicine, University of London. 891­93. Over, M. 1998. "The Effects of Societal Variables on Urban Rates of HIV Simms, I., and J. M. Stephenson. 2000. "Pelvic Inflammatory Disease Infection in Developing Countries: An Exploratory Analysis." In Epidemiology: What Do We Know and What Do We Need to Know?" Confronting AIDS: Evidence from the Developing World, ed. M. Sexually Transmitted Infections 76 (2): 80­87. Ainsworth, L. Fransen, and M. Over, 40­51. Brussels: European Stamm, W. E. 1999. "Chlamydia trachomatis Infections of the Adult." In Commission. Sexually Transmitted Diseases, 3rd ed., ed. K. K. Holmes, P. F. Sparling, ------. 1999. "The Public Interest in a Private Disease: An Economic P.-A. Mardh, S. M. Lemon, W. E. Stamm, P. Piot, and J. N. Wasserheit, Perspective on the Government Role in STD and HIV Control." In 407­22. New York: McGraw-Hill. Sexually Transmitted Diseases, 3rd ed., ed. K. K. Holmes, P. F. Sparling, Stanberry, L. R., S. L. Spruance, A. L. Cunningham, D. I. Bernstein, A. P.-A. Mardh, S. M. Lemon, W. E. Stamm, P. Piot, and J. N. Wasserheit, Mindel, S. Sacks, and others. 2002. "Glycoprotein-D-Adjuvant Vaccine 3­15. New York: McGraw-Hill. to Prevent Genital Herpes." New England Journal of Medicine 347 (21): ------. 2004. "Impact of the HIV/AIDS Epidemic on the Health Sectors 1652­61. of Developing Countries." In The Macroeconomics of HIV/AIDS, ed. M. Steen, R. 2001. "Eradicating Chancroid." Bulletin of the World Health Haacker. Washington, DC: International Monetary Fund. Organization 79 (9): 818­26. Over, M., P. Heywood, J. Gold, I. Gupta, S. Hira, and E. Marseille. 2004. Stone, K. M., K. L. Karem, M. R. Sternberg, G. M. McQuillan, A. D. Poon, HIV/AIDS Treatment and Prevention in India: Modeling the Costs and E. R. Unger, and others. 2002. "Seroprevalence of Human Consequences. Washington, DC: World Bank. Papillomavirus Type 16 in the United States." Journal of Infectious Over, M., and P. Piot. 1993. "HIV Infection and Sexually Transmitted Diseases 186 (10): 1369­402. Diseases." In Disease Control Priorities in Developing Countries, ed. Stoneburner, R. L., and D. Low-Beer. 2004."Population-Level HIV Declines D. T. Jamison, W. H. Mosley, A. R. Measham, and J. L. Bobadilla, and Behavioral Risk Avoidance in Uganda." Science 304 (5671): 714­18. 455­527. New York: Oxford University Press. Strand, A., R. Patel, H. C. Wulf, K. M. Coates, and International Peterman, T. A., L. S. Lin, D. R. Newman, M. L. Kamb, G. Bolan, J. Valacyclovir HSV Study Group. 2002. "Aborted Genital Herpes Zenilman, and others. 2000. "Does Measured Behavior Reflect STD Simplex Virus Lesions: Findings from a Randomized Controlled Risk? An Analysis of Data from a Randomized Controlled Behavioral Study." Sexually Transmitted Infections 78 (6): 435­39. Intervention Study: Project RESPECT Study Group." Sexually Turner, C. F., S. M. Rogers, H. G. Miller, W. C. Miller, J. N. Gribble, J. R. Transmitted Diseases 27 (8): 446­51. Chromy, and others. 2002. "Untreated Gonococcal and Chlamydial Pisani, E., G. P. Garnett, N. C. Grassly, T. Brown, J. Stover, C. Hankins, and Infection in a Probability Sample of Adults." Journal of the American others. 2003. "Back to Basics in HIV Prevention: Focus on Exposure." Medical Association 287 (6): 726­33. British Medical Journal 326 (7403): 1384­87. UNAIDS and WHO (Joint United Nations Programme on HIV/AIDS Rekart, M. L. 2002. "Sex in the City: Sexual Behavior, Societal Change, and and World Health Organization). 2000. Guidelines for Second STDs in Saigon." Sexually Transmitted Infections 78 (Suppl. 1): i47­54. Generation HIV Surveillance. WHO/CDS/EDC/2000.5, UNAIDS/ Samoff, E., E. Koumans, L. Markowitz, and others. 2004. "An Assessment 00.03E. Geneva: Working Group on Global HIV/AIDS and STI of Factors Associated with Type-Specific Human Papillomavirus Surveillance. Persistence in an Adolescent Clinic Population Using Generalized van Dam, C. J. 1995. "HIV, STD, and Their Current Impact on Estimating Equations--Atlanta, Georgia, 1999­2003." Oral presenta- Reproductive Health: The Need for Control of Sexually Transmitted tion at the Epidemic Intelligence Service National Conference, Atlanta, Diseases." International Journal of Gynecology and Obstetrics 50 Georgia, April 19. (Suppl. 2): S121­29. Sexually Transmitted Infections | 329 van Dam, C. J., G. Dallabetta, and P. Piot. 1999. "Prevention and Control Wawer, M. J., N. K. Sewankambo, D. Serwadda, T. C. Quinn, L. A. Paxton, of Sexually Transmitted Diseases in Developing Countries." In Sexually N. Kiwanuka, and others. 1999. "Control of Sexually Transmitted Transmitted Diseases, 3rd ed., ed. K. K. Holmes, P. F. Sparling, P.-A. Diseases for AIDS Prevention in Uganda: A Randomized Community Mardh, S. M. Lemon, W. E. Stamm, P. Piot, and J. N. Wasserheit, Trial--Rakai Project Study Group." Lancet 353 (9152): 525­35. 1381­90. New York: McGraw-Hill. Weir, S., C. Pailman, X. Mahlalela, N. Coetzee, F. Meidany, and J. T. Vos, T. 1994. "Attitudes to Sex and Sexual Behavior in Rural Matabeleland, Boerma. 2003. "From People to Places: Focusing AIDS Prevention Zimbabwe." AIDS Care 6 (2): 193­203. Where It Matters Most." AIDS 17 (6): 895­903. Walboomers, J. M. M., M. V. Jacobs, M. M. Manos, F. X. Bosch, J. A. WHO (World Health Organization). 1991. Management of Patients with Kummer, K. V. Shah, and others. 1999. "Human Papillomavirus Is a Sexually Transmitted Disease: Report of a WHO Study Group. Technical Necessary Cause of Invasive Cervical Cancer Worldwide." Journal of Report Series 810. Geneva: WHO. Pathology 189 (1): 12­19. ------. 2001a. Global Prevalence and Incidence of Selected Curable Sexually Wallin K. L., F. Wiklund, T. Luostarinen, T. Angstrom, T. Anttila, and F. Transmitted Infections: Overview and Estimates. Geneva: WHO. Bergman. 2002. "A Population-Based Prospective Study of Chlamydia ------. 2001b. Guidelines for the Management of Sexually Transmitted Trachomatis Infection and Cervical Carcinoma." International Journal Infections. Geneva: WHO. of Cancer 101 (4): 371­74. Wilson, D., P. Chiroro, S. Lavelle, and C. Mutero. 1989. "Sex Worker, Client Wasserheit, J. N. 1989. "The Significance and Scope of Reproductive Tract Sex Behaviour, and Condom Use in Harare, Zimbabwe." AIDS Care 1 Infections among Third World Women." International Journal of (3): 269­80. Gynecology and Obstetrics 3 (Suppl.): 145­68. ------. 1991. "Epidemiological Synergy: Interrelationships between HIV World Bank. 2003. World Development Report 2004: Making Services Work Infection and Other STDs." In AIDS and Women's Health: Science for for Poor People. New York: Oxford University Press. Policy and Action, ed. L. Chen, J. Sepulveda, and S. Segal, 47­72. New York: Plenum Press. Wasserheit, J. N., and S. O. Aral. 1996. "The Dynamic Topology of Sexually Transmitted Disease Epidemics: Implications for Prevention Strategies." Journal of Infectious Diseases 174 (Suppl. 2): S201­13. 330 | Disease Control Priorities in Developing Countries | Sevgi O. Aral and Mead Over with others Chapter 18 HIV/AIDS Prevention and Treatment Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, Lisa M. DeMaria, Becca Feldman, Helene Gayle, Julian Gold, Robert Grant, and Michael T. Isbell Although global commitment to control the HIV/AIDS pan- Enormous advances in HIV/AIDS treatment regimens have demic has increased significantly in recent years, the virus con- fundamentally altered the natural history of the disease and tinues to spread with alarming and increasing speed. By the end sharply reduced HIV-related morbidity and mortality in coun- of 2005, an estimated 40 million people worldwide were living tries where such treatments are accessible. The advent of anti- with HIV infection or disease, a notable rise from the 35 mil- retroviral drugs in the late 1980s began a revolution in the lion infected with HIV in 2001 (UNAIDS 2005). In 2005, close management of HIV, which can be seen as analogous to the to 5 million new HIV infections and 3 million AIDS deaths use of penicillin for treating bacterial infections in the 1940s. The occurred, more of both than in any previous year. Sub-Saharan most notable advance on the treatment front is the use of com- Africa remains the region most affected by HIV/AIDS; how- bination antiretroviral therapy, which is far more effective than ever, the virus is now spreading rapidly in Asia and parts of monotherapy (zidovudine or AZT), the standard of care when Eastern Europe. the first edition of this volume was published. Recent declines Despite the rapid spread of HIV, several countries have in the price of combination antiretroviral therapy in develop- achieved important success in curbing its transmission. The ing countries from US$15,000 per year to less than US$150 in extraordinary potential of HIV prevention is exemplified by some countries have prompted numerous developing countries such diverse efforts as Thailand's 100 percent condom pro- to introduce antiretroviral therapy through the public sector. gram, Uganda's remarkable decrease in HIV prevalence, and These declines also pose difficult questions regarding the optimal the community-based syndromic management of sexually allocation of limited resources for HIV/AIDS, as well as the transmitted infections (STIs) in Mwanza, Tanzania. Box 18.1 potential impact on already strained health care infrastructures. describes characteristics common to these programs. Successes also include the development and effective use of OBSTACLES TO HIV CONTROL highly sensitive and specific HIV screening tests, which have virtually eliminated infection from the blood supply in the Obstacles to effective HIV control include lack of prevention developed world and in most parts of the developing world and care coverage and lack of rigorous evaluations. Both are (WHO 2002a). In addition, the administration of a short course discussed below. of nevirapine to mothers during labor and to newborns post- partum reduces the risk of mother-to-child transmission Lack of Coverage and Access to Prevention Services (MTCT) by as much as 47 percent (Guay and others 1999). Notwithstanding these treatment strides, global efforts have However, recent data suggest that such short-term successes not proved sufficient to control the spread of the pandemic or may be at the expense of resistance and viral failure once treat- to extend the lives of the majority of those infected. The desired ment is introduced after delivery (Eshleman and others 2001). level of success has not yet been achieved for several reasons. 331 Box 18.1 Successful HIV Prevention Strategies The HIV prevention success stories highlighted in this · population-based programs designed to change social chapter stem in part from each country's unique cultural, norms historical, and infrastructural elements. Nevertheless, · increased open communication about sexual activities these successes share several common features, thereby and HIV/AIDS offering potential guidance for the development and · programs to combat stigma and discrimination implementation of prevention strategies in other settings. · condom promotion These features include: · STI surveillance and control · high-level political leadership · interventions targeting key "bridge" populations-- · active engagement of civil society and religious leaders populations that transmit the virus from high-risk to in a multisectoral approach low-risk groups. Source: Authors. Most people who could benefit from available control strate- Because antiretroviral therapy has historically been unavail- gies, including treatment, do not have access to them. Modelers able in most developing countries, national programs have commissioned by the World Health Organization (WHO) and lacked the means to undertake a comprehensive approach to the Joint United Nations Programme on HIV/AIDS (UNAIDS) HIV/AIDS (notable exceptions are Argentina, Brazil, and determined that existing interventions could prevent 63 per- Mexico, which provide universal coverage for antiretroviral cent of all infections projected to occur between 2002 and 2010 therapy). As discussed in chapter 8, control of the pandemic (Stover and others 2002). Nonetheless, a 2003 survey of cover- demands a two-front battle that emphasizes both prevention age revealed that fewer than one in five people at high risk of and care. Even though the prospect of greater access to treat- infection had access to the most basic prevention services, ment increases the feasibility of integrating prevention and care including condoms, AIDS education, MTCT prevention, vol- in resource-limited settings, it also raises new questions regard- untary counseling and testing (VCT), and harm reduction pro- ing the selection of optimal prevention programs to pair with grams (Global HIV Prevention Working Group 2003). WHO treatment programs. and UNAIDS estimate that only about 7 percent of the nearly 6 million people in need of treatment receive it and that the number of people who require antiretroviral therapy increases Lack of Rigorous Evaluations by 8,000 each day (UNAIDS 2004). In addition to poor coverage of key interventions, perhaps the Current coverage shortfalls, combined with the relentless greatest challenge to effective global control is the lack of expansion of the epidemic, underscore the acute need for rapid reliable evidence to guide the selection of interventions for spe- scale-up of prevention and treatment interventions--an imper- cific areas or populations. In the same way that global policy ative that the international community has acknowledged but makers are increasingly recognizing the need for rigorous eval- that remains to be realized after more than 15 years. However, uation of development programs to ensure their success and the activities of the Global Fund to Fight AIDS, Tuberculosis, and eliminate waste, the need for reliable scientific evaluations of Malaria and the U.S. President's Emergency Plan for AIDS AIDS control programs is equally paramount for the same rea- Relief (a five-year, US$15 billion initiative) suggest a growing sons. There are simply not enough resources to do everything commitment to tackle these issues. The latter aims to provide everywhere; choices must be made and priorities set. In the antiretroviral drugs for 2 million HIV-infected people, to pre- HIV/AIDS field, this information deficit is especially pro- vent 7 million new infections, to provide care for 10 million indi- nounced with respect to HIV prevention in general and preven- viduals, and to develop health system capacity in Vietnam and tion implemented on a population level in particular. Currently, in Africa and the Caribbean. Even though 15 countries are cur- the allocation of resources for HIV/AIDS prevention is seldom rently slated to receive support from the President's Emergency evidence based, primarily because of a lack of data on both the Plan, many of the countries most affected by HIV/AIDS-- effectiveness and the cost of interventions (Feachem 2004). including Lesotho, Malawi, Swaziland, and Zimbabwe--are not Few evaluations have collected data specifically on HIV included in the list of beneficiary countries. infection as an outcome (Fleming and DeMets 1996). In the 332 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others case of care and treatment, success and failure are more readily burden of disease, the determinants of transmission, and the and rapidly apparent, leading to a substantial degree of auto- effectiveness and cost-effectiveness of existing prevention correction of ineffective policies. In contrast, with respect to interventions. HIV prevention, it is unlikely that those infections that might have occurred in the absence of a prevention program would Burden of Disease be monitored, thus reducing the meaningfulness of the auto- As a result of large-scale implementation of data collection feedback cycle for prevention. This underscores the importance methods for surveillance worldwide and enhanced methods for of proactive, rigorous evaluation to differentiate success from validating and interpreting HIV-related data, the HIV/AIDS failure in a timely manner. Sound evidence on the effectiveness epidemic is probably one of the best documented epidemics of HIV prevention measures is especially important in light of in history. An increasing number of data sources contribute the tendency of many governments and international aid agen- to reasonably accurate estimates and a more nuanced under- cies to avoid programs that address sexual behaviors, drug use, standing of the epidemic's trends. Unfortunately, this relatively and highly stigmatized and vulnerable populations. accurate picture of where the epidemic is and has been is not In addition, prevention studies have rarely incorporated matched by similarly convincing maps of the factors that the well-defined control or comparison groups necessary to explain its spread. identify contextual factors that are essential for appropriately Although no single country has been spared the virus, the tailoring interventions to the diverse regional settings and the epidemic has affected certain regions of the world dispropor- myriad of microenvironments in which HIV transmission tionately, and Sub-Saharan Africa remains by far the hardest hit occurs (Grassly and others 2001). Contextual data are similarly region (table 18.1). With only 10 percent of the world's popu- critical for developing strategies to combat HIV/AIDS-related lation, it accounts for more than 75 percent of all HIV infec- stigma and restrictive social and gender norms, which often tions worldwide and more than 75 percent of AIDS-related frustrate attempts to address sexual and addictive behaviors deaths estimated for 2003. Asia and the Pacific, with several associated with HIV transmission. Even where national efforts large and populous countries, account for 7.4 million infec- have succeeded in curbing the spread of the epidemic, as in tions, or 19.5 percent of the current burden of disease. Senegal and Uganda, evidence often does not clearly indicate Prevention and treatment efforts in Sub-Saharan Africa and the specific, well-defined, contextual features that account for Asia--regions that together represent 85 percent of all current success. infections--have dictated, and will continue to dictate, global The lack of both contextual data and sound evidence trends in the burden of HIV- and AIDS-related mortality. regarding the effectiveness of HIV interventions hinders policy Between 1997 and 2001, the percentage of women living makers' ability to tailor HIV interventions to the nature and with HIV/AIDS increased from 41 to 50 percent. This trend is stage of national epidemics, something that the authors argue most apparent in Sub-Saharan Africa, where women represent is necessary to address HIV/AIDS effectively. In the absence of 57 percent of adults living with HIV and 75 percent of HIV- such data, HIV/AIDS expenditures undoubtedly incorporate infected young people. Even though women account for a an unacceptable degree of waste, people are unnecessarily smaller share of infections in Asia (28 percent), the disease bur- becoming infected with HIV, and HIV-infected individuals are den among women and girls is likely to rise as the epidemic dying prematurely. becomes generalized. More detailed information about the Why has this type of research not been more forthcoming? global burden of HIV/AIDS, regional differences, and trends In part it is because, by definition, such research is less innova- over time is available in the UNAIDS (2005) report on the tive scientifically and also typically less experimental than global AIDS epidemic. research to develop new interventions. It is handicapped both in competing for traditional research funding and in receiving academic recognition. The only way to redress the imbalance is Determinants of Infection through specific earmarking of significant research funds. HIV transmission predominantly occurs through three mech- anisms: sexual transmission, exposure to infected blood or ACTION UNDER UNCERTAINTY blood products, or perinatal transmission (including breast- feeding). The likelihood of transmission is heavily affected by Even though the current deficit in evaluation research is glar- social, cultural, and environmental factors that often differ ing, the magnitude and seriousness of the global pandemic markedly between and within regions and countries. There is means that action is nevertheless required. Moreover, despite also some indication that molecular, viral, immunological, or such gaps in knowledge, we can still improve control strategies other host factors might influence the likelihood of HIV trans- by tailoring interventions to the nature and scope of the epi- mission. For a more detailed discussion of sexual behaviors and demic. Summarized below is what is known with regard to the the contextual determinants of infection, see chapter 17. HIV/AIDS Prevention and Treatment | 333 Table 18.1 Deaths and Disability-Adjusted Life Years Attributed to AIDS by Region, Age, and Gender, 2001 Number (thousands) Region Total Both sexes, age 0­14 Both sexes, age 15 Percentage female Deaths World 2,576 439 2,133 46 High-income countries 22 0 21 23 Low- and middle-income countries 2,554 439 2,111 46 Sub-Saharan Africa 2,058 408 1,651 51 East Asia and the Pacific 107 5 100 25 Europe and Central Asia 28 0 27 14 Latin America and the Caribbean 83 8 73 36 Middle East and North Africa 4 0 2 25 Southeast Asia 272 18 255 23 Disability-adjusted life years World 71,460 13,586 57,875 47 High-income countries 665 7 660 23 Low- and middle-income countries 70,795 13,579 57,215 47 Sub-Saharan Africa 56,820 12,526 44,294 52 East Asia and the Pacific 3,121 195 2,927 25 Europe and Central Asia 982 25 957 18 Latin America and the Caribbean 2,354 260 2,092 36 Middle East and North Africa 105 20 84 39 Southeast Asia 7,413 553 6,861 25 Source: Mathers and others 2006. Table 18.2 Estimated HIV Transmission Risk per Exposure Type of exposure Estimated risk HIV transmission per exposure Receptive anal intercourse 3.0 percent (1/125 to 1/31) (DeGruttola and others 1989) Receptive vaginal intercourse 0.1 percent (1/2,000 to 1/667) (Mastro and others 1994; Wiley, Herschkorn, and Padian 1989) Insertive vaginal or anal intercourse 0.1 percent (1/3,333 to 1/1,111) (Nagachinta and others 1997; Peterman and others 1988) Needlestick injury 0.3 percent (1/313) (Henderson and others 1990) Use of contaminated injecting drug equipment 0.6 percent (1/149) (Kaplan and Heimer 1992) Mucous membrane 0.1 percent (1/1,111) (Ippolito, Puro, and De Carli 1993) Source: Authors. Sexual Transmission. Worldwide, sexual intercourse is the infected individual and by infectivity in the event of exposure. predominant mode of transmission, accounting for approxi- This also includes factors related to the infectiousness of the mately 80 percent of infections (Askew and Berer 2003). Sexual infected partner and the susceptibility of the uninfected partner. intercourse accounts for more than 90 percent of infections in Sub-Saharan Africa. Although many people who know they Infectivity The per contact infectivity of HIV from sexual trans- are infected reduce their risk behaviors, studies in developed mission varies depending on sexual activity (Royce and others countries suggest that a substantial percentage nevertheless 1997).Anal intercourse carries a higher transmission probability continue to engage in unprotected sex (Marks, Burris, and than penile-vaginal intercourse, and male-to-female transmis- Peterman 1999). The risk of sexual transmission is determined sion is more likely than female-to-male transmission. Data on by behaviors that influence the likelihood of exposure to an infectivity by transmission mode are shown in table 18.2. 334 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others Biological Mediators of Infectivity Untreated STIs increase The risk of sexual transmission is also strongly correlated the risk of sexual HIV transmission several-fold (Institute of with the plasma level of virus in the infected individual (Quinn Medicine 1997). Numerous epidemiological studies have sup- and others 2000); thus, infectivity varies over the natural ported the association of genital ulcers in general and of geni- progression of the disease. Individuals are most infectious sub- tal herpes (herpes simplex virus 2, or HSV-2) in particular with sequent to infection and again during the late stage of the HIV infection (Hook and others 1992). Not only does the bio- disease. Antiretroviral therapy significantly reduces the level of logical interaction between HSV-2 and HIV enhance the trans- virus, often to the point that standard tests cannot detect HIV mission and acquisition of HIV, but HIV infection is also asso- in the patient's blood (Palella and others 1998). Available data ciated with more frequent reactivation of HSV-2. The presence suggest that viral load reductions induced by antiretroviral of herpetic ulcers and lesions allows an entry point for HIV in therapy will lower infectiousness. Studies have shown a close the uninfected individual, and the presence of high copy num- relationship between the amount of viral suppression and the bers of HIV ribonucleic acid (RNA) in HSV-2 lesions in HIV- risk of vertical transmission (Garcia and others 1999). Quinn infected individuals underscores the importance for HIV pre- and others (2002) show that the risk of sexual transmission vention of controlling HSV-2 infections (Mbopi Keou and between couples in Africa was strongly related to the level of others 1999). viral load in the infected partner. Vaginal infections are also emerging as important risk factors for HIV. For example, infection with trichomonas Exposure to Infected Blood or Blood Products. Injection increases the risk for HIV seroconversion (Buve 2002). In addi- drug use and blood transfusion are two mechanisms of HIV tion, higher trichomonas rates have been detected in regions of exposure to infected blood. Determinants of each are discussed Sub-Saharan Africa that have higher HIV rates, and investiga- below. tors working throughout Sub-Saharan Africa report similar results, with odds ratios from 1.5 to 56.8 (Gregson and others Injection Because of the efficiency of HIV transmission 2001). In addition, studies have shown an increased risk of HIV through needle sharing, the introduction of HIV into an urban acquisition in patients who have bacterial vaginosis (Martin network of injecting drugs users can quickly lead to extraordi- and others 1999). narily high HIV prevalence in this population. Sharing of injec- Circumcision also affects HIV transmission. In a meta- tion equipment and frequency of injection are both important analysis of 27 studies (Weiss, Quigley, and Hayes 2000), uncir- correlates of HIV infection (Chaisson and others 1989). cumcised men were almost twice as likely to be infected with Attendance at shooting galleries, where sharing with anony- HIV as those who were circumcised. Studies that controlled mous injecting partners is likely to occur, is also an independent adequately for other risks and studies that separately assessed risk factor across many studies (Vlahov and others 1990). risk in high-risk populations, such as STI clinic attendees or Injecting cocaine (associated with "booting" or "kicking," where truck drivers, found an even stronger protective effect of cir- blood is drawn into the syringe and then injected) and having a cumcision. Similarly, an ecological study comparing two high- number of needle-sharing partners are also associated with HIV prevalence Sub-Saharan African cities with two low-prevalence infection (Anthony and others 1991). cities found that circumcised individuals were substantially less likely to be infected with HIV (Auvert and others 2001). Two Blood Transfusion The probability of becoming infected recent studies conducted in Kenya and India (Donnelly 2004; through an HIV-contaminated transfusion is estimated at Reynolds and others 2004) found that uncircumcised men had more than 90 percent (UNAIDS 1997), and the amount of HIV an HIV rate 7 to 11 times greater than circumcised men. More in a single contaminated blood transfusion is so large that indi- recently, results from a randomized controlled trial conducted viduals infected in this manner may rapidly develop AIDS. in South Africa indicated that the risk of HIV acquisition was Currently, between 5 and 10 percent of HIV infections world- reduced by more than 60 percent of men randomized for wide are transmitted through the transfusion of contaminated circumcision (controlling for sexual behavior, including con- blood products (WHO 2002a). Setting up and maintaining a dom use and health seeking behavior) in a community where safe blood supply will virtually eliminate HIV transmission more than 30 percent of the women were infected (Auvert and through transfusions. others 2005). Before circumcision among adult males becomes a wide- Perinatal Transmission. Perinatal HIV transmission includes spread policy recommendation, results are still pending in two both vertical transmission and transmission during breastfeed- similar trials. Obviously one issue is the acceptability of such a ing. Determinants of each are discussed below. procedure as well as the fact that some increase in high risk sex- ual activity was noted among the men who were circumcised, Vertical Transmission Perhaps the most compelling evidence although this did not offset the results of the intervention. of the significance of viral load and transmission risk has been HIV/AIDS Prevention and Treatment | 335 documented with respect to MTCT. Maternal viral load, as and how best to implement those interventions. We then dis- quantified by RNA polymerase chain reaction, is associated cuss the existing effectiveness and cost-effectiveness data. with increased risk in each mode of vertical transmission. A recent randomized clinical trial in Kenya found that maternal Essential Background Data for Any Intervention. Because plasma HIV RNA levels higher than 43,000 copies per milliliter the prioritization of prevention strategies for any epidemic were associated with a fourfold increase in vertical transmis- requires accurately identifying the epidemiological profile (dis- sion (John and others 2001). cussed below), maintaining a sound and reliable public health Independent of HIV RNA levels in maternal plasma, addi- surveillance system is a prerequisite for an effective prevention tional risk factors include cervical HIV deoxyribonucleic acid response. An understanding of HIV and STI prevalence and (DNA), vaginal HIV DNA, and cervical or vaginal ulcers. trends, as well as the prevalence and distribution of behaviors Chorioamnionitis has also been documented as a risk factor for that contribute to the epidemic's spread, should be supple- MTCT among African mothers (Ladner and others 1998), as mented by national monitoring systems that track sources and has exposure to maternal blood during labor and delivery. uses of funding to promote greater accountability. In addition, Newell (2003) estimates that for every hour an infant is data are needed to identify and characterize key contextual exposed to ruptured membranes, the risk of transmission issues that affect the selection of interventions. increases by 2 percent. Although surveillance is essential for an optimally strategic public health response, its utility depends on the degree to Breastfeeding Transmission through breastfeeding is likely which the information it yields is effectively deployed. As noted associated with an elevated viral load in the breast milk, which below, countries with concentrated epidemics should prioritize in turn is associated with maternal plasma viral load and CD4 interventions that are targeted to the populations at highest T cell levels. Mastitis has also been associated with increased risk. In Latin America, however, where information on national risk of vertical transmission. Meta-analyses suggest that the AIDS funding is strongest, the proportion of limited preven- cumulative probability of HIV infection increases from 0.6 per- tion resources that is not targeted to the populations at highest cent at age 6 months to 9.2 percent at age 3 (Read 2003). A risk of infection varies from less than 5 percent to more than study in Malawi, however, indicates that most transmission 50 percent (Saavedra 2000). This range strongly suggests that occurs in the early breastfeeding months, with an incidence per resource allocation is frequently not based on available epi- month of 0.7 percent at age 1 to 5 months, 0.6 percent at age 6 demiological and effectiveness data. to 11 months, and 0.3 percent at age 12 to 17 months (Miotti Table 18.3 summarizes information about the effectiveness and others 1999). In one study, infants who were breastfed in of the interventions discussed below. combination with receiving other supplementary foods were twice as likely to be infected at age 6 months than infants fed Cost-Effectiveness Estimates for Prevention Interventions. exclusively on breast milk or on formula (Coutsoudis and oth- How countries spend funds and which interventions they ers 2001). The hypothesis is that antigens and bacterial con- prioritize should be guided by estimates of the relative cost- taminants present in supplemental fluids and foods consumed effectiveness of such interventions. Unfortunately, reliable by infants who are not exclusively breastfed may cause inflam- estimates of cost-effectiveness are largely lacking, for a number mation and microtrauma to the infant's intestinal gut, thereby of reasons. The main reason is that HIV prevention interven- facilitating viral transmission. Another hypothesis is that mixed tions are difficult to force into a typology that clearly distin- feeding increases the risk of subclinical or clinical mastitis in guishes one intervention from another. For example, the the mother, which could increase milk viral load (Semba and counseling component of VCT has a strong information- others 1999). sharing element that overlaps with (a) information, education, Decisions about breastfeeding are further complicated by and communication (IEC) through the media; (b) peer recent data indicating possible increased mortality among interventions; and (c) the counseling component of STI treat- breastfeeding mothers (Nduati and others 2001) and by the ment. Similarly, the psychological support offered through stigma associated with not breastfeeding in countries where counseling is comparable to support provided through support abstaining from breastfeeding is tantamount to disclosing a groups or to interventions designed to increase social support. woman's HIV status. Such overlap and duplication among components of different interventions complicate efforts to estimate both the effective- ness and the cost-effectiveness of different interventions. Effectiveness and Cost-Effectiveness Several authors have recently reviewed estimates of cost- of Prevention Interventions effectiveness for the prevention interventions described here Below we discuss the need for ongoing surveillance and con- (Creese and others 2002; Jha and others 2001; Marseille and textual data to determine the effectiveness of HIV interventions others 2002; Walker 2003). These reviews address a number of 336 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others Table 18.3 Effectiveness of HIV Interventions Intervention Outcome Effect Citations School-based education Sexual debut The number of students reporting early Hayes and others 2003; Stanton and sexual debut was significantly lower in the others 1998 intervention group in both studies. Multiple sex partners The number of students reporting multiple sex Fawole and others 1999; Hayes and others partners was significantly lower in the 2003 intervention group in both studies. Condom use Condom use was significantly higher in the Fawole and others 1999; Harvey, Stuart, intervention group in three of the four studies and Swan 2000; Hayes and others 2003; and nonsignificantly higher in one study. Stanton and others 1998 HIV incidence The study found no significant differences Hayes and others 2003 in HIV incidence. STI prevalence and incidence The study found no significant differences Hayes and others 2003 in STI prevalence and incidence. Abstinence education Condom use The study found no significant differences Jemmott, Jemmott, and Fong 1998 in condom use. Early sexual debut The study found no significant differences Meekers 2000 in early sexual debut. VCTa Condom use Condom use was significantly higher in the Bentley and others 1998; Bhave and others intervention group in six of the seven studies 1995; Deschamps and others 1996; and unchanged in one study. Jackson and others 1997; Kamenga and others 1991; Levine and others 1998; Voluntary HIV-1 Counseling and Testing Efficacy Study Group 2000 Unprotected intercourse Unprotected intercourse was significantly Deschamps and others 1996; lower in the intervention group in both Voluntary HIV-1 Counseling and Testing studies. Efficacy Study Group 2000 HIV incidence HIV incidence was significantly lower in the Bhave and others 1995; Celentano and intervention group in one of the studies and others 2000 nonsignificantly lower in the other study. STI prevalence and incidence STI prevalence and incidence were Celentano and others 2000; Jackson and significantly lower in the intervention group others 1997; Levine and others 1998 in all three studies. Peer-based programs Condom use Condom use was significantly higher in the Kelly and others 1997; Norr and others intervention group in all four studies. 2004; Sikkema and others 2000; Stanton and others 1996 Unprotected intercourse Unprotected intercourse was significantly Basu and others 2004; Kegeles, Hays, and lower in the intervention group in all four Coates 1996; Kelly and others 1997; studies. Sikkema and others 2000 Communication about condoms Communication was significantly higher in Lauby and others 2000 with partner the intervention group. HIV incidence HIV incidence was significantly lower in the Ghys and others 2002; Katzenstein and intervention group in both studies. others 1998 STI prevalence and incidence STI prevalence and incidence were Ghys and others 2002 significantly lower in the intervention group. Condom promotion and Condom use Condom use was significantly higher in the Bentley and others 1998; Bhave and distribution and IECa intervention group in 10 of the 11 studies others 1995; Egger and others 2000; Ford and unchanged in 1 study. and others 1996; Jackson and others 1997; Jemmott, Jemmott, and Fong 1998; Kagimu and others 1998; Laga and others 1994; Levine and others 1998; Ngugi and others 1988; Pauw and others 1996 (Continues on the following page.) HIV/AIDS Prevention and Treatment | 337 Table 18.3. Continued Intervention Outcome Effect Citations HIV incidence HIV incidence was significantly lower in the Bhave and others 1995; Celentano and intervention group in two out of three studies others 2000; Laga and others 1994 and nonsignificantly lower in one study. STI prevalence and incidence STI prevalence and incidence were Bhave and others 1995; Celentano and significantly lower in the intervention group others 2000; Jackson and others 1997; in all four studies. Laga and others 1994; Levine and others 1998 Condom social marketing Condom use Condom use was significantly higher in Agha, Karlyn, and Meekers 2001; the intervention group in one study; no Meekers 2000 significant differences were found in the other study. Early sexual debut The study found no significant differences in Meekers 2000 early sexual debut. STI treatmenta HIV incidence HIV incidence was significantly lower in the Grosskurth and others 1995; Kamali and intervention group in two of the studies, but others 2003; Laga and others 1994; the other two studies found no significant Wawer and others 1999 differences. STI prevalence and incidence The prevalence and incidence of STIs were Jackson and others 1997; Kamali and oth- significantly lower in the intervention group ers 2003; Laga and others 1994; Mayaud in all six studies. and others 1997; Wawer and others 1999 Antiretroviral therapy to Mother-to-infant transmissionb Significant reduction in mother-to-infant HIV Ayouba and others 2003; Connor and others reduce MTCT transmission in the intervention group was 1994; Dabis and others 1999; Guay and found in all eight studies, with a range of others 1999; Jackson and others 2003; 33 to 67 percent reduction in transmission. PETRA Study Team 2002; Shaffer and others 1999; Wiktor and others 1999 MTCT feeding Mother-to-infant transmission Use of breast milk substitutes prevented Nduati and others 2000 substitutions 44 percent of infant infections and was associated with significantly improved HIV-1-free survival. Harm reduction in injecting HIV incidence Significant reduction in HIV incidence in the Des Jarlais and Friedman 1996; Hurley, drug users intervention group was found in both studies. Jolley, and Kaldor 1997 Reuse or sharing of syringes Significant reduction in needle sharing in the Jenkins and others 2001; Ksobiech 2003; intervention group was found in all three Peak and others 1995; Vlahov and others studies; correlation between needle 1997 exchange program attendance and lower needle sharing was found in one study. Drug substitution for Drug use This meta-analysis found significantly lower Metzger, Navaline, and Woody 1998 injecting drug users rates of drug use. Blood safety HIV infections averted HIV screening was associated with a Foster and Buve 1995; Laleman and reduction in HIV infections by both studies. others 1992 Units of HIV-positive blood HIV screening was associated with a Jacobs and Mercer 1999 averted reduction in units of HIV-positive blood. Universal precautions Blood volume transferred in Glove material reduced the transferred blood Mast, Woolwine, and Gerberding 1993 needlestick injury volume by 46 to 86 percent. Antiretroviral therapy for HIV seroconversion The study found a significant relationship Cardo and others 1997 prevention, postexposure between seroconversion and not having prophylaxis received antiretroviral therapy. Behavior change for those Condom use Condom use was significantly higher in the Kalichman and others 2001 HIV positive intervention group. Unprotected intercourse Unprotected intercourse was significantly Kalichman and others 2001 lower in the intervention group. Source: Authors. a. Studies examined may have included educational components, condom promotion and distribution components, HIV testing and counseling, or STI treatment. b. The types of MTCT antiretroviral therapy varied in theses studies. 338 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others Table 18.4 Cost-Effectiveness of Interventions by Epidemic Profile Epidemic profile (2001 US$) UNAIDS estimate of need for 2007 Concentrated epidemic (East Asia and the Pacific, Percentage Low-level epidemic Europe and Central Asia, Generalized low- Generalized high- of all (Middle East and Latin America and the level epidemic level epidemic prevention Intervention North Africa) Caribbean, South Asia) (Sub-Saharan Africa) (Sub-Saharan Africa) 2003 US$ millions needs Surveillance No CE studies found No CE studies found No CE studies found No CE studies found -- -- IEC No CE studies found No CE studies found No CE studies found No CE studies found 129 1 School-based No CE studies found India (E/D/no STIs) No CE studies found No CE studies found 100 1 education US$1,350 per HIV infection US$68 per DALY (World Bank 1999) Abstinence No CE studies found No CE studies found No CE studies found No CE studies found -- -- education VCT No CE studies found India Chad (M/S/no STIs) No CE studies found 2,175 22 US$196 per HIV infection US$891 to US$5,213 per HIV US$10 per DALY infection (World Bank 1999) US$45 to US$261 per DALY (Hutton, Wyss, and N'Diekhor 2003) Kenya and Tanzania (M/S/STI) US$270 to US$376 per HIV HIV/AIDS infection US$14 to US$19 per DALY (Sweat and others 2000) Prevention (Continues on the following page.) and Treatment | 339 340 | Disease Table 18.4 Continued Control Epidemic profile (2001 US$) UNAIDS estimate of need for 2007 Concentrated epidemic Priorities (East Asia and the Pacific, Percentage Low-level epidemic Europe and Central Asia, Generalized low- Generalized high- of all in (Middle East and Latin America and the level epidemic level epidemic prevention Developing Intervention North Africa) Caribbean, South Asia) (Sub-Saharan Africa) (Sub-Saharan Africa) 2003 US$ millions needs Peer-based No CE studies found United States (E/S/no STIs) Chad (sex workers) No CE studies found 3,696 37 Countries programs US$71,113 per HIV infection, US$6 to US$30 per HIV US$3,556 per DALY infection (Pinkerton and others 1998) US$0 to US$2 per DALY | Stefano United States (E/D/no STIs) (Hutton, Wyss, and N'Diekhor 2003) US$14,934 to US$18,719 per Bertozzi, HIV infection Chad (high-risk men) US$747 to US$936 per DALY US$24 to US$1,476 per HIV Nancy infection (Kahn and others 2001) US$1 to US$74 per DALY S. India (sex workers) Padian, (Hutton, Wyss, and US$52 per HIV infection N'Diekhor 2003) Jeny US$3 per DALY Chad (youths) (World Bank 1999) W US$129 to infinity per HIV egbreit, India (high-risk men) infection US$303 per HIV infection US$6 to infinity per DALY and US$15 per DALY others (Hutton, Wyss, and (World Bank 1999) N'Diekhor 2003) Cameroon (E/D/STIs) US$67 to US$137 per HIV infection US$3 to US$7 per DALY (Kumaranayake and others 1998) Condom promotion No CE studies found No CE studies found No CE studies found South Africa (female condom) 1,093 11 and distribution (M/D/STI) and IEC US$378 to US$4,094 per HIV infection US$19 to US$205 per DALY (Marseille and others 2001) Condom social No CE studies found No CE studies found Chad No CE studies found 198 2 marketing US$77 per HIV infection US$4 per DALY (Hutton, Wyss, and N'Diekhor 2003) STI treatment No CE studies found No CE studies found Chad South Africa (E/STI) 783 8 US$1,675 per HIV infection US$2,093 per HIV infection US$84 per DALY US$105 per DALY (Hutton, Wyss, and (Vickerman and others N'Diekhor 2003) forthcoming) Tanzania (E/S/STI) US$326 per HIV infection US$16 per DALY (Gilson and others 1997) Kenya (E/D/STI) US$11 to US$16 per HIV infection US$1 per DALY (Moses and others 1991) Antiretroviral No CE studies found Mexico (M) Zambia (E) South Africa (M) 320 3 therapy to reduce US$39,230 to US$42,528 per US$848 per HIV infection US$1,650 to US$3,844 per MTCT HIV infection US$34 per DALY HIV infection US$2,124 to US$2,303 per (Stringer and others 2003) US$66 to US$154 per DALY DALY Chad (AZT) (Wilkinson, Floyd, and Gilks (Rely and others 2003) 1998) US$924 to US$4,044 per HIV India infection Sub-Saharan Africa (M) $2,527 per HIV infection US$37 to US$162 per DALY US$5,279 to US$11,444 per $126 per DALY HIV infection (Hutton, Wyss, and (World Bank 1999) US$211 to US$458 per DALY HIV/AIDS N'Diekhor 2003) Chad (breastfeeding advice) (Marseille, Kahn, and Saba 1998) US$1,241 to US$4,382 per Prevention HIV infection Sub-Saharan Africa (nevira- pine) (M) US$50 to US$175 per DALY and US$142 to US$306 per HIV (Hutton, Wyss, and infection Treatment N'Diekhor 2003) US$6 to US$12 per DALY (Marseille and others 1999) | 341 (Continues on the following page.) Table 18.4 Continued 342 | Epidemic profile (2001 US$) UNAIDS estimate of need for 2007 Disease Concentrated epidemic Control (East Asia and the Pacific, Percentage Low-level epidemic Europe and Central Asia, Generalized low- Generalized high- of all (Middle East and Latin America and the level epidemic level epidemic prevention Priorities Intervention North Africa) Caribbean, South Asia) (Sub-Saharan Africa) (Sub-Saharan Africa) 2003 US$ millions needs in MTCT, feeding No CE studies found No CE studies found No CE studies found No CE studies found -- -- Developing substitution Harm reduction for No CE studies found Belarus (E) No CE studies found No CE studies found 241 2 injecting drug users US$353 per HIV infection Countries US$18 per DALY (Kumaranayake and others | 2004) Stefano Russia (E) Bertozzi, US$564 per HIV infection US$28 per DALY Nancy (Bobrik 2004) Drug substitution for No CE studies found No CE studies found No CE studies found No CE studies found -- -- S. Padian, injecting drug users Blood safety 0.01­1 percent HIV prevalence Chad Zimbabwe (E) 230 2 Jeny (M/D/STIs) US$75 to US$151 per HIV US$166 to US$1,010 per HIV W US$374 to US$45,173 per infection infection egbreit, HIV infection US$4 to US$8 per DALY US$8 to US$51 per DALY US$19 to US$2,259 per DALY and (Hutton, Wyss, and (McFarland and others 1995) others (Over and Piot 1996) N'Diekhor 2003) Zambia (E/D/STI) US$215 to US$262 per HIV infection US$11 to US$13 per DALY (Watts, Goodman, and Kumaranayake 2000) Zambia (E) US$41 per HIV infection US$2 per DALY (Foster and Buve 1995) 1­40 percent HIV prevalence US$9 to US$1,806 per HIV infection US$0.45 to $90 per DALY (Over and Piot 1996) Sterile injection Middle East (M) Southeast Asia Africa 94 1 US$393 per DALY US$143 to US$593 per DALY US$91 to US$230 per DALY (Dziekan and others 2003) Americas (Dziekan and others 2003) US$1,851 to US$56,642 per DALY Western Pacific US$953 per DALY (Dziekan and others 2003) Universal No CE studies found No CE studies found No CE studies found No CE studies found 663 7 precautions Antiretroviral No CE studies found United States (E/S/no STIs) No CE studies found No CE studies found 1 1 therapy for US$76,584 per HIV infection prevention and US$3,829 per DALY postexposure prophylaxis (Pinkerton, Holtgrave, and Bloom 1998) Vaccines No CE studies found No CE studies found No CE studies found No CE studies found -- -- Behavior change for No CE studies found No CE studies found No CE studies found No CE studies found 112 1 those who are HIV Source: Authors. -- not available. CE cost-effectiveness. Note: The authors have categorized each of the studies. The first time each study is mentioned, it is identified by whether it was modeled (M) or empirical (E); whether it calculated primary HIV infections averted (S, for static) or if it also showed secondary infections averted (D, for dynamic); and where appropriate, we indicate if the study also looked at the impact on STIs. The cost-effectiveness of these interventions will differ depending on the population to which they are targeted, (that is, mass interventions versus targeted interventions). In addition, the cost-effectiveness of each intervention may vary greatly by study, because each cost-effectiveness study is not uniform. No cost-effectiveness studies of male condom promotion were found, because condom promotion, distribution, and IEC are generally part of a larger program with many components and studies did not distinguish between the costs of individual components of such programs. HIV/AIDS Prevention and Treatment | 343 Box 18.2 Comprehensive Sex Education Versus Abstinence-Only Education The available data on sex education suggest the following: · Abstinence-only education is not effective in promoting · Sex education, including condom promotion, does not healthy sexual behaviors. Programs that promote both encourage or increase sexual activity (Kirby 2001). postponement of intercourse and contraceptive use · Sex education reduces risk and positively affects sexu- were more effective in changing behaviors than those al behaviors. In general, sex education programs that stressed abstinence alone. None of the abstinence- increase knowledge about AIDS and related issues, only programs that have been evaluated demonstrated increase intention to use condoms, and increase con- an overall positive effect on sexual behavior, nor did dom use among sexually active youths (Kim and they affect contraceptive use among sexually active others 1997). participants (Kirby 1997). Source: Authors. methodological issues that will not be repeated here. The the absence of studies to guide the level of investment in reviews agree that the availability of cost and cost-effectiveness IEC, the only reasonable alternative seems to be to imple- analyses for HIV/AIDS prevention strategies is limited and that ment IEC on the basis of data derived from relative levels of the need for such knowledge for planning and decision-making knowledge and understanding in the population. For exam- purposes is urgent. ple, if only 25 percent of the sexually active population were Table 18.4 summarizes available cost-effectiveness estimates able to describe how HIV is transmitted and prevented, for the four UNAIDS epidemic profiles that are described later clearly more IEC would be needed, but if 75 percent of the in table 18.5. The estimates of cost per disability-adjusted life population understood the basic facts about HIV/AIDS, the year (DALY) saved assume a uniform 20 DALYs lost per infected need for additional funding would be diminished. adult (Murray and Lopez 1996) and 25 DALYs lost per infected · School-based sex education. School-based sex education pro- child (Marseille and others 1999) and do not account for the grams, an aspect of IEC, provide information to young peo- increasing proportion of people living with HIV/AIDS in ple and reinforce healthy norms in a school setting developing countries who will have access to antiretroviral (Peersman and Levy 1998). Limited data have shown differ- therapy over the coming years. ences in students who have been exposed to school-based sex education (summarized in table 18.3). Box 18.2 reviews General Interventions Relevant for All Modes of Transmission the effectiveness of abstinence-only education and compre- The following are general interventions not specifically target- hensive sex education, subsets of school-based sex educa- ing the mode of transmission: tion. In light of more recent controlled studies that have not shown an effect on condom use, STIs, or HIV infection, any · Information, education, and communication. This interven- cost-effectiveness estimate is extremely speculative. tion includes education on HIV/AIDS and condom use · Voluntary counseling and testing. This intervention enables through pamphlets, brochures, and other promotional people to know their HIV status and provides counseling materials in classroom or clinic settings or through the support to help them cope with the outcome. Knowledge of radio, television, or press. In general, discerning the effec- serostatus may lead individuals to avoid engaging in risky tiveness of IEC alone is difficult, because IEC is often behaviors (Sweat and others 2000). Cost-effectiveness esti- included in condom promotion and distribution interven- mates of VCT vary widely,and as with many other prevention tions. Here we consider the effectiveness of IEC in concert interventions, these estimates are extremely sensitive to the with condom promotion and distribution. Of all available prevalence of HIV in the population that is seeking testing. prevention interventions, providing information and edu- · Peer-based programs. Peer interventions use influential cation about HIV/AIDS is perhaps the most difficult to members of a targeted community to disseminate informa- assess for cost-effectiveness. Numerous studies have shown tion or teach specific skills. Such interventions have gener- that information alone is typically insufficient to change risk ally been found to be effective in reducing unsafe behaviors. behavior. Accurate information, however, is indisputably the Work on the cost-effectiveness of peer-based interventions basis for informed policy discourse--a vital ingredient in in developing countries has been minimal. In Chad, Hutton, the fight against fear-based stigma and discrimination. In Wyss, and N'Diekhor (2003) reviewed data on 12 prevention 344 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others interventions and integrated them into a comparative analy- strategies to optimize the effectiveness and cost-effectiveness sis. Their findings suggest that peer education for sex work- of condom programs is urgently needed. ers is likely to be highly cost-effective and to entail one-fifth · STI screening and treatment. The latest analyses suggest that the cost of the next most favorable intervention, blood safe- STI control may be most effective as an HIV prevention ty. However, the estimated cost-effectiveness for the same strategy when initiated earlier in the course of national epi- intervention directed toward young people and high-risk demics and when sexual risk behaviors are high (Orroth and men is 33- to 36-fold lower. others 2003). In most developing countries, the greatest benefits from treating STIs almost certainly accrue from Interventions to Prevent Sexual Transmission Below we dis- averting the morbidity and mortality caused directly by STIs cuss the effectiveness and cost-effectiveness of interventions rather than indirectly because of reduced HIV transmission. that target sexual transmission of HIV: Estimates of the cost-effectiveness of STI treatment purely as a way to reduce HIV transmission vary widely. · Condom promotion, distribution, and social marketing. Condom promotion, distribution, and social marketing vary Prevention of Mother-to-Child Transmission The existing by epidemic profile.The evidence on condom promotion and data on the effectiveness and cost-effectiveness of HIV inter- distribution programs indicates that such programs result in ventions target MTCT in order of decreasing cost-effectiveness significantly higher condom use and significantly lower STI as follows: incidence (see table 18.3). Given the central role that condom promotion, distribution, and social marketing has played in · Avoidance of unwanted pregnancies among infected mothers. HIV prevention programs, the lack of data on the relative One of the most effective strategies to reduce HIV among cost-effectiveness of such programs 20 years into their imple- infants is to provide better contraception services.See box 18.3 mentation is striking. It is beyond dispute that the use of a for details. condom by sexual partners who are HIV-discordant is · Use of antiretroviral therapy. Evidence indicates that the pro- extraordinarily cost-effective, given the low cost and high vision of antiretroviral drugs to infected mothers signifi- effectiveness of the condom in preventing HIV transmission. cantly reduces vertical transmission (see table 18.4). The Information on the relative costs and effectiveness of differ- provision of antiretroviral therapy to prevent MTCT is ent approaches to increasing condom use by serodiscordant highly cost-effective, to the point of being cost-saving for sexual partners is not available, with the shortage of informa- women who already know that they are infected. When tion being far more acute for effectiveness than for costs. In screening of women is involved, cost-effectiveness declines the absence of empirical evidence, decision makers are as HIV prevalence falls, because of the larger number of reduced to formulating policy on the basis of theory and women who must be screened to identify an HIV-positive common sense. Even inefficient use of condoms by serocon- woman (Rely and others 2003). cordant couples is likely to be highly cost-effective because of · Feeding substitution. Whereas in high-income countries the the reduction in other STIs, cervical cancer, and unwanted health community recommends complete avoidance of pregnancies. However, more reliable information on breastfeeding for HIV-infected mothers to prevent postnatal Box 18.3 Preventing Mother-to-Child Transmission: Antiretroviral Therapy or Contraception? The differential effect of contraceptive delivery versus available, this number would be reduced to 49. If both antiretroviral therapy in preventing HIV can be shown by strategies were adopted, the number of infected infants comparing the provision of effective contraception and of would be further reduced to 25. nevirapine to a population of 1,000 HIV-infected The greatest difference between providing antiretrovi- women. In the absence of an intervention, approximately ral therapy and providing contraception is the number of 150 infants would be infected with HIV during delivery infants orphaned in the future because their mothers die (Cates 2004). If nevirapine were available, the number of of HIV infection. Three models all come to this conclu- infected infants would be reduced to 82 (the expected sion (Reynolds and others 2004; Stover and others 47 percent decline). If effective contraceptive services were forthcoming; Sweat and others 2004). Source: Authors. HIV/AIDS Prevention and Treatment | 345 HIV transmission, in developing countries the feasibility of through a blood safety program that ensures (a) a national this approach is often limited by such factors as cost, sus- blood transfusion service; (b) the recruitment of voluntary, tainability, lack of safe water, health, and child spacing and low-risk donors; (c) the screening of all donated blood for by sociocultural factors (Coutsoudis 2002). Prolonged HIV; and (d) the reduction of unnecessary and inappropri- breastfeeding more than doubles the likelihood of MTCT ate transfusions (UNAIDS 1997). Available evidence indi- (Nduati and others 2000). Because evidence indicates that cates that HIV screening is effective in reducing HIV infec- mixed feeding (breast milk and formula or other sub- tions (see table 18.4). Blood screening for HIV is costly but stance) has a higher risk of transmission than exclusive has been shown to be cost-effective in numerous studies in breastfeeding (Coutsoudis and others 1999), mothers developing countries (see table 18.3) (Foster and Buve 1995; should be counseled on the superiority of early weaning Hutton, Wyss, and N'Diekhor 2003; Watts, Goodman, and over mixed feeding. Even fewer data are available on the Kumaranayake 2000). The evidence appears to support the cost-effectiveness of feeding substitution. WHO and UNAIDS recommendations that all countries, regardless of the nature of the epidemic in the country, Prevention of Bloodborne Transmission Below we discuss the should implement a comprehensive blood safety program. effectiveness and cost-effectiveness of harm reduction for · Universal precautions. A critical component of standard injecting drug users, implementation of blood safety practices, infection control in health care settings is a prohibition on and provision of sterile injections: reusing needles and syringes. A controversy has recently · Harm reduction for injecting drug users. Harm reduction arisen among researchers who contend that HIV infections involves a combination of health promotion strategies for have been significantly misclassified because of the under- users, including needle and syringe exchange programs, counting of cases that result from unsafe injection practices ready access to effective drug treatment and substitution, by misattributing such cases to heterosexual transmission and provision of counseling and condoms. Brazil, which (Gisselquist and others 2003). However, after much investi- has reduced the incidence of HIV and kept HIV prevalence gation, WHO and the U.S. Department of Health and from reaching projected levels, has relied on strong official Human Services concluded that even though transmission support for harm reduction as a cornerstone of its national caused by unsafe injections may have been underreported, it prevention program (Mesquita and others 2003). A limited nevertheless does not account for an appreciable amount of number of studies have shown significant reductions in HIV transmission (WHO and UNAIDS 2003). Cost- HIV incidence among those exposed to needle exchange effectiveness analyses indicate that a combined policy strat- programs, and several studies have shown significant egy of single-use syringes and interventions to minimize reductions in needle sharing (see table 18.3). Methadone injection use could reduce injection-related infections by maintenance is both safe and effective as a treatment for as much as 96.5 percent, or 8.86 million DALYs between 2000 drug addiction (National Consensus Development Panel and 2030, at an average cost of US$102 per DALY. Additional on Effective Medical Treatment of Opiate Addiction 1998) cost-effectiveness studies are needed to guide decisions and may help reduce the risk of HIV transmission by regarding the optimal choice of technology in this area. enabling individuals to avoid the drug-using behaviors that To prevent bloodborne transmission of HIV and other can lead to HIV infection (Metzger, Navaline, and Woody diseases, health care workers, emergency personnel, and 1998; Needle and others 1998). However, the effect of drug others who might experience occupational exposure to treatment modalities on the rate of HIV transmission is blood or body fluids are advised to take universal precau- currently limited by laws in many countries that prohibit or tions. This approach, which treats all bodily fluids as poten- restrict the use of methadone maintenance or other drug tially infectious, includes the use of gloves, gowns, and gog- substitution strategies. The evidence supporting the cost- gles; the proper disposal of waste; and the use of sterile effectiveness of needle exchange programs in high-income injection and other infection control practices (CDC 1989). countries is strong. However, little has been published in Studies have demonstrated that the use of protective gear, relation to developing countries, partly because these pro- such as gloves, reduces the likelihood of blood exposure in grams have not been as widely implemented as hoped. health care settings. Given the low cost of syringes, the extremely high efficiency Although the cost-effectiveness of implementing universal of HIV transmission by this route, and the demonstrated precautions increases as HIV prevalence increases, universal effectiveness of harm reduction programs in changing precautions are unlikely to be cost-effective in resource- syringe-sharing behavior, needle exchange programs limited settings especially where HIV prevalence is low. should be one of the most cost-effective interventions. Postexposure prophylaxis with antiretroviral agents is · Implementation of blood safety practices. Transmission of considered the standard of care after occupational needle- HIV can be virtually eliminated in health care settings stick exposure to blood from an HIV-infected person. 346 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others Cost-effectiveness analyses of postexposure prophylaxis have people with lower levels of risk behavior. However, the difference been conducted only in high-income countries and have con- in the effectiveness between the two falls as epidemics become cluded that this intervention is not cost-effective (Low-Beer more generalized, and as the average and maximum size of the and others 2000; Pinkerton, Holtgrave, and Bloom 1998). connected components (number of people linked to each other directly or through others by their sexual or injecting risk behavior). Thus, in heavily affected countries, or those where PREVENTION IN THEORY AND PRACTICE: USING the virus has the potential to spread rapidly, prevention inter- EPIDEMIC PROFILES AND CONTEXTUAL FACTORS ventions are likely to become extremely cost-effective even when TO INFORM PREVENTION GUIDELINES targeted at individuals with relatively low levels of risk behavior. Consequently, countries with low-level and concentrated epi- Prevention studies and national experiences over the past demics should emphasize interventions that target individuals 20 years strongly suggest that prevention strategies are likely to at especially high risk of becoming infected or of transmitting be most effective when they are carefully tailored to the nature the virus, whereas countries with generalized epidemics should and stage of the epidemic in a specific country or community. also invest heavily in interventions that target entire populations UNAIDS has developed epidemiological categories for charac- or population subgroups. Thus, any determination of the likely terizing individual epidemics on the basis of prevalence of effectiveness and cost-effectiveness of specific interventions in infection in particular subpopulations and in the general pop- particular circumstances requires an accurate understanding of ulation (table 18.5). the stage and nature of the national epidemic. As a complement to the guidance provided by the epidemic The countrywide successes discussed in boxes 18.4 and 18.5 profile, Grassly and others (2001) recommend assessing the highlight population-level interventions that modify social prevalence of other STIs; estimating the extent of mixing norms as well as highlighting legislative and economic factors. between high- and low-risk groups (for example, men who Other examples include instituting government regulation of have sex with men who have sexual contact with female brothels and interventions to change social norms among sex partners); and estimating the prevalence of high-risk sexual workers in Thailand, implementing national sex education and behaviors in the population (such as lack of condom use with blood safety programs in Senegal in concert with creating a casual partners). They also cite two other critical contextual national registry of sex workers, and mandating involvement factors: the capacity of the health service and the social, eco- by women in politics in Uganda. nomic, and legislative context, including social norms and atti- tudes about sexual and drug use behaviors and the acceptance of breastfeeding. Contextual factors that may play a role in Low-Level Epidemic. Providing widespread VCT, screening the success of interventions include the status of women, the for STIs, universal precautions, and postexposure prophylaxis stigmatization of high-risk groups, and the presence of armed may not be cost-effective in a low-level epidemic. In this set- conflict and social upheaval. Together, the epidemic profile and ting, such as in the Middle East and North Africa, HIV/AIDS the context in which the epidemic occurs suggest various pre- control strategies should emphasize the following: vention strategies. · surveillance and individual-level interventions that target General Prevention Guidelines by Type of Epidemic key populations Generally, it is more important to change the behavior of people · IEC, including limited education through the mass media who have high levels of risk behavior than it is to change that of and sex education in schools Table 18.5 Epidemic Profiles Highest prevalence in a key Prevalence in the general Extent of HIV infection populationa (percent) population (percent) WHO region Low level 5 1 Middle East and North Africa Concentratedb 5 1 East Asia and the Pacific, Europe and Central Asia, Latin America and the Caribbean, South Asia Generalized low level 5 1­10 Sub-Saharan Africa Generalized high level 5 10 Sub-Saharan Africa Source: Adapted from UNAIDS 2004. a. Key populations include sex workers, men who have sex with men, and drug injecting users. b. We consider three types of concentrated epidemics depending on the key population most affected: sex workers, men who have sex with men, or drug injecting users. HIV/AIDS Prevention and Treatment | 347 Box 18.4 Thailand's 100 Percent Condom Program Thailand's HIV prevalence, fueled primarily by high rates · government-distributed condoms of commercial sex work and low levels of condom · STI testing and treatment use, began to rise rapidly in the late 1980s. Beginning in · surveillance and tracking of infections to points of 1989, the Thai government initiated a nationwide condom origin distribution and education campaign focusing on com- · strong political and financial commitment mercial sex workers and their clients to ensure 100 percent · active involvement of provincial and local governments. condom use in all commercial sex encounters. Elements thought to contribute to the program's success include Despite this unprecedented success, evidence indicates that enforcement of the 100 Percent Condom Program is · government-mandated 100 percent condom use in not as strong today as when it was initially implemented. commercial sex establishments A recent study in Bangkok found that 89 percent of sex · mass condom promotion advertising campaign workers used condoms, a decline from 96 percent in 2000 · education in commercial sex workplaces (UNDP 2004). Source: Authors. Box 18.5 Uganda HIV/AIDS Prevention Program Like many countries in Sub-Saharan Africa, Uganda expe- · implementing interventions to empower women and rienced a rapid increase in HIV incidence and a general- girls ization of the epidemic in the late 1980s and early 1990s. · having a strong focus on youths By 1991, overall HIV prevalence was 21 percent (Low-Beer · engaging in active efforts to fight stigma and discrimi- and Stoneburner 2003); however, the trajectory of nation Uganda's epidemic has differed markedly from that of its · emphasizing open communication about HIV/AIDS neighbors. By 2001, overall HIV prevalence had fallen to · engaging the religious leadership and faith-based 5 percent, with dramatic decreases in incidence among key organizations populations, such as soldiers, pregnant women, and young · creating Africa's first confidential VCT interventions women (USAID 2002). Critical components of Uganda's · emphasizing STI control and prevention. HIV prevention program include · having strong political support, especially from President Yoweri Museveni Source: Authors. · prevention programs for people living with HIV/AIDS and · responding to community attitudes toward sexual activity, as harm reduction for injecting drug users they may dictate people's response to sex education materials. · VCT that is available to key populations with the highest levels of risk behavior and infection rates Concentrated Epidemic. In a concentrated epidemic, as in · MTCT prevention to mothers known to be infected countries in East Asia and the Pacific, Europe and Central Asia, with HIV Latin America and the Caribbean, and South Asia, prevention · screening all blood for transfusions and providing sterile priorities should include the following: injections · addressing market inefficiencies in condom procurement · ongoing surveillance and distribution--including strategies such as bulk pur- · subsidized VCT and promotion of VCT among key chases and incentives populations 348 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others · HIV screening of pregnant women, guided by individuals' population-level interventions that can mobilize an entire soci- risk profiles ety so as to address prevention and care at all levels. Prevention · peer-based programs for key populations to educate should include the following: individuals at risk, promote safer behaviors, and distribute condoms · mapping and maintaining surveillance of risk behaviors, · harm reduction for injecting drug users, including needle STIs, and HIV infection exchange and drug substitution programs · offering routine, universal HIV testing and STI screening · STI screening and treatment for key risk groups and universal promotion of treatment · targeted distribution and promotion of condoms to key · promoting condom use and distributing condoms free in all populations with condom distribution linked to VCT and possible venues STI care. · providing VCT for couples seeking to have children · counseling pregnant women and new mothers to make In addition, contextual factors, such as government accept- informed and appropriate choices for breastfeeding. ance of needle exchange programs, incarceration of drug users, · implementing individual-level approaches to innovative and harassment of sex workers, will likely have a major impact mass strategies with accompanying evaluations of on the effectiveness of prevention efforts. Because HIV/AIDS is effectiveness typically concentrated in socially or economically marginalized · using the mass media as a tool for mobilizing society and populations in countries with concentrated epidemics, atten- changing social norms tion to socioeconomic factors and to the stigmatization of key · using other venues to reach large numbers of people effi- populations will also be vital to an effective response. ciently for a range of interventions--workplaces, transit venues, political rallies, schools and universities, and mili- Generalized Low-Level Epidemic. In a generalized low-level tary camps epidemic, such as in some countries in Sub-Saharan Africa (for · establishing official institutional policies to provide for example, Tanzania), the emphasis on targeted interventions harm reduction among injecting drug users. must be maintained or even strengthened. Interventions for In a generalized high-level epidemic, contextual factors-- broader populations must also be aggressively implemented. such as poverty and the fragility of the health care infrastruc- These prevention priorities should include the following: ture--will dramatically affect service provision at every level. · maintaining surveillance of STIs, risk behaviors, and HIV The status of women, an important factor in all epidemics, infections in the entire population, with a particular focus becomes an overriding concern in this setting, requiring priority on young people action to radically alter gender norms and reduce the economic, · extending mass media IEC beyond basic education social, legal, and physical vulnerability of girls and women. · providing routine voluntary and confidential HIV testing and STI screening and promoting treatment beyond key PREVENTION-CARE SYNERGY populations In addition to the benefits antiretroviral therapy has for the indi- · providing subsidized and social marketing of condoms and vidual being treated (Komanduri and others 1998; Ledergerber strengthened distribution to ensure universal access and others 2001), it almost certainly has other effects on popu- · offering HIV screening to all pregnant women lations where therapy is widely available. Effective antiretroviral · broadening peer approaches and targeted IEC to include all therapy appears to decrease the infectiousness of treated indi- populations with higher rates of STIs and risk behavior. viduals. Chemoprophylaxis in exposed, uninfected people may Contextual factors remain critical to the success of preven- reduce transmission. In addition, availability of treatment may tion efforts in generalized low-level epidemics, but population- destigmatize the disease and make prevention programs more level factors now have greater priority. The most important is effective (Castro and Farmer 2005). likely to be the status of women, especially with regard to their However, these benefits in relation to reduced transmission ability to control their sexual interactions, to negotiate VCT, to may be offset by a "disinhibition" of risk behavior that is asso- be protected from abuse, and to have property rights following ciated with greater availability of antiretroviral therapy, by the the death of a spouse. spread of drug-resistant HIV, or by increases in the incidence of exposure to partners with HIV infection because of increased Generalized High-Level Epidemic. In a generalized high-level survival. These sometimes opposing effects of offering therapy epidemic, such as in some countries in Sub-Saharan Africa (for may differ to such a degree that the net effects of widespread instance, Botswana and Zimbabwe), an attack on all fronts is therapy on transmission rates may vary among risk groups and required. Prevention efforts should focus on broadly based, across geographic regions. HIV/AIDS Prevention and Treatment | 349 Table 18.6 Effect of Antiretroviral Therapy on Transmission Dynamics Area or behavior affected Treatment effects expected to decrease transmission Treatment effects expected to increase transmission Viral load Decreased infectiousness of the treated partner is substantial even As survival increases, the incidence of exposure to with monotherapy (Musicco and others 1994). Transmission after partners with HIV infection may increase (Hammer exposure to individuals with a viral load of less than 1,500 copies and others 1997). per milliliter is extremely rare (Quinn and others 2000). No cases of sexual transmission from a partner with undetectable viremia have been reported. Prophylaxis Decreased susceptibility may occur during postexposure prophylaxis None. (Cardo and others 1997). Drug resistance Impaired fitness and decreased viral load during drug-resistant Impaired virological responses to therapy in the person viremia (Deeks and others 2000) appear to allow persistent who is infected by a resistant virus may partially offset decreases in infectiousness even after drug resistance has the beneficial effect on infectiousness (Little and others occurred (Leigh Brown and others 2003). 2002; Grant, Kahn, and others 2002). However, primary infection with a resistant virus may also be associated with slower progression of the disease (Grant, Hecht, and others 2002). Risk behavior Treatment may provide incentives for HIV testing and counseling, Decreased fear of HIV and disinhibition of risk behavior which has been associated with decreased risk behavior and HIV are possibilities (Katz and others 2002). Risk behavior by incidence. The availability of treatment may reduce stigma directly, people who are sick and who recover their health status and also indirectly by increasing the visibility of people living with may increase (Stolte and others 2001). HIV/AIDS. Risk reduction counseling during treatment programs may reduce risk behavior. Sexual networks Decreased fear of HIV may foster more informed risk behavior, Decreased fear of HIV may disinhibit risk behavior, reduce including increased use of testing and more thoughtful partner serosorting, and increase mixing between higher- and selection, including serosorting and sorting by risk level lower-risk groups in the population. (McConnell and Grant 2003). Epidemiological The effective prevalence of infectious people will decrease because Treatment-induced reduction in mortality may increase the of treatment effects on infectiousness or increased serosorting. prevalence of infection, although many being treated will be less infectious or better informed regarding risk reduction strategies. A rebound of viral load with treatment failure may mean that treatment postpones transmission rather than reducing it. Source: Authors. Table 18.6 reviews the information available on the popula- clinical care in resource-limited countries was confined to tion effects of antiretroviral therapy and makes suppositions managing the terminal stage of infection, including extremely about potential effects for those areas for which data and late diagnosis of opportunistic infections and cancers, use of research are lacking. The information in the table suggests that basic palliative symptom management, and short-term hospi- widespread therapy using currently available combination reg- talization just before death. Few people were aware of their HIV imens will provide a net benefit in relation to the transmission status until the onset of severe HIV-associated illness, and most of HIV. However, because confidence in this prediction is not did not seek help from the health care system until they were high, the population consequences of therapy programs must already terminally ill. be evaluated and monitored with active surveillance of pre- The advent of primary prophylaxis and treatment for scribing patterns, sexual risk behavior, STI prevalence, HIV opportunistic infections, including tuberculosis, prolonged incidence and prevalence, and prevalence of primary drug survival to a limited extent but did nothing to restore immune resistance and sexual networks of risk behavior. function. Such restoration was not possible until the advent of antiretroviral therapy. Because clinical intervention in HIV CARE AND TREATMENT is so recent in resource-limited settings, few cost-effectiveness studies are available. Those that are available on the treatment This section reviews evidence of the cost-effectiveness of of and prophylaxis for opportunistic infections were largely HIV/AIDS care and treatment interventions in resource- conducted before the availability of antiretroviral therapy and limited settings. Until relatively recently, the majority of HIV therefore need to be reestimated to be relevant for decision 350 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others making today. Fortunately, because the determinants of biolog- abundant, following up an initially positive ELISA with a sec- ical responses are better conserved across countries and cul- ond ELISA--and even a Western blot test if the second ELISA tural settings than the determinants of behavior, effectiveness is positive--may be appropriate (this is typically done in high- data from high-income countries can help inform decisions income countries). about treatment in resource-limited settings. However, in a high-prevalence environment where the prior Unlike drugs for many other high-burden health conditions probability is high and resources are scarce, such an approach in developing countries, antiretroviral therapy for HIV and is almost certainly not cost-effective. Each additional confirma- drugs for some of its associated opportunistic infections depend tory test decreases the number of false positive results, thereby on medications that are still under patent protection. averting the costs associated with such a result. The costs of Nevertheless, generic drug makers in India and Thailand have averting a false positive result range from US$425 with a single produced a range of effective antiretroviral therapies that com- confirmatory rapid test or ELISA to more than US$500,000 for bine multiple drugs into single tablets and reduce the pill burden a confirmatory Western blot test following two positive ELISAs to one tablet twice daily. These companies have made it possible as the prevalence of HIV in patients who are clinically suspected for prices to drop dramatically for some antiretroviral therapy of being infected is varied from 5 to 50 percent (these calcula- combinations--to less than US$250 per year, compared with tions are based on assumptions in John Snow, Inc. 2003 and more than US$4,000 for the same combinations (from the orig- WHO 2004). These results suggest that LMICs should not use inal manufacturers) in high-income countries. In response to a second confirmatory test unless the prevalence among this threat, some multinational pharmaceutical companies have patients is extremely low. introduced a system of price differentiation among countries depending on their per capita income and HIV/AIDS burden. Palliative Care In addition, the World Trade Organization's Agreement on Palliative care has traditionally focused on patients in the ter- Trade-Related Aspects of Intellectual Property Rights (TRIPS) minal stages of disease. More recent definitions of palliative includes a provision that permits compulsory licensing of care, including WHO's definition, have been broadened to pharmaceutical products in cases of national emergency and encompass quality-of-life issues of patients and their families other circumstances of extreme emergency, which is clearly the throughout the course of a life-threatening illness (WHO case for HIV/AIDS in much of the developing world. A 2003 2002b). The control of pain and other symptoms is the crux of World Trade Organization decision also made it easier for any palliative care model, but the WHO model also addresses low- and middle-income countries (LMICs) to import cheaper patients' and their families' psychological, social, and spiritual generics made under compulsory licensing if the countries are problems. Under this definition, in many developing countries, unable to manufacture the medicines themselves (WTO 2003). most people living with HIV/AIDS are not receiving the mini- As a result, some countries, including Brazil, India, and mum standard of palliative care. Of the 5 million people living Thailand, have begun to produce generic versions of antiretro- with HIV/AIDS in South Africa, one of the wealthiest countries viral drugs to be sold at greatly reduced prices. The TRIPS pro- in Sub-Saharan Africa, Carlisle (2003) estimates that only vision has also improved developing countries' bargaining 250,000 have access to palliative care services. In the face of a power with large pharmaceutical companies, to the point that growing epidemic of historic dimensions, the provision of some countries have been able to secure drugs from the origi- comprehensive palliative care represents a critical, but neg- nal manufacturers at substantially reduced prices. As a result, lected, global priority. the relative cost-effectiveness of different drug combinations Health care professionals have promoted community has been in rapid flux, increasing the importance of updating home-based care as an affordable way to expand the coverage recommendations frequently. of palliative care (Hansen and others 1998), but the great het- erogeneity among home-based care programs complicates Diagnostic HIV Testing comparisons. Most programs for which data are available A positive HIV test can be confirmed within one month of are community-based outreach programs administered by infection. Infection is diagnosed in two ways: by a biological local clinics or hospitals. These programs can consist of simple test that detects the presence of HIV antibodies or by diagnosis home visits to provide basic care for AIDS patients or may be of an opportunistic infection that is a clear sign of HIV disease. comprehensive schemes that provide care, palliative medica- The most widely used biological test in high-income countries, tions, meals, psychosocial support and counseling, and links to conducted in a laboratory on a blood sample, is called an primary and secondary health care. ELISA (enzyme-linked immunosorbent assay). Obtaining a Studies indicate that home-based care has considerable result may take several days. Rapid tests that can provide results potential to deal cost-effectively with the palliative care needs in 20 minutes are being used more widely as their costs fall. of HIV/AIDS patients (Ramsay 2003; UNAIDS 2001; Uys and When the prior probability of infection is low and resources are Hensher 2002; Wenk, Bertolino, and Pussetto 2000). Although HIV/AIDS Prevention and Treatment | 351 a Zimbabwe study found that home visits were associated with (2004) study of U.S. women demonstrated that the use of men- extensive travel time and costs (Hansen and others 1998), little tal health services was associated with reduced mortality and research has examined the extent to which home-based care that AIDS-related deaths were more likely among women who can be used to substitute for hospitalization, nor is evidence had symptoms of chronic depression. While results have not available to determine the most cost-effective combination of been replicated in resource-constrained countries, an assess- palliative care strategies. Most people living with HIV/AIDS do ment of clinic-based psychosocial support and counseling serv- incur some end-of-life costs in the formal health care sector. In ices in northern Thailand showed that 50% of PLWHA became one South African study, primary care clinic and hospital costs more positive about their lives and 40% stated that they learned accounted for 39 and 18 percent, respectively, of the costs of how to live with the disease (Tsunekawa and others 2004). care in the last year of life, whereas community home-based Although few data are available on the costs of various strate- care accounted for 42 percent (Uys and Hensher 2002). gies, interventions for psychosocial support appear to be cost- Higginson and others' (2003) meta-analysis concludes that effective--especially where innovative solutions, such as group overall evidence demonstrates a positive effect of home-based counseling sessions, are implemented. Although studies indi- palliative care, especially its effect on pain management and cate an improved quality of life for these patients, little infor- symptom control. Available data do not permit estimating a mation is available on the cost of the interventions. Additional cost per DALY of community-based palliative care programs, evaluation research is needed to guide decisions about how but a review of available studies suggests that palliative care much to invest in psychosocial support. provided by health professionals in the home is unlikely to be cost-effective in low-income countries. However, low-cost, Nutrition Programs and Food Security. Strong evidence community-based models have been developed that require indicates that malnutrition and AIDS work in tandem at both minimal external resources and function almost like care coop- the individual and the societal levels. Infection with HIV eratives among affected households. These models are likely to increases the risk of malnutrition in the individual, while mal- be highly cost-effective. nutrition worsens the impact of HIV and AIDS. Similarly, HIV/AIDS can both cause and be worsened by food insecurity. Symptom-Based Care. Pain management is extremely This reciprocity must be considered when planning specific important in HIV and is addressed in chapter 52. Diarrhea, nau- program responses. sea, vomiting, and skin problems are all symptoms that are tar- Protein deficiency is a well-known cause of cell-mediated geted for treatment in palliative care. Oral rehydration for diar- immunodeficiency (Vanek 1953). HIV-infected individuals rheal treatment costs pennies per episode. Nausea and vomiting need to consume more energy than uninfected individuals: as are prevalent in people with AIDS and can lead to anorexia and much as 10 percent greater consumption for asymptomatic weight loss (UNAIDS 2000). Treating nausea costs an estimated individuals and 20 to 30 percent more for symptomatic indi- US$1.75 per episode (Willbond and others 2001), and contin- viduals. Malnutrition alters the susceptibility of individuals to uous treatment of nausea and vomiting in end-stage patients HIV infection and their vulnerability to its various sequelae, costs about US$2 per day (World Bank 1997). increases the risk of HIV transmission from mothers to babies, Approximately 90 percent of people with HIV suffer from and accelerates the progression of HIV infection (Gillespie, some form of skin condition. These conditions include infec- Haddad, and Jackson 2001). tions, drug reactions, scabies, pressure sores, and cancers. Skin Small studies of adults with AIDS, including those on anti- often becomes dry in the middle and late stages of AIDS retroviral therapy, have shown that daily micronutrient supple- because of dehydration caused by persistent diarrhea, vomit- mentation increases bodyweight, reduces HIV RNA levels, ing, and malabsorption. The cost of treating an episode of skin improves CD4 counts, and reduces the incidence of rash is estimated to be US$2 (UNAIDS 2000). No estimates are opportunistic infections. Fawzi and others' (2004) large trial available on the benefits of providing such care in terms of among pregnant women infected with HIV in Tanzania DALYs, especially to terminally ill patients. demonstrates that multivitamin supplements (a) decrease the Psychosocial Support. Psychosocial support is an integral risk of progression to WHO stage 4 (progression from HIV to component of the multidisciplinary management strategies AIDS, the most advanced level of HIV infection) or death from that care providers regard as essential for people with HIV AIDS-related causes and (b) reduce many HIV-related symp- (Murphy and others 2004). Support for patients and families toms. The multivitamins used in the trial cost US$15 per can have a positive effect on adherence to therapies and can person per year (Fawzi and others 2004). contribute to the critical aim of integrating prevention with The World Food Program guidelines prioritize three nutri- treatment and care. tion interventions for people living with HIV/AIDS: counseling Psychosocial support and counseling has a positive effect on on specific behaviors, prescribed or targeted nutrition the quality of life of people living with HIV/AIDS. Cook's supplements, and links with food-based interventions and 352 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others programs. The guidelines cite three types of nutrition supple- activities (Van Liere 2002). Chapter 56 estimates that sustained ments: food rations to manage mild weight loss and nutrition- community nutrition programs would save US$200 to US$250 related side effects of antiretroviral therapy and to address per DALY. Such programs targeted at communities at especially nutritional needs in food-secure areas; micronutrient supple- high risk are likely to be even more cost-effective (World Food ments for specific HIV-positive risk groups; and therapeutic Programme 2001). foods for addressing moderate and severe malnutrition in HIV- positive adults and children. Cost-effectiveness data in support of these recommendations are not available, but the low costs Treatment of Opportunistic Infections of supplementation, coupled with the likely benefits to other and Secondary Prophylaxis malnourished household members, suggest that such interven- Even as the availability of antiretroviral therapy increases in tions will be highly cost-effective. many developing countries, appropriate diagnosis and man- Infection with HIV/AIDS can severely undermine an indi- agement of life-threatening opportunistic infections, including vidual's food security, affecting the availability, stability, access HIV-associated cancers, remain the most important aspects of to, and use of essential foods. The epidemic is stunting progress the care of patients with HIV disease. Opportunistic infections in rural development and causing significant increases in rural usually begin five to seven years after infection (Munoz, Sabin, poverty and destitution in the countries most affected by the and Phillips 1997) and occur progressively as uncontrolled epidemic (Bonnard 2002). Thus, interventions must consider HIV replication destroys the immune system (Colebunders the epidemic's impact on the broader community and not and Latif 1991). Figure 18.1 describes the cascade of infections solely on people living with the disease. Care-related household that occur as the immune system is depleted. Opportunistic and community-level interventions include school feeding infections are typically caused by organisms that exist in the with special take-home rations for families caring for orphans, environment of the body (on the skin, in the lungs and gas- food for training programs that promote income-generating trointestinal system) and remain latent until HIV has impaired activities, and food for work to support homestead production the immune system. Immune depletion: Early Intermediate Advanced (CD4 > 500) (500 > CD4 > 200) (CD4 < 200) CD4 cell count per L 1,000 Fever Guillain-Barré syndrome Myalgia Chronic demyelinating Arthralgia neuropathy Tinea Adenopathy Idiopathic thrombocytopenia Seborrhoeic dermatitis Malaise Reiter syndrome 500 Gingivitis Rash Polymyositis Warts, molluseum contaglosum Meningo- Bell palsy Tuberculosis encephalitis Sjogren syndrome Herpes zoster Sinusitis Herpes Oral candidiasis simplex Cryptosporidiosis Hairy leukoplakia 200 PCP Kaposi sarcoma Toxoplasmosis Non-Hodgkin Lymphoma Cryptococcal meningitis Cervical intraepithelial neoplasia Mycobacterium avium complex Primary CNS Non-Hodgkin Lymphoma Cytomegalovirus 0 0 10 weeks 5 years 10 years Source: Authors. Figure 18.1 Cascade of Infections and Cancers That Develop as Immune Function Is Depleted HIV/AIDS Prevention and Treatment | 353 The epidemiology of opportunistic infections is complex; it 1995) in some individuals, whereas some long-term survivors is related to the severity of individual immune depletion and remain disease free for longer than 15 years (Easterbrook shows considerable intercountry variation. Each infection has 1994). In developing countries, disease progression, though not its unique clinical expression, requiring specific diagnostic as well studied, appears to be more rapid (Morgan and others techniques and treatment. Many opportunistic infections can 1997). Once an AIDS-defining illness occurs, the average time be prevented by judicious use of chemoprophylaxis, ranging to death seems to be similar across countries, reported at from the low-cost (cotrimoxazole to prevent Pneumocystis approximately 12 to 18 months in Uganda and the United jiroveci pneumonia [PCP] at less than US$20 per year) to the States (Carre and others 1994). extremely expensive (ganciclovir to prevent cytomegalovirus at The time from presentation with an AIDS-defining oppor- more than US$10,000 per year) (Schneider and others 1995; tunistic infection to death depends on the type of infection, the Spector and others 1996). In high-income countries, antiretro- availability of care, and the patient's adherence to prescribed viral therapy has so effectively controlled viral replication that prophylaxis and treatment. Even as access to antiretroviral the process of HIV-related immune destruction has been therapy increases, prophylaxis for opportunistic infections slowed or halted, leading to marked declines in the incidence remains one of the most important ongoing and successful care of opportunistic infections and a dramatic reduction in their strategies for patients with advanced HIV disease. In high- resultant high death toll (McNaghten and others 1999). income countries, the widespread use of such simple inter- Unfortunately, the emerging problem of poor adherence to ventions as cotrimoxazole for PCP prophylaxis has had a drug regimes is now making HIV resistance to antiretroviral significant effect in delaying the onset of PCP, the most com- therapy more prevalent in high-income countries, triggering a mon initial AIDS-defining event, thus positively influencing resurgence of opportunistic infections. survival (Hoover and others 1993). However, prophylaxis for More than 20 infections and cancers have been associated opportunistic infections appears to be underused in LMICs. with severe immune depletion. The most common pathogens Prevention of PCP or any other opportunistic infection and cancers include bacteria such as Mycobacteria tuberculo- does not halt the relentless erosion of the immune system and sis and avium; protozoa such as Cryptosporidium, Strongyloides, provides only a short-term prolongation of life (Morgan and and Toxoplasma; fungi such as Candida, PCP, Cryptococcus, others 1997). The only way to halt or delay the progression of Aspergillis, and Penicillium (the latter largely restricted to South HIV disease is to interrupt viral replication. and Southeast Asia); viruses such as cytomegalovirus, herpes simplex, and herpes zoster; and cancers such as Kaposi sarcoma Role of Antiretroviral Therapy in Relation to Opportunistic and non-Hodgkin lymphoma. Infections. Antiretroviral therapy is effective in reducing viral The range of complications arising from continued HIV load and partially enabling immune restoration, thereby pre- infection varies from country to country, reflecting the differ- venting the onset and recurrence of opportunistic infections. ences in infectious agents that populations have encountered If taken strictly according to directions, antiretroviral therapy earlier in life or are exposed to when immunosuppressed. In can induce a sustained recovery of CD4 cell reactivity against high-income countries, the most common opportunistic opportunistic pathogens in severely immunosuppressed infections are PCP, esophageal candidiasis, cytomegalovirus patients (Li and others 1998). The effectiveness of antiretroviral retinitis, cryptococcal meningitis, toxoplasma encephalopathy, therapy is determined by its ability to rapidly reduce viral load cryptosporidium diarrhea, and human herpes virus­8 and and to sustain low levels of viral activity. This viral activity is Kaposi sarcoma (Bacellar and others 1994; Hoover and others what has an independent effect on increasing or decreasing sus- 1993; Lanjewar and others 1996; Selik, Starcher, and Curran ceptibility to opportunistic infections (Kaplan and others 2001). 1987). In resource-limited countries, because of the higher Initiating antiretroviral therapy can also have detrimental background prevalence of infectious agents, it is more com- effects by causing complications from latent or undiagnosed mon to encounter tuberculosis, cryptococcal meningitis, toxo- opportunistic infections, especially in resource-poor settings. plasma encephalopathy, infectious diarrhea, and nonspecific One of the challenges in initiating antiretroviral therapy in wasting (slim disease) (Hira and others 1998; Hira, Dore, and resource-limited settings is that patients tend to present late Sirisanthana 1998a; Sengupta, Lal, and Srinivas 1994). in their illness, usually when they have an opportunistic infec- The time from HIV infection to manifestation of the first tion that prompts them to seek medical care, or in the case of AIDS-defining illness varies within populations. In high- countries with lax pharmaceutical policy, when they buy anti- income countries, reports on the natural history of untreated retroviral therapy from a private pharmacy. It is well docu- HIV infection suggest that AIDS occurs between 7 and 10 years mented that initiating antiretroviral therapy in severely after infection (Alcabes and others 1993; Lui and others 1988). immunosuppressed patients can result in illnesses associated The time can be as short as 24 months (Anzala and others with reconstitution of the immune system (Shelburne and 354 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others others 2005). These illnesses can occur with all presenting risk of 10 percent in the general population to an annual risk of opportunistic infections and may be more serious than the 10 percent for those coinfected with HIV (Pape and others infection itself. The major problem with care of patients in this 1993). Hence, after five years, about 40 percent of HIV-infected situation is that they may believe the illness is a side effect of people with latent tuberculosis will have developed active their antiretroviral therapy and refrain from medicating. disease. Training clinicians to recognize and treat immune reconstitu- tion disease is therefore essential. Primary Prophylaxis for Opportunistic Infections Management of Opportunistic Infections. The three compo- Before the advent of antiretroviral therapy, the use of prophy- nents of effective management of oppportunistic infections are laxis to decrease the risk of acquiring opportunistic infections diagnosis, treatment, and secondary prophylaxis. As immune was the only intervention available to delay the onset of life- function continues to deteriorate, secondary prophylaxis is threatening infections (Kitahata and others 1996). With the required to prevent recurrence of the treated infection. Some of development of antiretroviral therapy in the 1990s, the preva- the most common infections, such as PCP, can be diagnosed lence of many opportunistic infections has been greatly with a reasonable degree of confidence by clinical history and reduced, and the use of prophylaxis has decreased correspond- treated empirically (Kaplan, Masur, and Holmes 2002). Less fre- ingly (Palella and others 2003). Nevertheless, prophylaxis for quently occurring infections often require sophisticated diag- opportunistic infections remains necessary in patients who nostic equipment and skilled clinicians to confirm a diagnosis lack access to antiretroviral therapy, in extremely immunosup- from a wide range of pathogenic possibilities before starting pressed patients until the therapy takes effect, in patients who complex and expensive treatment. For example, toxoplasmosis do not wish to or who cannot take antiretroviral therapy, in can be accurately diagnosed only by a lumbar puncture and CT patients for whom such therapy fails, and in the small group brain scan (and in some cases an MRI), and cryptosporidium of patients who are unable to recover sufficient CD4 cells diagnosis requires specialized laboratory techniques. despite good inhibition of viral replication (Berenguer and The full spectrum of options for treating opportunistic others 2004). Note that extensive clinical research is still being infections in developing countries has not been systematically carried out in relation to the withdrawal of secondary prophy- evaluated for cost-effectiveness. Because of the effect of anti- laxis following immune restoration with antiretroviral therapy. retroviral therapy on both the efficacy of treatment of individ- ual infections and on life expectancy (and therefore on potential DALYs gained from treating a life-threatening infection), the Treatment of HIV Infection with Antiretroviral Therapy limited economic evaluations conducted are already out of date. Combination therapy with multiple antiretroviral drugs is In particular, chronic infections such as Mycobacterium avium associated with prolonged survival. Whereas monotherapies complex and cytomegalovirus may be more effectively treated are associated with one year or less of additional survival, the over the medium term by reversing immunosuppression with survival benefit conferred by combination therapy appears to antiretroviral therapy than by directly treating the infectious be sustainable for extended periods (Palella and others 2003). agent. Other treatment regimens for opportunistic infections Long-term toxicities related to treatment may include athero- that were marginally cost-effective before antiretroviral therapy sclerosis, lipodystrophy, hepatic failure, and cardiac failure. may now become substantially more cost-effective if the patient Researchers are still evaluating the effects of these toxicities on can begin the therapy following treatment of the infection, HIV/AIDS mortality. thereby extending life expectancy. Table 18.7 shows the cost- effectiveness of care and treatment options for opportunistic infections and antiretroviral therapy. Cost-Effectiveness Considerations in the Choice and Initiation In most resource-limited settings, few specialized diagnostic of Antiretroviral Therapy. WHO has issued global guidelines facilities are available for opportunistic infections. Clinicians for scaling up antiretroviral therapy access; the guidelines pro- have little training in the diagnosis and management of com- mote a combination of stavudine, lamivudine, and nevirapine plex opportunistic infections, and laboratory backup is either (as a fixed-dose formulation) as initial therapy. A number of nonexistent or so expensive that end users cannot afford it. The clinical trials have produced results outlining differential efficacy spectrum of opportunistic infections in LMICs is such that for a number of antiretroviral therapy combinations, which pro- most require highly technical facilities for confirmation of vide guidance in the selection of appropriate drugs for treating diagnosis. Consider M. tuberculosis, the most prevalent such HIV (Yeni and others 2004). The preferred first-line medications infection in Thailand. The rate of latent tuberculosis becoming in developing countries are dictated by these considerations, in clinically active in the presence of HIV increases from a lifetime addition to pricing and patent concerns. HIV/AIDS Prevention and Treatment | 355 Table 18.7 Cost-Effectiveness of Care and Treatment for HIV/AIDS Cost-effectiveness (2001 US$/DALY) Before or when initiating Intervention Source antiretroviral therapy Failed or no antiretroviral therapy HIV testing and diagnosis Confirmatory ELISA, Western blot No cost-effectiveness studies found -- -- in developing countries Palliative care Pain alleviation Chapter 52 420/year of pain-free life added 420/year of pain-free life added Symptom-based care No cost-effectiveness studies found -- -- in developing countries Nutrition interventions Chapter 56 200­250 for HIV-negative individuals 200­250 for HIV-negative individuals End-of-life care No cost-effectiveness studies found -- -- in developing countries Treatment of opportunistic infections, per episode Oral candidias Modeling estimates based on 0.5­157 1­394 Esophageal candidiasis efficacy trials reported from 0.4­55 1­165 HIVInsite (CHI, 2005) and drug costs Histoplasmosis 12­77 81­539 (UNICEF and others 2004) Kaposi's sarcoma 6,236­63,700 12,460­127,400 Cryptococcal meningitis 3­86 21­546 Penicilliosis 11­72 76­483 Mycobacterium avium complex 31­51 87­320 Cytomegalovirus 586­995 4,875­5,120 PCP 0.4­5 3­35 Toxoplasmosis 5­44 31­291 Herpes simplex virus 3­32 7­80 Tuberculosis Chapter 16 200­370 50­450 South Africa (Floyd, Wilkinson, and Short-course ambulatory: 2­16 Short-course ambulatory: 2­16 Gilks 1997); Malawi, Mozambique, Short-course hospital: 3­8 Short-course hospital: 3­8 Tanzania (Murray and others 1991); Community-based directly observed Community-based directly observed Uganda (Saunderson 1995) therapy: 14­22 therapy: 14­22 Opportunistic infection prophylaxis PCP Modeling estimates based on 29­1487 590­29,817 Toxoplasmosis efficacy trials reported from 14­412 252­8,265 HIVInsite (CHI, 2005) and drug costs: Mycobacterium avium complex 786­3,604 2,247­18,020 (UNICEF and others 2004) Cytomegalovirus 151,855­972,955 976,209­4.5 million Tuberculosis preventive therapy Uganda (Bell, Rose, and Sacks 1999); 15­300 (Isoniazid, Rifampicin plus 15­300 (Isoniazid, Rifampicin plus Chapter 16 pyrazinamide, Isoniazid plus rifampicin) pyrazinamide, Isoniazid plus rifampicin) Early detection and screening for opportunistic infections HPV screening and treatment South Africa (Goldie and others Direct visual inspection using acetic Direct visual inspection using acetic 2001) acid: 4/years of life saved acid: 4/years of life saved Antiretroviral therapy First-line antiretroviral therapy Sub-Saharan Africa (Marseille, 350 350 Hofmann, and Kahn 2002) Second-line (and subsequent) India (Over and others 2004) 492/patient yeara 492/patient yeara antiretroviral therapy No cost-effectiveness studies found -- -- in developing countries Adherence interventions No cost-effectiveness studies found -- -- in developing countries Monitoring response to No cost-effectiveness studies found -- -- antiretroviral therapy in developing countries Source: Authors. -- not available. a. Antiretroviral therapy for the poorest HIV positive adults. The estimates include the cost of drugs, clinic visits, and laboratory tests for physician monitoring of treatment and assumes 50 percent condom use in the general population. Box 18.6 Antiretroviral Drugs Current antiretroviral drugs can be divided into three · Nonnucleoside reverse transcriptase inhibitors (NNRTIs) classes: started to be approved in 1997. Like nucleoside analogue reverse transcriptase inhibitors, nonnucleosides also · Nucleoside analogue reverse transcriptase inhibitors interfere with HIV's ability to infect cells by targeting (NRTIs) were the first type of drug available to treat reverse transcriptase. In contrast to nucleoside analogue HIV infection in 1987. When HIV infects a cell, it reverse transcriptase inhibitors, nonnucleosides bind copies its own genetic code into the cell's DNA, and the directly to the enzyme. This blocks the binding site of cell is then programmed to create new copies of HIV. the reverse transcriptase and inhibits the binding of To reproduce, HIV must first convert its RNA into nucleotides. DNA using the enzyme reverse transcriptase. · Protease inhibitors (PIs) were first approved in 1995. Nucleoside analogue reverse transcriptase inhibitors PIs interfere with viral replication by binding to the act like false building blocks and compete with the cell's viral protease enzyme and preventing it from process- nucleosides, thereby preventing DNA synthesis. This ing viral proteins into their functional forms and inhibits reverse transcriptase, which prevents HIV thereby rendering the resulting viral particles non- from infecting cells and duplicating itself. infectious (Peiperl, Coffey, and Volberding 2005). Source: Authors. In recent years, the most volatile parameter in cost- prescribed protease inhibitor­based regimen fails within a year effectiveness analyses for HIV/AIDS has been the prices of anti- (Deeks and others 1999). As a result, the cost-effectiveness of a retroviral drugs, which have dropped by about two orders of regimen is a function not only of its effectiveness in isolation, magnitude for some LMICs. Price reductions have not been but also of its impact on the effectiveness of future regimens. consistent across countries, nor have they necessarily been Thus, the comparative cost-effectiveness of different sequences larger for the poorest countries. This variability in pricing of regimens needs to be considered. greatly complicates the establishment of national guidelines The effectiveness of antiretrovirals depends on not only the regarding which regimens to prescribe under which circum- benefits conferred but also the associated side effects, the toxic- stances, because the ranking of regimens varies among and ity level of the drugs, and patients' adherence to the drug regi- within countries as relative prices change. Box 18.6 discusses men. The ability of care providers to detect incipient toxicity at the three classes of drugs used in antiretroviral therapy. an early stage also influences the magnitude of side effects and Because of their higher manufacturing costs and their toxicities. In low-income settings with limited laboratory more recent introduction into the market, protease inhibitors capacity, a greater proportion of side effects will not be detect- are more expensive than either nucleoside reverse transcrip- ed until they become severe. As a result, the relative cost- tase inhibitors or nonnucleoside reverse transcriptase effectiveness profiles will change depending on the availability inhibitors. They are also more difficult to manufacture, mak- of toxicity monitoring. ing them less attractive to generic manufacturers. Although Initiating antiretroviral therapy has a proven benefit for the difference is less marked, nucleoside reverse transcriptase patients with a CD4 count of fewer than 350 cells per cubic inhibitors tend to cost less than nonnucleoside reverse tran- millimeter (Palella and others 2003). In patients with a scriptase inhibitors. higher CD4 count, the benefits of antiretroviral therapy are Ranking different antiretroviral therapy regimens by their believed to be outweighed by the toxicities that may accrue cost-effectiveness is more complex than doing so for most ther- from continued drug exposure (Mallal and others 2000). apeutic situations, because a high proportion of patients will Concerted research efforts are needed to gauge both the aver- develop resistance to or intolerance of initial therapy and will age costs of care and the survival benefits of identifying need to stop their initial regimen and then initiate a second patients and initiating antiretroviral therapy while their (and perhaps a subsequent) regimen, if available. One U.S. immune function is still competent, compared with the costs cohort study suggests that for 50 percent of patients the and survival benefits associated with starting care late, on HIV/AIDS Prevention and Treatment | 357 presentation of an opportunistic infection--as is currently the ability to adhere to the prescribed therapy, including limited norm in LMICs. education and the consequent poorer understanding of their disease state, unstable housing and financial circumstances, a limited number of treatment options, and clinicians with lim- Drug Resistance. Drug resistance occurs as the virus evolves ited antiretroviral therapy treatment experience (Kitahata and to escape the inhibitory effects of antiretroviral drugs. The others 1996). Those factors, in addition to the toxicity of the capacity of HIV to mutate is extraordinary, as the wide therapy, influence adherence and future disease progression diversity of HIV variants that occurs worldwide demonstrates. rates (Duran and others 2001) and lead to an increase in drug Viral diversification is driven by low-fidelity enzymes (which resistance. Thus, poorly coordinated scale-up of antiretroviral have a high rate of mutation) that carry out replication of the therapy in some developing countries has the potential to viral genome. jeopardize both the duration of clinical benefit for the first Drug resistance resulting from being infected by a drug- wave of patients who receive substandard care and future resistant HIV strain is known as primary drug resistance. response rates as the prevalence of drug resistance increases Secondary drug resistance develops as a consequence of treat- (Harries and others 2001). ment. Primary HIV drug resistance to nucleoside reverse tran- Studies in India, Mexico, Senegal, and Uganda point to poor scriptase inhibitors, nonnucleoside reverse transcriptase adherence (which for some classes of drugs can be adherence inhibitors, and protease inhibitors has been reported of less than 95 percent), inadequate doses and regimes, and (Salomon and others 2000; Wegner and others 2000). The first poor monitoring as factors that contribute to more rapid reports of transmission of drug resistance have typically development of antiretroviral therapy resistance (Oyugi and occurred within a few years of a drug's introduction into clin- Bangsberg 2004, Laniece and others 2004, Bautista and others ical practice. The proportion of newly infected people who 2003, Liechty and Bangsberg 2003). By contrast, experiences in acquire drug-resistant HIV has implications for the choice of Haiti and Uganda suggest that it is possible to achieve adher- first-line regimen. Primary resistance in recently infected indi- ence rates in developing countries equal to or better than those viduals in high-income countries is stable or has been in observed in high-income countries (Farmer and others 2001; decline since 2000, following a rise between 1996 and 1999. Mitty and others 2002). Almost nothing is known regarding primary drug resistance among those recently infected in low-income countries, although this question will become more important with the Second-Line and Subsequent Therapies. Studies from high- increased availability of antiretroviral therapy in resource-lim- income countries have unequivocally demonstrated that the ited settings. probability that an antiretroviral therapy regimen will achieve Drug resistance is associated with increases in plasma viral viral suppression diminishes with each subsequent regimen RNA levels and attenuation of the responses of CD4 counts to (Deeks and others 1999). Similarly, the mean duration of viral therapy. Nonetheless, clinical and epidemiological observations suppression for those who achieve suppression is also lower for suggest that drug resistance does not completely offset the ben- subsequent regimens (Deeks and others 1999). This finding is efits of therapy (Deeks and others 1999; Ledergerber and oth- entirely expected because failing a previous regimen is associ- ers 1999). Individuals with drug-resistant HIV typically have ated with lower adherence, higher toxicity, or side effects and plasma viral RNA levels that remain 3- to 10-fold lower than increased resistance, all of which increase the probability of pretreatment levels. Furthermore, patients with drug resistance similar problems occurring with subsequent regimens. Thus, experience more rapid immunological decline and disease pro- the expected survival benefit per month of antiretroviral ther- gression if they discontinue their drugs (Nijhuis, Deeks, and apy declines with each change of regimen. In contrast, the Boucher 2001). monthly cost of therapy rises as a patient moves from first-line to more expensive protease inhibitor­based second-line and subsequent therapies. Given this steadily declining cost- Importance of Adherence to Prescribed Therapy. With cer- effectiveness, wealthier countries are likely to offer a greater tain drugs, resistance can develop in as little as two weeks if number of regimen changes than poorer countries. therapy is suboptimal (which can be less than 90 percent adherence). Conversely, patients who adhere to therapy can obtain continued viral suppression for many years without the need for second- or third-line options. Research has Laboratory Monitoring of Immune Function shown that drug adherence is one of the most important pre- to Guide Therapy dictors of continued treatment response (Mannheimer and Laboratory monitoring determines when antiretroviral therapy others 2002). Patients in resource-limited countries are likely should be initiated and when it should be changed because of to be subjected to a number of influences that challenge their toxicity, lack of efficacy, or resistance. The optimal frequency 358 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others and precision of monitoring depends on numerous factors, monitoring or body mass index monitoring rather than being principally the following: compared with no monitoring at all. · the expected rate of change of variables of interest Testing for Primary Resistance. Testing for resistance in indi- · the expected frequency of events, such as development of vidual patients is still costly, because of both the cost of the resistance, adherence failure, and side effects diagnostic kit and the sophisticated laboratory capacity · the relative cost of monitoring versus the cost of providing required to perform the tests. Because primary resistance is far ineffective treatment less prevalent in LMICs than in high-income countries, no · the magnitude of the secondary effects of monitoring serious consideration is being given at this time to initiating (motivating prevention, motivating adherence). individual resistance testing in the developing world. However, the choice of optimal first-line and subsequent treatment WHO has suggested a pragmatic approach to monitoring, strategies should be guided by information about the preva- with inexpensive, easy-to-measure parameters (bodyweight or lence of primary resistance to different antiretroviral drugs body mass index, body temperature, hemoglobin, liver in a particular country, which indicates that population- enzymes, and clinical symptoms) for monitoring in low- level monitoring of the prevalence of resistance among income countries. More specialized markers--namely, CD4 antiretroviral-naive people living with HIV/AIDS is important. count, viral load, and resistance genotyping--would be restricted to sentinel sites and tertiary care services (Gutierrez Monitoring Response to Therapy. Ideally, therapeutic failure and others 2004), at least initially. should be detected as soon as possible to permit the imple- The large price reductions for antiretroviral drugs are only mentation of clinical strategies to address toxicity, drug resist- now starting to be mirrored in the costs of monitoring tests as ance, or poor adherence. Therapeutic failure leads to rising new technologies are introduced, collective bargaining is under- viral load and falling immune competence and to the subse- taken, and international pressure mounts on diagnostic manu- quent development of opportunistic infections. Unfortunately, facturers to provide more favorable pricing for LMICs. earlier detection comes at a price: testing for increases in viral Commercial cytometric CD4 measurements are now available load, which can be detected soonest, is more expensive than to some developing countries at less than US$5 per test CD4 testing, which in turn is more expensive than the less sen- (R. Göhde, personal communication, 2004).Viral load testing is sitive monitoring of total lymphocyte count, which is more still significantly more expensive, but even those prices have expensive than monitoring body mass index or waiting until dropped to US$20 following negotiations on behalf of low- clinical signs of failure appear. Where facilities for detecting income countries by the William Jefferson Clinton Foundation. early failure are absent, first-line therapy should be replaced by Even when the potential savings become an operational reality a completely new combination at failure, usually a protease in developing countries, the costs of laboratory monitoring will inhibitor­based combination. still represent an important proportion of the costs of providing antiretroviral therapy. Monitoring Toxicity. Available antiretroviral drugs have sig- nificant toxicity. Such toxicity is often insidious, progressing Monitoring to Guide Initiation of Antiretroviral Therapy. If unnoticed until the patient's health has been seriously laboratory monitoring is performed, its optimal frequency impaired. Examples include zidovudine-associated anemia, must be determined. The closer patients get to an antiretrovi- nevirapine-associated impaired liver function, and didanosine- ral therapy threshold, the more often they must be tested to associated pancreatitis. Fortunately, the most commonly detect a CD4 decline that falls within a specific CD4 range. As encountered serious toxicities can be detected either on clinical use of antiretroviral therapy expands in LMICs and as the costs examination or with inexpensive laboratory tests. Data on the of drugs fall relative to the costs of laboratory monitoring, col- relative cost-effectiveness of different toxicity monitoring lecting empirical data and constructing models to compare dif- regimens are unavailable. Current guidelines identify what ferent monitoring strategies is becoming increasingly urgent. monitoring should be conducted in conjunction with specific In the absence of capacity to perform CD4 counts, several antiretroviral drugs, depending on whether laboratory capacity studies suggest that total lymphocyte count can be used as a is available (WHO 2004). proxy because of the correlation between the two counts (Badri Unfortunately, in the absence of a quantitative analysis of and Wood 2003). Research has also shown that falling body mass the costs of monitoring and the benefits associated with early index is highly predictive of disease progression (Pistone and detection of toxicity, it is difficult to provide guidance on the others 2002). In light of those findings, the cost-effectiveness of minimum laboratory capacity that should accompany the CD4 monitoring in developing countries must be considered delivery of specific treatment combinations. Clearly, extremely in terms of its incremental improvement over total lymphocyte low-cost monitoring tests are warranted for toxicities that HIV/AIDS Prevention and Treatment | 359 occur frequently. The preeminent example is anemia monitor- RESEARCH AGENDA ing for patients receiving zidovudine. Hemoglobin levels can be monitored for less than US$0.02 per test, which is almost As in many other areas of public health in developing coun- certainly cost-effective given that the incidence of anemia tries, a profound tension exists between (a) the need for with zidovudine therapy is approximately 10 percent in research to discover new technologies and interventions for advanced-stage patients and that anemia frequently progresses both prevention and care and (b) the need for research to learn to life-threatening levels if not detected. how to effectively apply the technologies that are currently available. The most important barrier to control is lack of Box 18.7 Interventions in the Pipeline or in Trial The following interventions are currently being developed conducted to assess whether acyclovir reduces HIV or evaluated: infectiousness in individuals infected with both HSV-2 and HIV. · Microbicides. Most microbicide products are currently · Tenofovir for preexposure use. Studies are now enrolling in preclinical development; however, 18 products are participants at three West African sites and will soon being evaluated in clinical research studies, most in begin in Botswana, Malawi, Thailand, and the United small phase 1 safety and acceptability trials. Three States. phase 3 effectiveness trials are currently under way. · Antiretroviral therapy to prevent sexual transmission. · Diaphragms. The safety and effectiveness of the A phase 3, randomized, controlled, multisite trial to diaphragm and Replens gel in preventing HIV and STIs assess whether antiretroviral therapy can prevent among women are being tested in an ongoing phase 3 sexual transmission of HIV in serodiscordant couples randomized controlled trial in South Africa and will begin in Brazil, India, Malawi, Thailand, and Zimbabwe. Two trials, in the Dominican Republic and Zimbabwe. Madagascar, are planned to test the diaphragm's effec- · Vaccines. Although preliminary results from a phase 3 tiveness against bacterial STIs. Several other trials in clinical trial in Thailand found that AIDSVAX failed Sub-Saharan Africa are planned to test the acceptability to protect against infection, several other vaccines are and safety of the diaphragms plus microbicides. being developed. Merck and GlaxoSmithKline have · Circumcision. Two randomized controlled trials are unveiled sizable vaccine programs and moved products under way in Kenya and Uganda to examine whether into human testing. An International AIDS Vaccine circumcision confers protection among adult men. Initiative U.K.-Kenya team is in the midst of interme- · Community-based VCT. Project Accept is a community- diate human trials of DNA/MVA (modified vaccinia based VCT trial in 32 communities in South Africa, virus Ankara), and Aventis Pasteur is taking ALVAC- Tanzania, and Zimbabwe and 14 communities in AIDSVAX into the final phase of trials. The South Thailand. Communities are randomized to receive either African AIDS Vaccine Initiative is preparing for the a community-based VCT intervention or a standard country's first trials, India's prime minister has pledged clinic-based VCT. The community-based VCT interven- national resources for vaccines, and the European tion has three major strategies: to make VCT more avail- Union is broadening its vaccine research for HIV. able in community settings, to engage the community · Behavior change programs for people with HIV. In recent through outreach, and to provide posttest support. years, a growing number of public health experts have · HSV-2 treatment. One study in six countries will deter- proposed implementing prevention interventions that mine the efficacy of twice-daily acyclovir in reducing target people with HIV (De Cock, Marum, and Mbori- susceptibility to HIV infection among high-risk, HIV- Ngacha 2003; Janssen and others 2001), although evi- negative, HSV-2 seropositive women and men who dence on the most effective strategies to encourage have sex with men. A companion study will also be safer behavior among people with HIV is lacking. Source: Authors. 360 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others knowledge about how best to implement packages of existing the prolongation of high-quality life from antiretroviral interventions at the appropriate scale to maximize the effect of therapy--all without damaging existing and often fragile prevention and care interventions and to protect the human health care infrastructure that must also address other health rights of those affected by the epidemic. Accurate surveillance concerns. Although simplified regimens, such as delivering data are needed on risk behaviors, and effectiveness research is multiple drugs in a single tablet and fewer doses per day, are needed to discern what interventions work where and how they desirable everywhere, they are especially important in low- do so. Unfortunately, few rigorous evaluations of new or exist- resource settings. Similarly, low-technology, low-cost monitor- ing interventions have been conducted using large prospective ing tests for antiretroviral therapy toxicity and for immunolog- cohorts, with the result that, for many interventions, convinc- ical and virological responses to treatment are especially ing data on effectiveness are not available. Finally, research on needed in low-income countries, which otherwise must cen- policy or structural interventions, which by definition must be tralize testing--an especially difficult prospect when transport conducted on a population level, is also insufficient. These and communications systems are poorly developed. interventions include the development and testing of such pol- icy tools as changing the tax structure, regulating the sex indus- try, and guaranteeing property rights and access to credit for CONCLUSION women. Box 18.7 lists new prevention interventions in the pipeline. Despite the glaring deficits in AIDS research, the magnitude Although numerous promising interventions are listed, results and seriousness of the global pandemic calls for action in the for most of these strategies are at best years away. Centuries absence of definitive data. The appropriate mix and distribu- hence, when future generations study the history of our time tion of prevention and treatment interventions depends on and the epidemic that killed 50 million or perhaps many more, the stage of the epidemic in a given country and the context in the most difficult question to answer may well be "why did they which it occurs. In the absence of firm data to guide program invest so little for so long in developing a vaccine?" Creating objectives, national strategies may not accurately reflect the such knowledge is about as close as one can get to a pure inter- priorities dictated by the particular epidemic profile, resulting national public good, and the lack of global cooperation in ade- in highly inefficient investments in HIV/AIDS prevention and quately funding such research is an indictment of global com- care. This waste undoubtedly exacerbates funding shortfalls mitment to multilateral cooperation. However, given both the and results in unnecessary HIV infections and premature uncertainty about whether developing an effective vaccine is deaths. The lack of good data--and thus the ability to tailor possible and the long delay until a new vaccine can be widely responses to epidemics--may be somewhat understandable applied, vaccine development efforts must be accompanied by when the burden of disease is minimal and the resources ded- the development of other new biomedical and behavioral pre- icated to it are similarly small. Neither is the case for vention technologies. HIV/AIDS. In contrast, research on care and treatment has been far more successful than research on prevention, and innovation in new therapies continues apace. The ability of HIV to rapidly evolve resistance to antiretroviral drugs, combined with the ACKNOWLEDGMENTS existence of an important market in high- and middle-income We are deeply indebted to Andrew Beggs, Susan Foster, James countries, appears to ensure continued investment in new Kahn, Lilani Kumaranayake, Elliot Marseille, Fern Terris- drug development. In addition, because treatment generally Prestholt, Seema Vyas, and Charlotte Watts for their back- has important commercial returns, HIV therapies, unlike ground papers that have informed this chapter. We also owe behavioral interventions, have benefited the most from private thanks to Sevgi Aral, Geoffrey Garnett, Robin Jackson, Marie sector investment. The paradox is that research on the behav- Laga, Meg Newman, Mead Over, and David Vlahov for their ioral aspects of adherence to drug regimens would improve the work on several sections of this chapter. Finally, we would like effectiveness of antiretroviral therapy, and thereby benefit both to thank Martin Gross and Phillip Machingura for their invalu- commercial and public interests. able contributions throughout this chapter. The greatest research challenges in relation to care and treat- ment in developing countries do not revolve around new drug development. They revolve around how to adapt care and treatment strategies to low-income, low-technology, low­ NOTE human resource capacity settings in ways that maximize adher- 1. See http://www.hivinsite.org/global?page=cr-00-04 for a compila- ence; minimize toxicity, monitoring, and costs; and maximize tion of international guidelines. HIV/AIDS Prevention and Treatment | 361 BIBLIOGRAPHY Bhave, G., C. P. Lindan, E. S. Hudes, S. Desai, U. Wagle, S. P. Tripathi, and J. S. Mandel. 1995. "Impact of an Intervention on HIV, Sexually Agha, S., A. Karlyn, and D. Meekers. 2001. "The Promotion of Condom Transmitted Diseases, and Condom Use among Sex Workers in Use in Non-Regular Sexual Partnerships in Urban Mozambique." Bombay, India." AIDS 9 (Suppl. 1): S21­30. Health Policy and Planning 16 (2): 144­51. Bobrik, A. 2004. "HIV Prevention among IDUs in Russia: A Cost- Alcabes, P., A. Munoz, D. Vlahov, and G. H. Friedland. 1993. "Incubation Effectiveness Analysis." Paper presented at the 14th International Period of Human Immunodeficiency Virus." Epidemiologic Reviews 15 Conference on the Reduction of Drug-Related Harm, April, Chiang (2): 303­18. Mai, Thailand. Anthony, J. C., D. Vlahov, K. E. Nelson, S. Cohn, J. Astemborski, and Bonnard, P. 2002. "HIV/AIDS Mitigation Using What We Already Know." L. Solomon. 1991. "New Evidence on Intravenous Cocaine Use and Technical Note 5, Food and Nutrition Technical Assistance (FANTA) the Risk of Infection with Human Immunodeficiency Virus Type 1." Project, Washington, DC. American Journal of Epidemiology 134: 1175­89. Buve, A. 2002. "HIV Epidemics in Africa: What Explains the Variations in Anzala, O. A., N. J. Nagelkerke, J. Bwayo, D. Holton, S. Moses, and E. HIV Prevalence?" International Union of Biochemistry and Molecular Ngugi. 1995. "Rapid Progression to Disease in African Sex Workers Biology Life 53 (4­5): 193­95. with Human Immunodeficiency Virus Type 1 Infection." Journal of Cardo, D. M., D. H. Culver, C. A. Ciesielski, P. U. Srivastava, R. Marcus, Infectious Diseases 171 (3): 686­89. D. Abiteboul, and others. 1997. "A Case-Control Study of HIV Askew, I., and M. Berer. 2003. "The Contribution of Sexual and Seroconversion in Health Care Workers after Percutaneous Exposure." Reproductive Health Services to the Fight against HIV/AIDS: A Centers for Disease Control and Prevention Needlestick Surveillance Review." Reproductive Health Matters 11 (22): 51­73. Group. New England Journal of Medicine 337 (21): 1485­90. Auvert, B., A. Buve, E. Lagarde, M. Kahindo, J. Chege, N. Rutenberg, and Carlisle, D. 2003. "Africans Are Dying of AIDS without Pain Relief." British others. 2001. "Male Circumcision and HIV Infection in Four Cities in Medical Journal 327 (7423): 1069. Sub-Saharan Africa." AIDS 15 (Suppl. 4): S31­40. Carre, N., C. Deveau, F. Belanger, F. Boufassa, A. Persoz, C. Jadand, and Auvert, B., A. Puren, D. Taljaard, E. Lagarde, R. Sitta, and J. Tambekou. others. 1994. "Effect of Age and Exposure Group on the Onset of AIDS 2005. "Impact of Male Circumcision on the Female-to-Male in Heterosexual and Homosexual HIV-infected Patients: SEROCO Transmission of HIV." Paper presented at the 3rd IAS Conference on Study Group." AIDS 8 (6): 797­802. HIV Pathegenosis and Treatment. Rio de Janeiro, July 24­27. Castro, A., and P. Farmer. 2005. "Understanding and Addressing AIDS- Ayouba, A., G. Tene, P. Cunin, Y. Foupouapouognigni, E. Menu, Related Stigma: From Anthropological Theory to Clinical Practice in A. Kfutwah, and others. 2003. "Low Rate of Mother-to-Child Haiti." American Journal of Public Health 95 (1): 53­59. Transmission of HIV-1 after Nevirapine Intervention in a Pilot Public Cates, W. 2004. "A Funny Thing Happened on the Way to FHI." Sexually Health Program in Yaounde, Cameroon." Journal of Acquired Immune Transmitted Diseases 31 (1): 3­7. Deficiency Syndrome 34 (3): 274­80. CDC (U.S. Centers for Disease Control and Prevention). 1989. Guidelines Bacellar, H., A. Munoz, E. N. Miller, B. A. Cohen, D. Besley, O. A. Selnes, for the Prevention of HIV and Hepatitis B. Atlanta: CDC. and others. 1994. "Incidence of Clinical AIDS Conditions in a Cohort Celentano, D. D., K. C. Bond, C. M. Lyles, S. Eiumtrakul, V. F. Go, C. Beyrer, of Homosexual Men with CD4+ Cell Counts <100/mm3: Multicenter and others. 2000. "Preventive Intervention to Reduce Sexually AIDS Cohort Study." Journal of Infectious Diseases 170 (5): 1284­87. Transmitted Infections: A Field Trial in the Royal Thai Army." Archives Badri, M., and R. Wood. 2003. "Usefulness of Total Lymphocyte Count in of Internal Medicine 160 (4): 535­40. Monitoring Highly Active Antiretroviral Therapy in Resource-Limited Chaisson, R. E., P. Bacchetti, D. Osmond, B. Brodie, M. A. Sande, and A. R. Settings." AIDS 17 (4): 541­45. Moss. 1989. "Cocaine Use and HIV Infection in Intravenous Drug Basu, I., S. Jana, M. J. Rotheram-Borus, D. Swendeman, S. J. Lee, Users in San Francisco." Journal of the American Medical Association P. Newman, and R. Weiss. 2004. "HIV Prevention among Sex Workers 261 (4): 561­65. in India." Journal of Acquired Immune Deficiency Syndrome 36 (3): CHI (Center for HIV Information). 2005. "HIV Insite." University of 845­52. California, San Francisco. http://hivinsite.org. Bautista, S., T. Dmytraczenko, G. Kombe, and S. Bertozzi. 2003. Costing of Colebunders, R. L., and A. S. Latif. 1991. "Natural History and Clinical HIV/AIDS Treatment in Mexico. In Technical Report No. 020. Edited by Presentation of HIV-1 Infection in Adults." AIDS 5 (Suppl. 1): Project PHRplus. Bethesda, MD: Abt Associates, Inc. S103­12. Bell, J. C., D. N. Rose, and H. S. Sacks. 1999. "Tuberculosis Preventive Connor, E. M., R. S. Sperling, R. Gelber, P. Kiselev, G. Scott, M. J. Therapy for HIV-Infected People in Sub-Saharan Africa Is Cost- O'Sullivan, and others. 1994. "Reduction of Maternal-Infant Effective." AIDS 13 (12): 1549­56. Transmission of Human Immunodeficiency Virus Type 1 with Bentley, M. E., K. Spratt, M. E. Shepherd, R. R. Gangakhedkar, Zidovudine Treatment." Pediatric AIDS Clinical Trials Group Protocol S. Thilikavathi, R. C. Bollinger, and S. M. Mehendale. 1998. "HIV 076 Study Group. New England Journal of Medicine 331 (18): 1173­80. Testing and Counseling among Men Attending Sexually Transmitted Cook, J. A., D. Grey, J. Burke, M. H. Cohen, A. C. Gurtman, J. L. Disease Clinics in Pune, India: Changes in Condom Use and Sexual Richardson, and others. 2004. "Depressive Symptoms and AIDS- Behavior over Time." AIDS 12 (14): 1869­77. Related Mortality among a Multisite Cohort of HIV-Positive Women." Berenguer, J., F. Laguna, J. Lopez-Aldeguer, S. Moreno, J. R. Arribas, American Journal of Public Health 94 (7): 1133­40. J. Arrizabalaga, and others. 2004. "Prevention of Opportunistic Coutsoudis, A. 2002. "Breastfeeding and HIV Transmission." In Public Infections in Adult and Adolescent Patients with HIV Infection: Health Issues in Infant and Child Nutrition, vol. 48, ed. R. E. Black GESIDA/National AIDS Plan Guidelines, 2004." Enfermedades and K. F. Michaelsen. Nestle Nutrition Workshop Series. Philadelphia: Infecciosas y Microbiologia Clinica 22 (3): 160­76. Lippincott Williams & Wilkins. Bertozzi, S. M., N. S. Padian, J. Wegbreit, B. Feldman, L. DeMaria, H. Gayle, Coutsoudis, A., K. Pillay, L. Kuhn, E. Spooner, W. Y. Tsai, and H. M. and others. Forthcoming. "HIV/AIDS Prevention and Treatment." Coovadia. 2001. "Method of Feeding and Transmission of HIV-1 from Disease Control Priorities Project Working Paper 39, Bethesda, MD. Mothers to Children by 15 Months of Age: Prospective Cohort Study http://www.fic.nih.gov/dcpp/wps.html. from Durban, South Africa." AIDS 15 (3): 379­87. 362 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others Coutsoudis, A., K. Pillay, E. Spooner, L. Kuhn, and H. M. Coovadia. 1999. Fawzi, W., G. Msamanga, G. Antelman, C. Xu, E. Hertzmark, D. "Influence of Infant-Feeding Patterns on Early Mother-to-Child Spiegelman, and others. 2004. "Effect of Prenatal Vitamin Transmission of HIV-1 in Durban, South Africa: A Prospective Supplementation on Lower-Genital Levels of HIV Type 1 and Cohort Study." South African Vitamin A Study Group. Lancet 354 Interleukin Type 1 Beta at 36 Weeks of Gestation." Clinical Infectious (9177): 471­76. Diseases 38 (5): 716­22. Creese, A., K. Floyd, A. Alban, and L. Guinness. 2002. "Cost-Effectiveness Feachem, R. G. 2004. "The Research Imperative: Fighting AIDS, TB, and of HIV/AIDS Interventions in Africa: A Systematic Review of the Malaria." Tropical Medicine and International Health 9 (11): 1139­41. Evidence." Lancet 359 (9318): 1635­43. Fleming, T. R., and D. L. DeMets. 1996. "Surrogate End Points in Clinical Dabis, F., P. Msellati, N. Meda, C. Welffens-Ekra, B. You, O. Manigart, and Trials: Are We Being Misled?" Annals of Internal Medicine 125: 605­13. others. 1999. "6-Month Efficacy, Tolerance, and Acceptability of a Floyd, K., D. Wilkinson, and C. Gilks. 1997. "Comparison of Cost Short Regimen of Oral Zidovudine to Reduce Vertical Transmission of Effectiveness of Directly Observed Treatment (DOT) and Con- HIV in Breastfed Children in Côte d'Ivoire and Burkina Faso: A ventionally Delivered Treatment for Tuberculosis: Experience from Double-Blind Placebo-Controlled Multicentre Trial." DITRAME Study Rural South Africa." British Medical Journal 315 (7120): 1407­11. Group. Lancet 353 (9155): 786­92. Ford, K., D. N. Wirawan, P. Fajans, P. Meliawan, K. MacDonald, and De Cock, K. M., E. Marum, and D. Mbori-Ngacha. 2003. "A Serostatus- L. Thorpe. 1996. "Behavioral Interventions for Reduction of Sexually Based Approach to HIV/AIDS Prevention and Care in Africa." Lancet Transmitted Disease/HIV Transmission among Female Commercial 362 (9398): 1847­49. Sex Workers and Clients in Bali, Indonesia." AIDS 10 (2): 213­22. Deeks,S.G.,J.D.Barbour,J.N.Martin,M.S.Swanson,and R.M.Grant.2000. Foster, S., and A. Buve. 1995. "Benefits of HIV Screening of Blood "Sustained CD4 T Cell Response after Virologic Failure of Protease Transfusions in Zambia." Lancet 346 (8969): 225­27. Inhibitor­Based Regimens in Patients with Human Immunodeficiency Garcia, P. M., L. A. Kalish, J. Pitt, H. Minkoff, T. C. Quinn, S. K. Burchett, Virus Infection." Journal of Infectious Diseases 181 (3): 946­53. and others. 1999. "Maternal Levels of Plasma Human Deeks, S. G., F. M. Hecht, M. Swanson, T. Elbeik, R. Loftus, P. T. Cohen, and Immunodeficiency Virus Type 1 RNA and the Risk of Perinatal others. 1999. "HIV RNA and CD4 Cell Count Response to Protease Transmission. Women and Infants Transmission Study Group." New Inhibitor Therapy in an Urban AIDS Clinic: Response to Both Initial England Journal of Medicine 341 (6): 394­402. and Salvage Therapy." AIDS 13 (6): F35­43. Ghys, P. D., M. O. Diallo, V. Ettiegne-Traore, K. Kale, O. Tawil, M. Carael, DeGruttola, V., G. R. Seage, K. H. Mayer, C. R. Horsburgh. 1989. and others. 2002. "Increase in Condom Use and Decline in HIV and "Infectiousness of HIV between Male Homosexual Partners." Journal Sexually Transmitted Diseases among Female Sex Workers in Abidjan, of Clinical Epidemiology 42 (9): 849­56. Côte d'Ivoire, 1991­1998." AIDS 16 (2): 251­58. Des Jarlais, D. C., and S. R. Friedman. 1996. "HIV Epidemiology and Gillespie, S., L. Haddad, and R. Jackson. 2001. "HIV/AIDS Food and Interventions among Injecting Drug Users." International Journal of Nutrition Security: Impacts and Actions." Paper prepared for the 29th Sexually Transmitted Diseases and AIDS 7 (Suppl. 2): 57­61. Session of the ACC/SCN Symposium on Nutrition and HIV/AIDS. Deschamps, M. M., J. W. Pape, A. Hafner, and W. D. Johnson. 1996. Gilson, L., R. Mkanje, H. Grosskurth, F. Mosha, J. Picard, A. Gavyole, and "Heterosexual Transmission of HIV in Haiti." Annals of Internal others. 1997. "Cost-Effectiveness of Improved Treatment Services for Medicine 125 (4): 324­30. Sexually Transmitted Diseases in Preventing HIV-1 Infection in Donnelly, J. 2004. "Circumcised Men Less Likely to Get AIDS." Boston Mwanza Region, Tanzania." Lancet 350 (9094): 1805­9. Globe, November 16. Gisselquist, D., J. J. Potterat, R. Rothenberg, E. M. Drucker, S. Brody, Duran, S., M. Saves, B. Spire, V. Cailleton, A. Sobel, P. Carrieri, and others. D. Brewe, and others. 2003. "Examining the Hypothesis That Sexual 2001. "Failure to Maintain Long-Term Adherence to Highly Active Transmission Drives Africa's HIV Epidemic." AIDScience 3 (10). Antiretroviral Therapy: The Role of Lipodystrophy." AIDS 15 (18): http://www.aidscience.org/Articles/AIDScience032.asp. 2441­44. Global HIV Prevention Working Group. 2003. Access to HIV Prevention-- Dziekan, G., D. Chisholm, B. Johns, J. Rovira, and Y. J. Hutin. 2003. "The Closing the Gap. Menlo Park, CA: Kaiser Family Foundation. Cost-Effectiveness of Policies for the Safe and Appropriate Use of Göhde, R. 2004. Personal Communication. HIV/AIDS Project Co- Injection in Healthcare Settings." Bulletin of the World Health ordinator, Müster, Germany. Organization 81 (4): 277­85. Easterbrook, P. J. 1994. "Non-Progression in HIV Infection." AIDS 8 (8): Goldie, S. J., L. Kuhn, L. Denny, A. Pollack, and T. C. Wright. 2001. "Policy 1179­82. Analysis of Cervical Cancer Screening Strategies in Low-Resource Settings: Clinical Benefits and Cost-Effectiveness. Journal of the Egger, M., J. Pauw, A. Lopatatzidis, D. Medrano, F. Paccaud, and G. D. American Medical Association 285 (24): 3107­15. Smith. 2000. "Promotion of Condom Use in a High-Risk Setting in Nicaragua: A Randomised Controlled Trial." Lancet 355 (9221): Grant, R. M., F. M. Hecht, M. Warmerdam, L. Liu, T. Liegler, C. J. 2101­05. Petropoulos, and others. 2002. "Time Trends in Primary HIV-1 Drug Resistance among Recently Infected Persons." Journal of the American Eshleman, S. H., M. Mracna, L. A. Guay, M. Deseyve, S. Cunningham, Medical Association 288 (2): 181­88. M. Mirochnick, and others. 2001. "Selection and Fading of Resistance Mutations in Women and Infants Receiving Nevirapine to Prevent Grant, R. M., J. Kahn, M. Warmerdam, L. Liu, C. J. Petropoulos, N. S. HIV-1 Vertical Transmission (HIVNET 012)." AIDS 15 (15): 1951­57. Hellman, and F. Hecht. 2002. "Transmission and Transmissibility of Drug Resistant HIV-1 (368-M)." Paper presented at 9th Conference on Farmer, P., F. Leandre, J. Mukherjee, R. Gupta, L. Tarter, and J. Y. Kim. 2001. Retroviruses and Opportunistic Infections, Seattle, February 24­28. "Community-Based Treatment of Advanced HIV Disease: Introducing DOT-HAART (Directly Observed Therapy with Highly Active Grassly, N. C., G. P. Garnett, B. Schwartlander, S. Gregson, and R. M. Antiretroviral Therapy)." Bulletin of the World Health Organization 79 Anderson. 2001. "The Effectiveness of HIV Prevention and the (12): 1145­51. Epidemiological Context." Bulletin of the World Health Organization Fawole, I. O., M. C. Asuzu, S. O. Oduntan, and W. R. Brieger. 1999. "A 79 (12): 1121­32. School-Based AIDS Education Programme for Secondary School Gregson, S., P. R. Mason, G. P. Garnett, T. Zhuwau, C. A. Nyamukapa, Students in Nigeria: A Review of Effectiveness." Health Education R. M. Anderson, and S. K. Chandiwana. 2001. "A Rural HIV Epidemic Resources 14 (5): 675­83. in Zimbabwe? Findings from a Population-Based Survey." HIV/AIDS Prevention and Treatment | 363 International Journal of Sexually Transmitted Diseases and AIDS 12 (3): Hutton, G., K. Wyss, and Y. N'Diekhor. 2003. "Prioritization of Prevention 189­96. Activities to Combat the Spread of HIV/AIDS in Resource- Grosskurth, H., F. Mosha, J. Todd, E. Mwijarubi, A. Klokke, K. Senkoro, Constrained Settings: A Cost-Effectiveness Analysis from Chad, and others. 1995. "Impact of Improved Treatment of Sexually Central Africa." International Journal of Health Planning and Transmitted Diseases on HIV Infection in Rural Tanzania: Management 18 (2): 117­36. Randomised Controlled Trial." Lancet 346 (8974): 530­36. Ippolito, G., V. Puro, and G. De Carli. 1993. "The Risk of Occupational Guay, L. A., P. Musoke, T. Fleming, D. Bagenda, M. Allen, C. Nakabiito, and Human Immunodeficiency Virus Infection in Health Care Workers: others. 1999. "Intrapartum and Neonatal Single-Dose Nevirapine Italian Multicenter Study." Italian Study Group on Occupational Risk Compared with Zidovudine for Prevention of Mother-to-Child of HIV Infection. Annals of Internal Medicine 153 (12): 1451­58. Transmission of HIV-1 in Kampala, Uganda: HIVNET 012 Institute of Medicine, ed. 1997. The Hidden Epidemic: Confronting Sexually Randomised Trial." Lancet 354 (9181): 795­802. Transmitted Diseases. Washington, DC: National Academy Press. Gutierrez, J. P., B. Johns, T. Adam, S. M. Bertozzi, T. T. Edejer, R. Greener, Jackson, D. J., J. P. Rakwar, B. A. Richardson, K. Mandaliya, B. H. Chohan, and others. 2004. "Achieving the WHO/UNAIDS Antiretroviral J. J. Bwayo, and others. 1997. "Decreased Incidence of Sexually Treatment 3 by 5 Goal: What Will It Cost?" Lancet 364 (9428): 63­64. Transmitted Diseases among Trucking Company Workers in Kenya: Hammer, S. M., K. E. Squires, M. D. Hughes, J. M. Grimes, L. M. Demeter, Results of a Behavioural Risk-Reduction Programme." AIDS 11 (7): J. S. Currier, and others. 1997. "A Controlled Trial of Two Nucleoside 903­09. Analogues Plus Indinavir in Persons with Human Immunodeficiency Jackson, J. B., P. Musoke, T. Fleming, L. A. Guay, D. Bagenda, M. Allen, and Virus Infection and CD4 Cell Counts of 200 Per Cubic Millimeter or others. 2003. "Intrapartum and Neonatal Single-Dose Nevirapine Less." New England Journal of Medicine 337 (11): 725­33. Compared with Zidovudine for Prevention of Mother-to-Child Hansen, K., G. Woelk, H. Jackson, R. Kerkhoven, N. Manjonjori, P. Transmission of HIV-1 in Kampala, Uganda: 18-Month Follow-up of Maramba, and others. 1998. "The Cost of Home-Based Care for the HIVNET 012 Randomised Trial." Lancet 362 (9387): 859­68. HIV/AIDS Patients in Zimbabwe." AIDS Care 10 (6): 751­59. Jacobs, B., and A. Mercer. 1999. "Feasibility of Hospital-Based Blood Harries, A. D., D. S. Nyangulu, N. J. Hargreaves, O. Kaluwa, and F. M. Banking: A Tanzanian Case Study." Health Policy and Planning 14 (4): Salaniponi. 2001. "Preventing Antiretroviral Anarchy in Sub-Saharan 354­62. Africa." Lancet 358 (9279): 410­14. Janssen, R. S., D. R. Holtgrave, R. O. Valdiserri, M. Shepherd, H. D. Gayle, Harvey, B., J. Stuart, and T. Swan. 2000. "Evaluation of a Drama-in- and K. M. De Cock. 2001. "The Serostatus Approach to Fighting the Education Programme to Increase AIDS Awareness in South African HIV Epidemic: Prevention Strategies for Infected Individuals." High Schools: A Randomized Community Intervention Trial." American Journal of Public Health 91 (7): 1019­24. International Journal of Sexually Transmitted Diseases and AIDS 11 (2): Jemmott, J. B., L. S. Jemmott, and G. T. Fong. 1998. "Abstinence and Safer 105­11. Sex: HIV Risk-Reduction for African American Adolescents." Journal Hayes, R., J. Chagalucha, H. Grosskurth, A. Obasi, J. Todd, B. Cleophas- of the American Medical Association 279 (19): 1529­36. Mazigr, and others. 2003. "Mema Kwa Vijana: A Randomised Jenkins, C., H. Rahman, T. Saidel, S. Jana, and A. M. Hussain. 2001. Controlled Trial of an Adolescents Sexual and Reproductive Health "Measuring the Impact of Needle Exchange Programs among Inject- Intervention Programme in Rural Mwanza, Tanzania: 2 Intervention ing Drug Users through the National Behavioural Surveillance in and Process Indicators." Paper presented at the International Society Bangladesh." AIDS Education and Prevention 13 (5): 452­61. of Sexually Transmitted Diseases Research Congress, Ottawa, July Jha, P., L. M. E. Vaz, F. Plummer, N. Nagelkerke, B. Willbond, E. Ngugi, 27­30. and others. 2001. "The Evidence Base for Interventions to Prevent Henderson, D. K., B. J. Fahey, M. Will, J. M. Schmitt, K. Carey, D. E. Koziol, HIV Infection in Low- and Middle-Income Countries." Commission and others. 1990. "Risk for Occupational Transmission of Human on Macroeonomics and Health Working Paper. WG5: 2. Commission Immunodeficiency Virus Type 1 (HIV-1) Associated with Clinical on Macroeonomics and Health, Geneva, Switzerland. Exposures: A Prospective Evaluation." Annals of Internal Medicine 113 John, G. C., R. W. Nduati, D. A. Mbori-Ngacha, B. A. Richardson, (1): 740­46. D. Panteleeff, A. Mwatha, and others. 2001. "Correlates of Mother-to- Higginson, I. J., I. G. Findlay, D. M. Goodwin, K. Hood, A. G. Edwards, Child Human Immunodeficiency Virus Type 1 (HIV-1) Transmission: A. Cook, and others. 2003. "Is There Evidence That Palliative Care Association with Maternal Plasma HIV-1 RNA Load, Genital HIV-1 Teams Alter End-of-Life Experiences of Patients and Their Caregivers?" DNA Shedding, and Breast Infections." Journal of Infectious Diseases Journal of Pain and Symptom Management 25 (2): 150­168. 183 (2): 206­12. Hira, S. K., G. J. Dore, and T. Sirisanthana. 1998. "Clinical Spectrum of John Snow Inc. 2003. Fact Sheets for Diagnostic Tests. http://deliver.jsi. HIV/AIDS in the Asia-Pacific Region." AIDS 12 (Suppl. B): S145­54. com/2002/archives/hivaids/test_kits/index.cfm. Hira, S. K., H. L. Dupont, D. N. Lanjewar, and Y. N. Dholakia. 1998."Severe Kagimu, M., E. Marum, F. Wabwire-Mangen, N. Nakyanjo, Y. Walakira, Weight Loss: The Predominant Clinical Presentation of Tuberculosis in and J. Hogle. 1998. "Evaluation of the Effectiveness of AIDS Health Patients with HIV Infection in India." National Medical Journal of India Education Interventions in the Muslim Community in Uganda." AIDS 11 (6): 256­58. Education and Prevention 10 (3): 215­28. Hook, E. W., R. O. Cannon, A. J. Nahmias, F. F. Lee, C. H. Campbell, Kahn, J. G., S. M. Kegeles, R. Hays, and N. Beltzer. 2001."Cost-Effectiveness D. Glasser, and others. 1992. "Herpes Simplex Virus Infection as a of the Mpowerment Project, a Community-Level Intervention for Risk Factor for Human Immunodeficiency Virus Infection in Young Gay Men." Journal of Acquired Immune Deficiency Syndrome Heterosexuals." Journal of Infectious Diseases 165 (2): 251­55. 27 (5): 482­91. Hoover, D. R., A. J. Saah, H. Bacellar, J. Phair, R. Detels, R. Anderson, Kalichman, S. C., D. Rompa, M. Cage, K. DiFonzo, D. Simpson, J. Austin, and others. 1993. "Clinical Manifestations of AIDS in the Era of and others. 2001. "Effectiveness of an Intervention to Reduce HIV Pneumocystis Prophylaxis: Multicenter AIDS Cohort Study." New Transmission Risks in HIV-Positive People." American Journal of England Journal of Medicine 329 (26): 1922­26. Preventive Medicine 21 (2): 84­92. Hurley, S. F., D. J. Jolley, and J. M. Kaldor. 1997. "Effectiveness of Needle- Kamali, A., M. Quigley, J. Nakiyingi, J. Kinsman, J. Kengeya-Kayondo, Exchange Programmes for Prevention of HIV Infection." Lancet 349 R. Gopal, and others. 2003. "Syndromic Management of Sexually (9068): 1797­800. Transmitted Infections and Behaviour Change Interventions on 364 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others Transmission of HIV-1 in Rural Uganda: A Community Randomised Preventive Measures among Injecting Drug Users in Svetlogorsk, Trial." Lancet 361 (9358): 645­52. Belarus." Addiction 99 (12): 1565­76. Kamenga, M., R. W. Ryder, M. Jingu, N. Mbuyi, L. Mbu, F. Behets, and Ladner, J., V. Leroy, P. Hoffman, M. Nyiraziraje, A. De Clercq, P. Van de others. 1991. "Evidence of Marked Sexual Behavior Change Associated Perre, and F. Dabis. 1998. "Chorioamnionitis and Pregnancy Outcome with Low HIV-1 Seroconversion in 149 Married Couples with in HIV-Infected African Women: Pregnancy and HIV Study Group." Discordant HIV-1 Serostatus: Experience at an HIV Counselling Journal of Acquired Immune Deficiency Syndrome and Human Center in Zaire." AIDS 5 (1): 61­67. Retrovirology 18 (3): 293­98. Kaplan, E. H., and R. Heimer. 1992. "A Model-Based Estimate of HIV Laga, M., M. Alary, N. Nzila, A. T. Manoka, M. Tuliza, F. Behets, and others. Infectivity via Needle-Sharing." Journal of Acquired Immune Deficiency 1994. "Condom Promotion, Sexually Transmitted Diseases Treatment, Syndrome 5 (11): 1116­18. and Declining Incidence of HIV-1 Infection in Female Zairian Sex Kaplan, J. E., D. L. Hanson, J. L. Jones, and M. S. Dworkin. 2001. "Viral Workers." Lancet 344 (8917): 246­48. Load as an Independent Risk Factor for Opportunistic Infections in Laleman, G., K. Magazani, J. H. Perriens, N. Badibanga, N. Kapila, HIV-Infected Adults and Adolescents." AIDS 15 (14): 1831­36. M. Konde, and others. 1992. "Prevention of Blood-borne HIV Kaplan, J. E., H. Masur, and K. K. Holmes. 2002. "Guidelines for Prevent- Transmission Using a Decentralized Approach in Shaba, Zaire." AIDS ing Opportunistic Infections among HIV-Infected Persons--2002: 6 (11): 1353­58. Recommendations of the U.S. Public Health Service and the Infectious Laniece, I., K. Diop, A. Desclaux, K. Sow, M. Ciss, B. Ndiaye, and I. Ndoye. Diseases Society of America." Morbidity and Mortality Weekly Report 2004. "Determinants of long-term adherence to antiretroviral drugs 51 (RR-8): 1­52. among adults followed over four years in Dakar, Senegal." Abstract Katz, M. H., S. K. Schwarcz, T. A. Kellogg, J. D. Klausner, J. W. Dilley, S. presented at the XV International AIDS Conference, July 11­16, Gibson, and others. 2002. "Impact of Highly Active Antiretroviral Bangkok, Thailand. Treatment on HIV Seroincidence among Men Who Have Sex with Lanjewar, D. N., B. S. Anand, R. Genta, M. B. Maheshwari, M. A. Ansari, Men: San Francisco." American Journal of Public Health 92 (3): 388­94. S. K. Hira, and others. 1996. "Major Differences in the Spectrum of Katzenstein, D. A., W. McFarland, M. Mbizvo, A. S. Latif, R. Machekano, J. Gastrointestinal Infections Associated with AIDS in India versus the Parsonnet, and others. 1998. "Peer Education among Factory Workers West: An Autopsy Study." Clinical Infectious Disease 23 (3): 482­85. in Zimbabwe: Providing a Sustainable HIV Prevention Intervention." Lauby, J. L., P. J. Smith, M. Stark, B. Person, and J. Adams. 2000. "A Paper presented at the 12th International Conference on AIDS, Community-Level HIV Prevention Intervention for Inner-City Geneva, June 28­July 3. Women: Results of the Women and Infants Demonstration Projects." Kegeles, S. M., R. B. Hays, and T. J. Coates. 1996. "The Mpowerment American Journal of Public Health 90 (2): 216­22. Project: A Community-Level HIV Prevention Intervention for Young Gay Men." American Journal of Public Health 86 (8): 1129­36. Ledergerber, B., M. Egger, M. Opravil, A. Telenti, B. Hirschel, M. Battegay, and others. 1999. "Clinical Progression and Virological Failure on Kelly, J. A., D. A. Murphy, K. J. Sikkema, T. L. McAuliffe, R. A. Roffman, L. J. Highly Active Antiretroviral Therapy in HIV-1 Patients: A Prospective Solomon, and others. 1997. "Randomised, Controlled, Community- Cohort Study." Swiss HIV Cohort Study. Lancet 353 (9156): 863­68. Level HIV-Prevention Intervention for Sexual-Risk Behaviour among Homosexual Men in U.S. Cities: Community HIV Prevention Research Ledergerber, B., A. Mocroft, P. Reiss, H. Furrer, O. Kirk, M. Bickel, and Collaborative." Lancet 350 (9090): 1500­5. others. 2001. "Discontinuation of Secondary Prophylaxis against Pneumocystis carinii Pneumonia in Patients with HIV Infection Who Kim, N., B. Stanton, X. Li, K. Dickersin, and J. Galbraith. 1997. "Effective- Have a Response to Antiretroviral Therapy: Eight European Study ness of the 40 Adolescent AIDS-Risk Reduction Interventions: A Groups." New England Journal of Medicine 344 (3): 168­74. Quantitative Review." Journal of Adolescent Health 20 (3): 204­15. Leigh Brown, A. J., S. D. Frost, W. C. Mathews, K. Dawson, N. S. Hellmann, Kirby, D. 1997. No Easy Answers: Research Findings on Programs to Reduce E. S. Daar, and others. 2003. "Transmission Fitness of Drug-Resistant Teen Pregnancy. Washington, DC: National Campaign to Prevent Teen Human Immunodeficiency Virus and the Prevalence of Resistance in Pregnancy. the Antiretroviral-Treated Population." Journal of Infectious Diseases ------. 2001. Emerging Answers: Research Findings on Programs to Reduce 187 (4): 683­86. Teen Pregnancy. Washington, DC: National Campaign to Prevent Teen Levine, W. C., R. Revollo, V. Kaune, J. Vega, F. Tinajeros, M. Garnica, and Pregnancy. others. 1998. "Decline in Sexually Transmitted Disease Prevalence in Kitahata, M. M., T. D. Koepsell, R. A. Deyo, C. L. Maxwell, W. T. Dodge, Female Bolivian Sex Workers: Impact of an HIV Prevention Project." and E. H. Wagner. 1996. "Physicians' Experience with the Acquired AIDS 12 (14): 1899­906. Immunodeficiency Syndrome as a Factor in Patients' Survival." New England Journal of Medicine 334: 701­6. Li, T. S., R. Tubiana, C. Katlama, V. Calvez, H. Ait Mohand, and B. Autran. 1998. "Long-Lasting Recovery in CD4 T-Cell Function and Viral-Load Komanduri, K. V., M. N. Viswanathan, E. D. Wieder, D. K. Schmidt, Reduction after Highly Active Antiretroviral Therapy in Advanced B. M. Bredt, M. A. Jacobson, and others. 1998. "Restoration of HIV-1 Disease." Lancet 351 (9117): 1682­86. Cytomegalovirus-Specific CD4 T-Lymphocyte Responses after Ganciclovir and Highly Active Antiretroviral Therapy in Individuals Liechty, C. A., and D. R. Bangsberg. 2003. "Doubts about DOT: antiretro- Infected with HIV-1." Nature Medicine 4 (8): 953­56. viral therapy for resource-poor countries." AIDS 17 (9): 1383­1387. Ksobiech, K. 2003. "A Meta-Analysis of Needle Sharing, Lending, and Little, S. J., S. Holte, J. P. Routy, E. S. Daar, M. Markowitz, A. C. Collier, and Borrowing Behaviors of Needle Exchange Program Attenders." AIDS others. 2002. "Antiretroviral-Drug Resistance among Patients Recently Education and Prevention 15 (3): 257­68. Infected with HIV." New England Journal of Medicine 347 (6): 385­94. Kumaranayake, L., P. Mangtani, A. Boupda-Duate, J. C. Foumena Abada, Low-Beer, D., and R. Stoneburner. 2003. "Behavior and Communication C. Cheta, Z. Njoumemi, and C. Watts. 1998. "Cost-Effectiveness of an Change in Reducing HIV: Is Uganda Unique?" African Journal of AIDS HIV/AIDS Peer Education Programme among Commercial Sex Research 2 (1): 9­21. Workers (CSW): Results from Cameroon [abstract no. 33592]." Paper Low-Beer, S., A. E. Weber, K. Bartholomew, M. Landolt, D. Oram, J. S. presented at the World AIDS Conference, Geneva, June 28­July 3. Montaner, and others. 2000. "A Reality Check: The Cost of Making Kumaranayake, L., P. Vickerman, D. Walker, S. Samoshkin, V. Romantzov, Post-Exposure Prophylaxis Available to Gay and Bisexual Men at High Z. Emelyanova, and others. 2004. "The Cost-Effectiveness of HIV Sexual Risk." AIDS 14 (3): 325­26. HIV/AIDS Prevention and Treatment | 365 Lui, K. J., W. W. Darrow, and G. W. Rutherford III. 1988. "A Model-Based Saves Lives and Money in Zimbabwe." Journal of Acquired Immune Estimate of the Mean Incubation Period for AIDS in Homosexual Deficiency Syndrome and Human Retrovirology 9 (2): 183­192. Men." Science 240 (4857): 1333­35. McNaghten, A. D., D. L. Hanson, J. L. Jones, M. S. Dworkin, and J. W. Ward. Mallal, S. A., M. John, C. B. Moore, I. R. James, and E. J. McKinnon. 2000. (1999). "Effects of Antiretroviral Therapy and Opportunistic Illness "Contribution of Nucleoside Analogue Reverse Transcriptase Primary Chemoprophylaxis on Survival after AIDS Diagnosis: Inhibitors to Subcutaneous Fat Wasting in Patients with HIV Adult/Adolescent Spectrum of Disease Group." AIDS 13 (13): 1687­95. Infection." AIDS 14 (10): 1309­16. Meekers, D. 2000. "The Effectiveness of Targeted Social Marketing to Mannheimer, S., G. Friedland, J. Matts, C. Child, and M. Chesney. 2002. Promote Adolescent Reproductive Health: The Case of Soweto, South "The Consistency of Adherence to Antiretroviral Therapy Predicts Africa." Journal of HIV/AIDS Prevention and Education for Adolescents Biologic Outcomes for Human Immunodeficiency Virus-Infected and Children 3 (4): 73­92. Persons in Clinical Trials." Clinical Infectious Diseases 34 (8): Mesquita,F.,D.Doneda,D.Gandolfi,M.I.Nemes,T.Andrade,R.Bueno,and 1115­21. others. 2003. "Brazilian Response to the Human Immunodeficiency Marks, G., S. Burris, and T. A. Peterman. 1999. "Reducing Sexual Virus/Acquired Immunodeficiency Syndrome Epidemic among Transmission of HIV from Those Who Know They Are Infected: Injection Drug Users."Clinical Infectious Diseases 37 (Suppl. 5): S382­85. The Need for Personal and Collective Responsibility." AIDS 13 (3): Metzger, D. S., H. Navaline, and G. E. Woody. 1998. "Drug Abuse 297­306. Treatment as AIDS Prevention." Public Health Reports 113 (Suppl. 1): Marseille, E., P. B. Hofmann, and J. G. Kahn. 2002. "HIV Prevention before 97­106. HAART in Sub-Saharan Africa." Lancet 359 (9320): 1851­56. Miotti, P. G., T. E. Taha, N. I. Kumwenda, R. Broadhead, L. A. Mtimavalye, Marseille, E., J. G. Kahn, K. Billinghurst, and J. Saba. 2001. "Cost- L. Van der Hoeven, and others. 1999. "HIV Transmission through Effectiveness of the Female Condom in Preventing HIV and STDs in Breastfeeding: A Study in Malawi." Journal of the American Medical Commercial Sex Workers in Rural South Africa." Social Science and Association 282 (8): 744­49. Medicine 52 (1): 135­48. Mitty, J. A., V. E. Stone, M. Sands, G. Macalino, and T. Flanigan. 2002. Marseille, E., J. G. Kahn, F. Mmiro, L. Guay, P. Musoke, M. G. Fowler, and "Directly Observed Therapy for the Treatment of People with Human others. 1999. "Cost-Effectiveness of Single-Dose Nevirapine Regimen Immunodeficiency Virus Infection: A Work in Progress." Clinical for Mothers and Babies to Decrease Vertical HIV-1 Transmission in Infectious Disease 34 (7): 984­90. Sub-Saharan Africa." Lancet 354 (9181): 803­9. Morgan, D., G. H. Maude, S. S. Malamba, M. J. Okongo, H. U. Wagner, Marseille, E., J. G. Kahn, and J. Saba. 1998. "Cost-Effectiveness of Antiviral D. W. Mulder, and others. 1997."HIV-1 Disease Progression and AIDS- Drug Therapy to Reduce Mother-to-Child HIV Transmission in Sub- Defining Disorders in Rural Uganda." Lancet 350 (9073): 245­50. Saharan Africa." AIDS 12 (8): 939­48. Moses, S., F. A. Plummer, E. N. Ngugi, N. J. Nagelkerke, A. O. Anzala, and Marseille, E., S. F. Morin, C. Collins, T. Summers, T. J. Coates, and J. G. J. O. Ndinya-Achola. 1991. "Controlling HIV in Africa: Effectiveness Kahn. 2002. "Cost-Effectiveness of HIV Prevention in Developing and Cost of an Intervention in a High-Frequency STD Transmitter Countries." In HIV InSite Knowledge Base. http://hivinsite.ucsf.edu/ Core Group." AIDS 5 (4): 407­11. InSite?page=kb-08-01-04. Munoz, A., C. A. Sabin, and A. N. Phillips. 1997. "The Incubation Period Martin, H. L., B. A. Richardson, P. M. Nyange, L. Lavreys, S. L. Hillier, of AIDS." AIDS 11 (Suppl. A): S69­76. B. Chohan, and others. 1999. "Vaginal Lactobacilli, Microbial Flora, Murphy, D. A., W. D. Marelich, D. Hoffman, and W. N. Steers. 2004. and Risk of Human Immunodeficiency Virus Type 1 and Sexually "Predictors of Antiretroviral Adherence." AIDS Care 16 (4): 471­84. Transmitted Disease Acquisition." Journal of Infectious Diseases 180 (6): Murray, C. J., E. DeJonghe, H. J. Chum, D. S. Nyangulu, A. Salomao, and 1863­68. K. Styblo. 1991. "Cost-Effectiveness of Chemotherapy for Pulmonary Mast, S. T., J. D. Woolwine, and J. L. Gerberding. 1993. "Efficacy of Gloves Tuberculosis in Three Sub-Saharan African Countries." Lancet 338 in Reducing Blood Volumes Transferred during Simulated Needlestick (8778): 1305­8. Injury." Journal of Infectious Diseases 168 (6): 1589­92. Murray, C. J., and A. D. Lopez. 1996. The Global Burden of Disease: A Mastro T. D., G. A. Satten, T. Nopkesorn, S. Sangkharomya, and I. M. Comprehensive Assessment of Mortality and Disability from Diseases, Longini. 1994. "Probability of Female-to-Male Transmission of HIV-1 Injuries, and Risk Factors in 1990 and Projected to 2020. Boston: in Thailand." Lancet 343 (8891): 204­7. Harvard University Press. Mathers, C. D., A. Lopez, C. Stein, D. Ma Fat, C. Rao, M. Inoue, and oth- Musicco, M., A. Lazzarin, A. Nicolosi, M. Gasparini, P. Costigliola, C. Arici, ers. 2006. "The Burden of Disease and Mortality by Condition: Data, and others. 1994. "Antiretroviral Treatment of Men Infected with Methods and Results for the Year 2001." In the Global Burden of Disease Human Immunodeficiency Virus Type 1 Reduces the Incidence of in 2001, ed. A. Lopez, C. Mathers, M. Ezzati, D. Jamison, and C. J. L. Heterosexual Transmission." Italian Study Group on HIV Heterosexual Murray. New York: Oxford University Press. Transmission. Archives of Internal Medicine 154 (17): 1971­76. Mayaud, P., F. Mosha, J. Todd, R. Balira, J. Mgara, B. West, and others. 1997. Nagachinta, T., A. Duerr, V. Suriyanon, N. Nantachit, S. Rugpao, "Improved Treatment Services Significantly Reduce the Prevalence of C. Wanapirak, and others. 1997. "Risk Factors for HIV-1 Transmission Sexually Transmitted Diseases in Rural Tanzania: Results of a from HIV-Seropositive Male Blood Donors to Their Regular Female Randomized Controlled Trial." AIDS 11 (15): 1873­80. Partners in Northern Thailand." AIDS 11 (14): 1765­72. Mbopi Keou, F. X., G. Gresenguet, P. Mayaud, H. A. Weiss, R. Gopal, D. W. National Consensus Development Panel on Effective Medical Treatment Brown, and others. 1999. "Genital Herpes Simplex Virus Type 2 of Opiate Addiction. 1998. "Effective Medical Treatment of Opiate Shedding Is Increased in HIV-Infected Women in Africa." AIDS 13: Addiction." Journal of the American Medical Association 280: 1936­43. 536­37. Nduati, R., G. John, D. Mbori-Ngacha, B. Richardson, J. Overbaugh, A. McConnell, J., and R. Grant. 2003. "Sorting out Serosorting Using Social Mwatha, and others. 2000. "Effect of Breastfeeding and Formula Network Methods. Paper presented at the 10th Conference on Feeding on Transmission of HIV-1: A Randomized Clinical Trial." Retroviruses and Opportunistic Infections, Boston, February 10­14. Journal of the American Medical Association 283 (9): 1167­74. McFarland, W., J. G. Kahn, D. A. Katzenstein, D. Mvere, and R. Shamu. Nduati, R., B. A. Richardson, G. John, D. Mbori-Ngacha, A. Mwatha, 1995. "Deferral of Blood Donors with Risk Factors for HIV Infection J. Ndinya-Achola, and others. 2001. "Effect of Breastfeeding on 366 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others Mortality among HIV-1 Infected Women: A Randomised Trial." Lancet PETRA Study Team. 2002. "Efficacy of Three Short-Course Regimens 357 (9269): 1651­55. of Zidovudine and Lamivudine in Preventing Early and Late Needle, R. H., S. L. Coyle, J. Normand, E. Lambert, and H. Cesari. 1998. Transmission of HIV-1 from Mother to Child in Tanzania, South "HIV Prevention with Drug-Using Populations--Current Status and Africa, and Uganda: A Randomised, Double-Blind, Placebo- Future Prospects: Introduction and Overview." Public Health Reports Controlled Trial." Lancet 359 (9313): 1178­86. 113 (Suppl. 1): 4­18. Pinkerton, S. D., D. R. Holtgrave, and F. R. Bloom. 1998. "Cost- Newell, M. L. 2003. "Antenatal and Perinatal Strategies to Prevent Mother- Effectiveness of Post-Exposure Prophylaxis Following Sexual Exposure to-Child Transmission of HIV Infection." Transactions of the Royal to HIV." AIDS 12 (9): 1067­78. Society of Tropical Medicine and Hygiene 97 (1): 22­24. Pinkerton, S. D., D. R. Holtgrave, W. J. DiFranceisco, L. Y. Stevenson, and J. A. Kelly. 1998. "Cost-Effectiveness of a Community-Level HIV Risk Ngugi, E. N., F. A. Plummer, J. N. Simonsen, D. W. Cameron, M. Bosire, Reduction Intervention." American Journal of Public Health 88 (8): P. Waiyaki, and others. 1988. "Prevention of Transmission of Human 1239­42. ImmunodeficiencyVirus inAfrica: Effectiveness of Condom Promotion and Health Education among Prostitutes." Lancet 2 (8616): 887­90. Pistone, T., S. Kony, M. A. Faye-Niang, C. T. Ndour, P. M. Gueye, D. Henzel, and others. 2002. "A Simple Clinical and Paraclinical Score Predictive Nijhuis, M., S. Deeks, and C. Boucher. 2001. "Implications of Anti- of CD4 Cell Counts below 400/mm3 in HIV-Infected Adults in retroviral Resistance on Viral Fitness." Current Opinions in Infectious Dakar University Hospital, Senegal." Transactions of the Royal Society of Diseases 14 (1): 23­28. Tropical Medicine and Hygiene 96 (2): 167­72. Norr, K. F., J. L. Norr, B. J. McElmurry, S. Tlou, and M. R. Moeti. 2004. Quinn, T. C., M. J. Wawer, N. Sewankambo, D. Serwadda, C. Li, F. Wabwire- "Impact of Peer Group Education on HIV Prevention among Women Mangen, and others. 2000. "Viral Load and Heterosexual Transmission in Botswana." Health Care for Women International 25 (3): 210­26. of Human Immunodeficiency Virus Type 1" Rakai Project Study Orroth, K. K., E. L. Korenromp, R. G. White, A. Gavyole, R. H. Gray, Group. New England Journal of Medicine 342 (13): 921­29. and L. Muhangi. 2003. "Higher-Risk Behaviour and Rates of Ramsay, S. 2003. "Leading the Way in African Home-Based Palliative Sexually Transmitted Diseases in Mwanza Compared to Uganda Care." Lancet 362 (9398): 1812­13. May Help Explain HIV Prevention Trial Outcomes." AIDS 17 (18): 2653­60. Read, J. S. 2003."Human Milk, Breastfeeding, and Transmission of Human Immunodeficiency Virus Type 1 in the United States." American Over, M., P. Heywood, J. Gold, I. Gupta, S. K. Hira, and E. Marseille. 2004. Academy of Pediatrics Committee on Pediatric AIDS. Pediatrics 112 HIV/AIDS Treatment and Prevention in India: Modeling the Cost and (5): 1196­205. Consequences. Health, Nutrition, and Population Series. Washington, DC: World Bank. Rely, K., S. Bertozzi, C. Avila-Figueroa, and M. T. Guijarro. 2003. "Cost- Effectiveness of Strategies to Reduce Mother-to-Child HIV Over, M., and P. Piot. 1996. "Human Immunodeficiency Virus Infection Transmission in Mexico, a Low-Prevalence Setting." Health Policy and and Other Sexually Transmitted Diseases in Developing Countries: Planning 18 (3): 290­98. Public Health Importance and Priorities for Resource Allocation." Journal of Infectious Diseases 174 (Suppl. 2): S162­75. Reynolds, H. W., B. Janowitz, R. Homan, and L. Johnson. 2004. "Cost- Effectiveness of Two Interventions to Avert HIV-Positive Births." Poster Oyugi, J., and D. Bangsberg. 2004 "Treatment outcomes and adherence to presentation at the XV International AIDS Conference, Bangkok, generic Triomune® and Maxivir® therapy in Kampala, Uganda." XV Thailand, July 11­16, 2004. International AIDS Conference. Bangkok, Thailand, July 10­16. Reynolds, S. J., M. E. Shepherd, A. R. Risbud, R. R. Gangakhedkar, R. S. Palella, F. J., K. M. Delaney, A. C. Moorman, M. O. Loveless, J. Fuhrer, G. A. Brookmeyer, A. D. Divekar, and others. 2004. "Male Circumcision and Satten, and others. 1998. "Declining Morbidity and Mortality among Risk of HIV-1 and Other Sexually Transmitted Infections in India." Patients with Advanced Human Immunodeficiency Virus Infection." Lancet 363 (9414): 1039­40. HIV Outpatient Study Investigators. New England Journal of Medicine Royce, R.A.,A. Sena,W. Cates, and M. S. Cohen. 1997."Sexual Transmission 338 (13): 853­60. of HIV." New England Journal of Medicine 336 (15): 1072­78. Palella, F. J., M. Deloria-Knoll, J. S. Chmiel, A. C. Moorman, K. C. Wood, Saavedra, J. 2000. "Economy and AIDS in Latin America." In AIDS in Latin A. E. Greenberg, and others. 2003. "Survival Benefit of Initiating America: A Multidisciplinary Vision, ed. by J. A. Izazola. Mexico City: Antiretroviral Therapy in HIV-Infected Persons in Different CD4 FUNSALUD. Cell Strata." Annals of Internal Medicine 138 (8): 620­26. Salomon, H., M. A. Wainberg, B. Brenner, Y. Quan, D. Rouleau, P. Cote, Pape, J. W., S. S. Jean, J. L. Ho, A. Hafner, and W. D. Johnson. 1993. "Effect and others. 2000. "Prevalence of HIV-1 Resistant to Antiretroviral of Isoniazid Prophylaxis on Incidence of Active Tuberculosis and Drugs in 81 Individuals Newly Infected by Sexual Contact or Injecting Progression of HIV Infection." Lancet 342 (8866): 268­72. Drug Use." Investigators of the Quebec Primary Infection Study. AIDS Pauw, J., J. Ferrie, R. Rivera-Villegas, J. Medrano-Martinez, A. Gorter, and 14 (2): F17­23. M. Egger. 1996. "A Controlled HIV/AIDS-Related Health Education Saunderson, P. R. 1995. "An Economic Evaluation of Alternative Programme in Managua, Nicaragua." AIDS 10 (5): 537­44. Programme Designs for Tuberculosis Control in Rural Uganda." Social Peak, A., S. Rana, S. H. Maharjan, D. Jolley, and N. Crofts. 1995. "Declining Science and Medicine 40 (9): 1203­12. Risk for HIV among Injecting Drug Users in Kathmandu, Nepal: The Schneider, M. M., T. L. Nielsen, S. Nelsing, A. I. Hoepelman, J. K. Eeftinck Impact of a Harm-Reduction Programme." AIDS 9 (9): 1067­70. Schattenkerk, Y. van der Graaf, and others. 1995. "Efficacy and Toxicity Peersman, G., and J. Levy. 1998. "Focus and Effectiveness of HIV- of Two Doses of Trimethoprim-Sulfamethoxazole as Primary Prevention Efforts for Young People." AIDS 12 (Suppl. A): S191­96. Prophylaxis against Pneumocystis carinii Pneumonia in Patients with Peipert, L., and Coffey, S. "About the Antiretroviral Drug Profiles." In HIV Human Immunodeficiency Virus: Dutch AIDS Treatment Group." InSite Knowledge Base, ed. L. Peipert, S. Coffey, P. Volberding. Journal of Infectious Diseases 171 (6): 1632­36. http://hivinsite.ucsf.edu. Selik, R. M., E. T. Starcher, and J. W. Curran. 1987."Opportunistic Diseases Peterman, T. A., R. L. Stoneburner, J. R. Allen, H. W. Jaffe, and J. W. Curran. Reported in AIDS Patients: Frequencies, Associations, and Trends." 1988. "Risk of Human Immunodeficiency Virus Transmission from AIDS 1 (3): 175­82. Heterosexual Adults with Transfusion-Associated Infections." Journal Semba, R. D., N. Kumwenda, D. R. Hoover, T. E. Taha, T. C. Quinn, of the American Medical Association 259 (1): 55­58. L. Mtimavalye, and others. 1999. "Human Immunodeficiency Virus HIV/AIDS Prevention and Treatment | 367 Load in Breast Milk, Mastitis, and Mother-to-Child Transmission of ------. 2000. AIDS: Palliative Care. UNAIDS Technical Update. Geneva: Human Immunodeficiency Virus Type 1." Journal of Infectious Diseases UNAIDS. 180 (1): 93­98. ------. 2001. Eight Case Studies of Home and Community Care for and by Sengupta, D., S. Lal, and Srinivas. 1994."Opportunistic Infection in AIDS." People with HIV/AIDS. Geneva: UNAIDS. Journal of the Indian Medical Association 92 (1): 24­26. ------. 2004. AIDS Epidemic Update: December 2004. Geneva: UNAIDS. Shaffer, N., R. Chuachoowong, P. A. Mock, C. Bhadrakom, W. Siriwasin, ------. 2005. AIDS Epidemic Update. December 2005. Geneva: N. L. Young, and others. 1999. "Short-Course Zidovudine for Perinatal UNAIDS. HIV-1 Transmission in Bangkok, Thailand: A Randomised Controlled Trial." Bangkok Collaborative Perinatal HIV Transmission Study UNDP (United Nations Development Programme). 2004. Thailand's Group. Lancet 353 (9155): 773­80. Response to HIV/AIDS: Progress and Challenges. Bangkok: UNDP. Shelburne, S. A., F. Visnegarwala, J. Darcourt, E. A. Graviss, T. P. Giordano, UNICEF (United Nations Children's Fund), UNAIDS (Joint United A. C. White Jr., and others. 2005. "Incidence and Risk Factors for Nations Programme on HIV/AIDS), WHO (World Health Immune Reconstitution Inflammatory Syndrome during Highly Organization), and Médécins sans Frontières. 2004. Sources and Prices Active Antiretroviral Therapy." AIDS 19 (4): 399­406. of Selected Medicines and Diagnostics for People Living with HIV/AIDS. Geneva: WHO. Sikkema, K. J., J. A. Kelly, R. A. Winett, L. J. Solomon, V. A. Cargill, R. A. Roffman, and others. 2000. "Outcomes of a Randomized Community- USAID (United States Agency for International Development). 2002. Level HIV Prevention Intervention for Women Living in 18 What Happened in Uganda? Washington, DC: USAID. Low-Income Housing Developments." American Journal of Public Uys, L., and M. Hensher. 2002. "The Cost of Home-Based Terminal Care Health 90 (1): 57­63. for People with AIDS in South Africa." South African Medical Journal Spector, S. A., G. F. McKinley, J. P. Lalezari, T. Samo, R. Andruczk, 92 (8): 624­28. S. Follansbee, and others. 1996. "Oral Ganciclovir for the Prevention of Van Liere, M. 2002."HIV/AIDS and Food Security in Sub-Saharan Africa." Cytomegalovirus Disease in Persons with AIDS: Roche Cooperative Paper presented at the Seventh Annual Economic Community of West Oral Ganciclovir Study Group." New England Journal of Medicine 334 African States Nutrition Forum. Banjul, The Gambia, September 2­6. (23): 1491­97. Vanek, J., O. Jírovec, and J. Lukes. 1953. "Interstitial Plasma Cell Stanton, B. F., X. Li, J. Kahihuata, A. M. Fitzgerald, S. Neumbo, Pneumonia in Infants." Annals of Paediatrics 180: 1­21. G. Kanduuombe, and others. 1998. "Increased Protected Sex and Vickerman, P., F. Terris-Prestholt, S. Delany, L. Kumaranayake, H. Rees, Abstinence among Namibian Youth Following a HIV Risk-Reduction and W. Watts. Forthcoming. "Are Targeted HIV Prevention Activities Intervention: A Randomized, Longitudinal Study." AIDS 12 (18): Still Cost-Effective in High Prevalence Settings? Results from an STI 2473­80. Treatment Intervention for Sex Workers in Hillbrow, South Africa." Stanton, B. F., X. Li, I. Ricardo, J. Galbraith, S. Feigelman, and L. Kaljee. Sexually Transmitted Diseases. 1996. "A Randomized, Controlled Effectiveness Trial of an AIDS Vlahov, D., B. Junge, R. Brookmeyer, S. Cohn, E. Riley, H. Armenian, and Prevention Program for Low-Income African-American Youths." others. 1997. "Reductions in High-Risk Drug Use Behaviors among Archives of Pediatric Adolescent Medicine 150 (4): 363­72. Participants in the Baltimore Needle Exchange Program." Journal of Stolte, I. G., N. H. Dukers, J. B. de Wit, J. S. Fennema, and R. A. Coutinho. Acquired Immune Deficiency Syndrome and Human Retrovirology 16 (5): 2001."Increase in Sexually Transmitted Infections among Homosexual 400­6. Men in Amsterdam in Relation to HAART." Sexually Transmitted Vlahov, D., A. Munoz, J. C. Anthony, S. Cohn, D. D. Celentano, and K. E. Infections 77 (3): 184­86. Nelson. 1990."Association of Drug Injection Patterns with Antibody to Stover, J., N. Fuchs, D. Halperin, A. Gibbons, and D. Gillespie. 2003. "Costs Human Immunodeficiency Virus Type 1 among Intravenous Drug and Benefits of Adding Family Planning to Services to Prevent Mother- Users in Baltimore, Maryland." American Journal of Epidemiology 132 to-Child Transmission of HIV (PMTCT)." Unpublished paper. The (5): 847­56. Futures Group. Voluntary HIV-1 Counseling and Testing Efficacy Study Group. 2000. Stover, J., N. Walker, G. P. Garnett, J. A. Salomon, K. A. Stanecki, P. D. Ghys, "Efficacy of Voluntary HIV-1 Counseling and Testing in Individuals and others. 2002. "Can We Reverse the HIV/AIDS Pandemic with an and Couples in Kenya, Tanzania, and Trinidad: A Randomized Trial." Expanded Response?" Lancet 360 (9326): 73­77. Lancet 356: 103­12. Stringer, E. M., M. Sinkala, J. S. Stringer, E. Mzyece, I. Makuka, R. L. Walker, D. 2003. "Cost and Cost-Effectiveness of HIV/AIDS Prevention Goldenberg, and others. 2003. "Prevention of Mother-to-Child Strategies in Developing Countries: Is There an Evidence Base?" Health Transmission of HIV in Africa: Successes and Challenges in Scaling Policy and Planning 18 (1): 4­17. Up a Nevirapine-Based Program in Lusaka, Zambia." AIDS 17 (9): Watts, C., H. Goodman, and L. Kumaranayake. 2000. "Improving the 1377­82. Efficiency and Impact of Blood Transfusion Services in the Context of Sweat, M. D., S. Gregorich, G. Sangiwa, C. Furlonge, D. Balmer, C. Increasing HIV Prevalence." Health Policy Unit, London. Kamenga, and others. 2000. "Cost-Effectiveness of Voluntary HIV-1 Wawer, M. J., N. K. Sewankambo, D. Serwadda, T. C. Quinn, L. A. Counseling and Testing in Reducing Sexual Transmission of HIV-1 in Paxton, N. Kiwanuka, and others. 1999. "Control of Sexually Kenya and Tanzania." Lancet 356 (9224): 113­21. Transmitted Diseases for AIDS Prevention in Uganda: A Randomised Sweat, M. D., K. R. O'Reilly, G. P. Schmid, J. Denison, and I. de Zoysa. 2004. Community Trial." Rakai Project Study Group. Lancet 353 (9152): "Cost-Effectiveness of Nevirapine to Prevent Mother-to-Child HIV 525­35. Transmission in Eight African Countries." AIDS 18 (12): 1661­71. Wegner, S. A., S. K. Brodine, J. R. Mascola, S. A. Tasker, R. A. Shaffer, M. J. Tsunekawa, K., S. Moolphate, H. Yanai, N. Yamada, S. Summanapan, and Starkey, and others. 2000. "Prevalence of Genotypic and Phenotypic J. Ngamvithayapong. 2004. "Care for People Living with HIV/AIDS: Resistance to Anti-Retroviral Drugs in a Cohort of Therapy-Naive An Assessment of Day Care Centers in Northern Thailand." AIDS HIV-1 Infected U.S. Military Personnel." AIDS 14 (8): 1009­15. Patient Care and Sexually Transmitted Diseases 18 (5): 305­14. Weiss, H. A., M. A. Quigley, and R. J. Hayes. 2000. "Male Circumcision and UNAIDS (Joint United Nations Programme on HIV/AIDS). 1997. Blood Risk of HIV Infection in Sub-Saharan Africa: A Systematic Review and Safety and AIDS: UNAIDS Point of View. Geneva: UNAIDS. Meta-Analysis." AIDS 14 (15): 2361­70. 368 | Disease Control Priorities in Developing Countries | Stefano Bertozzi, Nancy S. Padian, Jeny Wegbreit, and others Wenk, R., M. Bertolino, and J. Pussetto. 2000. "Direct Medical Costs of an Rural South Africa: An Issue of Cost-Effectiveness and Capacity." AIDS Argentinean Domiciliary Palliative Care Model." Journal of Pain and 12 (13): 1675­82. Symptom Management 20 (3): 162­65. Willbond, B., P. Thottingal, J. Kimani, L. M. E. Vaz, and F. A. Plummer. WHO (World Health Organization). 2002a. "Blood Safety: Aide-Memoire 2001. "The Evidence Base for Interventions in the Care and for National Blood Programmes." WHO, Geneva. Management of AIDS in Low and Middle Income Countries." ------. 2002b. "Definition of Palliative Care." WHO, Geneva. Commission on Macroeconomics and Health Working Paper Series, Paper WG5: 29. Commission on Macroeconomics and Health, ------. 2004. "Scaling Up Antiretroviral Therapy in Resource-Limited Geneva. Settings: Treatment Guidelines for a Public Health Approach." WHO, World Bank. 1997. Confronting AIDS: Public Priorities in a Global Geneva. Epidemic. New York: Oxford University Press. WHO (World Health Organization) and UNAIDS (Joint United Nations ------. 1999. "Project Appraisal Document on a Proposed Credit in the Programme on HIV/AIDS). 2003. "Expert Group Stresses That Unsafe Amount of SDR 140.82 Million to India for Second National Sex Is Primary Mode of HIV Transmission in Africa." Press release, HIV/AIDS Control Project." World Bank, Washington, DC. Geneva, March 14. World Food Programme. 2001. Food Security and HIV/AIDS: WFP Wiktor, S. Z., E. Ekpini, J. M. Karon, J. Nkengasong, C. Maurice, S. T. Executive Board Third Regular Session. Rome: WFP. Severin, and others. 1999. "Short-Course Oral Zidovudine for Prevention of Mother-to-Child Transmission of HIV-1 in Abidjan, WTO (World Trade Organization). 2003. "Decision Removes Final Patent Côte d'Ivoire: A Randomised Trial." Lancet 353 (9155): 781­85. Obstacle to Cheap Drug Imports." Press release, August 30. Wiley, J. A., S. J. Herschkorn, and N. S. Padian. 1989. "Heterogeneity in the Yeni, P. G., S. M. Hammer, M. S. Hirsch, M. S. Saag, M. Schechter, C. C. Probability of HIV Transmission per Sexual Contact: The Case of Carpenter, and others. 2004. "Treatment for Adult HIV Infection: 2004 Male-to-Female Transmission in Penile-Vaginal Intercourse." Statistics Recommendations of the International AIDS Society­USA Panel." in Medicine 8 (1): 93­102. Journal of the American Medical Association 292 (2): 251­65. Wilkinson, D., K. Floyd, and C. F. Gilks. 1998. "Antiretroviral Drugs as a Public Health Intervention for Pregnant HIV-Infected Women in HIV/AIDS Prevention and Treatment | 369 Chapter 19 Diarrheal Diseases Gerald T. Keusch, Olivier Fontaine, Alok Bhargava, Cynthia Boschi- Pinto, Zulfiqar A. Bhutta, Eduardo Gotuzzo, Juan Rivera, Jeffrey Chow, Sonbol A. Shahid-Salles, and Ramanan Laxminarayan Diarrheal diseases remain a leading cause of preventable death, fecal-oral transmission. Some are well known, others are recently especially among children under five in developing countries. discovered or emerging new agents, and presumably many This chapter reviews and prioritizes a number of available remain to be identified. They differ in the route from the stool interventions. to the mouth and in the number of organisms needed to cause The normal intestinal tract regulates the absorption and infection and illness. Among bacteria, the ability to survive secretion of electrolytes and water to meet the body's physio- stomach acid is an important determinant of the inoculum logical needs. More than 98 percent of the 10 liters per day of size required to cause illness. For example, Shigella bacteria are fluid entering the adult intestines are reabsorbed (Keusch resistant to low pH, and a few thousand organisms suffice, 2001). The remaining stool water, related primarily to the indi- which are readily transferred by direct person-to-person con- gestible fiber content, determines the consistency of normal tact or through contamination of inanimate objects, such as a feces from dry, hard pellets to mushy, bulky stools, varying cup. In contrast, bacteria readily killed by acid, such as Vibrio from person to person, day to day, and stool to stool. This vari- cholerae, require millions of organisms to cause illness, and ation complicates the definition of diarrhea, which by conven- therefore must first multiply in food or water to an infectious tion is present when three or more stools are passed in 24 hours dose. Some pathogens, such as rotavirus, display a sharp host that are sufficiently liquid to take the shape of the container in species preference, and others have a broad host range. Among which they are placed. The frequent passage of formed stool is Salmonella bacteria, certain bio-serotypes are adapted to infect not diarrhea (Black and Lanata 2002). Although young nursing animals and pose no threat to humans, and others are adapted infants tend to have five or more motions per day, mothers to humans and do not infect animals. The majority, however, know when the stooling pattern changes and their children are not adapted to a specific host and can infect either humans have diarrhea (Ronsmans, Bennish, and Wierzba 1988). The or domestic animals, thus facilitating transmission of these interval between two episodes is also arbitrarily defined as at organisms to humans. Less than a dozen of the more than 2,500 least 48 hours of normal stools. These definitions enable epi- individual Salmonella cause the majority of human infections, demiologists to count incidence, relapses, and new infections. reflecting the requirement for genes that encode essential virulence factors. The ability to identify virulence genes and their products TRANSMISSION has led to new molecular approaches to epidemiology and diagnosis, and undoubtedly will lead to new measures to pre- Diarrhea is caused by infectious organisms, including viruses, vent and treat diarrhea. Molecular methods also allow the bacteria, protozoa, and helminths, that are transmitted from separation of organisms that otherwise appear to be identical. the stool of one individual to the mouth of another, termed Nonpathogenic Escherichia coli in normal stool cannot be 371 separated from diarrhea-causing E. coli by standard methods; quantity that quickly exceeds total plasma and interstitial fluid however, identification of virulence genes or factors distin- volumes and is incompatible with life unless fluid therapy can guishes five groups of E. coli that cause illnesses ranging from keep up with losses. Such dramatic dehydration is usually due cholera-like watery diarrhea to neonatal diarrhea, persistent to rotavirus, enterotoxigenic E. coli, or V. cholerae (the cause of diarrhea, and bloody diarrhea (Nataro and Kaper 1998). cholera), and it is most dangerous in the very young. Persistent diarrhea is typically associated with malnutrition, either preceding or resulting from the illness itself (Ochoa, LABORATORY DIAGNOSIS Salazar-Lindo, and Cleary 2004). Even though persistent diar- rhea accounts for a small percentage of the total number of Etiologic diagnosis of diarrhea is valuable for public health diarrhea episodes, it is associated with a disproportionately interventions and case management. Microbiological culture increased risk of death. In India, persistent diarrhea accounted and microscopy remain the standard, despite their limited sen- for 5 percent of episodes but 14 percent of deaths, and a mor- sitivity. Their effectiveness is further reduced by antibiotic use, tality rate three times higher than briefer episodes (Bhan and and patients with severe illness are more likely both to be cul- others 1989). In Pakistan, persistent diarrhea accounted for 8 to tured and to have taken antibiotics. Even when cultures are 18 percent of episodes but 54 percent of deaths (Khan and oth- positive, the delay in laboratory identification limits their cost- ers 1993). In Bangladesh, persistent diarrhea associated with effectiveness for managing individual patients. The informa- malnutrition was responsible for nearly half of diarrhea deaths, tion is always epidemiologically and clinically important; how- and the relative risk for death among infants with persistent ever, during epidemics, culturing every patient is unnecessary diarrhea and severe malnutrition was 17 times greater than for when the causative organism is known. Antimicrobial resist- those with mild malnutrition (Fauveau and others 1992). ance data are essential to guide initial antibiotic choices. Persistent diarrhea occurs more often during an episode of New rapid tests to detect inflammatory mediators or white bloody diarrhea than an episode of watery diarrhea, and the or red blood cells in stool offer the promise of distinguishing mortality rate when bloody diarrhea progresses to persistent between secretory and inflammatory disease and optimizing diarrhea is 10 times greater than for bloody diarrhea without case management (Huicho and others 1996). High background persistent diarrhea. HIV infection is another risk factor for per- levels, probably from frequent infections, limits the use of such sistent diarrhea in both adults and children (Keusch and others tests in developing countries, where they would be most useful 1992). Management focuses on overcoming the nutritional (Gill and others 2003). alterations initiated by persistent diarrhea. Simple microscopy for protozoa or helminths can be quick Bloody diarrhea, defined as diarrhea with visible or micro- and effective when the proper sample is obtained and a well- scopic blood in the stool, is associated with intestinal damage trained technician is available to examine a fresh specimen, but and nutritional deterioration, often with secondary sepsis. Some these prerequisites are often not available in developing coun- dehydration--rarely severe--is common, as is fever. Clinicians tries. Newer immunological and nucleic acid­based tests to often use the term bloody diarrhea interchangeably with dysen- detect pathogen-specific factors hold great promise for all diar- tery; however,dysentery is a syndrome consisting of the frequent rhea agents, but they are too expensive or require specialized passage of characteristic, small-volume, bloody mucoid stools; instrumentation and trained technicians. For the foreseeable abdominal cramps; and tenesmus, a severe pain that accompa- future, then, syndromic diagnosis will be the norm. nies straining to pass stool. Those features show the severity of the inflammation. Agents that cause bloody diarrhea or dysen- tery can also provoke a form of diarrhea that clinically is not SYNDROMIC DIAGNOSIS bloody diarrhea, although mucosal damage and inflammation are present, and fecal blood and white blood cells are usually Three major diarrhea syndromes exist. They are acute watery detectable by microscopy. The release of host-derived cytokines diarrhea, which results in varying degrees of dehydration; per- causes fever, altering host metabolism and leading to the break- sistent diarrhea, which lasts 14 days or longer, manifested by down of body stores of protein, carbohydrate, and fat and the malabsorption, nutrient losses, and wasting; and bloody diar- loss of nitrogen and other nutrients. Those losses must be rhea, which is a sign of the intestinal damage caused by inflam- replenished during convalescence, which takes much longer mation. The three are physiologically different and require spe- than the illness does to develop. For these reasons, bloody diar- cific management. Syndromic diagnosis provides important rhea calls for management strategies that are markedly different clues to optimal management and is both programmatically than those for watery or persistent diarrhea. New bouts of infec- and epidemiologically relevant. tion that occur before complete restoration of nutrient stores Acute watery diarrhea can be rapidly dehydrating, with can initiate a downward spiral of nutritional status terminating stool losses of 250 milliliters per kilogram per day or more, a in fatal protein-energy malnutrition (Keusch 2003). 372 | Disease Control Priorities in Developing Countries | Gerald T. Keusch, Olivier Fontaine, Alok Bhargava, and others DIARRHEA, ENVIRONMENT, AND POVERTY Number of episodes per person per year Diarrheal disease affects rich and poor, old and young, and 6 those in developed and developing countries alike, yet a strong 5 relationship exists between poverty, an unhygienic environ- ment, and the number and severity of diarrheal episodes-- 4 especially for children under five. 3 Poverty is associated with poor housing, crowding, dirt floors, lack of access to sufficient clean water or to sanitary dis- 2 posal of fecal waste, cohabitation with domestic animals that 1 may carry human pathogens, and a lack of refrigerated storage for food--all of which increase the frequency of diarrhea. 0 0­5 6­11 1 year 2 years 3 years 4 years Poverty also restricts the ability to provide age-appropriate, months months nutritionally balanced diets or to modify diets when diarrhea Age group develops so as to mitigate and repair nutrient losses. The 1955­79 1980­90 1990­2000 impact is exacerbated by the lack of adequate, available, and affordable medical care. Thus, the young suffer from an appar- Source: Authors. ently never-ending sequence of infections, rarely receive appro- priate preventive care, and too often encounter the health care Figure 19.1 Median Age-Specific Incidences for Diarrheal Episodes system when they are already severely ill. per Child per Year from Three Reviews of Prospective Studies in Developing Areas, 1955­2000 Although the presence of blood in the stool is a recognized danger signal, prompting more urgent care seeking, even these patients either are not treated early or receive poor medical collection protocols (and only the later study includes data care. Ironically, the poor spend considerable amounts on inap- from China). Remarkably, the estimated median incidence of propriate care and useless drugs purchased from local shops diarrheal disease in children under five in developing countries and untrained practitioners. If antibiotics are properly pre- has not changed much since the early 1990s (figure 19.1): 3.2 scribed, poverty often limits the purchase of a full course of episodes per child per year in 2003 (Parashar and others 2003) treatment or leads to cessation of treatment as soon as symp- compared with 3.5 episodes per child per year in 1993 (Jamison toms improve, even though the infection has not been cured. and others 1993). However, many fewer surveys were available for the most recent review (31 in 20 countries) compared with the 1993 consensus (276 in 60 countries), reflecting diminished PUBLIC HEALTH SIGNIFICANCE OF support for the systematic collection of incidence data. DIARRHEAL ILLNESSES Incidence rates in Sub-Saharan Africa and Latin America are clearly greater than in Asia or the Western Pacific, while subject Continuing surveillance and longitudinal studies allow track- to greater data limitations from individual countries. Incidence ing of current levels and trends in diarrhea incidence and continues to show a peak in infants age 6 to 11 months, drop- mortality and provide the basis for future projections and for ping steadily thereafter. evaluations of different control strategies. The seemingly lower estimates of diarrheal incidence before 1980 (Snyder and Merson 1982) are likely due to methodolog- ical differences. These estimates are not precise or directly com- Morbidity parable; the trends are most relevant. The persistently high Comparisons over time of the global burden of diarrheal dis- rates of diarrhea throughout the 1990s despite intensive efforts eases have revealed secular trends and demonstrated the at control, particularly among children age 6 to 24 months, is of impact of public health interventions (Bern and others 1992; particular concern. Early childhood diarrhea during periods Kosek, Bern, and Guerrant 2003; Snyder and Merson 1982). of critical postnatal development may have long-term effects The long-term consequences of diarrhea are only now being on linear growth and on physical and cognitive functions. systematically assessed and are not reflected in earlier studies. Data on the incidence of shigellosis, the principal cause of Reviews in 1992 (Bern and others) and 2003 (Kosek, Bern, bloody diarrhea in developing countries, are even more lim- and Guerrant) are similar in many ways--for example, ited. Kotloff and others' (1999) review of studies on Shigella assessing morbidity at least twice weekly--but differ signifi- infection estimates that more than 113 million episodes occur cantly in the use of different sources for data on children under every year in children under five in developing countries, or 0.2 five and in the inclusion of studies differing in design and data episodes of bloody diarrhea per year caused by Shigella species. Diarrheal Diseases | 373 Millions of deaths per year · advising mothers to increase fluids and continue feeding during future episodes. 5 4 Victora and others' (2000) review provides evidence that this strategy, and especially oral rehydration therapy (ORT), has 3 influenced the outcome of dehydrating diarrhea. Data from 99 2 national surveys carried out in the mid 1990s and compiled by the United Nations Children's Fund (UNICEF) increasingly 1 show that diarrhea patients are appropriately managed in most parts of the world,with overall use rates of ORS or recommended 0 1975 1980 1985 1990 1995 2000 2005 home fluids reaching 49 percent. Country case studies in Brazil, the Arab Republic of Egypt, Mexico, and the Philippines showed Snyder and Merson 1982 Trend estimate a dramatic reduction of diarrhea mortality as ORT use rates Bern and others 1992 Parashar and others 2003 Kosek, Bern, and Guerrant 2003 Boschi-Pinto and Tomaskovic forthcoming increased from close to zero in the early 1980s to 35 percent in Brazil, 50 percent in Egypt, 81 percent in Mexico, and 33 percent Source: Authors. in the Philippines in the early 1990s. Hospital admissions for diarrhea also plummeted (Victora and others 2000). As mor- Figure 19.2 Estimates of Diarrhea Mortality, 1975­2000 tality attributable to acute dehydration decreased, the propor- tionate mortality associated with persistent diarrhea increased. Mortality Data from Brazil and Egypt suggest that even relatively low ORT use rates can positively affect mortality,because ORT use tends to Bern and others (1992); Kosek, Bern, and Guerrant (2003); and be much higher for severe illness (Victora and others 2000). Snyder and Merson (1982) also estimate diarrheal mortality Worldwide mortality caused by Shigella infection is esti- using data from longitudinal studies with active surveillance in mated to be 600,000 deaths per year among children under five, place (figure 19.2). The estimate before 1980 was 4.6 million or a quarter to a third of all diarrhea-related mortality in this deaths per year. This estimate dropped to 3.3 million per year age group (Kotloff and others 1999). Because mortality caused between 1980 and 1990 and to 2.6 million per year between by bloody diarrhea is not tracked separately, it is difficult to 1990 and 2000. Two other studies (Parashar and others 2003; assess the impact of standard case management recommenda- Boschi-Pinto and Tomaskovic forthcoming) report even lower tions, and disease-specific trends cannot be tracked. In the past figures for 1990­2000: 2.1 million and 1.6 million deaths per few years, however, data from the International Centre for year, respectively. Methodological variations (inclusion of stud- Diarrheal Disease Research, in Bangladesh, have shown a ies with different designs and data collection methods and marked decrease in the rate of hospitalization caused by inclusion of data from China, different sources for estimating Shigella, especially S. dysenteriae type 1, the most severe form of the number of children under five, and different strategies for shigellosis. Some investigators have suggested that this decrease calculating mortality for this age group) may account for some may be because Shigella infections are now in the low part of a of the striking differences. However, the end of the 20th cen- 10-year cycle (Legros 2004). The observed change could also be tury witnessed significant reductions in diarrheal deaths in explained by better case management with more efficacious children under five. antimicrobials. More comprehensive, syndrome-specific sur- This steady decline in diarrheal mortality, despite the lack of veillance data will be required if rational control priorities are significant changes in incidence, is most likely due to modern to be set, because the options for dehydrating and bloody diar- case management (introduced since the 1980s) and to the rheal diseases differ substantially. improved nutrition of infants and children. Major recommen- Despite national data that indicate a significant decline in dations include the following: mortality (Baltazar, Nadera, and Victora 2002; Miller and Hirschhorn 1995), diarrheal diseases remain among the five · counseling mothers to begin suitable home-prepared rehy- top preventable killers of children under five in developing dration fluids immediately on the onset of diarrhea countries and among the top two in many. · treating mild to moderate dehydration early with oral rehy- dration solution (ORS), reserving intravenous electrolytes for severe dehydration · continuing breastfeeding and complementary foods during Long-Term Consequences diarrhea and increasing intake afterward The long-term consequences of diarrheal diseases remain · limiting antibiotic use to cases of bloody diarrhea or dysen- poorly studied, and analyses of global trends have not consid- tery and avoiding antidiarrheal and antimotility drugs ered them. Niehaus and others (2002) recently evaluated the 374 | Disease Control Priorities in Developing Countries | Gerald T. Keusch, Olivier Fontaine, Alok Bhargava, and others long-term consequences of acute diarrheal disease on psy- Those data underpin the global campaign to promote exclu- chomotor and cognitive development in young children. sive breastfeeding for the first six months of life by increasing Following a cohort of 47 children in a poor urban community both the initiation and the duration of exclusive breastfeeding. in northeastern Brazil, they correlated the number of diarrheal The strategies include the following: episodes in the first two years of life with measures of cognitive function obtained four to seven years later. They found a sig- · hospital policies and actions to encourage breastfeeding and nificant inverse correlation (average decrease of 5.6 percent) discourage bottle feeding between episodes of early diarrheal disease and overall intellec- · counseling and education provided by peers or health tual capacity and concentration, even when controlling for workers maternal education or helminth infection, which are known · mass media and community education to be independent predictors of malnutrition and cognitive · mothers' support groups. defects. Test scores were also 25 to 65 percent lower in children with an earlier history of persistent diarrhea. Interventions focused on hospital practices apply where Recent evidence suggests that genetic factors may also be most women deliver in such facilities. Such interventions have involved in the developmental response to repeated diarrhea shown up to a 43 percent increase in exclusive breastfeeding (Oria and others 2005). Better and more sensitive assessment with good institutional policies and retraining of health staff tools are needed to define the relationships between diarrheal (Westphal and others 1995). Interventions focused on educa- diseases and developmental disorders and to calculate individ- tion and counseling increase exclusive breastfeeding by 4 to ual and societal costs and the cost-effectiveness of interven- 64 percent (Sikorski and others 2002). Peer-counseled women tions. In addition, early childhood malnutrition resulting from are less likely to stop exclusive breastfeeding than are those who any cause reduces physical fitness and work productivity in receive either professional support or no support, and their adults (Dobbing 1990). infants are 1.9 to 2.9 times less likely to have diarrhea (Barros and others 1995; Haider and others 1996). No large-scale peer counseling programs exist; therefore, feasibility is unknown. PREVENTIVE STRATEGIES Community-based mother's support groups are sustainable, but they have low coverage and are biased toward women who Strategies for controlling diarrheal diseases have remained are already motivated to breastfeed (Bhandari and others substantially unchanged since the 1993 edition of this volume 2003). Mass media can be effective where media coverage is (Martinez, Phillips, and Feachem 1993). The World Health high, where production skills are good, and where it addresses Organization (WHO 2004) recently reevaluated these inter- barriers to breastfeeding instead of just proclaiming its bene- ventions to determine the extent to which they have been effec- fits. We found no studies that examined the relationship tively implemented and their effect. between breastfeeding promotion and diarrheal disease mor- tality; however, estimates suggest such promotion could decrease all-cause mortality in children under five by 13 per- Promotion of Exclusive Breastfeeding cent (Jones and others 2003). Exclusive breastfeeding means no other food or drink, not Maternal HIV infection has put a new wrinkle in the "breast even water, is permitted, except for supplements of vitamins is best" dogma because of the risk of transmission of infection and minerals or necessary medicines. The optimal duration to the infant (De Cock and others 2000). There is a trade-off, of exclusive breastfeeding is six months (WHO 2001). A however, between the risk of mortality associated with replace- meta-analysis of three observational studies in developing ment feeding and the risk of HIV infection, especially where countries shows that breastfed children under age 6 months safe replacement feeding is difficult. For women who are HIV- are 6.1 times less likely to die of diarrhea than infants who are negative or whose status is unknown, WHO currently recom- not breastfed (WHO Collaborative Study Team 2000). mends exclusive breastfeeding for at least six months (WHO Exclusive breastfeeding protects very young infants from diar- 2000). The best option for HIV-positive women is acceptable, rheal disease in two ways: first, breast milk contains both affordable, sustainable, and safe replacement feeding. If this immune (specific) and nonimmune (nonspecific) antimicro- option is not possible, there are four alternatives: (a) heat- bial factors; second, exclusive breastfeeding eliminates the treated breast milk, (b) HIV-negative wet nurses, (c) unconta- intake of potentially contaminated food and water. Breast minated donor milk, or (d) exclusive breastfeeding for six milk also provides all the nutrients most infants need up to months and rapid discontinuation thereafter (WHO 2003). age 6 months. When exclusive breastfeeding is continued dur- A danger of promoting replacement feeding is that unin- ing diarrhea, it also diminishes the adverse impact on nutri- fected women or women with unknown HIV status will adopt tional status. the practice. Even in high-prevalence communities, the best Diarrheal Diseases | 375 option for women with unknown status for the overall health Brown, Dewey, and Allen (1998) reviewed experiences with of their children appears to be exclusive breastfeeding for six large-scale complementary feeding interventions in 14 coun- months. In Coutsoudis and others' (1999) cohort study in tries. They demonstrate that it is possible to provide nutrition- South Africa, the risk of mother-to-infant transmission of HIV ally improved complementary foods in diverse cultural settings after three months of exclusive breastfeeding was similar to that and that poor mothers are willing to prepare new foods their with no breastfeeding and significantly lower than that with children will eat. However, caregivers face considerable time mixed feeding. Providing antiretroviral therapy to the mother and resource constraints in providing such foods, especially should significantly extend the period of safe breastfeeding for during episodes of illness. A pilot study in Brazil that imple- the initially HIV-negative infants of HIV-positive mothers. mented nutritional counseling through the Integrated Management of Childhood Illness Program reported signifi- cant weight gain in children age one year or more, but not in Improved Complementary Feeding Practices younger children (Santos and others 2001). Ideally, complementary foods should be introduced at age Unfortified complementary foods do not meet all essential 6 months, and breastfeeding should continue for up to two micronutrient requirements. Although improvements in vita- years or even longer to increase birth intervals (WHO 2003). min A status do not significantly reduce the incidence of diar- There is a strong inverse association between appropriate, safe rhea and other common childhood illnesses, vitamin A supple- complementary feeding and mortality in children age 6 to 11 mentation can reduce the frequency of severe diarrhea (Barreto months. Malnutrition is an independent risk predictor for the and others 1994) and mortality (Ross and others 1995). frequency and severity of diarrheal illness. There is a vicious Chapter 28 describes interventions to promote vitamin A cycle in which sequential diarrheal disease leads to increasing intake. Zinc supplementation also reduces the incidence of nutritional deterioration, impaired immune function, and diarrhea. greater susceptibility to infection. The cycle may be broken by interventions to decrease infection incidence to reduce malnu- trition (Keusch and Scrimshaw 1986) or improving nutritional Rotavirus Immunization status to reduce the burden of infection (Victora and others Almost all infants acquire rotavirus diarrhea early in life, and 1999). rotavirus accounts for at least one-third of severe and poten- Improved feeding practices to prevent or treat malnutrition tially fatal watery diarrhea episodes--primarily in developing could save as many as 800,000 lives per year (Jones and others countries, where an estimated 440,000 vaccine-preventable 2003). Pediatricians have long been aware of an increase in rotavirus deaths per year occur (Parashar and others 2003), diarrhea incidence during weaning from exclusive breast milk compared with about a dozen in a developed country such as feeding. Microbial contamination of complementary foods France (Fourquet and others 2003). An effective rotavirus vac- (Mondal and others 1996) and nutritionally inadequate diets cine would have a major effect on diarrhea mortality in devel- during and after diarrhea episodes (Badruddin and others oping countries. 1991) increase the risk. Contamination of complementary In 1998, a quadrivalent Rhesus rotavirus­derived vaccine foods can potentially be reduced by educating caregivers on that reduced the frequency of severely dehydrating rotavirus-- hygienic practices (Guptill and others 1993), improving home but not the overall incidence of rotavirus infections--was food storage (English and others 1997), fermenting foods to licensed in the United States (Glass and others 1999). It was reduce pathogen multiplication (Kimmons and others 1999), cost-effective, even at US$100 for a full course of immuniza- or ingesting nonpathogenic probiotic microorganisms that col- tion, when direct economic losses resulting from health care onize the gut and help resist pathogens (Allen and others expenses and indirect costs of lost productivity and wages for 2004). These interventions have not been evaluated at scale in the caretakers were considered (Tucker and others 1998). The communities, and effectiveness trials are lacking. strategy was clear: use the high-priced vaccine routinely in We could not find any reports on the effects of complemen- industrial countries to subsidize its use in developing countries. tary feeding interventions on mortality. Five efficacy trials to However, postmarketing surveillance detected an apparent improve the intake of complementary foods noted a net increase in a relatively rare event, intussusception, a condition increase in energy intake of between 65 and 300 kilocalories a in which the intestine telescopes on itself, causing a potentially day and improvements of 0.25 to 0.46 standard deviations in serious obstruction (CDC 1999a). The relationship was weight-for-age and 0.04 to 0.35 standard deviations in height- strongest with the first dose of vaccine given with the first or for-age (Caulfield, Huffman, and Piwoz 1999). By extrapola- second dose of diphtheria-pertussis-tetanus vaccine (Peter and tion, this increment in growth should translate into a 2 to 13 others 2002), although this was counterbalanced by a decrease percent reduction in deaths associated with malnutrition in the incidence of intussusception in older children (Murphy (Black and others 1995). and others 2003). 376 | Disease Control Priorities in Developing Countries | Gerald T. Keusch, Olivier Fontaine, Alok Bhargava, and others The overall reduced incidence in immunized infants com- outbreak in Micronesia suggested that a single dose was useful pared with nonimmunized infants in these studies suggested in limiting the spread of cholera (Calain and others 2004). But that the vaccine may actually protect against later adverse because ORT is so inexpensive and useful in preventing death, events. Nonetheless, the ensuing controversy led to a reversal of immunization is not a high priority. Only Vietnam routinely the recommendation for universal immunization in the United deploys cholera vaccine. States and withdrawal of the vaccine from the market, preclud- Operational information on the costs, logistics, and avail- ing the possibility of its deployment in developing countries ability of vaccines for use by global programs and on the vul- (CDC 1999b). Because very young infants are less prone to nerable populations in high-risk settings who would benefit develop intussusception, initial immunization at birth might from cholera vaccine remains limited. Although scientific have been entirely safe. interest in a cholera vaccine remains high, its public health pri- Despite this setback, efforts to produce an effective and safe ority is less than that of a vaccine for rotavirus or Shigella. rotavirus vaccine continue. The Rhesus vaccine has been reli- censed to another manufacturer, and new vaccines derived Measles Immunization from human or bovine rotavirus are undergoing field trials in Measles is known to predispose to diarrheal disease secondary developing countries (Dennehy 2005). A monovalent human to measles-induced immunodeficiency. Feachem and Koblinsky rotavirus vaccine was introduced in Mexico in 2005. The entry (1983) estimate that measles vaccine given to 45 to 90 percent of both China and India into rotavirus vaccine development of infants would prevent 44 to 64 percent of measles cases, 0.6 and their potential for manufacturing quality vaccines at low to 3.8 percent of diarrheal episodes, and 6 to 26 percent of diar- cost will make it easier to deploy an effective vaccine where it is rheal deaths among children under five. Global measles immu- really needed. nization coverage is now approaching 80 percent, and the dis- ease has been eliminated from the Americas, raising hopes for Cholera Immunization global elimination in the near future (GAVI 2005), with a pre- dictable reduction in diarrhea as well. Endemic cholera is primarily a pediatric disease, although adult morbidity and mortality are significant, especially dur- ing epidemics. The lethality of cholera is due to the physio- Improved Water and Sanitary Facilities and Promotion logical consequences of rapid and profound dehydration. of Personal and Domestic Hygiene Oral rehydration therapy has dramatically improved survival Human feces are the primary source of diarrheal pathogens. and reduced the cost of treatment. Wherever parenteral and Poor sanitation, lack of access to clean water, and inadequate oral rehydration is readily available, even in epidemic situa- personal hygiene are responsible for an estimated 90 percent of tions, a cholera mortality rate above 1 percent indicates fail- childhood diarrhea (WHO 1997). Promotion of hand washing ure of the public health system to provide appropriate case reduces diarrhea incidence by an average of 33 percent (Huttly, management. Morris, and Pisani 1997); it works best when it is part of a A vaccine would further reduce the morbidity and mortal- package of behavior change interventions. Effects on mortality ity associated with cholera in endemic areas; however, devel- have not been demonstrated. However, the required behavior oping an effective, safe vaccine has proven difficult. The most change is complex, and significant resources are needed. immunogenic and protective vaccines tested thus far are Antiseptic soaps are more costly than plain hand soap and con- administered orally. Two such vaccines have been licensed: an fer little advantage. Washing hands after defecating or handling attenuated live vaccine and a heat-killed vaccine combined children's feces and before handling food is recommended, but with recombinant cholera toxin B subunit, which functions it entails an average of 32 hand washes a day and consumes as an immunoadjuvant (Graves and others 2000; Ryan and 20 liters of water (Graef, Elder, and Booth 1993). If soap is too Calderwood 2000). Many developing countries can produce costly, ash or mud can be used, but access to water remains the killed vaccine, especially without cholera toxin B. Current essential (Esrey 1996). oral cholera vaccines appear to be safe and offer reasonable Six rigorous observational studies demonstrated a median protection for a limited period; however, the main users have reduction of 55 percent in all-cause child mortality associated been individual travelers from industrial countries who may with improved access to sanitation facilities (Esrey, Feachem, be exposed to the risk of cholera while traveling in endemic and Hughes 1985). The greatest effect of improving sanitation areas. systems will be in areas of high population density and wher- The use of oral cholera vaccine in mass vaccination cam- ever the entire community, rather than single households, paigns as an adjunct to good case management, disposal of adopts the intervention. Current technology can be costly and fecal waste, and access to safe water during humanitarian dis- difficult to maintain, and in some settings it is simply not asters has recently been reviewed (WHO 1999). Analysis of an feasible. Diarrheal Diseases | 377 CASE MANAGEMENT antimicrobials decrease significantly (Baqui and others 2004). Large community-based studies are being implemented to cor- Two recent advances in managing diarrheal disease--(a) newly roborate these potentially important findings. formulated ORS containing lower concentrations of glucose and salts and (b) zinc supplementation--used in combination with promotion of exclusive breastfeeding, general nutritional Management of Bloody Diarrhea support, and selective and appropriate use of antibiotics, can The primary treatment for shigellosis, the most common and further reduce the number of diarrheal deaths among children. severe cause of bloody diarrhea, is antimicrobials. The choice of Families and communities are key to achieving case manage- effective, safe, and inexpensive oral drugs for use in developing ment goals by making these recommendations routine practice countries has, however, become problematic because of the in homes and health facilities. increasing prevalence of antimicrobial drug resistance (Salam 1998). Tetracycline, ampicillin, and the fixed-ratio combina- tion of trimethoprim and sulfmethoxazole, once used as first- New Oral Rehydration Solutions line treatment, are no longer reliably effective. When epidemic For more than 25 years, UNICEF and WHO have recom- dysentery caused by multidrug-resistant S. dysenteriae type 1 mended a single formulation of glucose-based ORS considered appeared in Africa and Asia in the 1980s and 1990s, nalidixic optimal for cholera, irrespective of cause or age group affected. acid was pressed into use (Salam and Bennish 1988). Nalidixic This formulation has proven effective and without significant acid is a drug used primarily for urinary tract infections, but it adverse effects (Ruxin 1994), but because watery stools persist is also effective against Shigella. Clinical responses were initially and duration of diarrhea is not reduced, mothers' and health excellent, but with continued use, resistance to nalidixic acid workers' acceptance of current ORSs has been suboptimal. has been increasing in many parts of the world (Dutta and During the past 20 years, efforts to improve ORS to treat others 2003). dehydration from all types of diarrhea and reduce stool output A number of other drugs have been tested and shown or duration have continued--for example, by reducing the effective, including ceftriaxone, azithromycin, pivmecillinam, sodium content in line with sodium losses for noncholera diar- and some new generation 5-fluoroquinolones, such as rhea. Compared with standard ORS, lower sodium and glucose ciprofloxacin (Salam 1998). Because of its effectiveness, safety, ORS reduces stool output, vomiting, and the need for intra- ease of administration by the oral route, short course, and low venous fluids (Hanh, Kim, and Garner 2001). If household use cost (US$0.10 for a three-day course for a 15-kilogram child), increases, new ORS can reduce childhood deaths from non- ciprofloxacin is the current drug of choice for shigellosis cholera diarrhea (Duggan and others 2004), and it appears to (Zimbasa Dysentery Study Group 2002). However, be as effective as standard ORS for children or adults with ciprofloxacin-resistant strains are already appearing (Pazhani cholera. A WHO expert group now recommends that ORS and others 2004), and it is only a matter of time before resist- containing 75 milliequivalents of sodium and 75 millimoles of ance becomes widespread, especially if the drug is readily avail- glucose per liter (total osmolarity, 245 milliosmoles per liter) be able and indiscriminately used. Because of these concerns, used everywhere (WHO and UNICEF 2004). development of a vaccine for Shigella is critical. The Diseases of the Most Impoverished initiative, supported by the Bill & Melinda Gates Foundation (Nossal 2003), which promotes vac- Zinc Supplementation cine development for Shigella, cholera, and typhoid, is a signif- A review of all relevant clinical trials indicates that zinc supple- icant advance since the previous edition of this volume. ments given during an episode of acute diarrhea reduce both duration and severity and could prevent 300,000 deaths in chil- dren each year (Black 2003). WHO and UNICEF now recom- COST-EFFECTIVENESS OF INTERVENTIONS mend that all children with acute diarrhea be given zinc in some form for 10 to 14 days during and after diarrhea (10 mil- Cost-effectiveness ratios of diarrheal disease interventions were ligrams per day for infants younger than 6 months and 20 mil- calculated by World Bank region in terms of disability-adjusted ligrams per day for those older than 6 months) (WHO and life years (DALYs) averted for a model population of 1 million, UNICEF 2004). following the standardized guidelines of the Disease Control Pilot studies in Brazil, Egypt, Ethiopia, India, Mali, Pakistan, Priorities Project for economic analyses (see chapter 15). and the Philippines that include zinc routinely in the manage- Europe and Central Asia were excluded because data were lack- ment of acute diarrhea not only show an improvement over ing owing to the low prevalence of disease. Input variables ORS alone but also suggest two important new effects: (a) use included (a) region-specific diarrhea morbidity rates adapted rates of ORS increase, and (b) use rates of antidiarrheals and from Kosek, Bern, and Guerrant (2003); (b) region-specific 378 | Disease Control Priorities in Developing Countries | Gerald T. Keusch, Olivier Fontaine, Alok Bhargava, and others underlying mortality rates and age structures provided by the for 2000. Infrastructure improvements for rural and urban Disease Control Priorities Project; (c) median intervention populations were considered separately because of differences effectiveness rates (that is, percentage of diarrheal morbidity in infrastructure type and cost, although the same effectiveness reduction and percentage of diarrheal mortality reduction); rates were used for both. and (d) median per capita intervention costs gathered from the The per child treatment costs and effectiveness rates used literature and from personal communications (table 19.1). are presented in table 19.1. Cost per treatment of ORT varied Because approximately 90 percent of all cases in the devel- widely depending on the type and method of ORT imple- oping world occur in children under five, the analysis focused mented. Oral rehydration therapy can be as inexpensive as on this age group alone. Uniform intervention effectiveness US$0.02 per child treated--the cost of a home remedy with rates were assumed for all regions because region-specific sugar and salt. However, treatment can become substantially information was not available. Regional variations in cost- more expensive if commercially manufactured ORS is used or effectiveness were due to regional variations in the prevalence if there are substantial personnel or infrastructure costs of diarrheal disease, in the diarrhea-attributable morbidity and (Martinez, Phillips, and Feachem 1993). Finally, our analysis mortality, and in the intervention cost, where region-specific considered only long-run marginal costs (which vary with the information was available. number of individuals treated) and did not include fixed costs Disability-adjusted life years are averted through the avoid- of initiating a program where none currently exists. ance of cotemporaneous disability and mortality attributable Figure 19.3 shows the cost-effectiveness of all interventions to diarrhea. We did not consider long-term developmental and over the first five years of life. Two interventions administered cognitive effects of childhood diarrhea or the external benefits during the first year of life--breastfeeding promotion (US$930 of interventions unrelated to diarrhea (for instance, benefits of per DALY) and measles immunization (US$981 per DALY)-- measles immunization unrelated to diarrhea or other health were the most cost-effective. ORT (US$1,062 per DALY) and benefits of improved public water and sanitation). Therefore, water and sanitation in rural areas (US$1,974 per DALY) were our estimates err on the conservative side. the next most cost-effective, but only if they were implemented We explored two general categories of interventions: early continuously for five years, thereby allowing an entire cohort of interventions that take place within the first year of life-- effectively treated children age zero to four to survive past the breastfeeding promotion and immunizations for rotavirus age at which they are most at risk for diarrheal infection, dis- (with the prototype Rhesus reassortant tetravalent vaccine), ability, and mortality. Rotavirus immunization (US$2,478 per cholera (with live oral vaccine), and measles--and other inter- DALY), cholera immunization (US$2,945 per DALY), and ventions that treat an entire cohort of children under five simul- water and sanitation in urban areas (US$6,396 per DALY) were taneously (improved water and sanitation). For early interven- the least cost-effective. tions, cost-effectiveness ratios were calculated by considering Among the early interventions, breastfeeding promotion was the cost of treating all newborns in a single year and the bene- less effective than other interventions but also less expensive fits (DALYs averted) from those treatments that occur over the than rotavirus and measles vaccination (table 19.1). Cholera first five years of life. These benefits include avoided mortality vaccination was less expensive than breastfeeding promotion, that allows individuals to live to the expected life expectancy for but it was also many times less effective because of the signifi- the region. Other interventions included ORT and improved cantly higher prevalence of diarrhea that is not related to water and sanitation infrastructure. Because a single year of cholera--making cholera vaccination the least cost-effective of these interventions yields only cotemporaneous benefits-- the early interventions considered. Oral rehydration therapy because effectively treated individuals do not necessarily live to and water and sanitation interventions were more effective than life expectancy given that they are likely to be reinfected the next breastfeeding and vaccination interventions in reducing mor- year--we calculated cost-effectiveness of a five-year interven- bidity and mortality caused by diarrhea, but they were also more tion. Analysis of a five-year intervention enabled us to consider expensive. However, our analysis for water and sanitation did the case in which an entire cohort of children age zero to four not consider the benefits of this intervention other than those avoids early childhood diarrheal mortality because of the inter- related to health, and the high cost-effectiveness ratio is more a vention and receives the benefit of living to life expectancy. limitation of our methodology than of the intervention itself. Disability and deaths averted for those benefiting from The high cost-effectiveness ratio for ORT is attributable to improved water and sanitation were calculated from only the the high variation in reported treatment costs, which may fraction of the model populations currently without access. For inflate the median cost used in this analysis (table 19.2). Given each region, the proportion of rural and urban children age the range of reported treatment costs (table 19.1), the cost- zero to four currently without access to improved water and effectiveness ratio of ORT could be as low as US$4 per DALY or sanitation was calculated using region-specific information as high as US$2,124 per DALY in low- and middle-income from World Bank Development Indicators (World Bank 2002) countries. High variation in reported treatment costs results in Diarrheal Diseases | 379 380 Table 19.1 Cost and Effectiveness Values Used to Calculate Cost-Effectiveness Ratios for Select Interventions for Diarrhea for Children under Age Five | Disease Median Median diarrhea Morbidity Median diarrhea Mortality Source regions cost/child Cost/child range morbidity reduction reduction range mortality reduction reduction range Control Model regions Sources or countries (2001 US$) (2001 US$) (percent) (percent) (percent) (percent) Priorities Breastfeeding promotion Costs in LMICs, EAP, MENA, Horton and others 1996; LMICs, Brazil, 8.98 0.46­17.50 Developing SA, SSA Martinez, Phillips, and Honduras, Mexico Feachem 1993 LAC Horton and others 1996 Brazil, Honduras, 1.86 0.46­3.26 Countries Mexico Effectiveness (age 0 to 5) | Gerald All Feachem and Koblinsky LMICs 4.5 1­8 10.5 4­17 1984 T.Keusch, Rotavirus immunization with RRV-TV Costs Olivier LMICs, EAP, LAC, Martinez, Phillips, and LMICs, India 53.80 3.33­104.30 MENA, SSA Feachem 1993; Narula, Fontaine, Tiwari, and Puliyeh 2004 SA Narula, Tiwari, and India 104.30 -- Alok Puliyeh 2004 Bhargava, Effectiveness (age 0 to 5) All Parashar and others Brazil, Peru, R. B. 8.54 -- 24.1 -- 1998a de Venezuela and Cholera immunization with live oral vaccine others Costs LMICs, EAP, MENA, Cookson and others 1997; LMICs, Argentina 3.65 1.70­5.60 SA, SSA Martinez, Phillips, and Feachem 1993 LAC Cookson and others 1997 Argentina 1.70 -- Effectiveness (age 0 to 5) All de Zoysa and Feachem Bangladesh 0.095 0.06­0.13 1.5 1­2 1985 Measles immunization Costs LMICs, LAC, Duke 1999; Feachem and LMICs, Côte d'Ivoire, 13.26 0.52­26.00 MENA, SA Koblinsky 1983; Phillips, Ghana, Indonesia, Feachem, and Mills 1987; Papua New Guinea, Martinez, Phillips, and Zambia Feachem 1993; Shann 2000 EAP Duke 1999; Phillips, Indonesia, 1.62 0.52­1.10 Feachem, and Mills 1987; Papua New Guinea Shann 2000 SSA Feachem and Koblinsky Côte d'Ivoire, 15.00 4.00­26.00 1983; Phillips, Feachem, Ghana, Zambia and Mills 1987 Effectiveness (age 0 to 5) All Feachem and LMICs 2.2 0.6­3.8 16 6.4­25.6 Koblinsky 1983 Water supply and sanitation improvement Costs (rural) All Esrey, Feachem, and LMICs 25.00 -- Hughes 1985 Costs (urban) All Esrey, Feachem, and LMICs 81.00 -- Hughes 1985 Effectiveness (age 0 to 5) All Esrey, Feachem, and LMICs 24 22­26 65 -- Hughes 1985; Esrey and others 1991 (Continues on the following page.) Diarrheal Diseases | 381 382 Table 19.1 Continued | Disease Median Median diarrhea Morbidity Median diarrhea Mortality Source regions cost/child Cost/child range morbidity reduction reduction range mortality reduction reduction range Control Model regions Sources or countries (2001 US$) (2001 US$) (percent) (percent) (percent) (percent) Priorities Oral rehydration therapy Costs in LMICs Horton and Claquin 1983; Bangladesh, Arab 5.50 0.02­11.00 Developing Islam, Mahalanabis, and Rep. of Egypt, The Majid 1994; Qualls and Gambia, Honduras, Robertson 1989; Indonesia, Malawi, Countries Shepard, Brenzel, and Swaziland, Turkey Nemeth 1986; WHO and UNICEF 2001 | Gerald EAP Shepard, Brenzel, and Indonesia 0.71 0.02­1.40 Nemeth 1986 T.Keusch, LAC Shepard, Brenzel, and Honduras 2.59 0.02­5.16 Nemeth 1986 Olivier MENA Shepard, Brenzel, and Arab Rep. of Egypt 4.89 0.02­9.75 Nemeth 1986 Fontaine, SA Horton and Claquin 1983; Bangladesh 2.91 0.02­5.80 Islam, Mahalanabis, and Alok Majid 1994 Bhargava, SSA Shepard, Brenzel, and The Gambia, 5.51 0.02­11.00 Nemeth 1986; Qualls and Malawi, Swaziland Robertson 1989 and Effectiveness (age 0 to 5) others All Boschi-Pinto and LMICs 0 -- 95 -- Tomaskovic forthcoming Source: Authors. LMICs low- and middle-income countries; EAP East Asia and the Pacific; LAC Latin America and the Caribbean; MENA Middle East and North Africa; SA South Asia; SSA Sub-Saharan Africa; -- not available. a. Effectiveness calculated based on vaccine efficacy reported in Parashar and others (1998) and under the assumption that rotavirus infection is responsible for 20 percent of all diarrheal morbidity and severe infection is responsible for 33.3 percent of all diarrheal mortality. Cost-effectiveness ratio (2001US$/DALY) 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 Low- and middle- East Asia and Latin America Middle East and South Asia Sub-Saharan income countries the Pacific and the Caribbean North Africa Africa Breastfeeding promotion Rotavirus immunization Cholera immunization Measles immunization Water and sanitation (urban) Water and sanitation (rural) ORT Source: Authors. Figure 19.3 Cost-Effectiveness: Intervention at Birth through Age 5 with Benefits that Occur over Five Years (age 0­4) Table 19.2 Cost-Effectiveness Ratios of Oral Rehydration Preventive strategies--such as breastfeeding, improving com- Therapy Interventions Based on Minimum, Median, and plementary feeding and using micronutrient supplementation Maximum per Capita Costs (2001 US$/DALY) or fortification, and increasing coverage with the full set of Expanded Programme on Immunization vaccines (especially Region Minimum cost Median cost Maximum cost measles vaccine)--are all useful and effective (GAVI 2005). Low- and middle-income Failure to separately track the full impact of bloody diarrhea-- countries 4 1,062 2,124 especially Shigella infection--on morbidity and mortality or East Asia and the Pacific 4 132 260 to effectively implement good clinical management (including Latin America and the guidelines for and control over the use of antibiotics) has con- Caribbean 20 2,570 5,120 tributed to the continuing burden of bloody diarrhea and Middle East and dysentery worldwide and the alarming increase in antibiotic North Africa 10 2,564 5,113 resistance. The challenges for the next decade will be to South Asia 4 642 1,279 increase or ensure universal appropriate implementation of Sub-Saharan Africa 4 988 1,972 these interventions in developing countries and to avoid a sit- Source: Authors. uation in which they compete for funding and staff time. Delivery of good-quality services is essential, and much high variation in cost-effectiveness for the other regions as well. remains to be learned through research before this require- There remains little doubt, however, about the effect of wide- ment can be met. spread use of ORT on diarrhea morbidity and mortality and Other interventions, such as vaccines against rotavirus, about the associated direct and indirect cost savings for treat- Shigella, or cholera, are either not yet available or not ready for ment and hospitalization. universal administration. Progress toward the development of these vaccines, with the highest priority for the first two, is RESEARCH AGENDA encouraging, but further investments in research and develop- ment will be required before large-scale implementation of Good evidence now supports the view that promoting ORT in these interventions can be considered. The cost of these vac- conjunction with other key interventions, preventive as well as cines will remain a major constraint for poor people, who can- curative, has had a large role in the marked reduction in deaths not pay for the costs of development and ensure reasonable of children caused by diarrhea (Victora and others 2000). profits for industry. However, increased public investment in Diarrheal Diseases | 383 fundamental and applied research, vaccine purchase schemes, CONCLUSIONS and development of low-cost, high-quality manufacturing capacity in developing countries may change the prevailing Existing interventions to prevent or treat diarrheal diseases dynamics. By creating public-private partnerships for vaccine have proven their efficacy in reducing mortality, but a major development, organized as targeted product development pro- challenge for the next 10 years will be to scale up these inter- grams, the public sector, private foundations, and industry are ventions to achieve universal utilization coverage. The United taking steps toward these goals. Nations Millennium Development Goal to reduce the mortality Because of the fecal-oral transmission of enteric pathogens, rate among children under five by two-thirds by 2015 will be improving the supply of safe water and the ability to safely dis- easier to attain if the scale-up goals are reached. New products pose of fecal waste are the best ways to reduce the burden of and tools could significantly improve the efficacy of these inter- morbidity and mortality. However, major investments and crit- ventions--for example, rapid specific diagnostics, new treat- ical improvements in water and sanitary waste disposal on the ment strategies based on reversing the pathophysiology of the necessary scale are unlikely to occur in the next decade or two. infection, simple and effective ways to produce clean water and Local low-tech solutions can be useful, and enhanced efforts to control human waste, and vaccines to prevent illness. However, find ways to improve water cleanliness at the point of use and these products and tools will not become widely available in to build simple latrines that will be used consistently are needed time to influence the achievement of the Millennium (chapter 41). However, in the face of HIV and the attention Development Goals. Continued investment in diarrheal disease being given to tuberculosis and malaria, coordinated efforts to research across the spectrum of basic, social and behavioral, build safe water and sanitation capacity at the local level, one and applied investigations is, therefore, essential, including village at a time, that are sufficient to significantly influence the expanded behavioral research to understand how parents assess burden of illness are unlikely--even though many more infants risk and how actionable health messages can be presented in and children die each year of preventable and treatable diar- different cultures and settings. rhea than of HIV/AIDS. The cycle of research, followed by implementation, followed by research has enabled the development of improved tools to REFERENCES manage diarrheal diseases--tools that have the potential to fur- Allen, S. J., B. Okoko, E. Martinez, G. Gregorio, and L. F. Dans. 2004. ther drive down diarrhea mortality. The challenge is to achieve "Probiotics for Treating Infectious Diarrhea." Cochrane Database high coverage and good practice with ORT and correct diar- Systematic Reviews (2): CD003048. rhea case management, including antimicrobial and nutrition Badruddin, S., A. Islam, K. H. Hendricks, Z. A. Bhutta, S. A. Shaikh, J. D. Snyder, and A. M. Molla. 1991. "Dietary Risk Factors Associated with interventions. Interventions to integrate health care through Acute and Persistent Diarrhea in Karachi, Pakistan." American Journal programmatic initiatives such as the Integrated Management of Clinical Nutrition 51: 745­49. of Childhood Illness program, critically evaluated elsewhere in Baltazar, J. C., D. P. Nadera, and C. G. Victora. 2002. "Evaluation of the this book (chapter 63), could be essential to ensure this high National Control of Diarrhoeal Diseases Programme in the coverage. Some concern remains that in low-resource settings Philippines, 1980­93." Bulletin of the World Health Organization 80: 637­43. such targeted vertical programs may be abandoned, to the Baqui, A. H., R. E. Black, S. El Arifeen, M. Yunus, K. Zaman, N. Begum, and detriment of the goals for disease burden reduction that they others. 2004. "Zinc Therapy for Diarrhoea Increased the Use of Oral were established to achieve. Rehydration Therapy and Reduced the Use of Antibiotics in The challenge posed by the case management of bloody Bangladeshi Children." Journal of Health, Population, and Nutrition 22 (4): 440­42. diarrhea is a different matter. Until a vaccine is available, the Barreto, M. L., L. M. P. Santos, A. M. O. Assis, M. P. N. Araujo, G. G. keystone for managing bloody diarrhea will continue to be the Farenzena, P. A. B. Santos, and R. L. Fiaccone. 1994. "Effect of Vitamin early use of effective antimicrobial agents. That is made diffi- A Supplementation on Diarrhoea and Acute Lower Respiratory-Tract cult by increasing drug resistance, aided by the widespread Infections in Young Children in Brazil." Lancet 344: 228­31. indiscriminate and inappropriate use of antimicrobials, and Barros, F. C., T. C. Semer, S. Tonioli Filho, E. Tomasi, and C. G. Victora. 1995. "The Impact of Lactation Centers on Breastfeeding Patterns, the increasingly difficult task of finding a safe, inexpensive, and Morbidity, and Growth: A Birth Cohort Study." Acta Paediatrica 84: effective oral agent and then ensuring that the drug is given in 1221­26. a clinically optimal manner. From a technical perspective, the Bern, C., J. Martines, I. de Zoysa, and R. I. Glass. 1992. "The Magnitude of development of a vaccine against Shigella infections is still in its the Problem of Diarrhoeal Disease: A Ten-Year Update." Bulletin of the infancy and in need of greater investment. For both watery and World Health Organization 70: 705­14. bloody diarrhea, the challenge of developing drugs to normal- Bhan, M. K., N. Bhandari, S. Sazawal, J. Clemens, and P. Raj. 1989. "Descriptive Epidemiology of Persistent Diarrhoea among Young ize the pathophysiology caused by the infection remains a sci- Children in Rural North India." Bulletin of the World Health entific challenge and a distant hope. Organization 67: 281­88. 384 | Disease Control Priorities in Developing Countries | Gerald T. Keusch, Olivier Fontaine, Alok Bhargava, and others Bhandari, N., R. Bahl, S. Mazumdar, J. Martines, R. E. Black, and M. K. for Child Health Interventions?" Papua and New Guinea Medical Bhan. 2003. "Infant Feeding Study Group: Effect of Community-Based Journal 42: 1­4. Promotion of Exclusive Breastfeeding on Diarrhoeal Illness Dutta, S., D. Dutta, P. Dutta, S. Matsushita, S. K. Bhattacharya, and and Growth: A Cluster Randomised Controlled Trial." Lancet 361: S. Yoshida. 2003. "Shigella dysenteriae Serotype 1, Kolkata, India." 1418­23. Emerging Infectious Diseases 9: 1471­74. Black, M. M., H. Dubowitz, J. Hutcheson, J. Berenson-Howard, and R. H. English, R. M., J. C. Badcock, T. Giay, T. Ngu, A. M. Waters, and S. A. Starr, Jr. 1995. "A Randomized Clinical Trial of Home Intervention for Bennett. 1997. "Effect of Nutrition Improvement Project on Morbidity Children with Failure to Thrive." Pediatrics 95: 807­14. from Infectious Diseases in Preschool Children in Vietnam: Black, R. E. 2003."Zinc Deficiency, Infectious Disease, and Mortality in the Comparison with Control Commune." British Medical Journal 315: Developing World." Journal of Nutrition 133 (Suppl. 1): 1485S­89S. 1122­25. Black, R. E., and C. F. Lanata. 2002. "Epidemiology of Diarrheal Diseases Esrey, S. A. 1996. "Water, Waste, and Well-Being: A Multicountry Study." in Developing Countries." In Infections of the Gastrointestinal Tract, American Journal of Epidemiology 143: 608­23. 2nd ed., ed. M. J. Blaser, P. D. Smith, J. I. Ravdin, H. B. Greenberg, and Esrey, S. A., R. Feachem, and J. M. Hughes. 1985. "Interventions for the R. L. Guerrant, 11­29. Philadelphia: Lippincott, Williams, and Wilkins. Control of Diarrhoeal Diseases among Young Children: Improving Boschi-Pinto, C., and L. Tomaskovic. Forthcoming. "Deaths from Water Supplies and Excreta Disposal Facilities." Bulletin of the World Diarrhoeal Diseases among Children under Five Years of Age in the Health Organization 63: 757­72. Developing World: A Review." Bulletin of the World Health Esrey, S. A., J. B. Potash, L. Roberts, and C. Shiff. 1991."Effects of Improved Organization. Water Supply and Sanitation on Ascariasis, Diarrhoea, Dracunculiasis, Brown, K., K. Dewey, and L. Allen. 1998. Complementary Feeding of Young Hookworm Infection, Schistosomiasis, and Trachoma." Bulletin of the Children in Developing Countries: A Review of Current Scientific World Health Organization 69: 609­21. Knowledge. WHO/NUT/98.1. Geneva: World Health Organization. Fauveau, V., F. J. Henry, A. Briend, M. Yunus, and J. Chakraborty. 1992. Calain, P., J. P. Chaine, E. Johnson, M. L. Hawley, M. J. O'Leary, H. "Persistent Diarrhea as a Cause of Childhood Mortality in Rural Oshitani, and C. L. Chaignat. 2004. "Can Oral Cholera Vaccination Bangladesh." Acta Paediatrica Supplement 381: 12­14. Play a Role in Controlling a Cholera Outbreak?" Vaccine 22: 2444­51. Feachem, R. G. A., and M. A. Koblinsky. 1983. "Interventions for the Caulfield, L. E., S. L. Huffman, and E. G. Piwoz. 1999. "Interventions to Control of Diarrhoeal Diseases among Young Children: Measles Improve Intake of Complementary Foods by Infants 6 to 12 Months Immunization." Bulletin of the World Health Organization 61: 641­52. of Age in Developing Countries: Impact on Growth and on the ------. 1984. "Interventions for the Control of Diarrhoeal Diseases Prevalence of Malnutrition and Potential Contribution to Child among Young Children: Promotion of Breast-Feeding." Bulletin of the Survival." Food and Nutrition Bulletin 20:183­200. World Health Organization 62: 271­91. CDC (U.S. Centers for Disease Control and Prevention). 1999a. Fourquet, F., J. C. Desenclos, C. Maurage, and S. Baron. 2003. "Acute "Intussusception among Recipients of Rotavirus Vaccine: United Gastroenteritis in Children in France: Estimates of Disease Burden States, 1998­1999." Morbidity and Mortality Weekly Reports 48: 577­81. through National Hospital Discharge Data." Archives of Pediatrics 10: ------. 1999b. "Suspension of Rotavirus Vaccine after Reports of 861­68. Intussusceptions: United States, 1999." Morbidity and Mortality Weekly GAVI (Global Alliance for Vaccines and Immunization). 2005. "Outcomes: Reports 53 (34): 786­89. Most Recent Data on the Impact of Support from GAVI/The Vaccine Cookson, S. T., D. Stamboulian, J. Demonte, L. Quero, C. M. De Arquiza, Fund and the Work of GAVI Partners." http://www.vaccinealliance. A. Aleman, and others. 1997. "A Cost-Benefit Analysis of org/General_Information/About_alliance/progupdate.php. Programmatic Use of CVD 103-Hgr Live Oral Cholera Vaccine in a Gill, C., J. Lau, S. L. Gorbach, and D. H. Hamer. 2003."Diagnostic Accuracy High-Risk Population." International Journal of Epidemiology 26: of Stool Assays for Inflammatory Bacterial Gastroenteritis in 212­19. Developed and Resource-Poor Countries." Clinical Infectious Diseases Coutsoudis, A., K. Pillay, E. Spooner, L. Kuhn, and H. M. Coovadia. 1999. 37: 365­75. "Influence of Infant-Feeding Patterns on Early Mother-to-Child Glass, R. I., J. S. Bresee, U. D. Parashar, R. C. Holman, and J. R. Gentsch. Transmission of HIV-1 in Durban, South Africa: A Prospective Cohort 1999. "First Rotavirus Vaccine License: Is There Really a Need?" Acta Study." Lancet 354: 471­76. Paediatrica Supplement 88: 2­8. De Cock, K. M., M. G. Fowler, E. Mercier, I. de Vincenzi, J. Saba, E. Hoff, Graeff, J. A., J. P. Elder, and E. M. Booth. 1993. Communication for Health and others. 2000. "Prevention of Mother-to-Child HIV Transmission and Behavior Change: A Developing Country Perspective. San Francisco, in Resource-Poor Countries: Translating Research into Policy and CA: Jossey Bass. Practice." Journal of the American Medical Association 283: 1175­82. Graves, P., J. Deeks, V. Demicheli, M. Pratt, and T. Jefferson. 2000."Vaccines Dennehy, P. H. 2005. "Rotavirus Vaccines: An Update." Current Opinion for Preventing Cholera." Cochrane Database Systematic Reviews (4): in Pediatrics 17: 88­92. CD000974. de Zoysa, I., and R. G. Feachem. 1985. "Interventions for the Control of Guptill, K. S., S. A. Esrey, G. A. Oni, and K. H. Brown. 1993. "Evaluation of Diarrhoeal Diseases among Young Children: Rotavirus and Cholera a Face-to-Face Weaning Food Intervention in Kwara State, Nigeria: Immunization." Bulletin of the World Health Organization 63: 569­83. Knowledge, Trial, and Adoption of a Home-Prepared Weaning Food." Dobbing, J. 1990. "Early Nutrition and Later Achievement." Proceedings of Social Science and Medicine 36: 665­72. the Nutrition Society 49: 103­18. Haider, R., A. Islam, J. Hamadani, N. J. Amin, I. Kabir, M. A. Malek, and Duggan, C., O. Fontaine, N. F. Pierce, R. I. Glass, D. Mahalanabis, N. H. others. 1996. "Breastfeeding Counselling in a Diarrhoeal Hospital." Alam, and others. 2004. "Scientific Rationale for a Change in the Bulletin of the World Health Organization 74: 173­79. Composition of Oral Rehydration Solution." Journal of the American Hanh, S. K., Y. J. Kim, and P. Garner. 2001. "Reduced Osmolarity Oral Medical Association 291: 2628­31. Rehydrations Solution for Treating Dehydration Due to Diarrhoea Duke, T. 1999. "Haemophilus influenzae Type B Vaccine in Papua New in Children: A Systematic Review." British Medical Journal 323: Guinea: What Can We Expect, and How Should We Determine Priority 81­85. Diarrheal Diseases | 385 Horton, S., and P. Claquin. 1983. "Cost-Effectiveness and User Murphy, B. R., D. M. Morens, L. Simonsen, R. M. Chanock, J. R. La Characteristics of Clinic-Based Services for the Treatment of Diarrhea: Montagne, and A. Z. Kapikian. 2003. "Reappraisal of the Association A Case Study in Bangladesh." Social Science and Medicine 17: 721­29. of Intussusception with the Licensed Live Rotavirus Vaccine Horton, S., T. Sanghvi, M. Phillips, J. Fielder, R. Perez-Escamilla, C. Lutter, Challenges Initial Conclusions." Journal of Infectious Diseases 187 (8): and others. 1996. "Breastfeeding Promotion and Priority Setting in 1301­8. Health." Health Policy and Planning 11: 156­68. Narula, D., L. Tiwari, and J. M. Puliyeh. 2004. "Rotavirus Vaccines." Lancet 364: 245­46. Huicho, L., M. Campos, J. Rivera, and R. L. Guerrant. 1996. "Fecal Screening Tests in the Approach to Acute Infectious Diarrhea: A Nataro, J., and J. B. Kaper. 1998. "Diarrheagenic Escherichia coli." Clinical Scientific Overview." Pediatric Infectious Disease 15: 486­94. Microbiological Reviews 11: 142­201. Huttly, S. R., S. S. Morris, and V. Pisani. 1997. "Prevention of Diarrhoea in Niehaus, M. D., S. R. Moore, P. D. Patrick, L. L. Derr, B. Lorntz, A. A. Lima, Young Children in Developing Countries." Bulletin of the World Health and R. L. Guerrant. 2002. "Early Childhood Diarrhea Is Associated Organization 75: 163­74. with Diminished Cognitive Function 4 to 7 Years Later in Children in a Northeast Brazilian Shantytown." American Journal of Tropical Islam, M. A., D. Mahalanabis, and N. Majid. 1994. "Use of Rice-Based Oral Medicine and Hygiene 66: 590­93. Rehydration Solution in a Large Diarrhea Treatment Centre in Bangladesh: In-House Production, Use, and Relative Cost." Journal of Nossal, G. J. 2003. "Gates, GAVI, the Glorious Global Funds, and More: All Tropical Medicine and Hygiene 97: 341­46. You Ever Wanted to Know." Immunology and Cell Biology 81: 20­22. Jamison, D. T., H. W. Mosley, A. R. Measham, and J. L. Bobadilla. 1993. Ochoa, T. J., E. Salazar-Lindo, and T. G. Cleary. 2004. "Management of Disease Control Priorities in Developing Countries. Oxford, U.K.: Children with Infection-Associated Persistent Diarrhea." Seminars in Oxford University Press. Pediatric Infectious Diseases 15: 229­36. Oria, R. B., P. D. Patrick, H. Zhang, B. Lorntz, C. M. de Castro Costa, G. A. Jones, G., R. W. Steketee, R. E. Black, Z. A. Bhutta, S. S. Morris, and the Brito, and others. 2005. "APOE4 Protects Cognitive Development in Bellagio Child Survival Study Group. 2003. "How Many Child Deaths Children with Heavy Diarrhea Burdens in Northeast Brazil." Pediatric Can We Prevent This Year?" Lancet 362: 65­71. Research 57: 310­16. Keusch, G. T. 2001. "Toxin-Associated Gastrointestinal Disease: A Clinical Parashar, U. D., J. S. Bresee, J. R. Gentsch, and R. I. Glass. 1998. "Rotavirus." Overview." In Molecular Medical Microbiology, ed. M. Sussman, Emerging Infectious Diseases 4: 561­70. 1083­88. New York: Academic Press. Parashar, U. D., E. G. Hummelman, J. S. Bresee, M. A. Miller, and R. I. ------. 2003. "The History of Nutrition: Malnutrition, Infection, and Glass. 2003. "Global Illness and Deaths Caused by Rotavirus Disease in Immunity." Journal of Nutrition 133: 336S­40S. Children." Emerging Infectious Diseases 9: 565­72. Keusch, G. T., and N. S. Scrimshaw. 1986. "Selective Primary Health Care: Pazhani, G. P., B. Sarkar, T. Ramamurthy, S. K. Bhattacharya, Y. Takeda, and Strategies for Control of Disease in the Developing World--XXIII. The S. K. Niyogi. 2004. "Clonal Multidrug-Resistant Shigella Dysenteriae Control of Infection to Reduce the Prevalence of Infantile and Type 1 Strains Associated with Epidemic and Sporadic Dysenteries in Childhood Malnutrition." Reviews of Infectious Diseases 8: 273­87. Eastern India." Antimicrobial Agents and Chemotherapy 48: 681­84. Keusch, G. T., D. M. Thea, M. Kamenga, K. Kakanda, M. Mbala, and Peter, G., M. G. Myers, the National Vaccine Advisory Committee, and the F. Davachi. 1992. "Persistent Diarrhea Associated with AIDS." Acta National Vaccine Program Office. 2002. "Intussusception, Rotavirus, Paediatrica Scandinavica 381 (Suppl.): 45­48. and Oral Vaccines: Summary of a Workshop." Pediatrics 110: e67. Khan, S. R., F. Jalil, S. Zaman, B. S. Lindblad, and J. Karlberg. 1993. "Early Phillips, M. A., R. G. A. Feachem, and A. Mills. 1987. Options for Diarrhoeal Child Health in Lahore, Pakistan: X--Mortality." Acta Paediatrica Disease Control: The Cost and Cost-Effectiveness of Selected Interventions Supplement 390: 109­17. for the Prevention of Diarrhea. London: Evaluation and Planning Kimmons, J. E., K. H. Brown, A. Lartey, E. Collison, P. P. Mensah, and K. G. Centre for Health Care. Dewey. 1999. "The Effects of Fermentation and/or Vacuum Flask Qualls, N., and R. Robertson. 1989. "Potential Uses of Cost Analyses in Storage on the Presence of Coliforms in Complementary Foods Child Survival Programs: Evidence from Africa." Health Policy and Prepared for Ghanaian Children." International Journal of Food Science Planning 4: 50­61. and Nutrition 50: 195­201. Ronsmans, C., M. L. Bennish, and T. Wierzba. 1988. "Diagnosis and Kosek, M., C. Bern, and R. L. Guerrant. 2003. "The Global Burden of Management of Dysentery by Community Health Workers." Lancet Diarrhoeal Disease, as Estimated from Studies Published between 1992 8610: 552­55. and 2000." Bulletin of the World Health Organization 81: 197­204. Ross, D. A., B. R. Kirkwood, F. N. Binka, P. Arthur, N. Dollimore, S. S. Kotloff, K. L., J. P. Winickoff, B. Ivanoff, J. D. Clemens, D. L. Swerdlow, P. J. Morris, and others. 1995. "Child Morbidity and Mortality Following Sansonetti, and others. 1999. "Global Burden of Shigella Infections: Vitamin A Supplementation in Ghana: Time since Dosing, Number Implications for Vaccine Development and Implementation of Control of Doses, and Time of Year." American Journal of Public Health Strategies." Bulletin of the World Health Organizaton 77: 651­66. 85:1246­51. Legros, D. 2004. "Shigellosis: Report of a Workshop." Journal of Health, Ruxin, J. N. 1994. "Magic Bullet: The History of Oral Rehydration Population and Nutrition 22: 445­49. Therapy." Medical History 38: 363­97. Martinez, J., M. Phillips, and R. G. A. Feachem. 1993. "Diarrheal Diseases." Ryan, E. T., and S. B. Calderwood. 2000. "Cholera Vaccines." Clinical In Disease Control Priorities in Developing Countries, ed. D. Jamison, Infectious Diseases 31: 561­65. W. H. Moseley, A. R. Measham, and J. S. Bobadilla, 91­115. Oxford, Salam, M. A. 1998. "Antimicrobial Therapy for Shigellosis: Issues on U.K.: Oxford University Press. Antimicrobial Resistance." Japanese Journal of Medical Science and Miller, P., and N. Hirschhorn. 1995. "The Effect of a National Control of Biology 51 (Suppl.): S43­62. Diarrheal Diseases Program on Mortality: The Case of Egypt." Social Salam, M. A., and M. L. Bennish. 1988. "Therapy for Shigellosis: I. Science and Medicine 40: S1­30. Randomized, Double-Blind Trial of Nalidixic Acid in Childhood Mondal, S. K., P. G. Gupta, D. N. Gupta, S. Ghosh, S. N. Sikder, K. Shigellosis." Journal of Pediatrics 113: 901­7. Rajendran, and others. 1996. "Occurrence of Diarrheal Diseases in Santos, I., C. G. Victora, J. Martines, H. Goncalves, D. P. Gigante, N. J. Valle, Relation to Infant Feeding Practices in a Rural Community in West and G. Pelto. 2001. "Nutrition Counseling Increases Weight Gain Bengal, India." Acta Paediatrica 85: 1159­62. among Brazilian Children." Journal of Nutrition 131: 2866­73. 386 | Disease Control Priorities in Developing Countries | Gerald T. Keusch, Olivier Fontaine, Alok Bhargava, and others Shann, F. 2000. "Immunization: Dramatic New Evidence." Papua and New ------. 2000. "New Data on the Prevention of Mother-to-Child Guinea Medical Journal 43: 24­29. Transmission of HIV and Their Policy Implications." Report of a Shepard, D. S., L. E. Brenzel, and K. T. Nemeth. 1986. "Cost-Effectiveness WHO technical consultation on behalf of a United Nations Population of Oral Rehydration Therapy for Diarrheal Diseases." Technical Note Fund, United Nations Children's Fund, and Joint United Nations 86­26, Population, Health and Nutrition Department, World Bank, Programme on HIV/AIDS interagency task team on mother-to-child Washington, DC. transmission of HIV, Geneva, October 11­13. Sikorski, J., M. J. Renfrew, S. Pindoria, and A. Wade. 2002. "Support for ------. 2001. "The Optimal Duration of Exclusive Breastfeeding: Results Breastfeeding Mothers." Cochrane Database of Systematic Reviews (1): of a WHO Systematic Review." http://www.who.int/inf-pr-2001/ CD001141. en/note2001-07.html. Snyder, J. D., and M. H. Merson. 1982. "The Magnitude of the Global ------. 2003. HIV and Infant Feeding--Framework for Priority Action. Problem of Acute Diarrhoeal Disease: A Review of Active Surveillance Geneva: WHO. Data." Bulletin of the World Health Organization 60: 604­13. ------. 2004. Family and Community Practices That Promote Child Tucker, A. W., A. C. Haddix, J. S. Bresee, R. C. Holman, U. D. Parashar, Survival, Growth, and Development--A Review of Evidence. Geneva: and R. I. Glass. 1998. "Cost-Effectiveness Analysis of a Rotavirus WHO. Immunization Program for the United States." Journal of the American WHO Collaborative Study Team. 2000. "Effect of Breastfeeding on Infant Medical Association 279: 1371­76. and Child Mortality Due to Infectious Diseases in Less Developed Victora, C. G., J. Bryce, O. Fontaine, and R. Monasch. 2000. "Reducing Countries: A Pooled Analysis." Lancet 355: 1104. Deaths from Diarrhoea through Oral Rehydration Therapy." Bulletin WHO and UNICEF (United Nations Children's Fund). 2001. "Reduced of the World Health Organization 78: 1246­55. Osmolarity Oral Rehydration Salts (ORS) Formulation." WHO/FCH/ Victora, C. G., B. R. Kirkwood, A. Ashworth, R. E. Black, S. Rogers, CAH/01.22. Report from a meeting of experts jointly organized by the S. Sazawal, and H. Campbell. 1999. "Potential Interventions for the United Nations Children's Fund and the World Health Organization, Prevention of Childhood Pneumonia in Developing Countries: Geneva. Improving Nutrition." American Journal of Clinical Nutrition 70: ------. 2004. Joint Statement: Clinical Management of Acute Diarrhoea. 309­20. WHO/FCH/CAH/04.7. Geneva: WHO; New York: UNICEF. Westphal, M. F., J. A. Taddei, S. I. Venancio, and C. M. Bogus. 1995. World Bank. 2002. World Development Indicators. CD-ROM. Washington, "Breastfeeding Training for Health Professionals and Resultant DC: World Bank. Institutional Changes." Bulletin of the World Health Organization 73: Zimbasa (Zimbabwe, Bangladesh, South Africa) Dysentery Study Group. 461­68. 2002. "Multicenter, Randomized, Double Blind Clinical Trial of Short WHO (World Health Organization). 1997. Health and Environment in Course versus Standard Course Oral Ciprofloxacin for Shigella Sustainable Development Five Years after the Health Summit. WHO/ Dysenteriae Type 1 Dysentery in Children." Pediatric Infectious Disease EHG/97.8. Geneva: WHO. Journal 21: 1136­41. ------. 1999. "Potential Use of Oral Cholera Vaccines in Emergency Situations." WHO/CDS/CSR/EDC/99.4. Report of a WHO meeting, Geneva, May 12­13. Diarrheal Diseases | 387 Chapter 20 Vaccine-Preventable Diseases Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, Mark Miller, and Neal A. Halsey Vaccination against childhood communicable diseases through pertussis, Hib, and Neisseria meningitis prevent respiratory dis- the Expanded Program on Immunization (EPI) is one of the eases. Some vaccines, such as those against measles and pertus- most cost-effective public health interventions available sis, prevent diseases that cause or contribute to malnutrition. (UNICEF 2002; World Bank 1993). By reducing mortality and Chapter 16 provides an in-depth review of tuberculosis and a morbidity, vaccination can contribute substantially to achiev- discussion of the potential impact of bacillus Calmette-Guérin ing the Millennium Development Goal of reducing the mortal- (BCG) vaccines. This chapter also does not discuss some new ity rate among children under five by two-thirds between 1990 vaccines, including conjugate Streptococcus pneumoniae, and 2015. Accelerated research into the development of new influenza, typhoid fever, and rotavirus, because other chapters vaccines has been made possible in part by innovative public- deal with those diseases and vaccines. Vaccines to prevent private partnerships, such as the Global Alliance for Vaccines mumps and varicella that are routinely used in some developed and Immunization (GAVI). GAVI focuses on expanding access countries are not included in most vaccination programs in by immunization programs in developing countries to new developing countries. Other interventions that can reduce the and underused vaccines, such as those for hepatitis B and burden of vaccine-preventable diseases and are not covered in Haemophilus influenzae type B (Hib). These newer, more this chapter include clean umbilical cord care to reduce the expensive vaccines are challenging previous notions of the incidence of neonatal tetanus, vitamin A therapy to reduce the cost-effectiveness of immunization. Analyses of their costs and case-fatality rate (CFR) from measles, and intensive clinical cost-effectiveness are particularly important because of the care that can reduce the mortality associated with most of the need to determine the level of resources required in the future vaccine-preventable diseases. to improve immunization programs, to cover the costs of new vaccines, and to allocate scarce public and external resources available for immunization in the most optimal manner. CAUSES AND EPIDEMIOLOGY OF DISEASES This chapter analyzes the costs and cost-effectiveness of PREVENTED BY VACCINES USED IN NATIONAL scaling up the EPI and introducing selected new vaccines into IMMUNIZATION PROGRAMS the program. It also summarizes the epidemiology of diseases preventable through immunization and estimates the disease The epidemiology and burden of vaccine-preventable diseases burden with and without immunization programs. In addi- vary by country and region partly because of differences in tion, the chapter discusses the organization, delivery, and vaccine uptake. Numerous other factors that contribute to the financing of immunization programs and highlights future disease burden include geography, seasonal patterns, crowding, prospects and areas for further study. nutritional status, travel to and from other countries, and pos- Several areas overlap with other chapters. For example, sibly genetic differences in populations that affect disease the vaccines that prevent measles, tuberculosis, diphtheria, severity. Table 20.1 summarizes the features of selected vaccines 389 390 Table 20.1 Selected Vaccine-Preventable Diseases and Vaccines | Disease Meningococcal Japanese Category Tuberculosis Diphtheria Tetanus Pertussis Poliomyelitis Measlesa Rubella Hib Hepatitis B Yellow fever disease encephalitis Control Causative Mycobacterium Toxin-producing Toxin-producing Bacterium Virus (serotypes Virus Virus Bacterium Virus Virus Neisseria menin- Virus Priorities agent tuberculosis bacterium bacterium (Bordetella 1, 2, and 3) (Haemophilus gitis groups A, B, (Corynebacterium (Clostridium pertussis) influenzae C, Y, W135 diphtheriae) tetani) type B) in Developing Reservoir Humans (some Humans Animal intes- Humans Humans Humans Humans Humans Humans Monkeys and Humans Birds and bovine) tines; soil humans mammals Spread Airborne droplet Close respiratory Spores enter Close Fecal-oral; Close respiratory Close respira- Close respira- Blood, perina- Bites by Close respiratory Bites by Countries nuclei from or cutaneous the body respiratory close respira- contact and tory contact and tory contact tal, household, infected contact infected sputum-positive contact through wounds contact tory contact aerosolized aerosolized occupational, mosquitoes mosquitoes persons or the umbilical droplets droplets or sexual | cord stump transmission Logan Trans- As long as spu- Usually under two No person-to- Usually under A few days Four days before A few days Chronic carriage Up to lifelong Infected individ- Chronic carriage Unknown, rare Brenzel, mission tum acid-fast weeks; some person trans- three weeks before and rash until two before to seven for months chronic carriage uals can trans- for months cases for sev- period bacilli are chronic carriers mission (starts before after acute days afterward days after rash; and transmis- mit the disease eral months Lara positive cough is symptoms up to one year sion when bitten by apparent) of age in a mosquito vec- J. congenitally tor during the W olfson, infected viremic phase (the first three Julia or four days of illness) Fox-Rushby Subclinical Common but Common No Mild illness More than 100 May occur in Common Common Common, espe- Common Common Common infection not important in common: may subclinical children under cially in infants transmission not be diag- infections for one, but relative ,and nosed each paralytic importance is others case minimal Duration of Not known; Lasting protective Lasting protec- Incomplete Lifelong type- Lifelong Lifelong Uncertain; no If develops, Lifelong Uncertain; no pro- Lifelong natural reactivation of immunity not pro- tive immunity and waning specific protection lifelong tection against immunity old infection duced by infec- not produced protection immunity against carriage carriage commonly tion; second by infection; and those previ- causes disease attack possible second attack ously infected possible may develop some disease (epiglottitis) Risk factors High population Crowding; low Wound contam- Young age; Poor environ- Highly transmis- Highly transmis- Failure to Carrier mother, Young age; for- Crowding; respi- Young age; for infection densities in socioeconomic inated by soil; crowding mental hygiene sible agent with sible; crowding; breastfeed; sibling, or sex est workers; ratory viral infec- forest workers; (for unvacci- regions with status umbilical cord; and sanitation nearly 100 per- low socioeco- crowding; low partner; multi- season (late tions, especially season nated indi- historically poor agricultural cent infectivity nomic status socioeconomic ple sex part- rainy season, influenza viduals) control; low work except for iso- status; immune ners; intra- early dry socioeconomic lated popula- deficiency, venous drug season) status; poor tions; crowding, including HIV use; unsafe access to care; low socioeco- injection immunodefi- nomic status practices ciency; malnu- trition; alco- holism; diabetes Case- See chapter 16 2 to 20 percent 25 to 90 per- Up to 10 per- 2 to 10 percent 0.05 to 10.0 per- Less than Meningitis, 5 to Acute, more 10 to 40 percent Untreated 90 to 5 to 30 per- fatality rateb cent cent in infants cent 0.1 percent 90 percent; than 1 percent; 100 percent; cent and children pneumonia 5 to chronic; 25 per- treated 5 to 25 percent cent (delayed) 20 percent Vaccine BCG attenuated Diphtheria toxoid Tetanus toxoid Killed whole- Live (OPV) Measles (two); Rubella (one Capsular poly- Hepatitis B sur- Yellow fever Vaccines for A, C, Live attenu- (number of Mycobacterium (three to five pri- (three to five in cell or acellular (three to four subcutaneous or two); saccharide face antigen attenuated live Y, W135 only; ated (two, doses); bovis (1); mary including children, includ- pertussis (three primary plus subcutaneous linked to protein (three to four); virus (1 plus unconjugated China only); route intradermal booster doses in ing booster to five, includ- campaigns);c Hib (three to intramuscular boosters); sub- polysaccharides killed (two); most countries); doses in many ing booster killed (IPV) five); intramus- cutaneous given subcuta- booster com- intramuscular countries; five doses in most (three to four) cular neously or intra- monly used for women of countries); muscularly: one but of uncer- childbearing intramuscular dose with repeat tain value age; adult three to five boosters years later for for injury high-risk persons; prevention); conjugated: for intramuscular C only or A, C, Y, + W135, one dose given intramuscularly Vaccine 0 to 80 percent More than More than 70 to 90 per- OPV: more than 95 percent at 12 95 percent (at More than 75 to 95 per- 90 to 98 percent Unconjugated Live attenu- efficacy for pulmonary 87 percent 95 percent cent 95 percent in months of age; 12 months 95 percent for cent; efficacy polysaccharides: ated: 90 per- tuberculosis; 75 (more than industrial coun- 85 percent at and up) invasive against chronic poor efficacy cent (after one to 86 percent 80 percent after tries; 72 to 98 9 months of age disease infection in under two years dose at one for meningitis two doses) in percent in from one dose, infants born to of age; conju- year); 94 to and miliary infants developing more than 98 carrier mothers; gated polysac- 100 percent Vaccine-Preventable tuberculosis countries; percent from more than charides: approxi- (after two lower protec- two doses 95 percent for mately 95 percent doses one to tion against exposure at and up serogroup two months type 3 than 1 older ages specific apart); inacti- and 2; IPV: vated: 80 per- more than cent (declining Diseases 95 percent to 55 percent after one year; no decrease in | another study) 391 (Continues on the following page.) 392 Table 20.1 Continued | Disease Meningococcal Japanese Category Tuberculosis Diphtheria Tetanus Pertussis Poliomyelitis Measlesa Rubella Hib Hepatitis B Yellow fever disease encephalitis Control Duration of Unknown; some Variable: probably 10 years or Unknown; Presumed life- Lifelong in most; Lifelong in Unknown, but More than 15 For at least 10 Unconjugated Unknown, may Priorities immunity evidence that around five years; more wanes with long for both rare cases of most; presumed lasts for at least years; further years and possi- wanes rapidly for be lifelong after pri- immunity longer in pres- time OPV and IPV, waning immu- rare cases of three years follow-up is bly for life children under mary series wanes with ence of natural but unknown nity after one waning immu- beyond period continuing five, more than in time boosting or dose, not two nity after one of greatest three to five Developing booster doses dose, not two exposure years for older children; conju- gated uncertain Countries Schedule Given at or near Three-dose sched- Normally given Usually given OPV: four doses First dose at First dose at 12 Three or four Several sched- One dose at 9 Unconjugated: Live: one year birth in popula- ule recommended as DTP vaccine in childhood as (birth, 6, 10, 9 or 12 to 15 to 15 months; doses; usually ules: at birth, 6, to 12 months one dose at two and two years; tions at high at 6, 10, and 14 to children; combination and 14 weeks) months); a sec- when given, a given during the and 14 weeks; with measles in years or older and killed: days 0, | Logan risk weeks in develop- unimmunized vaccine (DTP) in polio- ond opportunity second dose same visit as for with first three countries where second dose 7, and 30 fol- ing countries for pregnant endemic coun- to receive a with measles DTP doses of DTP; yellow fever three to five lowed by Brenzel, DTP vaccine; women should tries; birth dose dose of measles vaccine birth dose poses a risk years later for booster two other schedules be given two may be omitted vaccine (either needed if high risk; conju- years later and in common use; doses of elsewhere with through routine mother is a car- gate C: three then every Lara booster doses at tetanus toxoid fourth dose [18 months or rier and recom- doses at two, three years J. 18 months and or tetanus- given later; four to six years] mended if peri- three, and four or W olfson, four to six years reduced diph- supplemental or supplemental natal transmis- two, four, and six also suggested theria toxoid, doses (up to 10) immunization sion of months for and a total of given in activities) should hepatitis B is infants; one dose Julia five doses is national cam- be provided for frequent; four for older children Fox-Rushby required to pro- paigns for all children doses total can and adults; conju- vide protection eradication; IPV: be given gate A, C, Y, through all three to four although only W135 currently ,and childbearing doses: 2, 4, 6 to three are only approved for years 18 months, and required one dose at 11 others four to six years years or older Status as of 158 countries All countries; Childhood: all All countries; All countries; Routine first 110 countries in 89 countries; 147 countries; 29 of 43 coun- European coun- Southeast the end of using BCG; 78 percent countries; 78 percent 79 percent rou- dose all coun- 2003 global coverage global coverage tries at risk tries, Canada Asia 2001 85 percent coverage 78 percent coverage tine, plus sup- tries, 77 percent less than 42 percent using vaccine; (and United coverage coverage plemental coverage; sec- 18 percent 30 percent cov- States in 2005) coverage ond opportunity, erage in target 164 out of 192 population countries Comments Reasons for Recent trends to Five doses in Variability in Primary series Lower efficacy Lower efficacy None Efficacy lower None None None varying efficacy lower antibody adults provide whole cell vac- gives incom- when maternal when maternal if injected into are multifacto- levels in adults protection for cines; acellular plete protection antibody present antibody fat rial, including without booster more than vaccines used in developing present differences in doses because of 20 years in some devel- countries vaccines waning immunity oped countries and less natural boosting Sources: WHO 2002, 2004. DTP diphtheria-tetanus-pertussis; IPV inactivated polio vaccine; OPV oral polio vaccine. a. Measles vaccine is given as measles, measles-rubella, or measles-mumps-rubella vaccine. The latter two vaccines are routinely used in industrial countries and are increasingly being adopted in other countries. The World Health Organization recommends that the combination measles-rubella or measles-mumps-rubella vaccines be introduced only after careful evaluation of public health priorities within each country and following the establishment of an adequate program for measles control as demonstrated by high coverage rates as part of a well-functioning childhood immunization program. b. Note that variations in case-fatality rates are related to access to care, type of care administered, setting, age at onset of disease, and other factors. The ranges presented in this table reflect both uncertainty as to actual case-fatality rates and the variability of populations. c. As of 2003, an injected IPV is given alone or in combination with OPV in 31 countries. IPV is currently not recommended for routine use in developing countries because of its relatively high cost and uncertain efficacy when given at 6, 10, and 14 weeks. The usual recommended IPV schedule is 2, 4, and 6 to 18 months. Routine use of OPV is expected to cease following polio eradication. Stockpiles of monovalent OPV for each of the three virus types are under development to protect against vaccine-associated paralytic poliomyelitis and outbreaks of circulating vaccine-derived polioviruses. Vaccine-Preventable Diseases | 393 in use in childhood immunization programs throughout the especially in areas contaminated by animal feces (Cherry and world. Harrison 2004). The organism is usually transmitted through burns, cuts, and other penetrating injuries. Neonatal tetanus is the most common presentation in developing countries. The Burden of Vaccine-Preventable Diseases portal of entry is usually the umbilical stump but has been A number of vaccine-preventable diseases are not reportable associated with circumcision and other surgical procedures events in many countries. The estimates of the burden of dis- (Birmingham and others 2004; Stanfield and Galazka 1984). ease by the World Health Organization (WHO) are based on a Children born to women who do not have protective levels of combination of often incomplete vital registration data, mor- tetanus antibody are susceptible to neonatal tetanus. tality survey data, and mathematical models using numerous The estimated burden of neonatal tetanus assumes that in assumptions. Most models of vaccine-preventable diseases are areas with low rates of skilled delivery, all births not protected derived from the susceptible fraction of the population (calcu- by the immunization of pregnant women are subject to a preim- lated from natural immunity from presumed historical infec- munization era neonatal tetanus mortality rate expressed as tions in regions without previous vaccination and historical deaths per 1,000 live births (Birmingham and others 2004; immunization coverage rates), infectivity rates of disease, Griffiths and others 2004). In other areas, we assume that births sequelae of diseases, and estimates of local CFRs. The degree of not protected through immunization or skilled delivery are accuracy of these models is only as good as the data supporting subject to an incidence and CFR equal to 25 percent of the the assumptions. The disease burden is most appropriately rep- preimmunization era neonatal tetanus mortality rate.1 resented by a range of values reflecting uncertainty. In this CFRs are directly associated with the quality of medical care chapter, we estimate the burden of disease as the number of available. With the availability of secondary and tertiary care, deaths and DALYs per World Bank region in 2001. The follow- CFRs have declined to 25 percent or less (Cherry and Harrison ing description draws in part on discussion of methods for 2004; Wassilak and others 2004). The CFRs used to derive cases burden of disease calculations reflected in the Global from estimated deaths range from 40 percent in developed Immunization and Vision Strategy of WHO and the United countries to 80 percent in the poorest developing countries. We Nations Children's Fund (UNICEF) (Wolfson and Lydon 2005). estimate the tetanus burden other than for neonates by apply- ing an estimated age distribution of total tetanus to the esti- mated neonatal tetanus deaths (Galazka and others forthcom- Diphtheria ing) and region-specific CFRs, which indicate a range of from Diphtheria is caused by a toxin-producing strain of the bacte- 27 percent among children age one to four in developed coun- rium Corynebacterium diphtheriae, which is transmitted by tries to 65 percent among those age 80 or older in developing means of respiratory droplets. The 2001 WHO estimates of countries. diphtheria mortality are extrapolations from reported deaths in countries with full or partial vital registration systems. Before the widespread use of immunization,more than 5 per- Pertussis cent of people living in temperate climates suffered from clinical Bordetella pertussis is transmitted through respiratory excre- diphtheria at some point during their lifetimes (Griffith 1979). tions and occurs throughout the world. Most pertussis in Rates exceeding 100 cases per 100,000 population were seen developing countries occurs in school-age children. In devel- in Europe during World War II (Galazka, Robertson, and oped countries, mild or asymptomatic infections in adults are Oblapenko 1995). The CFRs from respiratory tract diphtheria believed to be common sources of transmission to very young have been 2 to 20 percent, with an average of 10 percent for infants (Edwards and Decker 2004). Clinical manifestations patients receiving good medical care (Feigin, Stechenberg, and include an initial 7 to 10 days of rhinorrhea progressing to a Hertel 2004).To estimate diphtheria deaths in the absence of vac- cough that becomes paroxysmal or spasmodic, usually associ- cination and to project future deaths with and without vaccina- ated with profuse rhinorrhea (Cherry and Heininger 2004). tion, we assumed an average incidence rate of 15 per 100,000 and Clinical pneumonia is seen in approximately 10 percent of CFRs of 2.5 percent in developed countries,5.0 percent in Europe infants. and Central Asia, and 10.0 percent elsewhere (Birmingham and Our estimates for the burden of pertussis followed the Stein 2003; Galazka and Robertson forthcoming). model described in Crowcroft and others (2003). We estimated that the proportion of susceptible children becoming infected in countries with vaccination coverage of less than 70 percent Tetanus over the previous five years was 30 percent by age 1, 80 percent Clostridium tetani is maintained in nature and is found in all by age 5, and 100 percent by age 15. For countries with cover- countries. Spores remain viable for many years in soil and dust, age of more than 70 percent in the past five years, we assumed 394 | Disease Control Priorities in Developing Countries | Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, and others that 10 percent of susceptible children were infected by age 1, younger than five, vitamin A deficient, or infected with HIV or 60 percent by age 5, and 100 percent by age 15. A vaccine effi- who have acquired measles from a household contact (Perry cacy of 80 percent was assumed for preventing infection and and Halsey 2004). Declines in CFRs in the past two decades are 95 percent for preventing deaths. The CFR was 0.20 percent in associated with the tendency of the disease to infect older chil- infants, 0.04 percent in children age one to four, and 0 percent dren, decreased crowding, and improved nutritional status in in those older than five in low-mortality countries; and 3.7 per- many developing countries (Perry and Halsey 2004). At the cent among infants, 1 percent among children age one to four, same time, recent studies indicate CFRs of 0.4 to 9.7 percent in and 0 percent in those older than five in high-mortality Sub-Saharan African countries with low immunization cover- countries. age (Perry forthcoming). Considerable controversy is associated with the number of Poliomyelitis deaths resulting from measles, because of difficulty in accu- Before the availability of polio vaccines, as many as 90 percent rately specifying the cause of death in children afflicted with of children in the developing world were infected with all three measles and in separating complications of measles from those types of the polio virus in the first two or three years of life of other conditions. In addition, CFRs, which have decreased (Sutter and Kew 2004). In developed countries, transmission rapidly in many countries, vary significantly. The natural his- occurred primarily in school-age children and more than 90 per- tory model used in this chapter is based on Stein and others cent of infections were asymptomatic; 4 to 8 percent of chil- (2003), modified to account for the effect of supplementary dren had nonspecific febrile illness and less than 1 percent immunization activities. developed acute flaccid paralysis (Sutter and Kew 2004). We derived estimates of the burden of disease in countries Children with residual paralysis require rehabilitation. with high-quality surveillance data and high sustained cover- Surgical intervention is necessary if contractures develop age of measles vaccine by adjusting the number of reported because of the lack of rehabilitative services following the acute cases by a reporting efficiency factor ranging from 5 to 40 per- illness. These children are at increased risk of premature death cent. In estimating the future burden of disease, the averted because of late onset postpolio muscle atrophy (postpolio syn- burden of disease, and the burden in countries without both drome), which occurs 20 to 40 or more years after acute illness. adequate surveillance and sustained high coverage, we assumed Disease burden estimates are based on actual active surveil- that the average number of cases per year is equal to the num- lance. The estimated 1,000 deaths a year caused by polio reflect ber of children in the current birth cohort who are not pro- past infections and current deaths. Following Robertson tected by either routine or supplemental vaccination. WHO (1993), we obtained the number of cases and deaths in the (2005a) estimates that in 2001, 611,000 deaths (approximately absence of immunization by applying an incidence rate of 1 per 5 percent of all childhood mortality) were attributable to 1,000 population under age five and CFRs of from 2.5 percent measles. in developed countries to 10.0 percent in Sub-Saharan Africa. An alternative proportional mortality approach, which is To determine current cases, we applied an estimate of notifica- based on retrospective verbal autopsy studies in 18 countries tion efficiency to reported cases. to derive the proportional causes of child deaths in 42 high- mortality countries, also has appeared in the literature (Morris, Measles Black, and Tomaskovic 2003). This model suggests that measles Measles is an acute respiratory viral infection. Children born to may have accounted for approximately 3 percent of all child- immune mothers are protected against clinical measles from hood deaths in 2000. passively acquired maternal antibodies until they are five to In countries with a high disease burden, the true number of nine months of age. More than 90 percent of infections are measles deaths may be somewhere between the proportional associated with clinical disease (Krugman 1963). Compli- mortality and natural history estimates. WHO (2005b) uses a cations include pneumonia, diarrhea, encephalitis, and blind- hybrid method that estimates that measles was responsible for ness, especially in children with vitamin A deficiency. In recent an average of 4 percent of mortality among children under five years, CFRs have been estimated at 3 percent in many develop- between 2000 and 2003, or approximately 400,000 deaths per ing countries, but historically they have been as high as 30 per- year. If the actual number of deaths in 2001 was 400,000, then cent in some community-based studies (Aaby 1988; Aaby and the cost per death averted will be lower than what has been esti- Clements 1989; Moss, Clements, and Halsey 2003; Perry and mated for this chapter, and the effect of increasing coverage will Halsey 2004). be overestimated because fewer deaths could be prevented. For a disease such as measles in which infection is almost Both of the approaches described have strengths and limita- universal in the absence of immunity, small changes in the CFR tions. We adopt the natural history approach for this analysis result in large changes in estimates of total mortality. Increased because the chapter includes deaths at all ages and the model complication and mortality rates occur in children who are can adapt to recent changes in CFRs and coverage rates. Vaccine-Preventable Diseases | 395 However, the natural history method is sensitive to the accu- average CFRs ranging from 1 percent among infants in devel- racy of parameter inputs such as CFRs and may underestimate oped countries to 12 percent in Sub-Saharan Africa. the effect of herd immunity. Further modeling efforts would need to incorporate sensitivity testing around a range of param- Hepatitis B eter estimates. In many developed countries, most transmission of hepatitis B In the absence of vaccination, the measles virus would infect occurs during or after adolescence, coinciding with the onset of almost 100 percent of the population, including most of the sexual activity and of drug abuse involving unsafe reuse of nee- 688 million children under five in the developing world. Using dles and syringes (McQuillan and others 1999). In many the methods described here, approximately 125 million African countries, transmission occurs primarily in early child- cases and 1.8 million to 2.0 million deaths per year would be hood through mucosal contact with infectious body fluids and expected in the absence of vaccination. unsafe injection practices (Margolis, Alter, and Hadler 1997). Some Asian countries have a high rate of chronic carrier states, Haemophilus influenzae Type b (Hib) and the primary mode of transmission is mothers to infants Hib is transmitted through the respiratory tract and causes (Beasley 1988; Mast and others 2004). The rate of symptomatic meningitis, pneumonia, septic arthritis, skin infections, disease is only about 1 percent in infancy and 10 percent in epiglottitis, osteomyelitis, and sepsis. Deaths caused by Hib early childhood, but it increases to 30 to 40 percent in adults. occur primarily from meningitis and pneumonia. In developed Serosurveys for carrier states of hepatitis B are available for countries, approximately half of diagnosed invasive infections almost all nations (WHO 1996). are meningitis (Wenger and Ward 2004). In developing coun- Models of hepatitis B disease burden are based on estimated tries, a larger proportion of identified cases is meningitis result- ratios between infected and carriage states at various ages or ing from underdiagnosis of other clinical syndromes (Martin estimates of the percentage of carriers that progress to and others 2004; Peltola 2000). Intervention studies have hepatoma, fulminant hepatitis, or cirrhosis at later stages of life demonstrated significant reductions in pneumonia in vacci- (Miller and McCann 2000). The model we used for estimating nated compared with unvaccinated children (Levine and oth- hepatitis B mortality estimates the age- and sex-specific pro- ers 1998; Mulholland and others 1997). Although infections gression of hepatitis B surface antigen infection to disease occur throughout the world, the incidence of Hib disease may incorporating competing mortality, particularly because indi- be lower in some Asian countries than in Africa and the viduals infected with HIV are more likely to perish from HIV Americas (Gessner and others 2005). before the full mortality impact from hepatitis B infection We derived estimates of Hib disease burden from incidence (Gay and others 2001; Griffiths, Hutton, and Pascoal 2005). rates and CFRs for meningitis and pneumonia. We derived Whereas most vaccine-preventable diseases that result in country-specific estimates of the incidence of Hib meningitis death occur at an early age shortly after the age of vaccination, from the literature on incidence in the prevaccine era (Bennett deaths from hepatitis B occur many years into the future. and others 2002). For countries without meningitis incidence Countries that introduce hepatitis B vaccines today will not data, we used the average incidence in countries with similar reap most of the benefits for many years. In the absence of epidemiological profiles. Regional averages ranged from 219 vaccination, we estimated approximately 1.4 million future cases per 100,000 to 3 per 100,000 population in children deaths attributable to hepatitis B for the 2001 birth cohort under one, and 1 to 15 per 100,000 population in children age after accounting for competing mortality. Global vaccination one to four. The CFR for meningitis is nearly 100 percent in the of 35 percent would prevent more than 500,000 of those absence of intensive antibiotic therapy, but it can be reduced to future deaths. Discounting the value of future hepatitis B 5 to 8 percent when appropriate therapy is available (Swartz deaths to their equivalent value in the present to make the 2004). We derived CFRs in a manner similar to that used for burden of disease prevented equivalent to that of other incidence rates and adjusted them on the basis of country- vaccine-preventable diseases results in approximately 87,000 specific data on access to care. Regional means ranged from 3 to deaths averted. 32 percent. Estimating the burden of Hib pneumonia is much more complex. A rapid assessment method assumes five pneumonia Yellow Fever cases for every meningitis case (WHO 2001). An alternative Yellow fever virus is transmitted by mosquitoes, primarily Aedes approach assumes that Hib is responsible for a fixed propor- eqypti, with a three- to six-day incubation period. Patients pre- tion (about 20 percent) of acute lower respiratory infection sent with intense headache, fever, chills, and myalgia, among deaths in the absence of immunization (Peltola 2000). We other symptoms. Although once much more widespread, derived pneumonia CFRs from a literature review of lower res- yellow fever is now limited to West and Central Africa, the piratory infections in children (Bennett and others 2002), with northern half of South America, and Panama. In approximately 396 | Disease Control Priorities in Developing Countries | Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, and others 15 to 20 percent of yellow fever patients, severe disease occurs, Turk 1982; WHO 1974). A standard immunization schedule with liver and kidney failure and cardiovascular collapse. The was established in 1984 on the basis of a review of immuno- CFR varies, with increased severity in older adults (Monath logical data for the original EPI vaccines: BCG, diphtheria- 2004). The average CFR in patients in Africa with jaundice is tetanus-pertussis (DTP), oral polio, and measles vaccines 20 percent (Monath and others 1980; Nasidi and others 1989). (Halsey and Galazka 1985). On the basis of surveillance data adjusted for underreport- Today, national immunization programs in developing ing, WHO (1992) estimates the global burden of yellow fever at countries are responsible for improving access to the tradi- 200,000 cases and 30,000 deaths in 1990. Most cases and deaths tional EPI antigens and introducing new vaccines. In 2002, the occur in 33 African countries, where 1 in 80 cases is assumed EPI introduced the Reaching Every District strategy, which to be reported. In South American countries, 1 in 10 cases is focused on achieving an 80 percent coverage rate of DTP3 in assumed to be reported. We use the implied incidence rate and 80 percent of districts and using immunization contacts to a CFR of 15 percent to project future mortality. Between 1990 deliver other high-priority child health interventions. In addi- and 2001, some improvement in routine coverage of yellow tion to delivering vaccinations, national immunization fever vaccine occurred, but the overall burden of yellow fever is programs are concerned with the quality and safety of immu- unlikely to have declined. nization through the adoption of safe injection technologies (autodisabled syringes, storage boxes, and incinerators) and ESTIMATES OF THE CURRENT BURDEN OF proper cold chain and vaccine stock maintenance. VACCINE-PREVENTABLE DISEASES AND OF In most developing countries, immunizations are provided THE BURDEN AVERTED BY VACCINATION through a system of fixed facilities at different levels of the health system. Immunization campaigns are discrete, time- Table 20.2 provides WHO estimates of deaths from selected limited efforts at national or subnational levels that usually vaccine-preventable diseases for 2001, taking immunization focus on specific antigens (for example, polio). Mobile strate- coverage rates into account. The greatest burden of disease is in gies rely on the use of specialized vehicles to transport health Sub-Saharan Africa, which accounts for 58 percent of pertussis professionals and vaccines to deliver services to remote or deaths, 41 percent of tetanus deaths, 59 percent of measles migrating populations. Outreach is a strategy by which staff deaths, and 80 percent of yellow fever deaths. East Asia and the members from a health facility travel to villages and surround- Pacific has the greatest burden from hepatitis B, with 62 per- ing areas to administer vaccines. Extended outreach refers to cent of deaths worldwide. South Asia also experienced a high more targeted and intensive efforts. disease burden, particularly for tetanus and measles. In 1999, the major international development partners Table 20.2 also shows the extent of mortality in the absence involved in immunization (for example, WHO, UNICEF, the of immunization and the estimated number of deaths averted World Bank, and bilateral donors) joined the Bill & Melinda by vaccination. In 2001, vaccination averted up to 52 percent of Gates and Rockefeller Foundations, the vaccine industry, and yellow fever deaths, 61 percent of measles deaths, 69 percent nongovernmental organizations to create GAVI (http://www. of tetanus deaths, 78 percent of pertussis deaths, 94 percent of vaccinealliance.org) to increase access to new and underused diphtheria deaths, and 98 percent of polio deaths that would vaccines in the world's poorest countries, improve access to have occurred in the absence of vaccination. These results basic immunization services, and accelerate research and demonstrate the significant effect that vaccination programs development pertaining to new vaccines and delivery technol- have had on worldwide disease burden. The figures also show ogy. Through the Vaccine Fund, GAVI raised more than that vaccination programs have been less successful in reducing US$1.3 billion to strengthen immunization systems, introduce the disease burden in Sub-Saharan Africa, where coverage rates new vaccines, and support safe injection practices. More than are lower. US$3 billion has been pledged for the next 10 years. Between Table 20.3 reports WHO estimates of disability-adjusted life 2000 and 2003, an additional 4 million children were vacci- years (DALYs) lost from vaccine-preventable diseases by region nated with DTP3, 42 million with hepatitis B, nearly 5 million for 2001, demonstrating the high burden of disease worldwide with Hib, and more than 3 million with yellow fever vaccine. from disability associated with sequelae of hepatitis B (liver cancer and cirrhosis), pertussis, and tetanus.2 COSTS AND COST-EFFECTIVENESS OF EXISTING EXPANDED PROGRAM ON IMMUNIZATION VACCINATION PROGRAMS WHO initiated the EPI in 1974 to provide countries with guid- Brenzel and Claquin (1994) and GAVI (2004) estimate the cost ance and support to improve vaccine delivery and to help make per fully immunized child (FIC) for the traditional six EPI vaccines available for all children (Hadler and others 2004; antigens as approximately US$20.3 We evaluated the cost per Vaccine-Preventable Diseases | 397 Table 20.2 Estimated Number of Deaths in the Absence of Vaccination, Deaths from Vaccine-Preventable Diseases, and Deaths Averted by Vaccination, All Ages, by Region and Vaccine, 2001 (thousands) East Europe Latin Middle Asia and and America East and Sub- High the Central and the North South Saharan Disease Total income Pacific Asia Caribbean Africa Asia Africa Diphtheria If no vaccination 78 3 28 4 8 5 21 10 Estimated deaths 5 1 1 1 1 1 3 1 Deaths averted 73 3 27 4 8 5 18 9 Pertussis If no vaccination 1,343 7 377 4 138 93 428 296 Estimated deaths 301 1 3 1 6 8 108 176 Deaths averted 1,042 7 374 4 132 85 320 120 Tetanus If no vaccination 936 1 110 1 20 23 543 239 Estimated deaths 293 1 27 1 1 4 140 121 Deaths averted 643 1 83 1 19 19 403 118 Poliomyelitis If no vaccination 52 1 15 1 3 4 17 11 Estimated deathsa 1 1 1 0 0 0 0 0 Deaths averted 51 n.a. 15 1 3 4 17 11 Measles If no vaccination 2,000 6 301 36 6 55 567 1,025 Estimated deathsb 676 1 77 4 1 7 239 348 Deaths averted 1,237 5 229 28 6 40 351 578 Hib If no vaccination 468 1 28 2 9 14 199 216 Estimated deaths 463 1 28 2 5 14 199 215 Deaths averted 5 1 1 1 4 1 1 1 Hepatitis B If no vaccination 600 34 370 36 11 17 75 58 Estimated deaths 600 34 370 36 11 17 75 58 Deaths avertedc 1 1 1 1 1 1 1 1 Yellow fever If no vaccination 63 n.a. n.a. n.a. 8 n.a. n.a. 54 Estimated deaths 30 n.a. n.a. n.a. 6 n.a. n.a. 24 Deaths averted 33 n.a. n.a. n.a. 2 n.a. n.a. 30 Source: Mathers and others 2006 and authors' calculations. n.a. not available. Note: Totals may not add due to rounding. a. Primarily deaths at older ages caused by delayed effect of poliomyelitis in childhood. b. See text for discussion of uncertainty regarding measles estimates. The values shown here are an updated version of the 2001 estimates. c. Deaths averted to date from the use of the hepatitis B vaccine in infant immunization programs are minimal, largely because of the long time period (20 to 40 years) to see mortality effects. 398 | Disease Control Priorities in Developing Countries | Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, and others Table 20.3 DALYs Lost from Vaccine-Preventable Diseases, All Ages By Region, 2001 (thousands) Europe Latin Middle East Asia and America East and Sub- High and the Central and the North South Saharan Disease Total income Pacific Asia Caribbean Africa Asia Africa Diphtheria 164 1 18 2 8 1 90 45 Tetanus 8,342 5 762 2 17 110 3,965 3,481 Pertussis 11,542 139 584 81 366 326 3,930 6,116 Poliomyelitis 145 8 49 2 6 8 55 17 Measles 23,129 23 2,318 236 13 470 6,527 13,539 Hepatitis Ba Acute hepatitis B 2,169 86 675 79 95 111 585 536 Liver cancer 9,168 1,223 5,925 379 277 138 464 762 Cirrhosis of the liver 15,780 2,146 3,890 2,084 1,513 686 4,249 1,212 Meningitisb 5,607 131 1,071 403 591 328 2,142 941 Lower respiratory infectionsc 85,920 2,314 10,827 2,111 3,043 2,974 34,196 30,455 Source: Mathers and others 2006 and Authors' calculations. Note: Totals may not add due to rounding. a. Includes all DALYs attributable to the three conditions. Hepatitis B is the underlying cause of only a portion of the liver cancer and cirrhosis of the liver DALYs. b. Includes all DALYs attributable to meningitis, including Hib, S. pneumococcus, and N. meningitides. c. Includes all DALYs attributable to lower respiratory infections, including Hib and S. pneumococcus. FIC for the childhood EPI cluster antigens by World Bank was for extended outreach services (US$5.81), perhaps because region on the basis of published and unpublished data. These the strategy is a more targeted approach. Routine facility-based studies used a standard costing approach that estimated the strategies had lower average costs (US$13.65 per FIC) than costs of labor, vaccines, supplies, transportation, communica- campaigns (US$26.82 per FIC) or mobile strategies (US$25.84 tion, training, maintenance, and overhead and included the per FIC). Higher unit costs associated with these strategies are annualized value of equipment, vehicles, and building space possibly attributable to a different mix of inputs as well as (Khaleghian 2001; USAID, Asia­Near East Region 1988; WHO greater expenses for per diems, fuel, and social mobilization. 1988). The number of FICs in these studies was measured The results also vary by World Bank region, with East Asia using community-based sample surveys (Henderson and and the Pacific (US$13.25) and Sub-Saharan Africa (US$14.21) Sundaresen 1982). having lower estimates of cost per FIC than Europe and Our literature review found 102 estimates of total and unit Central Asia (US$24.12) and the Middle East and North Africa immunization program costs from 27 countries between 1979 (US$22.15). and 2003 for different immunization delivery strategies The findings of our analysis are generally supported by the (Berman and others 1991;1 Beutels 1998, 2001; Brenzel 2005; literature (Creese 1986; Brenzel and Claquin 1994; Khaleghian Brenzel and Claquin 1994; Brinsmead, Hill, and Walker 2004; 2001), which has shown that variation in the cost per FIC is Creese 1986; Creese and Domínguez-Ugá 1987; Domínguez- related to the mix of delivery strategies, the prices of key inputs Ugá 1988; Edmunds and others 2000; Griffiths and others such as vaccines, and the overall scale of programs. In addition, 2004; Levin and others 2001; Pegurri, Fox-Rushby, and Walker an analysis of 13 national financial sustainability plans for 2005; Robertson and others 1992; Soucat and others 1997; immunization reveals a wide range in the cost per FIC by Steinglass, Brenzel, and Percy 1993). All costs were converted region and strategy.4 to 2001 U.S. dollar equivalents. Because total and unit costs are Recurrent costs are the lion's share of total immunization related to population size, table 20.4 reports population- costs (80 percent for fixed facility strategy and 92 percent for weighted results only. National immunization program refers campaigns), which has implications for the need for continu- to total national costs for all strategies. ous and predictable program financing. Labor costs account The population-weighted mean cost per FIC for all regions for the largest share (roughly 30 to 46 percent of total cost) for and all strategies is approximately US$17, with a range of US$3 all strategies except extended outreach. Vaccine costs range to US$31. The lowest mean population-weighted cost per FIC from 8 percent for mobile strategies to 29 percent for extended Vaccine-Preventable Diseases | 399 Table 20.4 Estimates of the Population-Weighted Annual Cost for the Traditional Vaccines per FIC, by Immunization Strategy and Region, 2001 (2001 US$) Europe Latin Middle East Asia and America East and Sub- and the Central and the North South Saharan All Pacific Asia Caribbean Africa Asia Africa regions Strategy (n 4) (n 1) (n 1) (n 1) (n 10) (n 15) (n 32) National immunization -- -- 18.10 22.15 24.82 21.05 23.52 program (23­27) (17­26) (17­27) n 2 n 2 n 6 Fixed facility 20.00 24.12 -- -- 13.79 6.31 13.65 (18­22) (6­24) (3­31) (3­31) n 2 n 7 n 6 n 16 Campaign -- -- -- -- -- 26.82 26.82 (13­28) (13­28) n 3 n 3 Mobile -- -- -- -- -- 25.84 25.84 n 1 n 1 Outreach 6.50 -- -- -- 7.11 -- 7.10 (4­9) n 1 (4­9) n 2 n 3 Extended outreach -- -- -- -- -- 5.81 5.81 (5.8­13) (5.8­13) n 3 n 3 Mean for all strategies 13.25 24.12 18.10 22.15 17.11 14.21 16.91 (4­22) (6­27) (3­31) (3­31) Source: Authors' calculations for the traditional vaccines based on the literature. -- not available. Note: Mean values are used in the analysis. Ranges for estimates are reported in parentheses. Europe and Central Asia, Latin America and the Caribbean, and the Middle East and North Africa are limited to one observation for each region, which may not be indicative of the cost per FIC in each region. However, in lieu of using region-specific estimates, the overall average (US$13.65) would be applied, which may underestimate the cost per FIC in these more developed regions, where higher unit costs for delivery of health services would be expected. outreach strategies. Transportation costs account for the Our analysis highlights the variation in cost per FIC by second-largest share of EPI costs for mobile strategies, while region and strategy and demonstrates the value of more disag- building costs account for a greater share of fixed facility gregated results for making policy decisions. However, given strategies. the limited sample of estimates available for the regions and Using data from table 20.4 on costs per FIC and multiplying strategies, the results should be used as an indicative guide for by the size of the population covered, we estimate US$1.17 bil- policy making and not as a substitute for country-specific cost- lion for the total cost of immunization programs in develop- effectiveness evaluations of strategies. In addition, our esti- ing countries in 2001, with a range of US$717 million to mates do not take into account household costs, such as time US$1.48 billion. At US$20 per FIC, the cost of the six tradi- spent seeking services, and other social costs. Our estimates tional vaccines in developing countries would have amounted also do not consider the direct and indirect costs of acute ill- to US$1.57 billion in 2001. Table 20.5 shows that the estimated nesses prevented by vaccination or the costs of long-term com- cost per death averted ranges from US$205 in South Asia and plications from disease and of adverse events associated with Sub-Saharan Africa to US$3,540 in Europe and Central Asia. vaccination (though the latter are unlikely to have a significant These results suggest that the cost per death averted rises with impact on costs because rates of serious complications are coverage rates. Europe and Central Asia, Latin America and the extremely low). Furthermore, the analysis focuses on FICs and Caribbean, and the Middle East and North Africa had higher underemphasizes the benefits of partial immunization. Future coverage rates in 2001, resulting in fewer deaths that could be economic evaluations of immunization program alternatives averted. The table also shows that the cost per DALY from the could consider these factors as a critical step in determining traditional EPI vaccines ranges from US$7 to US$438, depend- the allocation of scarce resources among high-priority health ing on region, mix of strategy, and levels of scale. interventions. 400 | Disease Control Priorities in Developing Countries | Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, and others Table 20.5 Average Cost per FIC, Total Immunization Cost, Cost per Death Averted and Cost per DALY for the Traditional Immunization Program by Region Latin Middle East Asia Europe America East and Sub- and the and and the North South Saharan Strategy Pacific Central Asia Caribbean Africa Asia Africa Cost/FIC (2001 US$) (from table 20.4) 13.25 24.12 18.10 22.15 17.11 14.21 Percentage of FIC 78.22 93.72 86.36 90.90 58.86 50.20 Estimated total immunization cost (2001 US$ millions) 316 131 174 152 227 172 Estimated deaths averted (thousands, from table 20.2) 728 37 174 153 1,109 867 Estimated cost/death averted (2001 US$) 434 3,540 1,030 993 205 205 Estimated cost/DALY (2001 US$) 85 395 438 166 16 7 Source: Authors' calculations. Note: DALY estimates are the sum total for diphtheria, pertussis, tetanus, polio, and measles from table 20.3. COST-EFFECTIVENESS OF INCREASING Percentage of FICs IMMUNIZATION COVERAGE FOR THE 100 TRADITIONAL EPI 90 80 WHO (2004) estimates that in 2001, 30 million children were 70 inadequately immunized with DTP. Achieving higher coverage rates by improving access for remote populations, accelerating 60 immunization delivery strategies, and introducing new vac- 50 cines will mean increasing the level of investment (Batt, Fox- 40 Rushby, and Castillo-Riquelme 2004). 30 We estimated the costs of scaling up EPI coverage for a hypo- 20 thetical population of 1 million in each region between 2002 and 10 2011. Costs were reported in 2001 dollars, and a 3 percent dis- count rate was applied. Brenzel (2005) provides details on the 0 methods. The costs of scaling up coverage are based on vaccine 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 and delivery costs per dose. We derived vaccine costs from the East Asia and the Pacific Europe and Central Asia unit price of each vaccine (provided by WHO, UNICEF, and the Latin America and the Caribbean Middle East and North Africa Vaccine Fund); wastage rates for vaccines and injection supplies South Asia Sub-Saharan Africa Developed countries by strategy; the required injection supplies; and the number of Source: Authors' estimation. doses per FIC. A 2 percent adjustment was made for inflation. We used data on the cost per FIC generated earlier to derive Figure 20.1 Coverage of FICs Projected to 2011 delivery costs per dose by strategy and region by subtracting the costs of vaccines, injection supplies, and fixed costs. · Finally, because the scale factor is derived from the unit costs Fixed costs were excluded from the scaling-up exercise of a health center visit, the assumption of constant fixed because they were assumed to remain constant during the pro- costs in the short run appears reasonable. jection period. Previous studies have found that the main cost drivers of · First, the largest projected coverage increase of 9 percentage immunization costs are the mix of strategies and the scale points (figure 20.1) may not require additional infrastruc- of immunization programs (Brenzel and Claquin 1994; ture investments. Domínguez-Ugá 1988; Robertson and others 1992; Soucat and · Second, how and to what extent fixed costs would change by others 1997). Countries are unlikely to achieve 90 percent or region is uncertain, and a conservative approach would be more coverage relying on fixed facilities alone because of to exclude them. limited population access. We estimate the proportion of FICs · Third, because most immunization costs are recurrent costs, obtained for each strategy by region.5 The best mix of strategies the analysis focuses on these. for increasing coverage will vary by country depending on the Vaccine-Preventable Diseases | 401 dispersion of the population, the access to health facilities, the For tetanus toxoid immunization, the additional discounted vaccines being delivered, and the effectiveness of various strate- cost per person vaccinated ranges from US$3.28 to US$4.06. gies in reaching target populations. Estimates are also adjusted The cost per death averted varies from US$271 to more than for the level of scale by a factor derived from the unit cost per US$190,000. The results of this analysis fall within the range of health center contact by coverage level (Mulligan and others estimates reported in the literature (Berman and others 1991; 2003), and details are provided in Brenzel (2005). Steinglass, Brenzel, and Percy 1993). Differences in coverage The total additional cost of reaching higher coverage levels levels and in protection against neonatal tetanus through was divided by the number of deaths averted. Coverage projec- skilled delivery contribute to the variation in results across tions for 2002­11 were based on statistical modeling of official regions. WHO and UNICEF estimates for the period between 1995 and The analysis shows both an increase in costs and potential 2002 for all developing countries. The model relates coverage in benefits from scaling up immunization programs. In practice, future years to that in the previous year, with the relationship the costs and benefits related to scaling up in any one region between past and future coverage differing for each region and will be highly dependent on a few countries or subnational economic status combination. areas within countries. Aggregate country- or region-level data Figure 20.1 shows historical and projected coverage rates do not reveal the efficiency that could best be obtained by by region. The figure shows that coverage increased in all the targeting immunization efforts on specific countries or geo- regions during the late 1980s under universal childhood immu- graphic areas rather than making diffuse investments across nization. After 1990, when funding for universal childhood regions. For instance, Miller and others (1998) show that India immunization waned, the figure indicates the subsequent stag- and Nigeria contribute the most to estimates of global measles nation and, in some cases, the declines in coverage rates. For the deaths; therefore, reducing transmission in those countries scaling-up period, we project that the coverage of FICs will would contribute the most to reducing the global disease bur- increase from 78 to 79 percent in East Asia and the Pacific, from den caused by measles. 92 to 95 percent in Europe and Central Asia, from 88 to 90 per- Despite its importance for policy, empirical and country- cent in Latin America and the Caribbean, from 91 to 95 percent specific evidence on how immunization program costs change in the Middle East and North Africa, from 70 to 79 percent in as coverage increases is lacking. Because scaling up im- South Asia, and from 52 to 61 percent in Sub-Saharan Africa. munization coverage will require more intensive efforts to find The projections show that three of the six regions are expected unvaccinated children, an extra cost for vaccinating each addi- to achieve 90 percent FIC by 2011. East Asia and the Pacific, tional child is generally expected. Nevertheless, most cost- South Asia, and Sub-Saharan Africa will require additional effectiveness studies assume constant returns to scale (Elbasha intensive efforts to achieve higher coverage rates. and Messonnier 2004; Karlsson and Johannesson 1998) even Table 20.6 reports the results of the scaling-up analysis for when emerging evidence suggests that the cost of vaccinating the traditional EPI vaccines, for tetanus toxoid vaccination for each additional child may rise with the size of delivery unit women of reproductive age, and for selected new vaccines. The (Valdmanis, Walker, and Fox-Rushby 2003). Box 20.1, which discounted incremental cost per child vaccinated with the tra- focuses on scaling up traditional immunization coverage, and ditional EPI vaccines ranges from US$10.89 in Latin America box 20.2, which focuses on new antigens, summarize the results and the Caribbean to US$12.84 in the Middle East and North of two studies that shed more light on this subject. Africa. The number of discounted deaths averted because of full immunization depends on incremental coverage rates and varies from 747 in Europe and Central Asia to 14,584 in Sub- COSTS AND COST-EFFECTIVENESS OF ADDING Saharan Africa, resulting in regional variations in the dis- NEW ANTIGENS TO THE CURRENT counted incremental cost per death averted from US$169 in IMMUNIZATION SCHEDULE Sub-Saharan Africa to US$1,754 in Europe and Central Asia.6 DALYs were estimated indirectly based on the ratio of deaths We also estimated the additional costs per person vaccinated to DALYs for each disease in 2001. This ratio is applied to the and cost per death averted of introducing new and underused hypothetical population in each World Bank region over the vaccines into the traditional EPI in a hypothetical population of projection period. Calculated this way, the number of DALYs 1 million in each region between 2002 and 2011. The new vac- averted will not account for changes in the average age of infec- cines considered protect against hepatitis B, yellow fever, Hib, tion that ordinarily results from expanding immunization cov- measles, rubella, Japanese encephalitis, and meningococcal A, erage. This method over-estimates the number of DALYs and as well as inactivated polio vaccine (IPV). For comparison pur- thus under-estimates the cost/DALY. Cost-effectiveness ratios poses, we assumed that new vaccines were introduced in 2002. should be treated as indicative only. The cost/DALY ranges from The additional cost of combination vaccines is net of the $2 to $20 for scaling up traditional immunizations. original cost of DTP vaccination to avoid duplication. The 402 | Disease Control Priorities in Developing Countries | Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, and others Table 20.6 Average Cost per Person Vaccinated and per Death Averted for Scaling Up Immunization Coverage and Adding in Selected New Vaccines in a Hypothetical Population of 1 million for 2002­11 (2001 US$, current vaccine prices) Europe Latin Middle East Asia and America East and Sub- and the Central and the North South Saharan Pacific Asia Caribbean Africa Asia Africa Traditional EPI (mix of strategies) Incremental discounted cost/person vaccinated 12.03 11.54 10.89 12.84 11.58 11.16 Incremental discounted deaths averted 3,165 747 2,552 4,576 7,584 14,584 Incremental discounted cost/death averted 478 1,754 791 698 274 169 Tetanus toxoid (mix of strategies) Incremental discounted cost/person vaccinated 4.06 3.34 3.28 3.34 3.98 3.88 Incremental discounted deaths averted 343 2 200 465 2,815 2,412 Incremental discounted cost/death averted 1,541 190,000 3,117 1,880 271 394 Second opportunity for measles (fixed facility) Incremental discounted cost/person vaccinated 1.08 1.05 0.98 1.19 1.04 1.00 Incremental discounted deaths averted 1,138 599 95 1,304 2,509 9,646 Incremental discounted cost/death averted 119 199 1,906 228 74 23 DTP-hepatitis B and Hib (pentavalent) vaccine (mix of strategies) Incremental discounted cost/person vaccinated 15.14 14.61 15.69 16.23 15.24 11.68 Incremental discounted deaths averted 47 19 116 274 192 1,796 Incremental discounted cost/death averted 40,000 85,000 25,000 14,000 14,000 1,433 Yellow fever (campaigns) Incremental discounted cost/person vaccinated n.a. n.a. 1.43 n.a. n.a. 1.42 Incremental discounted deaths averted n.a. n.a. 94 n.a. n.a. 376 Incremental discounted cost/death averted n.a. n.a. 2,810 n.a. n.a. 834 Incremental discounted cost/person vaccinated (mix of strategies) Hepatitis B monovalent (birth dose) 2.26 2.15 2.36 2.37 2.24 2.02 DTP-hepatitis B (tetravalent) 7.85 7.57 7.34 8.03 7.55 7.26 Rubella (campaigns) 1.20 1.19 1.20 1.07 1.19 1.19 Meningococcal A (fixed facilities) n.a. n.a. n.a. 2.73 n.a. 2.33 Japanese encephalitis (fixed facilities) 4.56 n.a. n.a. n.a. 4.37 n.a. Injectable polio vaccine (monovalent) 7.12 6.72 6.42 7.32 6.85 6.60 Injectable polio vaccine (combination with DTP) 13.88 14.84 14.62 15.28 14.77 14.19 Source: Authors' calculations. Note: n.a. refers to not applicable when a specific disease is not prevalent in a specific region. delivery cost per FIC was apportioned to individual antigens EPI schedule).Combination vaccines may be more cost-efficient on the basis of the share of number of doses per FIC for that because of potential savings in supplies, syringes, and health antigen (Brenzel 2005). Cost estimates are based on the num- workers'time, in addition to the overall health benefits of reduc- ber of doses required for full immunity (that is, hepatitis B, Hib, ing the number of required injections. However, if the combina- and IPV vaccines require three doses for full immunity, and tion vaccine does not reduce the number of visits a child would meningococcal A requires two doses for full immunity). Results ordinarily need to make to a health facility, any cost savings may are reported in table 20.6. be subsumed by the higher costs of increasing coverage. The analysis also assumes that an additional visit to a health The discounted incremental cost per person ranges from facility is required for new doses (depending on timing in the less than US$1 to US$16.23, depending on the unit price of Vaccine-Preventable Diseases | 403 Box 20.1 Marginal Costs of Immunization Services in India A study in Tamil Nadu evaluated immunization costs and The study used data from the health facility sample to coverage using a longitudinal panel dataset of immuniza- explain the determinants of immunization costs, which tion program costs (Brenzel 1995). Data were collected were modeled as a function of outputs, input prices, and from a stratified, random sample of facilities between 1989 other production-related variables that influence the cost and 1991 for the North Arcot District Polio Control function with respect to outputs. A random effects estima- Program.a The sample included 120 observations of 59 dif- tion was performed on the analysis sample relating the nat- ferent health centers: 17 followed for three years (29 per- ural logarithm of health facility costs to the type of polio cent), 27 followed for two years (46 percent), and 15 with vaccine in use, estimated target population, and size of a single observation (25 percent). Total immunization geographical area serviced by the health facility and natu- costs included the cost of labor, vaccines, injection sup- ral logarithms of the number of FICs per facility, the num- plies, transportation, and overhead and the value of equip- ber of hours spent by a village health nurse on immuniza- ment, vehicles, and buildings. tion services per facility, and the number of small pieces of During this period, coverage rates for FICs increased equipment used for immunization service delivery. from 5 to 77 percent. The table shows that the cost per The analysis revealed a significant association between dose and the cost per FIC increased during this period.b facility cost and the number of FICs, the hours worked Changes in the cost per dose were highly statistically sig- by village health nurses, the area served, and the type of nificant, whereas no statistical differences were apparent polio vaccine. When calculated using mean values, the in the cost per FIC during the study period. marginal cost per FIC was Rs 24.43 (US$1.30) lower than the associated average cost per FIC of Rs 183 (US$9.80), Comparison of Total Facility Immunization Costs, implying that the average cost curve lies above the mar- Immunization Activity, and Unit Costs by Year, North ginal cost curve for the sample of health facilities in India. Arcot District Polio Control Program, 1989­91 A declining relationship is apparent between costs and (2001 US$) coverage for this sample of facilities, calling into question Indicator Year 1 Year 2 Year 3 Overall assumptions of constant returns to scale. The results sug- gest that, in India, average cost-effectiveness ratios would Total costs 996 1,337 980 1,104 overestimate total resource needs. Using a single-point Variable costs 697 1,260 917 958 estimate of average unit costs to determine the use of Cost/dose 1.09 1.98 1.33 1.47 scarce public health resources will result in suboptimal Cost/FIC 13.11 27.92 17.07 19.37 resource allocations. Source: Brenzel (1995). a. The program was a joint effort by the Indian Council for Medical Research, the Centre for Advanced Research in Virology at the Christian Medical Centre and Hospital in Vellore, and the governments of Tamil Nadu and India. b. Higher costs in the second year reflect a change in the organization of the primary health care system in 1990 to improve access to basic services. vaccine, the type of vaccine, the delivery strategy, and the cov- facility strategies, and from US$65 to US$1,363 for campaigns erage levels. The results lead to several conclusions: These results are consistent with the literature. Foster, McFarland, and John (1993) find an incremental cost per death · First, the additional incremental cost per person vaccinated averted ranging from US$335 to US$552 in urban areas and is relatively small for some new vaccines. from US$327 to US$706 in rural areas. The Africa Measles · Second, because fixed costs are excluded, the results repre- Partnership (2004) estimates a cost per death averted of sent conservative estimates of additional costs. US$131 to US$393 in the African context, but these figures · Third, because of price uncertainty, cost variations are great- include the costs of infrastructure. est for newer vaccines, such as the DTP-IPV combination. In the hypothetical populations, the incremental cost per death averted for the pentavalent vaccine ranged from The second opportunity for measles has the lowest cost US$1,433 to more than US$85,000, depending mostly on the per death averted, ranging from US$23 to US$1,906 for fixed number of potential deaths that could be averted. Although a 404 | Disease Control Priorities in Developing Countries | Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, and others Box 20.2 An Immunization Costing Study of Adding New Vaccines to the EPI in Peru Data were collected from 19 government health facilities in US$ per year three districts in Peru, including five hospitals and 14 health 10 centers (Walker and others 2004). Total annual costs per 9 center included vaccines, supplies, personnel, cold chain, 8 overhead, and shared inputs. The average cost per dose for 7 Marginal cost traditional EPI antigens plus yellow fever varied from 6 US$1.50 to US$3.20 per dose as shown in the table, with 5 vaccines and personnel accounting for the bulk of costs. 4 At 2,000 doses, the marginal cost of delivering one 3 more dose is US$1.08, increasing to US$5.33 for 12,000 2 Average cost doses. Average and marginal costs are equal (US$1.18) 1 when 5,000 doses are provided per site. When an outlier 0 delivering many vaccines at a high cost was removed, cost- 0 5,000 10,000 15,000 20,000 Number of doses delivered minimizing output rose to 6,000 doses at US$1.11. Although each vaccination facility is likely to be associa- Source: Authors` calculations. ted with different average and marginal costs, considering Marginal and Average Vaccination in Sample Facilities in Peru vaccine provision across a range of providers is relevant, (2001 US$) because targets for vaccination can be set by site. Information about marginal costs can help determine what the most efficient size for vaccination facilities is in incremental cost per dose of vaccine of US$0.20, US$4.14, the long run and how to minimize costs across different and US$4.24, respectively. Adding these new vaccines size units in the short run, given targets (see figure). increased the total cost of providing 5,100 doses from When hepatitis B, Hib, and the pentavalent vaccine US$5,840 to US$9,415 and changed the minimum average (DTP-hepatitis B-Hib) are added to the delivery schedule, cost from US$1.18 per dose to US$1.68. Therefore, the the total annual additional cost increases to US$4,121, addition of new vaccines shifts both average and marginal US$11,886, and US$25,261, respectively, with an average costs upward. Mean Cost per Dose by Type of Facility, Selected Districts in Peru (2001 US$) Department Health Rural Provincial hospital National Cost items Health post center hospital hospital (Ayacucho) hospital Recurrent items Vaccines 0.59 0.87 1.39 1.03 0.60 0.31 Syringes 0.04 0.05 0.07 0.04 0.05 0.03 Personnel 0.46 0.28 1.17 0.33 0.76 0.29 Other 0.05 0.03 0.03 0.04 0.03 0.13 Capital items 0.01 0.02 0.09 0.01 0.03 0.02 Direct costs 1.15 1.25 2.75 1.45 1.47 0.78 Indirect costs 0.33 0.26 0.41 0.30 0.45 1.19 Average cost 1.48 1.51 3.17 1.79 1.92 1.98 Source: Walker and others (2004). Note: Totals may not sum exactly because of rounding. Vaccine-Preventable Diseases | 405 wide range of results was found, these estimates are supported introducing IPV into routine vaccination in the United States by the literature. Miller (1998) estimated between US$3,127 would cost an additional US$15 million to US$28 million de- and US$3.2 million per life saved for Hib vaccine. Brinsmead, pending on the type of schedule adopted, resulting in a cost per Hill, and Walker's (2004) systematic review of the literature on vaccine-associated paralytic poliomyelitis case prevented of the cost-effectiveness of Hib vaccine finds wide variations in approximately US$3 million. Sangrugee, Caceres, and Cochi results because of methodological differences and epidemio- (2004) found that the least costly option would be for pro- logical and health system characteristics. The discounted grams to stop providing OPV after postpolio eradication and incremental cost of introducing the pentavalent (DTP­ certification and that optionally introducing IPV with univer- hepatitis B­Hib) vaccine is roughly equal to the total mean cost sal IPV had the highest costs and the lowest expected number of the traditional vaccine package estimated earlier. This find- of vaccine-associated paralytic poliomyelitis cases. If the unit ing implies that introducing this combination vaccine may price of IPV fell to US$0.47, switching to IPV from OPV would double the financial requirements, an implication that is sup- be economically worthwhile. ported by data from national financial sustainability plans for immunization (Lydon 2004). FINANCIAL SUSTAINABILITY OF IMMUNIZATION The incremental discounted cost per person vaccinated with PROGRAMS a birth dose of hepatitis B is approximately US$2, and that for the tetravalent vaccine was between US$7 and US$8. The 10- Even though research has demonstrated that vaccination year time period for our analysis is too short to accumulate against childhood diseases is one of the most cost-effective deaths averted resulting from hepatitis B vaccination because health interventions, governments in many developing coun- deaths from liver cancer occur at older ages. Beutels's (1998, tries are considering how to meet the financing requirements 2001) reviews of studies of the cost-effectiveness of introducing of immunization programs, particularly as new vaccines are hepatitis B vaccine indicate that results vary depending on introduced and programs are scaled up. GAVI is working with assumptions of endemicity and the methodology used, with a countries to prepare for the transition from grant funding and cost per death averted ranging from US$3,500 to US$271,800. to secure the overall financial sustainability of national pro- Rubella vaccination had a low additional cost per person vac- grams. Approximately 55 countries have prepared national cinated, at slightly more than US$1. Golden and Shapiro (1984) financial sustainability plans for immunization. These plans found that vaccinating all prepubertal children with rubella vac- help countries evaluate the current and future costs and financ- cine had the highest benefit-cost ratio (ranging from US$1.70 to ing of national immunization programs and identify strategies US$1.96). Most benefits were future cost savings from long- to address future funding gaps (GAVI 2004; http://www.who. term institutional care. When rubella was delivered in combina- int/immunization_financing/en). tion with measles and mumps, the benefit-cost ratios varied According to a recent analysis of financial sustainability from US$4.70 to US$38.80 (Hinman and others 2002). plans, specific costs for immunization programs represent an The additional cost per person vaccinated with one dose average of 2 percent of total health spending and 6 percent of of Japanese encephalitis vaccine was between US$4.37 and government health spending and are equivalent to less than US$4.56. A study in Thailand using two doses showed a cost per 0.2 percent of gross domestic product on average. However, child ranging from US$2.31 to US$4.20, depending on the this profile changes after new and more expensive vaccines are mode of delivery (Siraprapasiri, Sawaddiwudhipong, and introduced. In some countries, program-specific costs for Rojanasuphot 1997). Ding and others (2003) estimate a cost per immunization can reach as high as 20 percent of government case averted of US$258 and a cost per DALY averted of US$16.80 health spending with introduction of combination vaccines for a five-dose inactivated Japanese encephalitis vaccine. (Lydon 2004). This share is related to the current unit price of Our analysis suggests an additional discounted cost per per- the vaccine, which is expected to decline. son vaccinated for injectable polio vaccine of between US$6.60 Governments and their development partners are chal- and US$7.32, depending on coverage levels and mix of delivery lenged to find ways to finance and sustain immunization strategy. The additional discounted unit cost of the combina- programs. In countries that are implementing reforms to tion DTP-IPV vaccine was higher, ranging from US$13.88 to achieve greater transparency and fiscal discipline through sec- US$15.28. These results are also sensitive to the current prices torwide approaches and medium-term expenditure frame- of the vaccine, which will probably decline in coming years. works, the additional financing requirements are compounded Brenzel (1995) finds that in India the cost per case prevented by the need to operate within a fixed budget for the health sec- for the combination DTP-polio vaccine was much lower than tor, so that increased funding needs for one program may for oral polio vaccine (OPV), primarily because the combina- necessitate budget cuts for others. This example illustrates the tion vaccine was associated with a greater reduction in the potential tradeoffs that exist at the country level, which create number of polio cases. Miller and others (1996) suggest that both opportunities for more open policy dialogue in relation to 406 | Disease Control Priorities in Developing Countries | Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, and others priority setting for the use of scarce public funds and risks that tion of routine BCG vaccination, given the low risk of the cost of new vaccines may not be readily integrated into acquiring tuberculosis in early childhood. If the BCG were not national plans and budgets. Because of the financial implica- administered during the first month of life, program costs tions of reaching higher coverage levels and simultaneously would be reduced by the value of one visit and by the costs introducing new vaccines, policy makers will not only have to associated with vaccine purchase, shipping, storage, and weigh the cost-worthiness of alternative investments but also administration. have to understand their long-term budgetary implications. RESEARCH AGENDA IMPROVING THE COSTS AND COST- EFFECTIVENESS OF IMMUNIZATION PROGRAMS Private and public sector investment in research and develop- ment pertaining to new vaccines and improved use of existing The cost-effectiveness of immunization programs could be im- vaccines is considerable. Most research and development is proved by either reducing costs or improving programs' health focusing on vaccines likely to have the greatest effect in the benefits. Programs could reduce costs by using a more efficient developed world and the best financial return; however, by mix of delivery strategies, reducing vaccine wastage, and using means of public-private partnerships for product develop- lower-cost inputs while maintaining the same quality of serv- ment, foundations have stepped in to support vaccine research ice. Reductions in the price of vaccines in the near future will and development for diseases for which the greatest burden also reduce costs. Innovations in vaccine technology may result occurs in developing countries. in more widespread use of vaccine vial monitors, and increased New vaccines are being developed that could be incorpo- use of heat-stable vaccines could potentially reduce the cost of rated into EPI schedules, including vaccines that protect against the cold chain, although these innovations may themselves add rotavirus, S. pneumoniae, malaria, cervical cancer associated to costs. The number of children and adults immunized can with human papilloma virus, HIV/AIDS, and dengue. New and be increased by creating additional demand for vaccination; improved vaccines are also being developed to protect against reducing missed opportunities; and reducing the dropout rate meningococcal infections in infancy and Japanese encephalitis between the first and third doses of DTP, hepatitis B, and other (NIH 2000). WHO recently created the Initiative for Vaccine vaccines. Finally, changes in the EPI schedule could affect total Research Department to facilitate global coordination of re- costs by reducing the number of doses required to achieve search and development efforts for these and other vaccines. immunity and thereby reducing the number of visits, resulting In compiling data for this chapter, we noted a number of key in savings in the costs of labor, supplies, transport, and perhaps gaps in knowledge that could usefully drive a research agenda overhead. and contribute to more evidence-based policy making in the The EPI schedule was established in 1984 based on a review future (Fox-Rushby and others 2004). First, little is currently of immune responses to diphtheria, tetanus, pertussis, polio, known about how and why delivery costs change with increas- and measles vaccines starting at different ages and with varying ing numbers of vaccinations and at higher coverage rates and intervals between doses (Halsey and Galazka 1985). The EPI whether economies of scale can be achieved. Little is known schedule administers three doses of DTP at the shortest possi- about the relative cost-effectiveness of different strategies to ble intervals to complete the immunization series as early in life increase coverage given different baseline coverage rates. This as possible. However, if the primary series could be reduced to issue relates to other questions of the optimal timing for intro- two doses with a booster dose at 12 to 15 months of age, the ducing new vaccines and of how decisions should vary given cost savings from reduced visits and one fewer dose of DTP in different epidemiological and economic settings. Future countries that administer a fourth dose of DTP would be con- research should therefore consider the extent to which cost- siderable. Additional serological studies would be needed to effectiveness analyses need to be repeated for every country or compare the existing EPI schedule with the theoretical sched- context or whether (and how) estimating and validating rela- ule before a new schedule could be adopted. Also, other vac- tionships across countries and accounting for uncertainty in cines to be introduced into immunization programs would estimates of costs and effects are possible. need to be revaluated in this schedule. Two doses of IPV Second, more attention needs to be given to measuring administered beginning at two months of age induce protective effect. For example, even though the coverage of single antigens levels of antibodies between 95 and 100 percent for each of the required to reach particular levels of FICs should be accounted three polio types (Halsey and others 1997; Plotkin and Vidor for, economic evaluations need to move beyond such indica- 2004). tors of output to measuring effect on the quantity and qual- Some countries with a low incidence of tuberculosis (such ity of life. In evaluating different schedules, methodological as those of Eastern Europe) are considering the discontinua- research needs to focus on how to incorporate the combined Vaccine-Preventable Diseases | 407 effects of multiple vaccinations in this respect. Remarkably few Financing and sustaining immunization programs are chal- studies have considered the effect on nonhealth benefits, such lenges that governments in developing countries and their as economic growth and welfare. The larger the package of development partners will face. The financial implications of vaccinations considered, the more important this question reaching higher coverage levels and the simultaneous desire to becomes. introduce new vaccines will require policy makers to consider both the relative cost-effectiveness of interventions and the long-term budgetary implications. CONCLUSIONS Although global and regional estimates of cost-effectiveness of interventions are useful guides, further analytical work will This chapter confirms that vaccination of children and women be needed to evaluate the relative benefits (deaths and cases with the traditional EPI vaccines is a highly cost-effective pub- averted and DALYs) and costs (delivery and treatment) of vac- lic health intervention, although cost-effectiveness ratios vary cines for different delivery strategies and higher coverage rates, by region, delivery strategy, and level of scale. Overall, vaccina- particularly at the country level. tion has had a significant effect on reducing mortality and morbidity from childhood diseases and will be a priority intervention for achieving the child health Millennium ACKNOWLEDGMENTS Development Goals. Improving and sustaining measles control The authors thank Tina Proveaux for editorial and technical are among the most cost-effective interventions in high- assistance and Howard Barnum, Mariam Claeson, Felicity mortality regions. Cutts, Tony Measham, and Philip Musgrove for reviewing Establishing and maintaining high immunization coverage drafts of the manuscript. Santiago Cornejo and Ravi rates in many of the poorest developing countries have proven Cheerukupalli provided valuable inputs. challenging for those with high population growth rates, limited infrastructure and resources, and fluctuating demand for services. According to historical coverage rate trends, NOTES Europe and Central Asia, Latin America and the Caribbean, and the Middle East and North Africa are expected to achieve 1. For our analysis, the preimmunization era neonatal tetanus mortal- ity rate per 1,000 live births is used: developed countries, 0.1; East Asia and 90 percent coverage of FIC by 2011, with East Asia and the the Pacific, 4.7; Europe and Central Asia, 0.4; Latin America and the Pacific, South Asia, and Sub-Saharan Africa lagging behind. Caribbean, 4.4; Middle East and North Africa, 4.7; South Asia, 15.3; and Increasing and sustaining higher immunization coverage Sub-Saharan Africa, 10.2. 2. Because disease classification does not have a one-to-one correspon- rates will require further efforts so that disease control can be dence with those prevented by vaccine, according to table 20.3 is based on maintained, particularly when a perception exists at the com- estimates of the proportion of these illnesses that may be preventable by munity level that vaccine-preventable diseases are no longer a specific vaccines. For example, some meningitis and acute lower respira- major public health issue. At higher coverage rates, further tory infections are caused by Hib or S. pneumoniae, and some cirrhosis is caused by hepatitis B. disease burden reductions will be smaller, which will affect rel- 3. A fully immunized child is a standard term that refers to a child who ative cost-effectiveness. Targeted approaches in countries or at has received one dose of BCG vaccine, three doses each of oral polio vac- subnational levels could potentially yield high returns, espe- cine and DTP vaccines, and one dose of measles vaccine. The number of FICs does not include children who have been partially immunized, so this cially in those areas with poor control of vaccine-preventable measure underestimates the total effect on the disease burden. However, diseases. the number of FICs is representative of the effectiveness of the delivery Our analysis shows that the cost per FIC will increase as system in providing access to immunization services to children. The authors are aware that fully vaccinating a child does not correspond to full countries scale up immunization coverage and introduce new immunity. vaccines. Adding more antigens to traditional EPIs has been 4. The mean population-weighted cost per FIC for the financial sustain- successfully accomplished in many countries, especially for Hib ability plans for immunization was US$21.06. The plans use DTP3 coverage and hepatitis B vaccines. Although many of the new vaccines as a proxy for FICs rather than coverage measured through population- based surveys (http://www.who.int/immunization_financing/en). under consideration are more expensive than those for the 5. Assumptions about the relative distribution of FICs by strategy and original six targeted EPI diseases, they may still be relatively region were based loosely on such factors as the proportion of the popu- cost-effective compared with other interventions and with lation with access to health services for fixed facilities and the likelihood of active mobile strategies. treatment costs. Our analysis shows a wide range of cost- 6. A proxy for the total number of deaths averted is the sum of the indi- effectiveness estimates depending on the type of vaccine, vidual deaths averted for each antigen in the traditional EPI. This figure vaccine prices, coverage levels, and delivery strategy, with the may overestimate the actual number of deaths averted by fully immuniz- ing children and therefore underestimate the cost per death averted. additional incremental cost per person being relatively small However, the values estimated by region appear to support previously for some new vaccines. Declines in unit prices of new vaccines reported estimates, and direct estimation of deaths averted was impossible also will affect cost-effectiveness results. given data and model limitations. 408 | Disease Control Priorities in Developing Countries | Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, and others REFERENCES Creese, A. L., and M. A. Domínguez-Ugá. 1987. "Cost-Effectiveness of Immunization Programs in Colombia."Pan American Health Organiza- Aaby, P. 1988. "Malnutrition and Overcrowding/Intensive Exposure in tion Bulletin 21: 377­94. Severe Measles Infection: Review of Community Studies." Reviews of Crowcroft, N. S., C. Stein, P. Duclos, and M. Birmingham. 2003. "How Best Infectious Diseases 10: 478­91. to Estimate the Global Burden of Pertussis?" Lancet Infectious Diseases Aaby, P., and C. J. Clements. 1989. "Measles Immunization Research: A 3: 413­18. Review." Bulletin of the World Health Organization 67: 443­48. Ding, D., P. E. Kilgore, J. D. Clemens, L. Wei, and X. Zhi-Yi. 2003. "Cost- Africa Measles Partnership. 2004. "Measles Investment Case." Board of Effectiveness of Routine Immunization to Control Japanese the Global Alliance for Vaccines and Immunization. http://www. Encephalitis in Shanghai, China." Bulletin of the World Health vaccinealliance.org/resources/Measles_Investment_Case_FINAL_w_ Organization 81: 334­42. addendum.pdf. Domínguez-Ugá, M. A. 1988. "Economic Analysis of the Vaccination Batt, K., J. Fox-Rushby, and M. Castillo-Riquelme. 2004."The Costs, Effect, Strategies Adopted in Brazil in 1982." Bulletin of the World Health and Cost-Effectiveness of Strategies to Increase Coverage of Routine Organization 22: 250­68. Immunizations in Low- and Middle-Income Countries: Systematic Edmunds, D. J., A. Dejene, Y. Mekkonen, M. Haile, W. Alemnu, and Review of the Grey Literature." Bulletin of the World Health Organiza- D. J. Nokes. 2000. "The Cost of Integrating Hepatitis B Virus Vaccine tion 82: 689­96. into National Immunization Programs: A Case Study from Addis Beasley, R. P. 1988. "Hepatitis B Virus: The Major Etiology of Ababa." Health Policy and Planning 15: 408­16. Hepatocellular Carcinoma." Cancer 61: 1942­56. Edwards, K. M., and M. D. Decker. 2004. "Pertussis Vaccine." In Vaccines, Bennett, J. V., A. E. Platonov, M. P. E. Slack, P. Mala, A. H. Burton, and ed. S. A. Plotkin and W. A. Orenstein, 471­528. Philadelphia: Saunders. S. E. Robertson. 2002. Haemophilus influenzae Type B (Hib) Meningitis Elbasha, E. H., and M. L. Messonnier. 2004. "Cost-Effectiveness Analysis in the Pre-vaccine Era: A Global Review of Incidence, Age Distributions, and Health Care Resource Allocation: Decision Rules under Variable and Case-Fatality Rates. WHO/V&B/02.18. Geneva: World Health Returns to Scale." Health Economics 13: 21­35. Organization. http://www.who.int/vaccines-documents/DocsPDF02/ Feigin, R. D., B. W. Stechenberg, and P. Hertel. 2004. "Diphtheria." In www696.pdf. Textbook of Pediatric Infectious Diseases, ed. R. D. Feigin, J. D. Cherry, Berman, P., J. Quinley, B. Yusuf, S. Anwar, U. Mustaini, A. Azof, and G. J. Demmler, and S. L. Kaplan, 1305­13. Philadelphia: Elsevier. B. Iskandar. 1991. "Maternal Tetanus Immunization in Aceh Province, Foster, S. O., D. A. McFarland, and A. M. John. 1993. "Measles." In Disease Sumatra: The Cost-Effectiveness of Alternative Strategies." Social Control Priorities in Developing Countries, ed. D. T. Jamison, W. H. Science and Medicine 33: 185­92. Mosley, A. R. Measham, and J. L. Bobadilla, 161­87. New York: Oxford Beutels, P. 1998. "Economic Evaluations Applied to HB Vaccination: University Press and World Bank. http://www.fic.nih.gov/dcpp/ General Observations." Vaccine 16: S84­92. dcp1/dcp1-ch8.pdf. . 2001. "Economic Evaluations of Hepatitis B Immunization: A Fox-Rushby, J. A., M. Kaddar, R. Levine, and L. Brenzel. 2004. "The Global Review of Recent Studies (1994­2000)." Health Economics 10: Economics of Vaccination in Low- and Middle-Income Countries." 751­74. Bulletin of the World Health Organization 82: 640. Birmingham, M., and C. Stein. 2003. "The Burden of Vaccine-Preventable Galazka, A., M. Birmingham, M. Kurian, and F. Gasse. Forthcoming. Diseases." In The Vaccine Book, ed. B. R. Bloom and P.-H. Lambert, "Tetanus." In The Global Epidemiology of Infectious Disease, ed. C. J. L. 1­21. San Diego, CA: Elsevier. Murray and A. D. Lopez. Cambridge, MA: Harvard University Press. Birmingham, M., L. Wolfson, M. Kurian, U. Griffiths, F. Gasse, and Galazka, A. M., and S. E. Robertson. Forthcoming. "Diphtheria." In The J. Vandelaer. 2004. "Estimating the Burden of Neonatal Tetanus." Global Epidemiology of Infectious Diseases, ed. C. J. L. Murray and Unpublished manuscript. A. D. Lopez, Cambridge, MA: Harvard University Press. Brenzel, L. 1995. "Final Report on the Longitudinal Cost-Effectiveness Galazka, A. M., S. E. Robertson, and G. P. Oblapenko. 1995. "Resurgence Study of the North Arcot District Polio Control Program (NADPCP)." of Diphtheria." European Journal of Epidemiology 11: 95­105. Resources for Child Health Project, U.S. Agency for International GAVI (Global Alliance for Vaccines and Immunization). 2004. "Guidelines Development, Arlington, VA. for Preparing a National Immunization Financial Sustainability Plan." . 2005. "Methods Used to Estimate the Costs of Scaling Up Geneva, GAVI. http://www.who.int/immunization_financing/tools/ Immunization Services for the Vaccine Preventable Disease Chapter, en/FSP_Guidelines_April%202004_En.pdf. Disease Control Priorities Project." World Bank, Washington, DC. Gay, N. J., W. J. Edmunds, E. Bah, and C. B. Nelson. 2001. Estimating the Brenzel, L., and P. Claquin. 1994. "Immunization Programs and Their Global Burden of Hepatitis B. Geneva: World Health Organization. Costs." Social Science and Medicine 39: 527­36. Gessner, B. D., A. Sutanto, M. Linehan, I. G. G. Djelantik, T. Fletcher, Brinsmead, R., S. Hill, and D. Walker. 2004. "Are Economic Evaluations of K. Ingerani, and others. 2005. "The Incidence of Vaccine-Preventable Vaccines Useful to Decision Makers? Case Study of Haemophilus Haemophilus influenzae Type B Pneumonia and Meningitis in Influenzae Type B Vaccines." Pediatric Infectious Disease Journal 23: Indonesian Children Using a Hamlet-Randomized Vaccine Probe 32­37. Design." Lancet 365 (9453): 43­52. Cherry, J. D., and R. E. Harrison. 2004. "Tetanus." In Textbook of Pediatric Golden, M., and G. L. Shapiro. 1984. "Cost-Benefit Analysis of Alternative Infectious Diseases, ed. R. D. Feigin, J. D. Cherry, G. J. Demmler, and Programs of Vaccination against Rubella in Israel." Public Health 98: S. L. Kaplan, 1766­76. Philadelphia: Elsevier. 179­90. Cherry, J. D., and U. Heininger. 2004. "Pertussis and Other Bordetella Griffith, A. H. 1979. "The Role of Immunization in the Control of Infections." In Textbook of Pediatric Infectious Diseases, ed. R. D. Feigin, Diphtheria." Developments in Biological Standards 43: 3­13. J. D. Cherry, G. J. Demmler, and S. L. Kaplan, 1588­1608. Philadelphia: Griffiths, U. K., G. Hutton, and E. D. Pascoal. 2005."The Cost-Effectiveness Elsevier. of Introducing Hepatitis B Vaccine into Infant Immunization Services Creese, A. L. 1986. "Cost-Effectiveness of Potential Immunization in Mozambique." Health Policy Plan 20 (1): 50­59. Interventions against Diarrheal Disease." Social Science and Medicine Griffiths, U. K., L. J. Wolfson, A. Quddus, M. Younus, and R. A. Hafiz. 2004. 23: 231­40. "Incremental Cost-Effectiveness of Supplementary Immunization Vaccine-Preventable Diseases | 409 Activities to Prevent Neo-natal Tetanus in Pakistan." Bulletin of the Conjugate, and Rotavirus Vaccines in National Immunization World Health Organization 82: 643­51. Schedules." Health Economics 9: 19­35. Hadler, S. C., S. L. Cochi, J. Bilous, and F. T. Cutts. 2004. "Vaccination Miller, M. A., S. C. Redd, S. Hadler, and A. Hinman. 1998. "A Model to Programs in Developing Countries." In Vaccines, ed. S. A. Plotkin and Estimate the Potential Economic Benefits of Measles Eradication for W. A. Orenstein, 1407­42. Philadelphia: Saunders. the United States." Vaccine 20: 1917­22. Halsey, N. A., M. Blatter, G. Bader, M. L. Thoms, F. F. Willingham, Miller, M. A., R. W. Sutter, P. M. Strebel, and S. C. Hadler. 1996. "Cost- J. C. O'Donovan, and others. 1997. "Inactivated Poliovirus Vaccine Effectiveness of Incorporating Inactivated Poliovirus Vaccine into the Alone or Sequential Inactivated and Oral Poliovirus Vaccine in Two-, Routine Childhood Immunization Schedule." Journal of the American Four-, and Six-Month-Old Infants with Combination Haemophilus Medical Association 276: 967­71. influenzae Type B/Hepatitis B Vaccine." Pediatric Infectious Disease Monath, T. P. 2004. "Yellow Fever Vaccine." In Vaccines, ed. S. A. Plotkin Journal 16: 675­79. and W. A. Orenstein, 1095­176. Philadelphia: Saunders. Halsey, N., and A. Galazka. 1985. "The Efficacy of DPT and Oral Monath T. P., R. B. Craven, A. Adjukiewicz, M. Germain, D. B. Francy, Poliomyelitis Immunization Schedules Initiated from Birth to 12 Weeks L. Ferrara, and others. 1980. "Yellow Fever in The Gambia, 1978­79: of Age." Bulletin of the World Health Organization 63 (6): 1151­69. Epidemiologic Aspects with Observations on the Occurrence of Henderson, R. H., and T. Sundaresan. 1982. "Cluster Sampling to Assess Orungo Virus Infections." American Journal of Tropical Medicine and Immunization Coverage: A Review of Experience with a Simplified Hygiene 29: 912­28. Sampling Method." Bulletin of the World Health Organization 60: Morris, S. S., R. E. Black, and L. Tomaskovic. 2003. "Predicting the 253­60. Distribution of Under-Five Deaths by Cause in Countries without Hinman, A. R., B. Irons, M. Lewis, and K. Kandola. 2002. "Economic Adequate Vital Registration Systems." International Journal of Analyses of Rubella and Rubella Vaccines: A Global Review." Bulletin of Epidemiology 32: 1041­51. the World Health Organization 80: 264­70. Moss, W. J., C. J. Clements, and N. A. Halsey. 2003. "Immunization of Karlsson, G., and M. Johannesson. 1998. "Cost-Effectiveness Analysis and Children at Risk of Infection with Human Immunodeficiency Virus." Capital Costs." Social Science and Medicine 46: 1183­91. Bulletin of the World Health Organization 81: 61­70. Khaleghian, P. 2001. "Immunization Financing and Sustainability: A Mulholland, K., S. Hilton, R. Adegbola, S. Usen, A. Oparaugo, Review of the Literature." Special Initiatives Report 40. Bethesda, MD: C. Omosigho, and others. 1997. "Randomised Trial of Haemophilus Partnerships for Health Reform Project, Abt Associates. influenzae Type B Tetanus Protein Conjugate Vaccine [Corrected] for Krugman, S. 1963. "Measles and Poliomyelitis Vaccines." New York State Prevention of Pneumonia and Meningitis in Gambian Infants." Journal of Medicine 63: 2973­77. Lancet 349: 1191­97. Levin, A., S. England, J. Jorissen, B. Garshong, and J. Teprey. 2001. Case Mulligan, J.-A., J. A. Fox-Rushby, T. Adam, B. Johns, and A. Mills. 2003. Study on the Costs and Financing of Immunization Services in Ghana. "Unit Costs of Health Care Inputs in Low- and Middle-Income Report by PHR plus. Bethesda, MD: Abt Associates. Regions." Disease Control Priorities Project Working Paper 9. DCPP, National Institutes of Health, Bethesda, MD. http://www.fic.nih.gov/ Levine, M. M., R. Lagos, O. S. Levine, I. Heitmann, N. Enriquez, M. E. Pinto, dcpp/wpb9.pdf. and others. 1998. "Epidemiology of Invasive Pneumococcal Infections in Infants and Young Children in Metropolitan Santiago, Chile, a Nasidi, A., T. P. Monath, K. DeCock, O. Tomori, R. Cordellier, O. D. Olaleye, Newly Industrializing Country." Pediatric Infectious Disease Journal 17: and others. 1989. "Urban Yellow Fever Epidemic in Western Nigeria, 287­93. 1987." Transactions of the Royal Society of Tropical Medicine and Hygiene 83: 401­6. Lydon, P. 2004. "Financial Sustainability Plan Analysis: A Look across 22 GAVI Countries." World Health Organization, Geneva. NIH (National Institutes of Health). 2000. "Jordan Report 20th Anniversary: Accelerated Development of Vaccines." http://www. Margolis, H. S., M. J. Alter, and S. C. Hadler. 1997. "Viral Hepatitis." In niaid.nih.gov/dmid/vaccines/jordan20/. NIH, Bethesda. Viral Infections of Humans: Epidemiology and Control, ed. A. S. Evans and R. A. Kaslow, 363­418. New York: Plenum. Pegurri, E., Fox-Rushby, J., and Walker, D. 2005. "The Effects and Costs of Expanding Coverage of Immunization Services in Developing Martin, M., J. M. Casellas, S. A. Madhi, T. J. Urquhart, S. D. Delport, Countries: A Systematic Literature Review." Vaccine 23: 1624­35. F. Ferrero, and others. 2004. "Impact of Haemophilus influenzae Type B Conjugate Vaccine in South Africa and Argentina." Pediatric Infectious Peltola, H. 2000."Worldwide Haemophilus influenzae Type B Disease at the Disease Journal 23: 842­47. Beginning of the 21st Century: Global Analysis of the Disease Burden 25 Years after the Use of the Polysaccharide Vaccine and a Decade Mast, E., F. Mahoney, M. A. Kane, and H. S. Margolis. 2004. "Hepatitis B after the Advent of Conjugates." Clinical Microbiology Reviews 13: Vaccine." In Vaccines, ed. S. A. Plotkin and W. A. Orenstein, 299­338. 302­17. Philadelphia: Saunders. Perry, R. T. Forthcoming. Mathers, C. D., A. D. Lopez, and C. J. L. Murray. 2006. "The Burden of Disease and Mortality by Condition: Data, Methods, and Results for Perry, R. T., and N. A. Halsey. 2004. "The Clinical Significance of Measles: the Year 2001." In Global Burden of Disease and Risk Factors. ed. Alan D. A Review." Journal of Infectious Diseases 189 (Suppl. 1): S4­16. Lopez, Colin D. Mathers, Majid Ezzati, Dean T. Jamison, and Plotkin, S. A., and E. Vidor. 2004. "Poliovirus Vaccine: Inactivated." In Christopher J. L. Murray. New York: Oxford University Press. Vaccines, ed. S. A. Plotkin and W. A. Orenstein, 625­50. Philadelphia: McQuillan, G. M., P. J. Coleman, D. Kruszon-Moran, L. A. Moyer, Saunders. S. B. Lambert, and H. S. Margolis. 1999. "Prevalence of Hepatitis B Robertson, R. L., A. J. Hall, P. E. Crivelli, Y. Lowe, H. M. Inskip, and Virus Infection in the United States: The National Health and S. K. Snow. 1992. "Cost-Effectiveness of Immunizations: The Gambia Nutrition Examination Surveys, 1976 through 1994." American Journal Revisited." Health Policy and Planning 7: 111­22. of Public Health 89: 14­18. Robertson, S. E. 1993. The Immunological Basis for Immunization Series Miller, M. A. 1998. "An Assessment of the Value of Haemophilus influenzae Module 6: Poliomyelitis. WHO/EPI/GEN/93.16. Geneva: World Health Type B Conjugate Vaccine in Asia." Pediatric Infectious Disease Journal Organization. 17: S152­59. Sangrugee N, V. Caceres, and S. Cochi. 2004. "Cost Analysis of Post-polio Miller, M. A., and L. McCann. 2000. "Policy Analysis of the Use of Certification Immunization Policies." Bulletin of the World Health Hepatitis B,Haemophilus influenzae Type B-,Streptococcus pneumoniae- Organization 82: 9­15. 410 | Disease Control Priorities in Developing Countries | Logan Brenzel, Lara J. Wolfson, Julia Fox-Rushby, and others Siraprapasiri T., W. Sawaddiwudhipong, and S. Rojanasuphot. 1997. Delivering Routine Immunization Services in Peru." Bulletin of the "Cost-Benefit Analysis of Japanese Encephalitis Vaccination Program World Health Organization 82: 676­82. in Thailand." Southeast Asian Journal of Tropical Medicine and Public Wassilak, S, G. F. Trudy, V. Murphy, M. H. Roper, and W. A. Orenstein. Health 28: 143­48. 2004."Tetanus Toxoid."In Vaccines,ed.S.A.Plotkin andW.A.Orenstein, Soucat, A., D. Levy-Bruhl, X. De Bethune, P. Gbedonou, J.-P. Lamarque, 745­82. Philadelphia: Saunders. O. Bangoura, and others. 1997. "Affordability, Cost-Effectiveness, and Wenger, J. D., and J. Ward. 2004. "Haemophilus influenzae Vaccine." In Efficiency of Primary Health Care: The Bamako Initiative Experience Vaccines, ed. S. A. Plotkin and W. A. Orenstein, 229­68. Philadelphia: in Benin and Guinea." International Journal of Health Planning and Saunders. Management 12: S81­108. WHO (World Health Organization). 1974. Handbook of Resolutions. Vol. 1, Stanfield, J. P., and A. Galazka. 1984. "Neonatal Tetanus in the World 1.8. World Health Assembly, Fourteenth plenary meeting, 23 May Today." Bulletin of the World Health Organization 62: 647­69. 1974. Geneva: WHO. Stein, C. E., M. Birmingham, M. Kurian, P. Duclos, and P. Strebel. 2003. . 1988. EPICost. Geneva: WHO. "The Global Burden of Measles in the Year 2000: A Model That Uses . 1992. Global Health Situation and Projections: Estimates. Country-Specific Indicators." Journal of Infectious Diseases 187 WHO/HST/92.1. Geneva: WHO. whqlibdoc.who.int/hq/1992/ (Suppl. 1): S8­15. WHO_HST_92.1.pdf. Steinglass, R., L. Brenzel, and A. Percy. 1993. "Tetanus." In Disease Control . 1996. "HBsAG Endemicity." http://wwwstage/vaccines- Priorities in Developing Countries, ed. D. T. Jamison, W. H. Mosley, A. R. surveillance/graphics/htmls/hepbprev.htm. WHO, Geneva. Measham, and J. L. Bobadilla, 189­220. New York: Oxford University Press and World Bank. . 2001. Estimating the Local Burden of Haemophilus influenzae Type b (Hib) Disease Preventable by Vaccination: A Rapid Assessment Sutter, R. W., and O. M. Kew. 2004. "Poliovirus Vaccine: Live." In Vaccines, Tool. WHO/V&B/01.27. Geneva: WHO. ed. S. A. Plotkin and W. A. Orenstein, 651­706. Philadelphia: Saunders. . 2002. Core Information for the Development of Immunization Swartz, M. N. 2004. "Bacterial Meningitis: A View of the Past 90 Years." Policy, 2002 Update. WHO/V&B/02.28. Geneva: WHO. http://www. New England Journal of Medicine 351: 1826­28. who.int/vaccines-documents/DocsPDF02/www557.pdf. Turk, D. C. 1982. "Clinical Importance of Haemophilus influenzae: 1981." . 2004. "Progress toward Global Immunization Goals, 2001." In Haemophilus influenzae, ed. S. H. Sell and P. G. Wright, 3­9. New http://www.who.int/vaccines/. WHO, Geneva. York: Elsevier. . 2005a. "Progress in Reducing Global Measles Deaths: 1999­2003." UNICEF (United Nations Children's Fund). 2002. State of the World's Weekly Epidemiological Record 80 (9): 78­81. Vaccines and Immunization. New York: United Nations. . 2005b. World Health Report 2005: Make Every Mother and Child USAID (U.S. Agency for International Development), Asia­Near East Count. Geneva: WHO. Region. 1988. "Resources for Child Health Project." Asia-Near East Bureau Guidance for Costing of Health Services Delivery Projects, Wolfson, L., and P. Lydon. 2005. "Methodology for Estimating Baseline Arlington, VA. and Future Levels of Costing (and Impact) for the Global Immunization Vision and Strategy 2005­2015, Draft 1.1." World Valdmanis, V., D. Walker, and J. Fox-Rushby. 2003. "Are Vaccination Sites Health Organization, Geneva. in Bangladesh Scale Efficient?" International Journal of Technology Assessment in Health Care 19: 692­97. World Bank. 1993. Investing in Health: World Development Report, 1993. New York: Oxford University Press. Walker, D., N. R. Mosqueira, M. E. Penny, C. F. Lanata, A. D. Clark, C. F. B. Sanderson, and J. Fox-Rushby. 2004. "Variation in the Costs of Vaccine-Preventable Diseases | 411 Chapter 21 Conquering Malaria Joel G. Breman, Anne Mills, Robert W. Snow, Jo-Ann Mulligan, Christian Lengeler, Kamini Mendis, Brian Sharp, Chantal Morel, Paola Marchesini, Nicholas J. White, Richard W. Steketee, and Ogobara K. Doumbo Malaria is the most important of the parasitic diseases of CAUSES, EPIDEMIOLOGY, MANIFESTATIONS, humans, with 107 countries and territories having areas at risk AND DIAGNOSIS of transmission containing close to 50 percent of the world's population (Hay and others 2004; WHO 2005). More than Four species of the genus Plasmodium cause malarial infections 3 billion people live in malarious areas and the disease causes in humans: P. falciparum, P. vivax, P. ovale, and P. malariae. between 1 million and 3 million deaths each year (Breman, Virtually all deaths are caused by falciparum malaria. Human Alilio, and Mills 2004; Snow and others 2003). Recent esti- infection begins when the malaria vector, a female anopheline mates of the global falciparum malaria morbidity burden have mosquito, inoculates plasmodial sporozoites from its salivary increased the number to 515 million cases, with Africa suffer- gland into humans during a blood meal. The sporozoites ing the vast majority of this toll (Snow and others 2005). In mature in the liver and are released into the bloodstream as addition, almost 5 billion clinical episodes resembling malaria merozoites. These invade red blood cells, causing malaria occur in endemic areas annually, with more than 90 percent of fevers. Some forms of the parasites (gametocytes) are ingested this burden occurring in Africa (Breman 2001; Breman, Alilio, by anopheline mosquitoes during feeding and develop into and Mills 2004; Carter and Mendis 2002; Snow and others sporozoites, restarting the cycle. 1999, 2003; Snow, Trape, and Marsh 2001). P. falciparum predominates in Haiti, Papua New Guinea, and The disease has resurged in many parts of the tropics, and Sub-Saharan Africa, while P. vivax is more common in Central nonmalarious countries face continual danger from importa- America and the Indian subcontinent and causes more than tion. Contributing to this resurgence are the increasing prob- 80 million clinical episodes of illness yearly (Mendis and others lems of Plasmodium falciparum resistance to drugs and of the 2001). The prevalence of these two species is approximately Anopheles vectors' resistance to insecticides. The recent findings equal in the Indian subcontinent, eastern Asia, Oceania, and that insecticide-treated nets (ITNs) are extremely cost-effective South America. P. malariae is found in most endemic areas, in preventing malaria and overall deaths and that intermittent especially throughout Sub-SaharanAfrica,but is much less com- preventive therapy (IPT) (treatment doses given during mon than the other species. P. ovale is unusual outside Africa, periods of vulnerability) is effective for protecting pregnant and where it is found accounts for less than 1 percent of isolates. women and their fetuses, along with the discovery of new drugs While more than 40 anophelines can transmit malaria, the (artemisinins) and their use in combination with other most effective are those such as Anopheles gambiae, which are antimalarials and promising vaccine trials, have given great long-lived, occur in high densities in tropical climates, breed impetus to the battle against this scourge (Alonso and others readily, and bite humans in preference to other animals. The 2004; Armstrong-Schellenberg and others 2001; Lengeler 2004; entomological inoculation rate (EIR)--that is, the number of Newman and others 2003; Yeung and others 2004). sporozoite-positive mosquito bites per person per year--is the 413 most useful measure of malarial transmission and varies from less than 1 in some parts of Latin America and Southeast Asia Extrinsic to more than 300 in parts of tropical Africa. The epidemiology of malaria may vary considerably within relatively small geographic areas. In tropical Africa or coastal Papua New Guinea, with P. falciparum transmission, more than Human Control Social, behavioral, one human bite per infected mosquito can occur per day and and prevention economic, and people are infected repeatedly throughout their lives. In such measures political factors areas, morbidity and mortality during early childhood are con- Intrinsic siderable. For survivors, some immunity against disease devel- ops in these areas, and by adulthood, most malarial infections are asymptomatic. This situation, with frequent, intense, year- Parasite Mosquito round transmission, is termed stable malaria. In areas where Environmental conditions transmission is low, erratic, or focal, full protective immunity is not acquired and symptomatic disease may occur at all ages. This situation is termed unstable malaria. An epidemic or complex emergency can develop when changes in environmen- Source: Breman 2001. tal, economic, or social conditions occur, such as heavy rains Figure 21.1 Determinants of the Malaria Burden: Intrinsic and following drought or migrations of refugees or workers from a Extrinsic Factors nonmalarious to an endemic region. A breakdown in malaria control and prevention services intensifies epidemic condi- tions. Epidemics occur most often in areas with unstable malaria, such as Ethiopia, northern India, Madagascar, Sri seizures, and deep coma--have some residual neurological Lanka, and southern Africa. Many other African countries sit- deficit when they regain consciousness. Protein-calorie under- uated in the Sahelian and Sub-Saharan areas are susceptible to nutrition and micronutrient deficiencies, particularly zinc and epidemics (Djimdé and others 2004; Worrall, Rietveld, and vitamin A, contribute substantially to the malaria burden Delacollette 2004). Public health specialists have only recently (Caulfield, Richard, and Black 2004). begun to appreciate the considerable contribution of urban In areas with intense and stable transmission, falciparum malaria, with up to 28 percent of the burden in Africa occur- malaria in primigravid and secundigravid women is associated ring in rapidly growing urban centers (Keiser and others 2004). with low birthweight (average reduction of about 170 grams) The determinants of malaria and of risk factors for patients and consequently with increased infant and childhood mor- and communities relate to intrinsic (human, parasite, and tality (Steketee and others 2001). HIV infection predisposes all vector) and extrinsic (environmental, control, and socioeco- pregnant women to more frequent and severe malaria, and the nomic) factors (figure 21.1)(Breman 2001). Both humoral and reverse may also be true (Ter Kuile and others 2004). P. vivax cellular immunity are necessary for protection. malaria in pregnancy is also associated with a reduction in Anemia may be quite common among young children living birthweight (average reduction of some 100 grams), and this in areas with stable transmission, particularly where the para- effect is greater in multigravid than in primigravid women site is resistant to chloroquine, sulfadoxine-pyrimethamine (Nosten and others 1999). (SP), or other drugs. Correctly and promptly treated, uncom- The confirmatory diagnosis of clinical malaria rests on the plicated falciparum malaria has a mortality rate of approxi- microscopic demonstration of asexual forms of the parasite in mately 0.1 percent (Sudre and others 1992). Once vital organ stained peripheral blood smears. Newer diagnostic tests using dysfunction occurs or the proportion of erythrocytes infected antigen and nucleic acid detection methods are being evalu- increases to more than 3 percent, mortality rises steeply. Coma ated. These tests are promising, but their limitations in relation is a characteristic and ominous feature of falciparum malaria, to species sensitivity (except for P. falciparum), parasite quanti- and despite treatment it is associated with death rates of some tation, field feasibility, and costs necessitate further develop- 20 percent among adults and 15 percent among children. ment and evaluation (Warhurst and Williams 2004). Convulsions, usually generalized and often repeated, occur in up to 50 percent of children with cerebral malaria (CM) BURDEN OF DISEASE (Mung'Ala-Odera, Snow, and Newton 2004). Whereas less than 3 percent of adults suffer neurological sequelae, roughly 10 to In 2001, the World Health Organization (WHO) ranked malaria 15 percent of children surviving CM--especially those with as the eighth-highest contributor to the global disease burden hypoglycemia, severe malarial anemia (SMA), repeated as reflected in disability-adjusted life years (DALYs), and the 414 | Disease Control Priorities in Developing Countries | Joel G. Breman, Anne Mills, Robert W. Snow, and others Table 21.1 Deaths and DALYs from Deaths Attributable to All Causes and to Malaria by WHO Region, 2000 Malaria Malaria DALYs Deaths, 2000 deaths DALYs from deaths, 2000 as a as a per- All causes Malaria percent- All causes Malaria centage age of all of all Region Population Thousands Percent Thousands Percent deaths Thousands Percent Thousands Percent DALYs World 6,122,211 56,554 100.0 1,124 100.0 2.00 1,467,257 100.0 42,279 100.0 2.90 Africa 655,476 10,681 18.9 963 85.7 9.00 357,884 24.4 36,012 85.2 10.10 Americas 837,967 5,911 10.5 1 0.1 0.02 145,217 9.9 108 0.2 0.07 Eastern 493,091 4,156 7.3 55 4.9 1.30 136,221 9.3 2,050 4.8 1.50 Mediterranean Europe 874,178 9,703 17.2 1 0.1 0.010 151,223 10.3 20 0.04 0.01 Southeast Asia 1,559,810 14,467 25.6 95 8.5 0.70 418,844 28.5 3,680 8.7 0.90 Western Pacific 1,701,689 11,636 20.6 10 0.9 0.09 257,868 17.6 409 1.0 0.20 Source: Breman, Alilio, and Mills 2004; WHO 2002b. Percentages may not add up to 100 percent because of rounding. second highest in Africa (WHO 2002a). The DALYs attributa- Table 21.2 Population at Risk for Falciparum Malaria, Cases ble to malaria were estimated largely from the effects of P. falci- and Attack Rates by World Health Organization Region, parum infection as a direct cause of death and the much smaller 2002* contributions of short-duration, self-limiting, or treated mild Population Cases in Falciparum attack febrile events, including malaria-specific mild anemia and neu- at risk in millions rate (1,000 persons rological disability following CM (Murray and Lopez 1996, Region millions (percent) (IQR) (percent) at risk per year) 1997). The estimate assumes that each illness event or death World 2,211 (100) 515 (100) 236 can be attributed only to a single cause that can be measured [298­659] reliably. Table 21.1 shows deaths and DALYs from deaths attrib- Africa 521 (24) 365 (71) 701 utable to malaria and to all causes by WHO region (WHO [216­374] 2002a). It does not include the considerable toll caused by the Americas 55 (3) 4 (1) 73 burden of malaria-related moderate and severe anemia, low [2­8] birthweight, and comorbid events (Snow and others 2003). Eastern 176 (8) 12 (2) 68 Sub-Saharan African children under four represent 82 percent Mediterranean [5­25] of all malaria-related deaths and DALYs. Malaria accounts for Europe 4 ( 1) 1( 1) 157 2.0 percent of global deaths and 2.9 percent of global DALYs. In [ 1­1] the African region of WHO, 9.0 percent of deaths and 10.1 per- Southeast Asia 1,314 (59) 119 (23) 91 cent of DALYs are attributable to malaria. [66­224] Recent analysis of falciparum malaria morbidity concludes Western Pacific 142 (6) 15 (3) 106 that 515 (interquartile range 298 to 659) million cases occur [9­26] yearly (table 21.2). This figure is 92 percent higher than the Source: Modified from Snow and others 2005. 278 million malaria cases estimated by WHO for 1998, which IQR interquartile range, urban-adjusted. also includes those attributable to P. vivax, and 200 percent higher than previous estimates for areas outside of Africa Mills 2004; Snow and others 2003). Of the 10.6 million yearly (Snow and others 2005). While the malaria incidence globally deaths in children younger than 5 years 8 percent are ascribed is 236 episodes per 1,000 persons per year in all endemic areas, to malaria (Bryce and others 2005). In 1998, an empirical it ranges from about 400 to 2,000 (median 830) episodes per analysis of malaria mortality undertaken on behalf of WHO 1,000 persons per year in areas with intense, stable (hyperen- used malaria risk maps to capture measures of disability, mor- demic and holoendemic) transmission; these areas represent bidity, and mortality associated with P. falciparum prevalence 38 percent of all falciparum-endemic areas. rates among African populations and yielded an estimate of about 1 million (Korenromp and others 2003; Snow and Marsh Background 2002; Snow and others 2003; Snow, Trape, and Marsh 2001). Recent estimates of malaria deaths have varied from 0.5 million Each malarious country must now measure its own burden and to 3.0 million per year (Breman 2001; Breman, Alilio, and progress toward decreasing that burden (WHO 2005). Conquering Malaria | 415 Geographic and climate-driven (mainly rainfall) models of significantly to malaria mortality (tables 21.3 and 21.4). The suitability for malaria transmission characterize the diversity of differences in numbers derive from the different methods of cal- malaria transmission across the African continent (Craig, culating the burden. Children presenting with an acute febrile Snow, and le Sueur 1999; Snow and others 1999). Four distinct disease, peripheral parasitemia, and low hemoglobin concen- areas can be identified: trations account for the majority of inpatient admissions in · class 1, no transmission (northern and parts of southern Africa) Table 21.4 Deaths from Malaria in Children Under Five, · class 2, marginal risk (mainly in some areas of southern Africa, 2001 Africa and in high-altitude [ 1500 meters] settings) Cause of malaria-related death Number of malaria deaths · class 3, seasonal transmission with epidemic potential (along the Sahara fringe and in highlands) Cerebral malaria 110,000 · class 4, stable and unstable malarious areas (most areas Severe malarial anemia 190,000­974,000 south of the Sahara to southern Africa and below an altitude Respiratory distress 110,000 of around 1,500 meters). Hypoglycemia 153,000­267,000 Low birthweight 62,000­363,000 Direct Consequences of P. falciparum Infection Total deaths from malaria 625,000­1,824,000 Two major syndromes, CM and moderate (hemoglobin less All-cause deathsa 962,000­2,806,000 than 11 grams per deciliter) and severe (hemoglobin less than Sources: Breman, Alilio, and Mills 2004; Murphy and Breman 2001. 8 grams per deciliter) malarial anemia, contribute directly and a. Children under five represent 65 percent of all deaths in Africa as per Snow and others (2003). Table 21.3 Malaria-Related Mortality and Morbidity, Africa, 2000 Percentage affected by age in years Condition Number (range) 0­4 5­14 15 Mortality and Morbidity Malaria-specific mortality 1,144,572a 65 19 16 (702,957­1,605,448) Maternal mortality (anemia) 5,300 n.a. n.a. 100 Infant mortality (pregnancy related) 71,000­190,000 100 n.a. n.a. Fatal adverse events from malaria drugs 2,300 100 Unknown Unknown HIV from blood transfusion necessitated by malaria 5,300­8,500 100 -- Epilepsy-related mortality after cerebral malaria Unknown Unknown Unknown Unknown Malaria-related anemia, undernutrition, and HIV mortality Unknown Unknown Unknown Unknown Episodes of malaria (thousands) 213,549 51 35 14 (134,322­324,617) Illness days from malaria (thousands) 803,699 69 21 10 (494,416­1,298,872) Neurocognitive sequelae after CM Hemiparesis 360­400 100 Unknown Quadriparesis 770­860 100 Unknown Hearing impairment 650­730 100 Unknown Visual impairment 300­330 100 Unknown Behavioral difficulties 1,540­1,720 100 Unknown Language deficits 7,000­7,800 100 Unknown Epilepsy 2,700­3,000 100 Unknown Effects on cognition Unknown Unknown Unknown Unknown Source: Breman, Alilio, and Mills 2004; Snow and others 2003. n.a. not applicable. -- not available. Note: Figures in parentheses are interquartile ranges. a. WHO (2002b) reports 1,124,000 deaths from malaria. 416 | Disease Control Priorities in Developing Countries | Joel G. Breman, Anne Mills, Robert W. Snow, and others areas with stable transmission. The arbitrary definition of sis, quadriparesis, hearing and visual impairments, speech and 5 grams of hemoglobin per deciliter is prognostic for a fatal out- language difficulties, behavioral problems, epilepsy, and other come and proves useful clinically as a criterion for transfusion. problems (table 21.3). The incidence of neurocognitive sequelae Lactic acidosis commonly coexists with hypoglycemia and is following severe malaria is only a fraction of the true residual (with coma, repeated convulsions, shock, and hyperpara- burden, and the impact of milder illness is unknown. sitemia) an important predictor of death from severe malaria Studies of children presenting to hospital with malaria and (White and Breman 2005; WHO 2000b). hemoglobin of less than or equal to 5 grams per deciliter indi- The vast majority of deaths in developing countries occur cate a median transfusion rate of 80.1 percent. Thus, 275,400 to outside the formal health service, and in Africa, most govern- 442,290 surviving SMA admissions, newborn through 14 years, ment civil registration systems are incomplete (Breman 2001; will be exposed to blood transfusion each year in Sub-Saharan Greenberg and others 1989). Newer demographic and disease- Africa. As a result, each year 5,300 to 8,500 children, age birth tracking systems are being used globally and should help rectify through 14 years, living in stable endemic areas of Africa are the woefully inadequate vital statistics available for malaria and likely to acquire HIV infection because of exposure to blood other diseases (INDEPTH 2002). transfusion to manage SMA (Colebunders and others 1991; Health personnel usually attribute causes of death during Savarit and others 1992). demographic surveillance system surveys through a verbal autopsy interview with relatives of the deceased about the Indirect and Comorbid Risks symptoms and signs associated with the terminal illness. Both The DALY model of malaria does not sufficiently take it into the specificity and the sensitivity of verbal autopsy vary account as an indirect cause of broader morbid risks. Some considerably depending on the background spectrum of other consider anemia to be caused indirectly unless linked to acute, common diseases, such as acute respiratory infection, gas- high-density parasitemia. Similarly, low birthweight may also troenteritis, and meningitis, which share common clinical be indirectly attributable to malaria, and a child's later under- features with malaria (Korenromp and others 2003). nutrition and growth retardation linked to malaria infection In malarious Africa, some 30 to 60 percent of outpatients enhances the severity of other concomitant or comorbid with fever may have parasitemia. Monthly surveillance of infectious diseases through immune suppression. Thus, households will detect a quarter of the medical events that are malaria infection contributes to broad causes of mortality detected through weekly surveillance, and weekly contacts with beyond the direct fatal consequences of infection and is prob- cohorts identify approximately 75 percent of events detected ably underestimated (Breman, Alilio, and Mills 2004; Snow through daily surveillance (Snow, Menon, and Greenwood and others 2003). 1989). Given the predominance of fevers, malaria case man- In Africa, pregnant women experience few malaria-specific agement in Africa and other endemic areas usually centers on fever episodes but have an increased risk of anemia and presumptive diagnosis. placental sequestration of the parasite. Maternal clinical mani- Estimates of the frequency of fever among children suggest festations are more apparent in areas with less intense one episode every 40 days. If we assume that the perceived transmission, particularly in Asia. Estimates indicate that in frequency of fever in Africa is similar across all transmission Sub-Saharan Africa, malaria-associated anemia is responsible areas (and possibly all ages), African countries would witness for 3.7 percent of maternal mortality, or approximately 5,300 approximately 4.9 billion febrile events each year. Estimates maternal deaths annually. indicate that in areas of stable malaria risk, a minimum of Prematurity and intrauterine growth retardation resulting 2.7 billion exposures to antimalarial treatment will occur each in low birthweight associated with maternal malaria account year for parasitemic persons, or 4.93 per person per year (Snow for 3 to 8 percent of infant mortality in Africa (Steketee and and others 2003). While these diagnostic, patient management, others 1996, 2001). Assuming an infant mortality rate of 105 and drug delivery assumptions are debatable, they indicate the per 1,000 live births, Snow and others (2003) calculate that in magnitude of the challenges malaria presents. 2000, 71,000 to 190,000 infant deaths were attributable to The case-fatality rates of CM are high, even with optimal malaria in pregnancy (table 21.3). Other studies indicate that management. Murphy and Breman (2001) report a mean case- malaria-associated low birthweight accounted for 62,000 to fatality rate of 19.2 percent and Snow and others (2003) cite a 363,000 infant deaths (Murphy and Breman 2001). figure of 17.5 percent. Those who succumb at home without Anemia among African children is caused by a combination optimal treatment will have higher case-fatality rates. of nutritional deficiencies and iron loss through helminth Studies of neurological sequelae after severe malaria indicated infection, red cell destruction, decreased red cell production that 3 to 28 percent of survivors suffered from such sequelae, as a result of infectious diseases, and genetically determined including prolonged coma and seizures (Mung' Ala-Odera, hemoglobinopathies. Chronic or repeated infections, often Snow, and Newton 2004). CM is associated with hemipare- associated with parasite resistance to drugs, are more likely to Conquering Malaria | 417 involve bone marrow suppression (Menendez, Fleming, and absenteeism (Fernando and others 2003) As noted earlier, Alonso 2000). malaria may result in low birthweight, and low birthweight can Murphy and Breman (2001) estimate that 190,000 to lead to a range of persistent impaired outcomes, predomi- 974,000 deaths per year in Sub-Saharan Africa are attributable nantly behavioral difficulties, cerebral palsy, mental retarda- to SMA. Children residing in areas where the prevalence of tion, blindness, and deafness. The recently launched studies of P. falciparum was more than 25 percent had a 75 percent preva- intermittent preventive treatments during infancy (IPTi) lence of anemia. By modeling the relationship between anemia should provide a more precise means of examining the benefits and parasite prevalence, Snow and others (2004) found that of IPTi and consequences on learning and performance of mild anemia rose 6 percent for every 10 percent increase in the infection early in life (Holding and Kitsao-Wekulo 2004; Rosen prevalence of infection. Reducing the incidence of new infec- and Breman 2004; Schellenberg and others 2005). tions through ITNs or the prevalence of blood-stage infections through chemoprophylaxis or IPT for children halved the risk of anemia. INTERVENTIONS AND THEIR EFFECTIVENESS Caulfield, Richard, and Black (2004) report that iron, zinc, Malaria will be conquered only by full coverage, access to, and and protein-calorie deficits are responsible for a considerable use of antimalarial services by priority groups; rapid, accurate amount of malaria-related mortality and morbidity and indi- diagnosis; prompt and effective patient management (diagno- cate that 57.3 percent of deaths of underweight children under sis, treatment, counseling and education, referral); judicious five are attributable to nutritional deficiencies. One striking use of insecticides to kill and repel the mosquito vector, includ- feature of the global distribution of anthropometric markers of ing the use of ITNs; and control of epidemics. Eliminating undernutrition is its congruence with the distribution of malaria from most endemic areas remains a distant, huge, but endemic malaria. Improved growth among young children has surmountable challenge because of the widespread Anopheles more recently been demonstrated in The Gambia and Kenya in breeding sites; the large number of infected people; the use of a comparison of those protected and unprotected by ITNs ineffective antimalarial drugs; and the inadequacies of (Ter Kuile and others 2003). resources, infrastructure, and control programs. The Roll Back Early during the HIV epidemic, Greenberg and others Malaria Partnership, which began in 1998, aims to halve the (1988) and Greenberg (1992) demonstrated that malaria- burden of malaria by 2010 and has developed strategies and associated anemia treated with unscreened blood transfusions targets for 2005 (box 21.1). While ambitious, the initiative is contributed to HIV transmission. At the same time, two longi- making substantial progress by means of effective and efficient tudinal cohort studies in Kenya and Uganda and one hospital- deployment of currently available interventions (WHO 2003, based case-control study in Uganda demonstrated that HIV 2005). Indeed, Brazil, Eritrea, India, and Vietnam are reporting infection approximately doubles the risk of malaria parasitemia recent successes in reducing the malaria burden (Barat 2005). and clinical malaria in nonpregnant adults and that increased Despite the enormous investment in developing a malaria vac- HIV immunosuppression is associated with higher-density cine administered by means of a simple schedule and recent parasitemias (French and others 2001; Whitworth and others promising results in the laboratory and in field trials in Africa, 2000). In pregnant women, the presence of HIV increases the no effective, long-lasting vaccine is likely to be available for rate and intensity of parasitemia and frequency of anemia general use in the near future (Alonso and others 2004; Ballou (Ter Kuile and others 2004). The increasing incidence of HIV- and others 2004). associated febrile illnesses may lead to increased use of antimalarials. Some believe that the recommended use of trimethoprim-sulfamethoxazole for prophylaxis of bacterial Drug Use pneumonia and other infections in HIV/AIDS patients may Proper use of drugs is essential. Early diagnosis and effective contribute to SP resistance and that monitoring is required. Yet, treatment of patients lends credibility to the malaria program, evaluation of trimethoprim-sulfamethoxazole for malaria pro- strengthens confidence in the health care system by families phylaxis in Mali did not show any increases in parasite resistance and communities, and raises the esprit of clinicians and public mutations specific for these drugs (Thera and others 2005). health workers. Malaria accounts for 13 to 15 percent of medical reasons for absenteeism from school, but little information is available on Early Diagnosis and Treatment. Early diagnosis and effective the performance of parasitized schoolchildren (Holding and treatment can cure infection, prevent further morbidity Kitsao-Wekulo 2004). A randomized placebo control study of and progression to severe disease and death, and arrest chloroquine prophylaxis in Sri Lankan schoolchildren demon- transmission. This intervention requires timely and accurate strated an improvement in mathematics and language scores diagnosis; use of efficacious drugs; education of patients and by those who received chloroquine but found no difference in their families about the disease, home management, and 418 | Disease Control Priorities in Developing Countries | Joel G. Breman, Anne Mills, Robert W. Snow, and others Box 21.1 Roll Back Malaria's Strategy and Goals for 2005 The goal of Roll Back Malaria Partnership is to halve the burden of malaria by 2010. The following targets for specific intervention strategies were established at the Abuja Malaria Summit in April 2000. Strategy Abuja target (by 2005) · Prompt access to effective treatment · 60 percent of those suffering with malaria should have access to and be able to use correct, affordable, and appropriate treatment within 24 hours of the onset of symptoms. · Provision of ITNs · 60 percent of those at risk for malaria, particularly chil- dren under five and pregnant women, will benefit from a suitable combination of personal and community pro- tective measures, such as ITNs. · Prevention and control of malaria in pregnant women · 60 percent of pregnant women at risk of malaria will have access to IPTa. · Epidemic and emergency response · 60 percent of epidemics are detected within two weeks of onset. · 60 percent of epidemics are responded to within two weeks of detection. Source: WHO 2003b, 2005. a. The original Abuja declaration included the recommendation for chemoprophylaxis as well, but current WHO and Roll Back Malaria policy strongly recommends IPT and not chemopro- phylaxis for preventing malaria during pregnancy. prevention; and referral to higher levels of the health system. in areas with resistance to single drugs are combination The following are critical to the effectiveness of this treatments, preferably artemisinin combination therapy intervention: (ACT) (WHO 2001a, 2001b, 2003a, 2005). While ACT is a welcome, life-saving approach, more information on the · Timeliness. A febrile malaria attack warrants early treat- cost-effectiveness of this new strategy is needed (Arrow, ment. If left untreated, a proportion of P. falciparum malaria Panosian, and Gellband 2004; Yeung and others 2004). infections, perhaps 1 in 250 (and to a far lesser extent infec- Baird (2005) reports that ACT costs range from US$2.00 tions with other malaria species) will progress to severe (artesunate-amodiaquine, three doses in 48 hours) to disease within a few hours to a few days (Greenwood and US$9.12 (artemether-lumefantrine, six doses in 48 hours) others 2005). The globally agreed goal is that diagnosis per adult treatment; WHO has obtained the latter drug for and effective treatment should be provided within 24 hours US$2.40 per adult treatment for qualified purchasers, mean- of the onset of symptoms and signs. ing those from low-income malarious countries. · Diagnosis and effective drug treatment. An accurate diagno- · Location of clinical management. Effective management of sis of malaria is based on detection of the parasite and, if patients requires skilled and well-equipped personnel at all laboratory diagnosis is not feasible, on clinical grounds. levels of the health system. The two strategies for delivering Health workers must monitor the therapeutic efficacy of antimalarials effectively are through health facilities and in drugs closely and change treatment policies when parasite or near the home when access to health facilities is limited. resistance to chloroquine (figure 21.2), SP, and other drugs Given the pervasiveness of malaria infections, more infor- emerges (Baird 2005; Laxminarayan and others 2006; WHO mation on the most effective and efficient ways to promote 2002a). Concerns include unreliable and inaccurate home treatment is urgently needed. microscopy and the disadvantages of alternative tests plus the widespread distribution and use of substandard and Evidence for the effectiveness of early diagnosis and treat- counterfeit drugs. The recommended treatments for malaria ment is available from two different epidemiological conditions Conquering Malaria | 419 % Failure 100 80 60 40 20 0 of of CAR Togo Faso Mali EthiopiaBurundi EritreaRwandaKenya Chad Tanzania ZambiaUganda GabonRep. Rep. GhanaLiberia NigerGuinea Benind'Ivoire LeoneGambiaNigeria Botswana Senegal Zimbabwe Cameroon Mauritania Mozambique Dem. Côte Sierra The Congo, Burkina Congo, Eastern, Southern, Great Lakes block Central block Western block Source: WHO 2005. Note: WHO has established 126 sentinel sites in 36 African countries that monitor the efficacy of locally used antimalarial drugs by following patients in clinics. The box indicates the 25th­75th percentile, the vertical line the lower and upper values, and where the lines cross the median. Figure 21.2 Chloroquine Treatment Failure in Africa, 1997­2002 and health systems. In areas of low to moderate transmission, In primigravid and secundigravid women, the incidence of for example, southern Africa, the Americas, and Southeast Asia, severe maternal anemia, the incidence and density of placental where health care systems are relatively effective, the two major parasitemia, and the incidence of low birthweight were 25 to consequences of prompt and effective intervention are 95 percent lower when mothers were given IPT than when they reduction of the period of infectivity of infected persons, were not (Kayentao and others 2005; Steketee and others 2001). and thus reduced transmission and incidence, and a lowered While initial studies indicate that IPT in infancy reduces ane- case-fatality rate and overall mortality. In Vietnam and the mia and febrile episodes, more research is needed. SP is becom- KwaZulu-Natal province of South Africa, P. falciparum malaria ing ineffective for IPT, and health experts are suggesting ACT as incidence and mortality rates fell when effective treatment a replacement (White 2005). policies (artesunate and ACT) replaced failing monotherapies (Hung and others 2002). The effective drug policies were Chemoprophylaxis. Chemoprophylaxis is advised by travel implemented in conjunction with enhanced vector control; medicine specialists for nonresidents of endemic areas who are thus, effective treatment alone does not account for the fall in exposed to malaria for short periods. WHO does not recom- malaria incidence and mortality. mend chemoprophylaxis for permanent residents of endemic In areas with stable, high transmission of P. falciparum areas because of low feasibility and compliance and high costs malaria--for instance, Papua New Guinea and Sub-Saharan relative to the public health benefits (WHO 1996, 2000a, Africa--where access to treatment is poor, little is known about 2000b). The choice of chemoprophylaxis will depend on the the impact of early and effective treatment on malaria trans- drug-sensitivity profile, tolerance, side effects, costs, regimen, mission. With a high EIR (10 to 1,000), a reduction in trans- and compliance by patients (Baird 2005; Kain, Shanks, and mission intensity is unlikely to affect the incidence of disease Keystone 2001; White and Breman 2005). The effectiveness of until low EIRs (less than 10) are reached. chemoprophylaxis will depend mainly on patient compliance and parasite susceptibility. IPT in Pregnancy and Infancy. IPT is recommended in pregnancy in areas with high and stable transmission of P. falci- Supervision and Policy Change. In relation to drug use, all parum malaria; IPT usually consists of two curative doses of control activities must be supervised and evaluated rigorously antimalarial treatment. The recommended drug is SP given dur- to ensure high-quality care for patients and program effective- ing the second and third trimesters of pregnancy during prena- ness and efficiency (Bryce and others 1994; Jha, Bangoura, and tal care visits. IPT in infancy involves giving infants treatment Ranson 1998). Studies in 37 countries indicate the need for a doses during vaccination or well-baby visits to health clinics. treatment policy change to ACT (WHO 2002a). Several 420 | Disease Control Priorities in Developing Countries | Joel G. Breman, Anne Mills, Robert W. Snow, and others countries in Africa, Asia, and South America are now imple- required to inform national initiatives to scale up ITN use menting ACT as the first-line treatment for uncomplicated (Lengeler and Sharp 2003); yet, recent encouraging reports malaria (WHO 2005). Many countries are also trying to show that Eritrea, Malawi, Togo, Zambia, and other countries improve the time lag before treatment. More than 83 percent of in Africa are already scaling up nationally with high coverage. the population in the three malarious provinces in South Africa is within 10 kilometers of a health facility (Sharp and le Sueur Vector Control 1996). In Burkina Faso, Ethiopia, and Uganda, where access to The reduction of Anopheles breeding and biting of humans clinics was poor and difficult, mothers and community health involves different methods of insecticide and repellent applica- workers were empowered to dispense treatment, which result- tion, environmental management, and behavioral change of ed in major reductions in mortality and morbidity in children populations at risk. (Kidane and Morrow 2000; Pagnoni and others 1997; Sirima and others 2003). Indoor Residual Spraying of Dwellings with Insecticide. Indoor residual spraying (IRS) is the application of long-lasting insecticides (up to six months) on the walls of dwellings. Insecticide-Treated Nets Insecticides repel mosquitoes from entering houses or impart The use of ITNs (bednets, curtains, and other materials) to a lethal dose of the insecticide on the female mosquito when it provide personal protection by killing or repelling mosquitoes rests on a sprayed surface, thereby preventing subsequent is one of the major strategies of malaria control (RBM 2002). transmission. IRS is most effective against indoor-biting Pyrethroids are recommended for the periodic treatment or re- (endophilic) mosquito vectors. Vector susceptibility and post- treatment of the protective materials. feeding behavior are the main criteria to be considered when The effectiveness of ITNs depends on their acceptability by choosing an insecticide: organophosphates, carbamates, and the population at risk and their affordability. It is contingent pyrethroids are the main compounds used although some on the habits, biology, and susceptibility of the mosquito countris still rely on organochlorines (dichlorodiphenyl- vector; the compliance of the human population; and the con- trichloroethane, or DDT). centration of insecticide on or in the fiber, which has to be The effectiveness of this intervention depends on cost, toxi- maintained by regular re-treatment or by incorporating the city, acceptability of the insecticide, its residual effects, and local insecticide in the fiber for long duration. political and international partnership commitment. Malaria Over 20 studies in Africa and Asia have demonstrated more incidence decreased sharply following the use of IRS in large- than 50 percent protective efficacy for individual users of ITNs scale programs in many parts of Africa, the Americas, Asia, and in reducing malaria episodes, 29 percent protection against Europe (Lengeler and Sharp 2003). severe malarial disease, and substantial protection against In southern Africa, more than 13 million people in six coun- anemia (Lengeler 2004). Most importantly, the use of ITNs tries are protected from malaria (WHO 2002c). Control was reduced child mortality by 18 percent in five sites in Sub- initiated through the use of IRS and supported by engineering Saharan Africa (Lengeler 2004). approaches, larviciding, prompt diagnosis, and effective treat- Lengeler's (2004) review demonstrates the efficacy of ITNs ment. Examples of successful and sustained malaria elimina- in both stable and unstable transmission areas. Widespread tion using IRS with effective drug treatment are available from use of ITNs resulted in an overall reduction in mortality of Cyprus, Greece, Portugal, Spain, and the former Yugoslavia and 19 percent, protected against anemia, and had a substantial its successor states (Curtis and Mnzava 2000). The most suc- impact on mild disease episodes. One large-scale rural study cessful malaria control efforts were linked closely to research in Tanzania found that ITNs and untreated nets reduced mor- and took place in many parts of Asia, where the notable exam- tality of children one month to four years, with protective ple is the near eradication of malaria from Sri Lanka in the efficacies of 27 and 19 percent, respectively (Armstrong- early 1960s, and in Central and South America from the late Schellenberg and others 2001). Re-treating ITNs semiannually 1950s to early 1970s (Alilio, Bygbjerg, and Breman 2004; Carter or just before the annual peak in transmission is essential for and Mendis 2002; Gilles 2002). effective vector control and is proving a major logistical and financial challenge. Fortunately, new types of nets with a long- Larviciding and Fogging. Larviciding is the application of lasting insecticidal property are now available, and re- chemical insecticides, including those of biological origin and treatment will soon cease to be an issue. The salutary impact insect growth regulators, to breeding sites. These may be of large-scale ITN programs has been demonstrated in China applied to all mosquito breeding sites or targeted to the breed- (Tang 2000), Tanzania (Abdulla and others 2001; Armstrong- ing sites of specific vectors. Recommended compounds, for- Schellenberg and others 2001; WHO 2005), and Vietnam mulations, and dosages for larviciding are available through the (Hung and others 2002). More operational experience is WHO Pesticides Evaluation Scheme. Fogging or space spraying Conquering Malaria | 421 with insecticides requires specialized equipment, because the These persons are an extension of the health system and work particle size of the insecticide determines its suspension quali- under the direct supervision of health facility staff or non- ties in the air, the number of droplets, and the penetration of governmental organizations and in conformity with standards space. The insecticide is not effective for as long as with IRS and norms established by the national government (Gilles or ITNs, and application must occur during periods of peak 2002). Such information can help to increase the standard of target mosquito activity, generally at night. patient care and prevention programs by promoting citizen Larviciding is not generally indicated for large-scale vector and community advocacy and demand for control. control in rural endemic areas because of the difficulty of locat- ing all breeding sites, their often temporary nature, and the required frequency of application. Thus, larviciding is usually ECONOMICS OF MALARIA CONTROL limited to urban areas, refugee camps, and industrial and INTERVENTIONS development projects. Despite its impact on mosquito density and its contribution to reducing transmission, larviciding is not Goodman, Coleman, and Mills's (2000) study represents the as effective as IRS and ITNs in reducing mosquito longevity most thorough attempt to compare the cost-effectiveness of a (Najera and Zaim 2002). wide range of malaria control interventions. They find that in a very low-income country, the cost-effectiveness range Civil Engineering. Draining and filling larval breeding sites is per DALY averted was US$19 to US$85 for ITNs (nets plus one of the oldest methods of mosquito control and must be insecticide), US$32 to US$58 for residual spraying (two rounds targeted to the breeding sites of locally important malaria per year), US$3 to US$12 for chemoprophylaxis for children vectors. Civil engineering strategies can require costly heavy (assuming an existing delivery system), US$4 to US$29 for IPT equipment and materials and are useful for eliminating perma- for pregnant women, and US$1 to US$8 for case-management nent breeding sites in urban areas, which are increasing improvements. Goodman, Coleman, and Mills (2000) find that globally, and at development project sites where earth removal even though some interventions are relatively cheap, achieving has occurred. high coverage may require a level of expenditure currently out of reach for many African countries and that overcoming oper- ational barriers to achieving widespread coverage is likely to Home Repellents and Insecticide Use require substantial assistance from external donors. Commercially available mosquito repellents are applied directly on the skin or clothing as aerosols, lotions, or creams and contain active ingredients that protect the individual from mos- Analysis quito bites. Commercially available mosquito coils containing The following analysis incorporates new knowledge on the pyrethroids can be burnt to repel mosquitoes, and electrically effects of interventions and on their costs for a low-income, heated dispensers serve a similar function. Some communities Sub-Saharan African population living in an area of high, in endemic regions use smoke, burning herbs, or plants to deter stable transmission. The modeling draws on a wide range of mosquitoes from entering the home. sources on the costs and effects of each intervention, extrapo- N,N-diethyl-meta-toluamide (DEET) is the most widely lated across settings and operational conditions. The approach used and effective ingredient in commercially available repel- allows for changing cost-effectiveness over time, for example, as lents (Curtis and others 1991). While several studies have shown resistance to antimalarial drugs or insecticides increases. To that mosquito coils are effective at repelling mosquitoes, they are address problems of uncertainty in relation to many of the not as effective as ITNs (Charlwood and Jolly 1984). parameters, we used probabilistic sensitivity analysis, which allows for multivariate uncertainty by assigning ranges rather than point estimates to input variables. We assumed that cost Health Education and Counseling and effectiveness input variables follow uniform triangular or Health education is the provision of information via news- normal continuous probability distributions (Mulligan, Morel, papers, radio, or television, and health counseling is interactive, and Mills 2005). is individual, and involves the transfer of skills. The provision We consider the cost-effectiveness of a limited subset of of information to households on ways to prevent malaria is interventions: ITNs, IRS, IPT during pregnancy, and patient needed in all endemic communities. It should cover the impor- management with a change of the first-line drug. We include tance of early treatment and where to access it, the use of costs to the provider and the community and incremental out- referral services, and the significance of full compliance with of-pocket expenses for households, but because of major treatment and other interventions. The necessary information valuation and measurement problems, we do not include the can be provided by community and voluntary health workers. indirect costs of patients' time to seek care and of the lost 422 | Disease Control Priorities in Developing Countries | Joel G. Breman, Anne Mills, Robert W. Snow, and others productivity resulting from morbidity. We consider only gross US$34. These results should be interpreted with caution costs for all interventions except patient management, given because of uncertainty in relation to the estimates of effective- the uncertainty inherent in estimating savings. ness in children under five. The model was based on one round of spraying per year in areas of seasonal transmission and two Insecticide-Treated Nets. We based our analysis of ITNs on rounds per year in areas of high, intense (perennial) transmis- the delivery mechanism used in the WHO Special Programme sion. Effectiveness will depend on the length of the transmis- for Research and Training in Tropical Diseases (WHO/TDR) sion seasons and on the insecticide. DDT lasts for six months trials, where householders, community health workers, and or more, lambda-cyhalothrin for three to six months, and program staff worked together to treat the nets. In relation to malathion and deltamethrin for only two to three months insecticide, we considered permethrin and deltamethrin. (Lengeler 2004; Lengeler and Sharp 2003). Deltamethrin is effective for a year; thus, re-treatment is annual. Permethrin lasts for six months; thus, we assumed two IPT during Pregnancy. We analyzed the cost-effectiveness of treatments per year if the transmission season is longer than IPT assuming that primigravid women are given two or three six months. The activities undertaken were the training of doses of SP at a prenatal clinic. We analyzed benefits to the child staff and community health workers, a campaign to inform the by decreased mortality and benefits to the mother resulting community about the intervention, the procurement and from changes in the incidence of severe anemia. We estimated transport of the insecticide and nets, and the initial treatment the effect of IPT on the neonatal mortality rate as a function of and the re-treatment of the nets. We calculated cost- increased birthweight based on the Cochrane meta-analysis effectiveness for each intervention for two scenarios: one of malaria prevention in pregnancy (Gülmezoglu and Garner whereby nets were distributed to households and the second 1998). The model allowed for level of drug resistance, prob- whereby treatment was arranged for existing nets. We drew ability of initial attendance at a prenatal clinic, probability of estimates of the effectiveness of ITNS from a recent meta- returning for a second visit, probability of returning for a third analysis of WHO/TDR-sponsored trials conducted in Sub- visit, and compliance with the drug regimen. We estimated both Saharan Africa (Lengeler 2004). We adjusted the key parameter incremental and average costs. and effectiveness estimates to account for the proportion of The incremental CER for IPT in pregnancy using SP had a children sleeping and not sleeping under a recently treated net. 90 percent range from US$9 to US$21 with a mean of US$13. With one treatment per year using deltamethrin, the mean Average total cost-effectiveness had a mean of US$24 (90 per- cost per DALY averted was US$11 (90 percent range of US$5 to cent range of US$16 to US$35). US$21). With one treatment of permethrin per year the cost- effectiveness ratio (CER) increased slightly to a mean of Change in First-Line Drug. We analyzed the cost-effectiveness US$12 (90 percent range of US$6 to US$20). Two treatments of of changing first-line therapies to SP and to ACT using a permethrin per year increased the mean CER to US$17 (90 per- patient-management model with a decision-tree framework in cent range of US$9 to US$31). Even if net coverage is low and which a patient presents with uncomplicated malaria at an out- nets have to be distributed and treated, the intervention patient facility and progresses to full recovery, recovery with remains an extremely attractive use of resources. Moreover, the neurological sequelae, or death. Three potential drug policy model includes health benefits only for children under five. If changes were considered (see table 21.5). For current drug poli- we included benefits for other household members, the CERs cies we assumed either chloroquine or SP as the first-line drug, would be lower. with SP or amodiaquine as a second-line drug and quinine as the third-line drug. We then examined the cost-effectiveness of Insecticide Residual Spraying. We considered four insecti- policy switches to either SP or ACT as first-line drug (with cides for our analysis of the cost-effectiveness of IRS: DDT; amodiaquine as second line and quinine as third line). malathion; and two pyrethroids, deltamethrin and lambda- Amodiaquine was chosen as the second-line drug in the new cyhalothrin. We used the results of the Cochrane Review drug policy because, like chloroquine and SP, it is relatively meta-analysis of ITN trials conducted in Africa to approxi- cheap. Low compliance, adverse effects, and potential cross- mate the results of spraying on morbidity and mortality and resistance between amodiaquine and chloroquine excluded it adjusted effectiveness estimates to account for noncompliance from first-line selection. We omitted other potential drugs to (Lengeler 2004). limit the scope of analysis. We found little difference between the CERs for the four We used commonly found levels of drug resistance to create insecticides when one round of spraying was done per year: likely ranges. When used as second-line treatment, we assumed they ranged from US$5 to US$18. With two rounds per year, that a drug faced half the level of resistance than when it was costs increased, but we assumed that effectiveness remained the used as a first-line drug. We estimated the growth rate of resist- same, so all the CERs approximately doubled to US$11 to ance to each drug based on its expected current location along Conquering Malaria | 423 Table 21.5 Change in First-Line Drug Current drug policy New drug policy Scenario First-line drug Second-line drug Third-line drug First-line drug Second-line drug Third-line drug A Chloroquine SP Quinine SP Amodiaquine Quinine B Chloroquine SP Quinine ACT Amodiaquine Quinine C SP Amodiaquine Quinine ACT Amodiaquine Quinine Source: Authors. Resistance to current drug (percent) Table 21.6 CERs for ITNs, IRS, and IPT 100 (2001 US$) 90 Mean cost per 90 per- 80 Intervention DALY averted cent range 70 ITNs (net insecticide treatment) 60 Deltamethrin 11 5­21 50 Permethrin (1 treatment) 12 6­20 40 Permethrin (2 treatments) 17 9­31 30 20 ITNs (without provision of nets) 10 Deltamethrin 5 2­7 0 Permethrin (1 treatment) 6 3­9 Chloroquine Chloroquine SP to ACT Permethrin (2 treatments) 11 6­17 to SP to ACT IRS (1 round) switch is dominant switch is cost-effective (less costly, more effective) ( $150/DALY averted) Melathion 12 8­18 switch is cost-effective current therapy is dominant DDT 9 5­13 ( $150/DALY averted) (less costly, more effective) Deltamethrin 10 6­14 Source: Authors. Lambda-cyhalothrin 10 6­14 IRS (2 rounds) Figure 21.3 Cost-Effectiveness of Switching the First-Line Drug Malathion 24 15­34 DDT 17 11­24 a sigmoid growth curve. We assessed compliance with each Deltamethrin 18 12­27 drug based on the length and complexity of each regimen. We Lambda-cyhalothrin 19 12­28 gave the widest range (20 to 70 percent) to compliance with ACT, which usually requires a three-day treatment, to account IPT for the different lengths of regimen and variety of formulations. Incremental costs 13 9­21 Figure 21.3 shows that a switch from chloroquine to SP is Average costs 24 16­35 cost-effective (less than $150 per DALY averted) when chloro- Source: Authors' calculations. quine resistance is above 35 percent. Switching from chloro- quine to ACT becomes cost-effective as chloroquine resistance reaches around 37 percent. Switching from SP to ACT becomes net coverage is already high, and IPT is even more cost-effective cost-effective as SP resistance reaches 12 percent. This low if prenatal care coverage is good. However, even if levels of threshold is due to the high growth rate of resistance to SP infrastructure are poor, these interventions are still attractive when it is used as a first-line therapy. based on cost-effectiveness criteria. Curtis and Mnzava's (2000) review comparing worldwide Results and Interpretation. Table 21.6 presents the average trials of ITNs and IRS for malaria control suggests that they CERs for the interventions reviewed. All interventions can be were of equivalent effectiveness. Lengeler and Sharp (2003 21,) considered attractive using a cutoff of US$150 per DALY also conclude that choosing between IRS and ITNs is "large- averted. For the childhood preventive interventions, the level ly a matter of operational feasibility and availability of existing infrastructure is a key factor in influencing of local resources, rather than one of malaria epidemiology or cost-effectiveness. ITNs are an even better use of resources if cost-effectiveness." DDT is the cheapest insecticide but is 424 | Disease Control Priorities in Developing Countries | Joel G. Breman, Anne Mills, Robert W. Snow, and others seldom used because of concerns about its environmental at the main high-risk groups. For example, a national voucher impact. An updated systematic review of the health effects of scheme is currently being implemented in Tanzania to provide IRS as well as a full economic comparison between spraying every pregnant woman in the country with a free ITN, and a with DDT and other insecticides and ITNs is urgently required. net is being distributed to children at the time of measles vac- Such reviews should include the environmental benefits of cination in Togo. reducing DDT levels and the costs of alternative interventions. Re-treatment of ITNs on a large-scale remains a formidable As effective patient management, IPT, and other approaches operational issue, and free distribution of insecticide in the toward drug use become more widespread, drug resistance will way it is done in Vietnam and China is probably the best way increase, which will affect cost-effectiveness. In terms of first- forward. This should occur while waiting the availability of line treatment, while a switch from chloroquine to SP is unlikely long-lasting insecticidal nets that do not require retreatment. to be costly, cost-effectiveness depends on the initial level of Curtis and Maxwell (2002) point to the successful experi- severe resistance to each drug. A switch from chloroquine to ence of Vietnam, where ITNs now protect 10 million people ACT is likely to be costly but more effective, given that resist- and the public health service provides free insecticide. This ance to ACT is essentially nonexistent and the growth rate of approach in Africa would require a substantial commitment by resistance to ACT is likely to be low (Yeung and others 2004). donors and governments given that most African governments Given the high growth rate of resistance to SP, switching to ACT spend only around US$4 per capita per year on health (World becomes cost-effective when SP resistance surpasses 12 percent, Bank 2003). Eritrea, with 65 percent ITN coverage, and other a relatively low threshold. A switch from chloroquine to ACT African countries are beginning to make great strides in getting appears to be highly cost-effective at all initial levels of chloro- ITNs to their populations (Barat 2006; WHO 2005; World quine resistance above 37 percent. Recent studies indicate the Bank 2005). remarkable effectiveness of the ACT artemether-lumefantrine The move toward ACT poses the most difficult questions in in East African areas of chloroquine, amodiaquine, and SP relation to the long-term affordability of malaria control inter- resistance (Mutabingwa and others 2005; Piola and others ventions. After scale-up, estimates indicate that the additional 2005); it is expected that the availability of ACTs will increase annual costs of ACT versus current failing drugs range from and the costs will decrease in the near future. US$300 million to US$500 million globally (Arrow, Panosian, and Gellband 2004), with the precise amount depending on the Affordability and Scaling Up extent to which all fevers are treated. This figure does not Cost-effectiveness analyses can identify which interventions include drugs for nonmalarial fevers in endemic areas and the are the most efficient to implement, but information is also costs for the substantial health system strengthening required needed on affordability. Some interventions are cheap, such as to make the most effective use of ACT. These include the costs prevention in pregnancy. Achievement of high coverage with an of improved drug regulation, pharmacovigilance, diagnostics, intervention to prevent childhood malaria (for example, ITNs) and implementation of different drug policies for different has a high total cost, which needs to be borne in part by exter- population groups. The introduction of ACT has permanently nal funding. For example, full coverage of children (assuming changed the economic landscape of malaria control. Innovative 22 percent of the population is under five years of age) would funding solutions at the global level are required to ensure that cost US$2.81 million per 1 million persons of the general pop- effective drugs are made available to the most vulnerable ulation covered (assuming the provision of nets plus two groups. The Global Fund to Fight AIDS, Tuberculosis, and rounds of permethrin). The same coverage with IRS would cost Malaria is providing the major leadership and resources for about US$4.01 million using deltamethrin (two rounds). securing ACTs and other commodities, and many other agen- In addition to the difficulties of financial feasibility, the cies and organizations are joining coalitions to combat malaria implementation of ITN interventions poses operational and through control, research, training, funding, and advocacy logistical challenges. A number of strategic approaches are activities (Breman, Alilio, and Mills 2004; Feacham 2004: available for national ITN upscaling (RBM 2002). The main http://www. theglobalfund.org/en). ones are social marketing (for example, in Malawi and Kenya), assisted commercial sector development (Senegal, Mali, Tanzania), and totally free distribution (Togo). An important Economic Benefits of Malaria Control feature of all the approaches is that their cost per net distrib- Given the substantial total costs of several malaria control inter- uted decreases significantly as the scale of the undertaking ventions, the case for introducing them can be strengthened by increases, so it is hard to use available estimates to project cost evidence on non-health-related benefits, especially evidence of in national programs. income gains or prevention of income losses. Three different Independent of the main ITN distribution strategy, a num- approaches to measuring such benefits are assessing the direct ber of countries are implementing additional actions targeted and indirect costs of malaria, studying the relationship between Conquering Malaria | 425 malaria and the output of agricultural or industrial activities, alent to 2.6 percent of annual household income (Ettling and and exploring the macroeconomic impact of malaria. Only the others 1994), and Russell (2004) reports that malaria generally third approach sheds much light on the benefits of control as consumes less than 10 percent of family income. Leighton and opposed to the burden of the disease in the absence of control. Foster (1993) estimate that the total annual value of malaria- This point is important, because in places facing the most related production losses was 2 to 6 percent of gross domestic severe malaria problems, direct information on the economic product in Kenya and 1 to 5 percent in Nigeria. The productiv- benefits of control is largely unavailable. ity consequences of mortality have received relatively little attention. Direct and Indirect Costs of Malaria. The standard approach The direct and indirect cost approach involves two method- has been to view the two key determinants of the economic ological problems. First, calculations are based mainly on days costs of malaria as the direct costs of expenditure on preven- of illness and neglect mortality, debility (usually from anemia), tion and treatment and the indirect costs of productive labor and neurological and other long-term sequelae. Second, house- time lost because of malaria morbidity and mortality, and to holds' coping strategies are likely to reduce the immediate estimate the total economic impact by adding the direct and impact of illness, although in the long term they may impose indirect costs (Mills 1992). Households use a range of preven- costs through the sale of assets such as livestock, which jeop- tive measures (mosquito coils, aerosol sprays, bednets, and ardizes a household's asset base. mosquito repellents) to differing degrees. A review of evidence for Sub-Saharan Africa found that monthly per capita house- Relationship between Malaria and Output. With the excep- hold expenditures ranged from US$0.05 per person in rural tion of a study of cotton production in Côte d'Ivoire, where the Malawi to US$2.10 in urban Cameroon, equivalent to US$0.24 prevalence of parasitemia above a cutoff density had a major and US$15 per household in 1999 U.S. dollars (Chima, impact on labor efficiency (Audibert, Mathonnat, and Henry Goodman, and Mills 2003). The costs of treatment for malaria 2003), studies have not succeeded in identifying impacts on include out-of-pocket expenditures for consultation fees, output. One possibility is that the risk of malaria may discour- drugs, transport, and subsistence at a distant health facility. For age the planting of crops that require intensive cultivation, the Sub-Saharan Africa, these costs ranged between US$0.41 and settlement and cultivation of fertile land, or the development of US$3.88 per person, equivalent to between US$1.88 and US$26 tourism and industry in suitable locations, but good evidence per household per year. Household expenditure on treatment is is lacking. usually highly regressive, consuming a much larger proportion of income in the poorest households. Macroeconomic Impact of Malaria. Recent empirical cross- Computations of public expenditures on malaria preven- country comparisons of economic growth indicate that tion and treatment are imprecise because most fall within eliminating malaria would have a strong positive impact on general health service expenditures. About 20 to 40 percent of economic development. Gallup and Sachs (2001) use cross- outpatient visits in malarious Africa are for fever, and suspect- country regression analysis to relate the growth in gross domes- ed malaria among inpatients ranges from 0.5 to 50.0 percent of tic product per capita between 1965 and 1990 to initial income admissions. Kirigia and others (1998) found that inpatient levels, initial human capital stock, policy variables, geographi- treatment for pediatric malaria absorbed 15 percent of the cal variables, and a malaria index calculated as the product of annual recurrent costs of inpatient care in one Kenyan hospital the fraction of land area with endemic malaria in 1965 and the and 9 percent in another. Ettling and Shepard (1991) estimated fraction of malaria cases that were due to P. falciparum in 1990. that Rwanda's Ministry of Health spent 19 percent of its oper- Their results suggest that countries with a substantial amount ating budget on malaria treatment. Because of the dominance of malaria grew 1.3 percent per year less than countries with of out-of-pocket spending by households, public expenditures little or no malaria between 1965 and 1990 (controlling for on malaria generally account for a minority of total malaria other influences on growth) and that a 10 percent reduction in expenditure. malaria was associated with 0.3 percent higher growth per year. The methods used to measure and value time lost and a McCarthy, Wolf, and Wu (2000) employ a similar approach to day's work vary considerably between studies; the average time explore the impact of malaria on average per capita growth lost per episode for a sick adult and for an adult caring for sick rates during three five-year periods. They find a significant children ranged from one to five days. The average indirect negative association between malaria and economic growth, cost per episode ranged from US$0.68 for children under although the estimated impact differed sharply across coun- 10 years of age in Malawi to US$23 per adult episode in tries. The impact was smaller than that found by Gallup and Ethiopia. Authors have concluded that aggregate productivity Sachs (2001), exceeding 0.25 percent per year in only a quarter losses can be significant for households and for the economy of the sample countries and averaging 0.55 percent for those in as a whole. In Malawi, the indirect costs of malaria were equiv- Sub-Saharan Africa. 426 | Disease Control Priorities in Developing Countries | Joel G. Breman, Anne Mills, Robert W. Snow, and others Benefit-Cost Ratios · carry out strategic and basic research on vector-parasite- Mills and Shillcutt 2004 related the evidence on the macroeco- host interactions nomic benefits of malaria control to information on the costs · assess mechanisms for addressing drug and insecticide of reducing malaria to calculate a benefit-cost ratio (Mills and resistance Shillcutt 2004). Depending on the assumptions, the benefit- · develop and carry out field evaluations of transgenic meth- cost ratio ranged between 1.9 and 4.7 (using a 3 percent dis- ods for interrupting malaria transmission. count rate). In terms of economic growth alone malaria control is extremely cost beneficial. Policy Research Social, economic, and policy research should focus on the RESEARCH PRIORITIES following: Current interventions to combat malaria remain inadequate for · developing and applying a common methodology for meas- achieving the increased levels of successful patient management uring socioeconomic status and prevention to which all malarious countries aspire and · carrying out policy and operational research on the impact, for which ambitious targets have been set (box 21.1). Greatly viability, sustainability, and optimal balance of public- increased support for malaria research and for developing insti- private partnerships tutional capacity must occur to make advances and to bring them · investigating ethical, legal, and social issues pertaining to to populations in need. The WHO Scientific Working Group on new malaria-related tools. Malaria and others have identified four major areas of research as follows (Remme and others 2002; WHO 1996, 2003c). Capacity strengthening for research and operations (includ- ing clinical trials) is urgently needed. This strengthening will Patient Management result in the ability to better evaluate new drugs and vaccines Patient management, including treatment, should address the and existing malaria control tools (ITNs, IPT) and to tackle the following: scaling up of malaria strategies. A classification of research priorities by time frame includes · evaluation of treatment effectiveness and of access to treat- the three-year, five-year, and ten-year targeting. ment for uncomplicated malaria in children and during pregnancy, with an emphasis on home management and Three-year completion targeting includes the following: evaluation of alternative delivery systems · investigation of the pathogenesis of malaria, in particular · apply to the U.S. Food and Drug Administration (FDA) anemia and immune response mechanisms with evidence for two separate fixed-dose artemisinin com- · evaluation of new approaches, for example, rectal drug binations at a target adult treatment price of US$1.00, or administration, for managing severe illness US$0.60 or less for children; two combinations are neces- · evaluation of ACT, including delivery approaches via the sary, as one may fail in testing or have other unforeseen public and private sectors problems · development of new drugs with novel targets; the Medicines · evaluate the two best candidate ACT drugs for IPT in preg- for Malaria Ventura (MMV) is spearheading this activity. nant women and in infancy and early childhood (IPTi) · launch studies to reach an evidence-based conclusion on the Prevention Research costs and benefits of long-lasting ITNs versus IRS in Sub- Saharan Africa and other endemic areas Prevention research should focus on the following: · carry out operational studies to determine the best methods · new approaches to drug-based malaria prevention, includ- of deploying ACT through the public and private sectors so ing IPT in children and during pregnancy that first-level patient management can occur at the home · strategies for scaling up the use of ITNs. and at the village levels; this involves studying packaging and distribution networks, assessing adherence, and ad- Innovative Approaches dressing how to deploy and supervise use of artesunate rec- Innovative approaches should use new technologies, including tal suppositories recent advances in sequencing the DNA of P. falciparum and · conduct economic reviews and predictive modeling to make A. gambiae, to achieve the following: an economic case for increased international investment in malaria control, including collecting detailed data from sev- · discover and develop drugs, diagnostics, vaccines, insecti- eral scaled-up, national, fully supported control programs cides, and antiparasite effector molecules using genomics employing the best available strategies and interventions Conquering Malaria | 427 · conduct in Africa and other endemic continents a maxi- Alonso, P. L., J. Sacarlal, J. J. Aponte, A. Leach, E. Macete, J. Milman, mum number of FDA-compliant phase I and phase II and others. 2004. "Efficacy of the RTS,S/AS02A Vaccine against Plasmodium falciparum Infection and Disease in Young African malaria vaccine trials to select candidates for clinical trials. Children: Randomised Controlled Trial." Lancet 364: 1411­20. Armstrong-Schellenberg, J. R., S. Abdulla, R. Nathan, O. Mukasa, T. J. The five-year completion target is as follows: Marchant, N. Kikumbih, and others. 2001."Effect of Large-Scale Social Marketing of Insecticide-Treated Nets on Child Survival in Rural · develop and deploy more sensitive, specific, and predictive Tanzania." Lancet 357: 1241­47. diagnostic tests that are inexpensive and practical Arrow, K. J., C. B. Panosian, and H. Gellband, eds. 2004. Saving Lives, Buying Time: Economics of Malaria Drugs in an Age of · carry out operational studies on how to deploy diagnostics Resistance. Washington, DC: National Academy Press for Institute of in areas of low, intermediate, and high transmission. Medicine. Audibert, M., J. Mathonnat, and M. C. Henry. 2003. "Malaria and Property The ten-year completion target has these focuses: Accumulation in Rice Production Systems in the Savannah Zone of Côte d'Ivoire." Tropical Medicine and International Health 8 (5): · develop an inexpensive, safe, and synthetic trioxane anti- 471­83. malarial drug Baird, J. K. 2005. "Effectiveness of Antimalarial Drugs." New England Journal of Medicine 352: 1565­77. · develop another new drug with a target and resistance Ballou, R., M. Arevalo-Herrera, D. Carucci, T. L. Richie, G. Corradin, mechanism that is unrelated to existing drugs C. Diggs, and others. 2004. "Update on the Clinical Development of · carry out basic insecticide research to develop new Candidate Malaria Vaccines." American Journal of Tropical Medicine approaches to both personal protection and residual house and Hygiene 71 (Suppl. 2): 239­47. spraying Barat, L. M. 2006."Four Malaria Success Stories: How Malaria Burden Was Successfully Reduced in Brazil, Eritrea, India, and Vietnam." American · launch successful phase III and phase IV field testing of Journal of Tropical Medicine and Hygiene 74 (1): 12­16. malaria vaccines to prevent clinical illness and transmission Breman, J. G. 2001. "The Ears of the Hippopotamus: Manifestations, with licensure by regulatory bodies Determinants, and Estimates of the Malaria Burden." American Journal · use genomic, proteomic, and bioinformatic tools to better of Tropical Medicine and Hygiene 64 (Suppl. 1­2): 1­11. understand the pathogenesis of malaria, to design new Breman, J. G., M. S. Alilio, and A. Mills. 2004. "Conquering the Intolerable Burden of Malaria: What's New, What's Needed: A Summary." drugs and vaccine candidates for training scientists, and to American Journal of Tropical Medicine and Hygiene 71 (Suppl. 2): 1­15. transfer technologies Bryce, J., C. Boschi-Pinto, K. Shibuya, R. E. Black, and the WHO Child · train a critical mass of leaders in science and operations to Health Epidemiology Reference Group. 2005. "WHO Estimates of the carry out the required research in support of control. Causes of Deaths of Children." Lancet 365 (9465): 1114­6. Bryce, J., J. B. Roungou, P. Nguyen-Dinh, J. F. Naimoli, and J. G. Breman. 1994. "Evaluation of National Malaria Control Programmes in Africa." Bulletin of the World Health Organization 72 (3): 371­81. CONCLUSION Carter, R., and K. Mendis. 2002."Evolutionary and Historical Aspects of the Burden of Malaria." Clinical Microbiological Reviews 15 (4): 564­94. Given the heavy burden of malaria, the need to use existing Caulfield, L., S. A. Richard, and R. Black. 2004. "Undernutrition as strategies and interventions in scaled-up programs more an Underlying Cause of Malaria Morbidity and Mortality." effectively and to deploy them more widely is urgent and mer- American Journal of Tropical Medicine and Hygiene 71 (Suppl. 2): its the highest priority, especially in Africa. While existing tools 55­63. can be improved, newer tools are required. The history of Charlwood, D., and D. Jolly. 1984. "The Coil Works against Mosquitoes in Papua, New Guinea." Transactions of the Royal Society of Tropical malaria research and control shows that they are synergistic. Medicine and Hygiene 78: 678. Integrating research and control activities has resulted in suc- Chima, R., C. Goodman, and A. Mills. 2003. "The Economic Impact of cess in several areas of the world, and will result in vanquishing Malaria in Africa: A Critical Review of the Evidence." Health Policy malaria early in the 21st century. 63 (1): 17­36. Colebunders, R., R. Ryder, H. Francis, W. Nekwei, Y. Bahwe, I. Lebughe, and others. 1991. "Seroconversion Rate, Mortality, and Clinical Manifestations Associated with the Receipt of a Human REFERENCES Immunodeficiency Virus-Infected Blood Transfusion in Kinshasa, Zaire." Journal of Infectious Diseases 164 (3): 450­56. Abdulla, S., J. A. Schellenberg, R. Nathan, O. Mukasa, T. Marchant, Craig, M. H., R. W. Snow, and D. le Sueur. 1999. "A Climate-Based T. Smith, and others. 2001. "Impact on Malaria Morbidity of a Distribution Model of Malaria Transmission in Sub-Saharan Africa." Programme Supplying Insecticide Treated Nets in Children Age under Parasitology Today 15: 105­11. Two Years in Tanzania: Community Cross-Sectional Study." British Medical Journal 322: 270­73. Curtis, C. F., J. D. Lines, B. Lu, and A. Renz. 1991. "Natural and Synthetic Repellents." In Control of Disease Vectors in the Community, ed. C. F. Alilio, M. S., I. Bygbjerg, and J. G. Breman. 2004. "Are Multilateral Malaria Curtis, 75­92. London: Wolfe. Research and Control Programs the Most Successful? Lessons from the Past 100 Years." American Journal of Tropical Medicine and Hygiene 70 Curtis, C. F., and C. Maxwell. 2002. "Free Insecticide for Nets Is Cost- (Suppl. 2): 268­78. Effective." Trends in Parasitology 18: 204­5. 428 | Disease Control Priorities in Developing Countries | Joel G. Breman, Anne Mills, Robert W. Snow, and others Curtis, C. F., and A. E. P. Mnzava. 2000. "Comparison of House Spraying INDEPTH Sites. Vol. 1 of Population and Health in Developing and Insecticide-Treated Nets for Malaria Control." Bulletin of the World Countries. Ottawa: International Development Research Centre. Health Organization 78 (12): 1389­1401. Jha, P., O. Bangoura, and R. Ranson. 1998."The Cost-Effectiveness of Forty Djimdé, A. A., A. Dolo, A. Quattara, S. Diakité, C. V. Plowe, and O. K. Health Interventions in Guinea." Health Policy and Planning 13 (3): Doumbo. 2004. "Molecular Diagnosis of Resistance to Antimalarial 249­62. Drugs during Epidemics and in War Zones." Journal of Infectious Kain, K. C., G. D. Shanks, and J. S. Keystone. 2001. "Malaria Diseases 190 (4): 853­55. Chemoprophylaxis in the Age of Drug Resistance: I. Currently Ettling, M. B., D. A. McFarland, L. J. Schultz, and L. Chitsulo. 1994. Recommended Drug Regimens." Clinical Infectious Diseases 33 (2): "Economic Impact of Malaria in Malawian Households." Tropical 226­34. Medicine and Parasitology 45: 74­79. Kayentao, K., M. Kodio, R. D. Newman, H. Maiga, D. Doumtabe, A. Ettling, M. B., and D. S. Shepard. 1991. "Economic Cost of Malaria in Ongoiba, and others. 2005. "Comparison of Intermittent Preventive Rwanda." Tropical Medicine and Parasitology 42 (3): 214­18. Treatment with Chemoprophylaxis for the Prevention of Malaria Feacham, R. G. A. 2004."The Research Imperative: Fighting AIDS, TB, and during Pregnancy in Mali." Journal of Infectious Diseases 191 (1): Malaria." Tropical Medicine and International Health 9 (11): 1139­41. 109­16. Fernando, D., R. Wickremasinghe, K. N. Mendis, and A. R. Keiser, J., J. Utzinger, M. Caldas de Castro, T. A. Smith, M. Tanner, and Wickremasinghe. 2003. "Cognitive Performance at School Entry of B. H. Singer. 2004. "Urbanization in Sub-Saharan Africa and Malaria Children Living in Malaria-Endemic Areas of Sri Lanka." Transactions Control." American Journal of Tropical Medicine and Hygiene 71 of the Royal Society of Tropical Medicine and Hygiene 97 (3): 161­65. (Suppl. 2): 118­27. French, N., J. Nakiyingi, E. Lugada, C. Watera, J. A. Whitworth, and C. F. Kidane, G., and R. H. Morrow. 2000. "Teaching Mothers to Provide Home Gilks. 2001. "Increasing Rates of Malarial Fever with Deteriorating Treatment of Malaria in Tigray, Ethiopia: A Randomised Trial." Lancet Immune Status in HIV-1 Infected Ugandan Adults." AIDS 15 (7): 356 (9229): 550­55. 899­906. Kirigia, J. M., R. W. Snow, J. Fox-Rushby, and A. Mills. 1998. "The Gallup, J. L., and J. D. Sachs. 2001. "The Economic Burden of Malaria." Cost of Treating Pediatric Malaria Admissions and the Potential American Journal of Tropical Medicine and Hygiene 64 (Suppl. 1): Impact of Insecticide-Treated Mosquito Nets on Hospital 85­96. Expenditure." Tropical Medicine and International Health 3: Gilles, H. M., ed. 2002. "Historical Outline." In Essential Malariology, 4th 145­50. ed., ed. D. A. Warrell and H. M. Gilles, 1­7. New York: Arnold. Korenromp, E. L., B. G. Williams, E. Gouws, C. Dye, and R. W. Snow. Goodman, C., P. Coleman, and A. Mills. 2000. Economic Analysis of 2003. "Measuring Trends in Childhood Malaria Mortality in Africa: Malaria Control in Sub-Saharan Africa. Geneva: Global Forum for A New Assessment of Progress toward Targets Based on Verbal Health Research. Autopsy." Lancet Infectious Diseases 3: 349­58. Greenberg, A. E. 1992. "HIV and Malaria." In AIDS in the World, ed. J. M. Laxminarayan, R., Z. Bhutta, A. Duse, P. Jenkins, T. O'Brien, I. N. Mann, D. J. Tarantola, and T. W. Netter, 143­48. Cambridge, MA: Okeke, A. Pablo-Mendez, K. P. Klugman. 2006. "Drug Resistance." Harvard University Press. In D. T. Jamison, J. G. Breman, A. Measham, and others, Disease Greenberg, A. E., P. Nguyen-Dinh, J. M. Mann, N. Kabote, R. L. Control Priorities in Developing Countries, eds. 2nd. ed. New York: Colebunders, H. Francis, and others. 1988. "The Association Between Oxford University Press, ch. 55. Malaria, Blood Transfusions, and HIV Seropositivity in a Pediatric Leighton, C., and R. Foster. 1993. "Economic Impacts of Malaria in Kenya Population in Kinshasa, Zaire." Journal of the American Medical and Nigeria." Abt Associates, Health Financing and Sustainability Association 259: 545­49. Project, Bethesda, Maryland. Greenberg, A. E., M. Ntumbanzondo, N. Ntula, L. Mawa, J. Howell, and Lengeler, C. 2004. "Insecticide-Treated Bed Nets and Curtains for F. Davachi. 1989. "Hospital-Based Surveillance of Malaria-Related Preventing Malaria." Cochrane Database Systematic Reviews (2) Paediatric Morbidity and Mortality in Kinshasa, Zaire." Bulletin of the CD000363. World Health Organization 67 (2): 189­96. Lengeler, C., and B. Sharp. 2003. "Indoor Residual Spraying and Greenwood, B. M., K. Bojang, C. J. M. Whitty, and G. A. T. Targett. 2005. Insecticide-Treated Nets." In Reducing Malaria's Burden. Evidence of "Malaria." Lancet 365 (9469): 1487­98. Effectiveness for Decision Makers. Washington DC: Global Health Gülmezoglu, A. M., and P. Garner. 1998. "Malaria in Pregnancy in En- Council, 17­24. http://www.globalhealth.org demic Areas (Cochrane Review)." Cochrane Library 3, Oxford, Update McCarthy, F. D., H. Wolf, and Y. Wu. 2000. "The Growth Costs of Malaria." Software. In NBER Working Paper 7541, National Bureau of Economic Research, Hay, S. I., D. J. Rogers, J. F. Toomer, and R. W. Snow. 2004. "Annual Cambridge, MA. Plasmodium falciparum Entomological Inoculation Rates (EIR) Mendis, K., B. J. Sina, P. Marchesini, and R. Carter. 2001. "The Neglected across Africa: Literature Survey, Internet Access, and Review." Burden of Plasmodium vivax Malaria." American Journal of Tropical Transactions of the Royal Society of Tropical Medicine and Hygiene 94: Medicine and Hygiene 64 (Suppl. 1): 97­105. 113­27. Menendez, C., A. F. Fleming, and P. L. Alonso. 2000. "Malaria-Related Holding, P. A., and P. K. Kitsao-Wekulo. 2004. "Describing the Burden of Anemia." Parasitology Today 16: 469­76. Malaria on Child Development: What Should We Be Measuring and How Should We Be Measuring It?" American Journal of Tropical Mills,A. 1992."The Economic Evaluation of Malaria Control Technologies: Medicine and Hygiene 71 (Suppl. 2): 71­79. The Case of Nepal." Social Science and Medicine 34: 965­72. Hung, I. Q., P. J. Vries, P. T. Giao, N. V. Nam, T. Q. Binh, M. T. Chong, Mills, A., and S. Shillcutt. 2004. "The Challenge of Communicable and others. 2002. "Control of Malaria: A Successful Experience Disease." In Global Crises, Global Solutions, ed. B. Lomborg. from Viet Nam." Bulletin of the World Health Organization 80 (8): Cambridge, U.K.: Cambridge University Press. 660­66. Mulligan, J., C. Morel, and A. Mills. 2005. "Cost-Effectiveness of INDEPTH (International Network of Field Sites with Continuous Malaria Control Interventions." Disease Control Priorities Demographic Evaluation of Populations and Their Health in Project Background Paper, Disease Control Priorities Project, Developing Countries). 2002. Population, Health, and Survival at Bethesda, MD. Conquering Malaria | 429 Mung'Ala-Odera, V., R. W. Snow, and C. R. J. C. Newton. 2004. "The Sharp, B. L., and D. le Sueur. 1996. "Malaria in South Africa--The Past, the Burden of the Neurocognitive Impairment Associated with Present, and Selected Implications for the Future." South African Falciparum Malaria in Sub-Saharan Africa." American Journal of Medical Journal 86 (1): 83­89. Tropical Medicine and Hygiene 71 (Suppl. 2): 64­70. Sirima, S. B., A. Konaté, A. B. Tiono, N. Convelho, S. Cousins, and Murphy, S. C., and J. G. Breman. 2001. "Gaps in the Childhood Malaria F. Pagnoni. 2003. "Early Treatment of Childhood Fevers with Pre- Burden in Africa: Cerebral Malaria, Neurologic Sequelae, Anemia, packaged Antimalarial Drugs in the Home Reduces Severe Malaria Respiratory Distress, Hypoglycemia, and Complications of Morbidity in Burkina Faso." Tropical Medicine and International Health Pregnancy." American Journal of Tropical Medicine and Hygiene 64 8 (2): 1­7. (Suppl. 1): 57­56. Snow, R. W., M. H. Craig, U. Deichmann, and K. Marsh. 1999. "Estimating Murray, C. J. L., and A. D. Lopez. 1996. The Global Burden of Disease: A Mortality, Morbidity, and Disability Due to Malaria among Africa's Comprehensive Assessment of Mortality and Disability from Diseases, Non-pregnant Population." Bulletin of the World Health Organization Injuries, and Risk Factors in 1990 and Projected to 2020. Cambridge, 77: 624­40. MA: Harvard University Press. Snow, R. W., M. H. Craig, C. R. J. C. Newton, and R. W. Steketee. 2003. . 1997. "Mortality by Cause for Eight Regions of the World: "The Public Health Burden of Plasmodium falciparum Malaria in Global Burden of Disease Study." Lancet 349: 1269­76. Africa: Deriving the Numbers." Working Paper 11, Disease Control Mutabingwa, T. K., D. Anthony, A. Heller, R. Hallett, J. Ahmed, C. Priorities Project, Bethesda, MD. Drakeley, and others. 2005. "Amodiaquine Aone, Amodiaquine Snow, R. W., C. A. Guerra, A. M. Noor, H. Y. Myint, and S. I. Hay. 2005. Sulfadoxine-Pyrimethamine, Amodiaquine Artesunate, and "The Global Distribution of Clinical Episodes of Plasmodium falci- Artemether-Lumefantrine for Outpatient Treatment of Malaria in parum Malaria." Nature 434: 214­17. Tanzanian Children: a Four-Armed Randomized Effectiveness Trial." Snow, R. W., E. Korenromp, C. Drakely, and E. Gouws. 2004. "Pediatric Lancet 365 (9469): 1474­80. Mortality in Africa: Plasmodium falciparum Malaria as a Cause or Risk?" Najera, J. A., and M. Zaim. 2002. Malaria Vector Control: Decision Making American Journal of Tropical Medicine and Hygiene 70 (Suppl. 2): Criteria and Procedures for Judicious Use of Insecticides. WHO/ 16­24. CDS/WHOPES/2002.5. Geneva: World Health Organization. Snow, R. W., and K. Marsh. 2002. "The Consequences of Reducing Newman, R. D., M. E. Parise, L. Slutsker, B. Nahlen, and R. W. Steketee. Plasmodium falciparum Transmission in Africa." Advances in 2003. "Safety, Efficacy, and Determinants of Effectiveness of Parasitology 52: 235­64. Antimalarial Drugs during Pregnancy: Implications for Prevention Snow, R. W., A. Menon, and B. M. Greenwood. 1989. "Measuring Programs in Plasmodium falciparum­Endemic Sub-Saharan Africa." Morbidity from Malaria." Annals of Tropical Medicine and Parasitology Tropical Medicine and International Health 6: 488­506. 83: 321­23. Nosten, F., R. McGready, J. A. Simpson, K. L. Thwai, S. Balkan, T. Cho, and others. 1999. "Effects of Plasmodium vivax Malaria in Pregnancy." Snow, R. W., J. F. Trape, and K. Marsh. 2001. "The Past, Present, and Future Lancet 354 (9178): 546­49. of Childhood Malaria Mortality in Africa." Trends in Parasitology 17: 593­97. Pagnoni, F., N. Convelbo, J. Tiendrebeogo, S. Cousens, and F. Esposito. 1997. "A Community-Based Programme to Provide Prompt and Steketee, R. W., B. L. Nahlen, M. E. Parise, and C. Menendez. 2001. "The Adequate Treatment of Presumptive Malaria in Children." Trans- Burden of Malaria in Pregnancy in Malaria-Endemic Areas." American actions of the Royal Society of Tropical Medicine and Hygiene 91 (5): Journal of Tropical Medicine and Hygiene 64 (Suppl.): 28­35. 512­17. Steketee, R. W., J. J. Wirima, A. W. Hightower, L. Slutsker, D. L. Heymann, Piola, P., C. Fagg, F. Bajunirwe, S. Biraro, F. Grandesso, E. Ruzagira, and and J. G. Breman. 1996. "The Effect of Malaria and Malaria Prevention others. 2005. "Supervised Versus Unsupervised Intake of Six-Dose in Pregnancy on Offspring Birthweight, Prematurity, and Intrauterine Artemether-Lumefantrine for Treatment of Acute, Uncomplicated Growth Retardation in Rural Malawi." American Journal of Tropical Plasmodium falciparum Malaria in Mbarara, Uganda: A Randomized Medicine and Hygiene 55 (Suppl.): 33­41. Trial." Lancet 365 (9469): 1467­73. Sudre, P., J. G. Breman, D. McFarland, and J. P. Koplan. 1992. "Treatment RBM (Roll Back Malaria). 2002. Scaling-up Insecticide-Treated Netting of Chloroquine Resistant Malaria in African Children: A Cost- Programmes in Africa: A Strategic Framework for Coordinated National Effectiveness Analysis." International Journal of Epidemiology 21: Action. Geneva: WHO/CDS/RBM/2002. 146­54. Remme, J. H., E. Blas, L. Chitsulo, P. M. Desjeux, H. D. Engers, T. P. Kanyok, Tang, L. 2000. "Progress in Malaria Control in China." Chinese Medical and others. 2002."Strategic Emphasis for Tropical Diseases Research: A Journal 113 (1): 89­92. TDR Perspective." Trends in Parasitology 18 (10): 421­26. Ter Kuile, F. O., M. Parise, F. Verhoef, V. Udhayakumar, R. D. Newman, Rosen, J. B., and J. G. Breman. 2004. "Malaria Intermittent Preventive A. M. Van Eijk, and others. 2004. "The Burden of Co-infection with Treatment in Infants (ITPi), Chemoprophylaxis, and Childhood Human Immunodeficiency Virus Type 1 and Malaria in Pregnant Vaccinations." Lancet 363 (9418): 1386­88. Women in Sub-Saharan Africa." American Journal of Tropical Medicine Russell, S. 2004. "The Economic Burden of Illness for Households in and Hygiene 70 (Suppl. 2): 41­54. Developing Countries: Catastrophic or Manageable? A Review of Ter Kuile, F. O., D. J. Terlouw, S. K. Kariuki, P. A. Phillips-Howard, L. B. Studies Focusing on Malaria, TB, and HIV/AIDS." American Journal of Mirel, W. A. Hawley, and others. 2003. "Impact of Permethrin-Treated Tropical Medicine and Hygiene 71 (Suppl. 2): 147­55. Bednets on Malaria, Anemia, and Growth in Infants in an Area of Savarit, D., K. M. De Cock, R. Shutz, S. Konate, E. Lackritz, and Intense Perennial Malaria Transmission in Western Kenya." American A. Bondurand. 1992. "Risk of HIV Infection from Transfusion with Journal of Tropical Medicine and Hygiene 68 (Suppl. 4): 68­77. Blood Negative for HIV Antibodies in a West African City." British Thera, M. A., P. S. Sehdev, D. Coulibaly, K. Traore, M. N. Garba, and others. Medical Journal 305 (6852): 498­502. 2005. "Impact of Trimethoprim-Sulfamethoxa-Infectious Diseases." Schellenberg, D. S., C. Menendez, J. J. Aponte, E. Kahigwa, M. Tanner, Journal of Infections Diseases 192 (10): 1823­29. H. Mshinda, and P. Alonso. 2005. "Intermittent Preventive Treatment Warhurst, D. C., and J. E. Williams. 2004. "Laboratory Procedures for for Tanzanian Infants: Follow-up to Age 2 Years of a Randomized Diagnosis of Malaria." In Malaria: A Hematological Perspective, ed. Placebo-Controlled Trial." Lancet 365 (9469): 1481­83. S. H. Abdalla and G. Pasvol, 1­27, London: Imperial College Press. 430 | Disease Control Priorities in Developing Countries | Joel G. Breman, Anne Mills, Robert W. Snow, and others White, N. J. 2005. "Intermittent Presumptive Treatment for Malaria." PLoS . 2002c. South African Malaria Control Programme. Roll Back Medicine 2 (1): 63. Malaria in Southern Africa. Baseline Survey 2001. Harare, Zimbabwe: White, N. J., and J. G. Breman. 2005. "Malaria and Babesiosis: Diseases WHO. Caused by Red Blood Cell Parasites." In Harrison's Principles of Internal . 2003a. Access to Antimalarial Medicines: Improving the Medicine, 16th ed., ed. D. Kasper, E. Braunwald, A. S. Fauci, S. L. Affordability and Financing of Artemisinin-Based Combination Hauser, D. L. Longo, and J. L. Jameson, 1218­33. McGraw-Hill: Therapies. Geneva: WHO. New York. . 2003b. The Africa Malaria Report 2003. Geneva: WHO; New Whitworth, J., D. Morgan, M. Quigley, A. Smith, B. Mayania, H. Eotu, and York: United Nations Children's Fund. others. 2000. "Effect of HIV-1 and Increasing Immunosuppression on . 2003c. The Scientific Working Group on Malaria Research. Malaria Parasitaemia and Clinical Episodes in Adults in Rural Uganda: TWR/SWG/03.03. Geneva: WHO. A Cohort Study." Lancet 356 (9235): 1051­56. . 2005. The World Malaria Report 2005: Roll Back Malaria. WHO (World Health Organization). 1996. Investing in Health Research Geneva: WHO. and Development: Report of the Ad Hoc Committee on Health Research World Bank. 2003. 2003 World Development Indicators. Washington, DC: Relating to Future Intervention Options. TDR/Gen/96.1. Geneva: WHO. World Bank. . 2000a. Twentieth Report of the WHO Expert Committee on Malaria. Technical Report Series 892. Geneva: WHO. . 2005. Rolling Back Malaria, Global Strategy and Booster Program. Washington, DC: World Bank. . 2000b. "Severe and Complicated Malaria." Transactions of the Worrall, E., A. Rietveld, and C. Delacollette. 2004. "The Burden of Malaria Royal Society of Tropical Medicine and Hygiene 95 (Suppl.): 51­90. Epidemics and Cost-Effectiveness of Interventions in Epidemic . 2001a. Antimalarial Drug Combination Therapy. Report of Situations in Africa." American Journal of Tropical Medicine and a WHO Technical Consultation 4­5 April, 2001. WHO/CDS/RBM/ Hygiene 71 (Suppl. 2): 136­40. 2001.35. Geneva: WHO. Yeung, S., W. Pongtavornpinyo, I. M. Hastings, A. J. Mills, and N. J. White. . 2001b. "The Use of Antimalarial Drugs." Report of a WHO 2004. "Antimalarial Drug Resistance, Artemisinin-Based Combination Informal Consultation, WHO, Geneva. Therapy (ACT), and the Contribution of Modeling to Elucidating . 2002a. Monitoring Antimalarial Drug Resistance. WHO/ Policy Choices." American Journal of Tropical Medicine and Hygiene 71 CDS/CSR/EPH/2002.1 WHO/CDS/RBM/2002.39. Geneva: WHO. (Suppl. 2): 179­86. . 2002b. World Health Report 2002: Reducing Risks, Promoting Healthy Life. Geneva: WHO. Conquering Malaria | 431 Chapter 22 Tropical Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Onchocerciasis, and Leprosy Jan H. F. Remme, Piet Feenstra, P. R. Lever, André Médici, Carlos Morel, Mounkaila Noma, K. D. Ramaiah, Frank Richards, A. Seketeli, Gabriel Schmunis, W. H. van Brakel, and Anna Vassall Tropical diseases are infectious diseases that are found pre- and geographic distribution differ. Chagas disease is caused by dominantly in the tropics, where ecological and socioeconomic infection with a protozoan, leprosy by a mycobacterium, and conditions facilitate their propagation. Climatic, social, and LF and onchocerciasis by filarial nematodes. Three are vector- economic factors create environmental conditions that facili- borne diseases, but leprosy is transmitted directly from person tate transmission, and the lack of resources prevents affected to person. Chagas disease occurs only in the Americas, populations from obtaining effective prevention and adequate onchocerciasis is found predominantly in Africa, and LF and care. Tropical diseases are diseases of the poor, and investments leprosy occur in all tropical regions. in control and research to develop more effective intervention tools and strategies have been minimal (Gwatkin, Guillot, and Heuveline 1999; Remme and others 2002). For some, however, Chagas Disease effective intervention methods have been developed, and suc- Chagas disease--also known as American trypanosomiasis-- cessful control has been achieved. is a zoonotic disease caused by the protozoan hemoflagellate This chapter focuses on four tropical diseases--Chagas dis- Trypanosoma cruzi that is mainly transmitted by large, blood- ease, lymphatic filariasis (LF), onchocerciasis, and leprosy--for sucking, reduviid bugs of the subfamily Triatominae (known as which effective means of control are available. All four diseases kissing bugs). Infection with this blood parasite has been are targeted for elimination as a public health problem. Control recorded in more than 150 species of 24 families of domestic strategies are being implemented at scale and have already and wild mammals as well as in humans. In the vertebrate host, achieved a major reduction in the burden of disease, and the T. cruzi usually infects macrophage, muscle, and nerve cells. causative agent has even been eliminated in some previously Human infection with T. cruzi most commonly originates endemic areas. Those successes have not come easily, and much through contact of broken skin or mucosa with the excretion of remains to be done to ensure complete and sustained control of infected insect vectors. The incubation period ranges from 7 to the diseases. 15 days, leading to the acute phase of infection--characterized by patent parasitemia--which may last up to four weeks. The DISEASE CHARACTERISTICS AND TRANSMISSION acute phase may be without obvious symptoms. Romaña's sign--that is, uniocular, bipalpebral edema with regional Chagas disease, LF, onchocerciasis, and leprosy are all parasitic lymphadenopathy--is diagnostic of the acute infection but infections, but their causative agents, modes of transmission, occurs in less than 5 percent of infections. 433 If a recent infection is untreated, the individual will remain more rarely, through contaminated food or infection in the infected for life. After an asymptomatic period of 10 years or laboratory (WHO 2002a). more, some 10 to 40 percent of those infected will develop cardiac or digestive complications that are characteristic of Lymphatic Filariasis the chronic stage of the disease. In chagasic myocardiopathy LF is caused by species of nematode parasites--Wuchereria the most common symptoms are dyspnea and arrhythmias. bancrofti, Brugia malayi, and Brugia timori--and is transmitted Electrocardiographic alterations can occur, such as right bun- by mosquitoes (WHO 2002c). The adult filarial parasites live dle branch block, left anterior hemiblock, or both, which may in the lymphatics of humans. After mating, each female worm require a pacemaker implant. Apical aneurisms are also typical produces several thousand offspring, microfilariae, during its of advanced chagasic cardiopathy, which may rupture on exces- lifetime. The microfilariae are found in humans' internal organs sive exercise, leading to sudden death. Chagas disease can also and appear in peripheral blood at times that coincide with the involve intestinal complications characterized by severe dilata- vector's biting activity. The biting mosquito ingests the micro- tions of parts of the digestive tract known as megasyndromes. filariae along with the blood meal, and they develop into Megaesophagus and megacolon are the most common. infective-stage larvae in 10 to 12 days. When an infective mos- Symptoms of megaesophagus are dysphagia and odinophagia quito bites a human, the infective-stage larvae are transmitted and subsequent malnutrition. Chagasic megacolon is charac- to the human host and develop into the adult stage in about terized by constipation and meteorism. As a result of colon one year. The adult parasites live 5 to 10 years, of which the distension and contractions, abdominal pain is frequent, and fecund life span is 4 to 6 years. Several hundreds to thousands fecalomas are a complication. of infective mosquito bites are necessary to establish infection. More than 120 species of Triatominae and three transmis- Of the three parasite species, W. bancrofti accounts for nearly sion cycles are recognized. The domestic cycle, responsible for 90 percent of LF infections worldwide. B. malayi is prevalent maintaining infection in humans, occurs mostly in rural or only in some parts of South and Southeast Asia, and B. timori is periurban areas where houses have adobe walls and thatched found only in Indonesia. Several species of Culex, Anopheles, roofs. Humans, dogs, cats, and in some countries guinea pigs Aedes, and Mansonia mosquitoes are involved in the transmis- are the main parasite reservoirs in this cycle. The vector lives sion of LF. C. quinquefasciatus is the major vector in Africa, and multiplies in cracks in the walls, holes in the roof, under Asia, and South America and transmits nocturnally periodic and behind furniture and pictures, and so on. The sylvatic cycle W. bancrofti. Among anophelines, An. gambiae and An. funestus involves sylvatic triatomine bugs that become infected and in play a significant role in Africa. Several Aedes species, particu- turn infect rodents, marsupials, and other wild animals. The larly Ae. polynesiensis, are the major vectors in the South Pacific third is the peridomestic cycle in which mammals participate islands, where diurnally subperiodic W. bancrofti is common. B. (domestic rodents, marsupials, livestock, cats, dogs) by moving malayi is primarily transmitted by Mansonia and Anopheles freely in and out of human dwellings, and sylvatic bugs are species. attracted to lights in houses and to food. This peridomestic Infected people can harbor microfilaremia without overt cycle acts as a link between the domestic and sylvatic cycles. clinical manifestations. The disease process is determined pri- Occasionally, infected sylvatic species of Triatominae fly into marily by living adult worms, inflammatory responses caused houses and contribute to transmission either by feeding and by the death of adult worms, and secondary bacterial infec- defecating on the people or their domestic animals or (indi- tions. The inflammatory response begins with the death of or rectly) by contaminating food and drink in which the parasites damage to adult worms, which leads to host reaction and acute can survive. In the Amazon region, cases of acute Chagas dis- filarial lymphangitis. A heavy worm burden and the presence ease have been associated with sylvatic Triatominae contami- of worms in the scrotal area precipitate the development of nating sugarcane or fruit juice. hydrocele, chyluria, chylocele, and lymph scrotum. Lymphatic Transmission by blood transfusion is the second-most com- dysfunction caused by dilatation of the lymphatic vessels mon way of acquiring T. cruzi infection. The true incidence of makes the patient more prone to repeated secondary bacterial infection through blood transfusion is unknown, because most infection, which precipitates lymphedema and elephantiasis. cases are not recognized. In transfusionally acquired T. cruzi Microfilariae play an important role in the pathogenesis of infection, the incubation period is 30 or more days, and the tropical pulmonary eosinophilia (Dreyer and others 2000). most common symptoms are fever, general lymph node enlargement, and splenomegaly (Schmunis and others 2001). Transplacental transmission of T. cruzi can occur, and esti- Onchocerciasis mates indicate that 5 percent of newborns born to chagasic Onchocerciasis is an infection with the filarial parasite mothers will become infected. Less common routes of trans- Onchocerca volvulus. The main complications are severe eye dis- mission are by transplantation with an infected organ or, ease that can lead to blindness and severe skin disease with 434 | Disease Control Priorities in Developing Countries | Jan H. F. Remme, Piet Feenstra, P. R. Lever, and others unsightly lesions and intense itching (WHO 1995a). O. volvulus intracellular parasite that resides predominantly in is transmitted by vector blackflies of the genus Simulium, whose macrophages. It is the only bacterium that infects peripheral larvae and pupae develop in rapidly flowing, well-oxygenated nerves, showing a preference for Schwann cells, particularly of streams and rivers. As a result, onchocerciasis is often known as unmyelinated fibers. river blindness. The most important vectors are members of the The disease spectrum of leprosy ranges from a single self- S. damnosum complex inAfrica and the Middle East and S. neavei healing, hypopigmented macule to a generalized illness causing in parts of East Africa. Of the many vectors in the Americas, the widespread peripheral nerve damage and affecting even bones most important are S. ochraceum, S. metallicum, S. oyapockense, and internal organs. Skin lesions may be well- or ill-defined S. guianense, and S. exiguum. hypopigmented macules, plaques, or nodules that are localized When taking a blood meal,infected Simulium vectors deposit or distributed over the whole skin. They may be hypaesthetic, one or more infective (third-stage) O. volvulus larvae, which anesthetic, hyperaesthetic, or have normal sensibility. Nerve reach adulthood in the human host after about a year but may lesions occur in dermal nerves as well as in superficial sensory live as long as 14 years. The adult worms typically entwine in nerves and mixed nerve trunks. One or more nerves may be nodules where they mate, producing microfilariae that migrate enlarged on palpation. Signs such as clawing of fingers and into the skin, eyes, and other organs. These microfilariae are toes, "absorption" of digits caused by repeated injury, and dry unable to develop into adult worms without first being ingested skin are secondary to impairment of motor, sensory, and auto- in the blood meal of a blackfly vector. The microfilariae trans- nomic nerve function. form in the vector over a period of 6 to 12 days to produce the A diagnosis of leprosy is based on finding at least one of third-stage larvae that are infective to humans. three so-called cardinal signs (ILA 2002): The thousands of microfilariae that do not succeed in reach- ing a blackfly vector die in the human body, provoking inflam- · diminished sensibility in a typical macule or plaque in the matory reactions in tissues. Inflammation in the eyes leads to skin irreversible ocular lesions, resulting first in impaired vision and · palpable enlargement of one or more peripheral nerve finally in total blindness (WHO 1995a). The death of microfi- trunks at specific sites lariae in the skin gives rise to intense itching, dermatitis, depig- · demonstration of acid-fast mycobacteria in a slit skin smear. mentation, and atrophy of the skin (Murdoch and others 2002). A less common complication is lymphadenitis, which Currently, patients are classified based on clinical signs only, may lead to hanging groin and elephantiasis of the genitals, and but skin smear results are taken into account when available. increasing evidence indicates that onchocerciasis is a risk factor Patients who have more than five skin lesions or who have a for epilepsy and hyposexual dwarfism in certain areas positive skin smear are classified as multibacillary; others are (Boussinesq and others 2002). The greater is the body load of classified as paucibacillary. adult worms and microfilariae, the greater is the risk of devel- The skin signs of leprosy are relatively harmless, but com- oping skin and eye disease. plications of the disease may lead to severe consequences, such The disease pattern of onchocerciasis--in particular the as blindness, infertility, disfigurement, and severe sensory and severity of ocular disease--varies considerably between geo- motor disability. Reactions--that is, episodes of acute inflam- graphic zones. Onchocercal blindness can be extensive in hyper- mation caused by hypersensitivity to bacterial antigens--can endemic communities of the West African savannas, whereas in be particularly severe. Patients can develop nerve damage with- forest villages with a comparable intensity of infection, the skin out any obvious sign of these reactions, but after neuropathy manifestations tend to be the main complications of the disease has become irreversible, it may lead to secondary impairments, (Dadzie and others 1989; Murdoch and others 2002). These dif- such as wounds, contractures, and shortening of digits. As a ferences may reflect the existence of different vector-parasite result of visible impairments or activity limitations--or simply complexes, with strains of O. volvulus that differ in pathogenic- because of the diagnosis of leprosy--many people experience ity (Zimmerman and others 1992). The vector-parasite com- psychosocial problems (van Brakel 2000). plex in the West African savanna is responsible for the most The exact mode of transmission of M. leprae is still not fully severe form of ocular onchocerciasis in the world: in the most understood, but the respiratory tract seems to play an impor- affected villages, more than 10 percent of the population may be tant role. The primary reservoir of infection is the human host. blind because of onchocerciasis. Untreated multibacillary leprosy patients are able to shed large amounts of M. leprae from the nose, and household and social contacts of such patients are at a higher risk of developing Leprosy leprosy than the general population (van Beers, Hatta, and Leprosy is caused by Mycobacterium leprae, a gram-positive, Klatser 1999). M. leprae­specific DNA sequences have been iso- strongly acid-fast bacterium. M. leprae is an obligate, lated from the noses of apparently healthy individuals, and Tropical Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Onchocerciasis, and Leprosy | 435 widespread seropositivity against M. leprae­specific antigens year per patient, and disability awards, which in one state has been demonstrated in endemic areas, although the role of accounted for US$399,600 (Dias 1987; Schofield and Dias these individuals in transmitting leprosy is not fully under- 1991). In Bolivia, in 1992, aggregate treatment costs were esti- stood. Effective antileprosy treatment usually renders a patient mated at US$21 million. In Chile, in 1997, aggregate treatment noninfectious within a few days. costs for Chagas disease were estimated at US$14 million to US$19 million (Schenone 1998), and in Uruguay, in 1996, costs were estimated at US$15 million (Salvatella and Vignolo 1996). DISEASE BURDEN Information on the number of people infected is often difficult Lymphatic Filariasis to obtain for tropical diseases. Many infected people may be LF is endemic in 83 countries, with 1.1 billion people living in without obvious symptoms, those with symptoms may not known endemic areas. In 1992, the WHO Expert Committee seek care at public health facilities, and those who do may not estimated that 78 million people were infected (WHO 2002c). be reported. Routine health information systems provide little This estimate was later revised to 119 million, and current esti- information on the number of people infected in the popula- mates indicate that LF is responsible for the loss of 4.6 million tion. Surveys are more informative but are rarely done. A bet- DALYs per year. Many endemic areas lack reliable data on the ter picture emerges only when control programs need to map prevalence of LF, and estimates of the number infected may the distribution of the disease as a basis for targeting large-scale increase when more precise data become available from epi- interventions. Hence, the apparent paradox is that intensifica- demiological mapping. Nationwide mapping in four neighbor- tion of disease control may result in a significant initial increase ing countries in West Africa showed that LF was endemic in a in estimates of the burden of disease through better epidemio- much wider area than expected, and the findings resulted in a logical data. dramatic increase in the estimated number infected (Gyapong and others 2002). Chagas Disease Epidemiological trends have varied widely among different Chagas disease is an important public health problem in 17 regions in recent decades. LF was controlled or eliminated from countries in Latin America. Estimates from the 1980s indi- several islands in the Pacific, and China has seen a dramatic re- cated that some 16 million to 18 million individuals were infect- duction in infection levels. Unfortunately, in India and Africa, ed (WHO 1991), and in the 1990s, a series of multinational the most endemic areas of the world, recent decades have control initiatives was launched that was designed to interrupt witnessed little change (WHO 2002c). transmission by eliminating domestic insect vectors The acute form of the disease is common and causes severe and improving the serological screening of blood donors. As a hardship in endemic communities. Infected individuals suffer result, estimates of the number of infected people were revised from one to eight acute episodes per year, and during each to 9.8 million in 2001 (Schmunis 2000). The estimated burden episode, affected patients are bedridden for three to five days. of disease in terms of disability-adjusted life years (DALYs) Morbidity caused by chronic LF is mostly lifelong, and the declined from 2.7 million in 1990 (World Bank 1993) to 586,000 disease is considered the second leading cause of disability in in 2001 (Mathers, Murray, and Lopez 2006). Because of the world (WHO 1995b). Patients affected by elephantiasis or migration, T. cruzi­infected individuals can be found outside hydrocele are often victims of societal discrimination, and the Latin America (for example, in Spain or the United States). disease impairs their educational and employment opportuni- Estimates from the 1980s suggested that 5 million people in ties, marriage prospects, and sexual life. Case-control studies in the Americas had symptoms of Chagas disease (WHO 1991). India revealed that affected individuals are 27 percent less pro- These estimates decreased to 1.2 million to 2.8 million in the ductive than their uninfected counterparts (Ramu and others 1990s. The World Health Organization (WHO) attributed 1996). The patients work less and often switch to lighter jobs, 45,000 yearly deaths to Chagas disease (WHO 1991). WHO leading to a loss of more than 1 billion person-days per year in decreased its mortality estimates to 13,000 in 2001 (WHO India alone (Ramaiah and others 2000), which translates into 2002d). an annual economic loss equivalent to 0.63 percent of gross In all affected countries, Chagas disease has been responsi- national product. ble for a high burden of disease and significant direct and indi- rect costs. Reports from Brazil in the late 1980s suggested that the aggregate costs for pacemakers and intestinal surgeries Onchocerciasis for Chagas disease were US$250 million per year, excluding More than 99 percent of those infected with O. volvulus reside the costs of consultations, care, and supportive treatment for in 30 endemic countries in Africa, with the remainder living chronic chagasic patients, which amounted to US$1,000 per in the Republic of Yemen and six countries of the Americas. 436 | Disease Control Priorities in Developing Countries | Jan H. F. Remme, Piet Feenstra, P. R. Lever, and others In 1995, the WHO Expert Committee on Onchocerciasis treatment duration for multibacillary patients and the cleaning estimated that 17.7 million people were infected, of whom up of patient registers. about 270,000 were blind and another 500,000 were severely Leprosy is reported from all regions of the world, but the visually impaired (WHO 1995a). However, more recent burden of disease, which is estimated at 192,000 DALYs, is con- information from rapid epidemiological mapping of centrated in a few countries. During 2003, 513,798 new cases onchocerciasis (Noma and others 2002) by the African were detected, of which more than 80 percent were in Brazil, Programme for Onchocerciasis Control (APOC) indicates India, Madagascar, Mozambique, Nepal, and Tanzania (WHO that the number of those infected is twice as high and that 2004a). India alone accounted for about 75 percent of the new some 37 million people were infected in 1995. This revised cases. Case detection has remained remarkably stable over the estimate corresponds to an estimated 1.99 million DALYs lost past decade. Trends in case detection rates should be analyzed because of onchocerciasis in 1995. in conjunction with the proportion of new patients with grade Using the most recent rapid epidemiological mapping data 2 impairment (an indicator of the delay between onset of the and the latest APOC data on treatment coverage and assum- disease and diagnosis) and the proportion of children among ing that four rounds of ivermectin treatment will reduce the new cases (an indicator of recent transmission). prevalence of troublesome itching by 85 percent and the bur- Virtually all published data on leprosy-related disability den of visual impairment and blindness by 35 percent give a concern impairments. In 1997, WHO estimated the global DALY estimate of 1.49 million DALYS lost for 2003 (see prevalence of patients with visible impairments (disability table 22.1). grade 2) as 2 million. A similar number may have sensory In addition to the burden of blindness and severe itching, impairment without deformity. Sensory and motor impair- onchocerciasis has important socioeconomic consequences. In ment that are already present at diagnosis are important risk the West African savanna, fear of blindness has resulted in the factors for developing additional impairment and disability. depopulation of fertile river valleys, severely affecting agricul- Evidence indicates that sensory impairment itself causes signif- tural production. It was this socioeconomic impact, and not icant functional disability. just the health impact, that led to the creation of the The prevalence of activity limitations among people Onchocerciasis Control Program (OCP) in West Africa in 1975 affected by leprosy is unknown. Van Brakel and Anderson's (Remme 2004b). (1998) survey in Nepal finds that among those with any Even though the importance of onchocercal blindness has impairment, about 20 percent had limitations in relation to long been recognized, only in 1995 did research demonstrate one or more indoor activities and up to 34 percent had signifi- that the public health importance of onchocercal skin disease cant limitations in relation to common outdoor activities. Even was even greater. Troublesome itching associated with dermal less is known about the prevalence of restrictions on social and onchocerciasis makes working, studying, or interacting socially economic participation. Surveys are urgently needed to assess difficult (Murdoch and others 2002; Vlassoff and others 2000). the extent of patients with leprosy-related disabilities who Onchocercal itching now accounts for 60 percent of DALYs lost require intervention. (Remme 2004a). Other skin manifestations, such as reactive Two difficulties affect the validity of DALY estimates for lep- skin lesions, are not included in the DALY estimates, even rosy. The first is the lack of data, particularly on the burden of though they are highly prevalent and have major psychosocial functional and psychosocial disability caused by leprosy. The and economic impacts. Onchocercal skin disease also dimin- second is that the effect of leprosy often goes well beyond the ishes people's income-generating capacity, and the school affected individual; the psychosocial consequences may affect dropout rate is twice as high among children from households the whole family. People without any visible signs of leprosy in which the head of household is affected by onchocercal skin may be stigmatized simply because they are known to be a lep- disease (Benton 1998). rosy patient. Even after completing treatment, people may remain stigmatized. Leprosy In May 2001, 10 years after the World Health Assembly had Summary of DALY Estimates adopted a resolution to eliminate leprosy by the end of the Table 22.1 summarizes the DALY estimates for each of the four millennium, the target--a prevalence rate of less than 1 per diseases by World Bank region. The high estimate for LF 10,000--had been achieved at the global level. The number of reflects not only its wider distribution and the larger number cases registered for treatment worldwide fell from 5.4 million of people affected, but also the reduction in the burden for the in 1985 to 460,000 by the end of 2003 (WHO 2004a); however, other three diseases as a result of control efforts. For those dis- this trend should not be taken at face value because the reduc- eases for which there has been significant progress toward tion is attributable mainly to such factors as the shortening of elimination, public health officials should remain aware of the Tropical Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Onchocerciasis, and Leprosy | 437 Table 22.1 DALYs Lost, by Disease and World Bank Region (thousands) Disease (date East Asia and Europe and Latin America Middle East and Sub-Saharan High-income of information) the Pacific Central Asia and the Caribbean North Africa South Asia Africa countries Total Chagas disease 0 1 583 0 0 0 1 585 (2001) LF (2001) 373 1 9 4 2,412 1,656 212 4,667 Onchocerciasis 0 0 2 0.4 0 481 0 484 (2003) Onchocerciasis 0 0 2 0.4 0 1,487 0 1,490 (latest APOC data) Leprosy (2001) 34 0 18 2 113 24 1 192 Source: Mathers, Lopez, and Murray 2006; WHO 2004b; authors' calculations. burden of disease that is currently averted but that might Pact regions, targeted primarily against Rhodnius prolixus, fol- return if control were not to be sustained before transmission lowed in 1997. has been completely eliminated. The results in the Southern Cone region have been impres- sive, with vast areas now free of domestic infestation with T. infestans and other vector species. In Argentina, seropreva- INTERVENTIONS AND THEIR EFFECTIVENESS lence rates among men age 18 to 20 drafted for military service decreased from 5.8 percent in 1981 to 1.2 percent in 1993. The For each of the four diseases in this chapter, effective interven- number of cases of Chagasic cardiomyopathy, when compared tions are available. with the number expected in the absence of control, indicates a decrease of 81 percent in the population up to 18 years of age. In 2001, a WHO commission certified that 4 of the 18 endemic Chagas Disease provinces were free of vectorial transmission. In Brazil, The primary approaches to control of Chagas disease are halt- domestic infestation rates decreased by 98.3 percent between ing transmission and providing adequate treatment for those 1991 and 2000. Of the 11 Brazilian states that were originally infected. The two most important routes of transmission are endemic for T. infestans, 9 have been certified as free of vecto- insect vectors and blood transfusion from infected donors; rial transmission. In Chile, house infestation rates decreased thus, control programs focus on eliminating domestic vector from 28.80 percent in 1982 to less than 0.01 percent in 1999, populations and improving the serological screening of blood when the country was certified free of vectorial transmis- donors. sion. Uruguay also achieved a dramatic reduction in house infestation rates, from 5.7 percent in 1983 to 0.3 percent in Vector Control. Triatoma infestans lives only inside houses 1997, when it too was certified as free of vectorial transmission. and in the peridomestic area. Work during the late 1940s sug- Bolivia and Paraguay have not yet eliminated transmission, but gested that spraying houses with residual insecticides could thousands of houses have been sprayed since 1991. eliminate vectors' domestic populations. The effect and sus- tainability of such vector control programs can be enhanced Blood Transfusion Control. The purpose of screening for when they are combined with improved housing and when T. cruzi in blood banks is to eliminate all units of potentially communities are well informed and closely involved in vector infected blood. Argentina and Brazil require screening to be surveillance activities. done using two serological tests to reduce the risk of false neg- Argentina and Brazil initiated programs for nationwide atives; however, the cost-benefit ratio of the two-test approach vector control in the 1960s, and Chile and Uruguay did so in may be questionable in countries where prevalence is low and the 1970s. These programs were strengthened in 1991 by the the reagents used for diagnosis are highly sensitive. Southern Cone Initiative, a multinational effort to eliminate In 1993, the national coverage of blood donor screening infestation by T. infestans launched by the ministries of health was analyzed in four Central American and six South American of Argentina, Bolivia, Brazil, Chile, Paraguay, and Uruguay and countries (Schmunis and others 1998). At that time, only coordinated by the WHO Regional Office for the Americas. Honduras, Uruguay, and República Bolivariana de Venezuela Similar regional initiatives for Central America and the Andean screened 100 percent of donors, and even in those countries 438 | Disease Control Priorities in Developing Countries | Jan H. F. Remme, Piet Feenstra, P. R. Lever, and others infected transfusions were possible because of the lack of among patients and their communities are the principal strate- sensitivity of the reagents used. Since then, the sensitivity and gic approaches (Dreyer, Dreyer, and Noroes 2002). specificity of serological tests have improved, and more coun- tries have passed legislation requiring the screening of all blood Mass Treatment. It is not yet known how many years of MDA donors. By 2001, seven endemic countries were screening are needed to eliminate LF transmission, but empirical evi- 100 percent of blood donors for T. cruzi, four were screening dence on the effect of MDA on transmission is progressively more than 99 percent of donors, and two were screening about becoming available. In Anopheles-transmitted W. bancrofti in 90 percent; but four countries were still screening fewer than Papua New Guinea, four rounds of MDA with diethylcarba- 25 percent of donors. In countries with a high number of mazine or diethylcarbamazine plus ivermectin that reached immigrants from Latin America, such as Spain and the United about 88 percent of the target population reduced the annual States, thousands of individuals are potentially infected, and transmission potential (the estimated number of infective- screening of blood donors for T. cruzi infection may be indi- stage larvae inoculated per person per year) by 97 percent and cated in these countries. 84 percent in low- and high-transmission areas, respectively. In India, where W. bancrofti is transmitted by C. quinquefasciatus, Treatment. If untreated, most individuals infected with T. six rounds of MDA that reached 54 to 75 percent of the target cruzi will remain infected for life. Spontaneous cure is rare. population reduced the annual transmission potential by 95 Only two drugs, nifurtimox and benznidazole, are effective for and 80 percent in diethylcarbamazine- and ivermectin-treated treating T. cruzi. Both are highly effective for acute infections villages, respectively (Ramaiah and others 2003). Modeling and can be used in cases of congenital Chagas disease. Their studies indicate that the required duration of treatment will effectiveness for treating chronic cases remains unclear, but depend largely on the treatment coverage achieved and the increasing evidence indicates that they are effective in clearing extent of systematic noncompliance to treatment--that is, parasitemia when administered to young cases, which may noncompliance by the same individuals during successive impede the development of chronic lesions. Both drugs may treatment rounds (Stolk and others 2003). The physiology of cause serious side effects and should be administered under the vectors also plays a role, because with Culex-transmitted LF, medical supervision. the critical microfilariae density required to interrupt trans- mission is thought to be lower than in areas where Anopheles is the vector. Lymphatic Filariasis The addition of albendazole to the two established anti- In recent years, new control tools and strategies have become filarial drugs--diethylcarbamazine and ivermectin--is based available for LF (Ottesen and others 1997), and the World on clinical trials indicating that the combination therapy is as Health Assembly has adopted a resolution on the global elimi- good as or better than single-drug therapy and that albendazole nation of LF. The Global Programme for the Elimination of may enhance the macrofilaricidal action of diethylcarbamazine. Lymphatic Filariasis was launched in 2000 with the primary Albendazole is also effective and safe against intestinal helminth goals of interrupting transmission and preventing suffering infections, and its inclusion may enhance compliance with and disability caused by the disease (Ottesen 2000). MDA. However, clinical trials have not yet been conclusive, and The core strategy for interrupting transmission is annual more robust evidence on the advantages of combination thera- mass drug administration (MDA) to treat the entire at-risk py is needed (Addiss and others 2004; Gyapong and others population for a period long enough to ensure that levels of 2005). Community trials are ongoing in India and Africa, and blood microfilariae remain below those necessary to sustain preliminary results of a trial in south India suggest that the transmission. Two annual, single-dose, two-drug regimens are combination of diethylcarbamazine and albendazole may recommended for MDA: ivermectin plus albendazole in indeed achieve greater reduction in the prevalence of antigene- African countries that are coendemic for onchocerciasis, and mia than diethylcarbamazine alone (Rajendran and others diethylcarbamazine plus albendazole for all other endemic 2002). A study in Nigeria showed that the addition of albenda- countries. Where feasible, diethylcarbamazine-fortified salt as zole to ivermectin had an additive effect on reducing LF mos- the only source of domestic salt for a period of at least six quito infection rates. (Richards and others 2005). months would be an alternative strategy to MDA. The principal strategy for alleviating suffering and decreas- Vector Control. Vector control has sometimes been extremely ing the disability caused by LF focuses on decreasing secondary effective against LF. In the Solomon Islands, 9 to 10 years of bacterial and fungal infection of limbs or genitals whose lym- vector control virtually eliminated LF. In India, five years of phatic function has already been compromised by filarial infec- integrated vector control in an urban area reduced the overall tion. Operationally, a regimen of meticulous local hygiene of prevalence of microfilariae by 28 percent and the prevalence in affected areas and the creation of hope and understanding children by 92 percent. Studies suggest that 11 to 12 years of Tropical Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Onchocerciasis, and Leprosy | 439 effective vector control may eliminate LF (Ramaiah, Das, and interrupt transmission and eliminate the parasite reservoir. Dhanda 1994). Vector control combined with chemotherapy Despite initial problems with reinvasion by infective flies, the produced the best results. The introduction of polystyrene strategy proved effective, eliminating onchocerciasis as a public beads in vector breeding habitats and treatment with diethyl- health problem throughout the OCP area. The OCP was suc- carbamazine reduced the annual infective biting rate in cessfully concluded in 2002, but concerns remain about the Tanzania by 99.7 percent (Maxwell and others 1990). In India, possible recrudescence of onchocerciasis through reinvasion by vector control combined with single-dose treatment with infected blackflies or migration of infected persons into OCP diethylcarbamazine plus ivermectin reduced the annual trans- areas. The OCP countries, therefore, need to maintain effective mission potential by 96 percent, compared with 60 percent surveillance to identify any recurrences of infection (Richards using chemotherapy alone (Reuben and others 2001). and others 2001). Such results, along with the limitations of MDA for completely eliminating microfilariae in some situations, have Treatment with Ivermectin. In 1987, Merck & Co., the reactivated the debate on the role of vector control in LF elim- manufacturer of ivermectin, agreed to donate the drug for ination (Burkot and Ichimori 2002). However, few endemic onchocerciasis control for as long as needed (Peters and countries have an adequate vector control infrastructure. Phillips 2004). Clinical and community trials involving more In some African countries, the same vector species transmit than 70,000 people showed that annual ivermectin treatment both LF and malaria. In such situations, the effect of malaria was safe, prevented ocular and dermal morbidity, and signifi- control measures, particularly insecticide-treated bednets, on cantly reduced transmission; however, ivermectin is a microfi- LF vector densities and transmission needs further evaluation. laricide and does not kill the adult worms, and long-term A review of the role and feasibility of community-based vector treatment is needed to sustain suppression of the microfilarial control strategies and large-scale application of biological con- load (Remme 2004b). Additional research is needed to deter- trol agents is also needed. mine the extent to which repeated treatments reduce the reproductive capacity of the adult worm population over time. Morbidity Management. The second objective of the Global The introduction of ivermectin allowed the OCP to achieve Programme for the Elimination of Lymphatic Filariasis is to its objective in 12 years instead of 14 by combining vector con- decrease the disability caused by LF. Simple and cheap methods trol with ivermectin treatment, but most important, it also pro- have been developed for managing lymphedema, using water vided an opportunity to control onchocerciasis in endemic and soap occasionally supplemented with antibiotics. Studies areas outside the OCP where vector control was not feasible. in India, Africa, and the Americas have shown that such This ability led to the creation of two other regional programs methods can significantly improve the quality of life of those for controlling onchocerciasis in endemic areas of Africa and affected, but implementation of this strategy has greatly lagged the Americas: APOC (Remme 1995) and the Onchocerciasis behind the MDA campaigns. Elimination Program for the Americas (OEPA) (Richards and others 2001). Onchocerciasis The World Bank and WHO launched APOC in 1995 to Onchocerciasis control is based on vector control and large- serve 19 onchocerciasis-endemic countries outside the OCP. scale ivermectin treatment. APOC's principal strategy is to establish annual ivermectin distribution in highly endemic areas to prevent eye and skin Vector Control. Vector control used to be the only feasible morbidity. In partnership with ministries of health and non- intervention when available drugs were too toxic for large- governmental organizations, APOC currently provides more scale use. Following success with vector control in Kenya, where than 35 million ivermectin treatments per year and aims to the application of larvicides resulted in local elimination of reach 65 million treatments per year before its scheduled the vector S. neavei, and in selected locations in West Africa, termination in 2010. APOC uses an approach referred to as where the application of larvicides effectively stopped local vec- community-directed treatment with ivermectin, whereby local tor breeding but could not prevent reinvasion of infective communities rather than health services direct the treatment vectors from elsewhere, vector control was considered feasible process (Amazigo and others 2002). A community decides col- in the West African savanna if carried out on a large scale. In lectively whether it wants ivermectin treatment, how it will col- 1975, the OCP started large-scale vector control operations lect ivermectin tablets from the medical supply entity, when using helicopters for weekly spraying of larvicides over the vec- and how the tablets will be distributed, who will be responsible tor breeding sites in river rapids (Molyneux 1995). The opera- for distribution and recordkeeping, and how the community tion ultimately covered some 50,000 kilometers of rivers over a will monitor the process. Health workers provide only the nec- geographic area of 1,235,000 square kilometers. The OCP's essary training and supervision. To date, communities have strategy was to maintain vector control for at least 14 years to responded enthusiastically to this approach (Seketeli and 440 | Disease Control Priorities in Developing Countries | Jan H. F. Remme, Piet Feenstra, P. R. Lever, and others others 2002), and interest is now growing in exploring this of rifampicin in preventing leprosy, with preliminary results strategy for interventions against other diseases (Homeida and from the Indonesian trial indicating a significant protective others 2002). effect. In the Americas, O. volvulus transmission occurs only in a few small areas in six endemic countries. Accordingly, OEPA's Prevention of Disabilities. As concerns primary prevention, strategy is based on intense ivermectin treatment twice a year leprosy-related disability is preventable, but when peripheral that should allow eventual cessation of ivermectin delivery neuropathy has become established, it is irreversible and leads without the risk of recrudescence (Richards and others 2001). to lifelong morbidity and disability (Bekri and others 1998; OEPA was launched in 1992 and is currently reaching more Meima and others 1999). Early case detection and treatment than 85 percent of its intended target population. are therefore likely to be the most effective interventions in relation to preventing disability. When detected and treated Leprosy in time with corticosteroids, primary impairments may be The objectives of leprosy control are to interrupt transmission, reversible, but because many patients present late, some 11 to to cure patients, to prevent the development of associated 51 percent do not recover or get worse. deformities, and to rehabilitate those patients already afflicted In relation to secondary prevention, the main strategy is with deformities. The strategy involves early case detection self-care to prevent worsening of impairments in people who and the provision of adequate chemotherapy and comprehen- already have irreversible neural impairment or secondary sive patient care (ILA 2002). impairments such as wounds and contractures. The role of health care workers is to educate patients so that they can be Multidrug Therapy. Dapsone therapy for leprosy was intro- responsible for the daily management of the effects of nerve duced in the late 1940s and successfully used as monotherapy function impairment. An essential part of secondary preven- for two decades. In the 1970s, resistance to dapsone emerged, tion is the use of protective footwear by people with anesthetic and WHO introduced multidrug therapy (MDT) in 1982. feet. Several studies have shown that the use of locally Paucibacillary patients were to be given a six-month regimen acceptable, appropriate footwear is a cost-effective interven- of daily dapsone and supervised monthly rifampicin. tion for those with a loss of plantar sensation (ILA 2002). Multibacillary patients were to be treated with a three-drug Reconstructive surgery, protective footwear for people with regimen for two years or, where feasible, until the skin smear insensitive feet, and assistive devices to correct or prevent activ- had become negative. This regimen followed the paucibacillary ity limitations are also used in secondary prevention. regimen, adding a smaller daily dose of clofazimine and a larger As concerns tertiary prevention, the stigma attached to supervised dose once a month. leprosy often prevents patients from participating in normal These regimens have had good results, with a relapse community activities. Strategies include counseling of those incidence of less than 0.1 percent per year (ILA 2002). No affected and their families, neighbors, and communities; voca- multidrug-resistant leprosy has been reported so far, and tional training; and advocacy work. reports of rifampicin-resistant M. leprae have been few. In 1998, the standard multibacillary MDT regimen was shortened Rehabilitation. Impairments often lead to activity limitations to 12 months. Long-term relapse rates for the 12-month regi- and restrictions on social participation, which can be prevented men are not yet available. by correcting the underlying impairment if it is not yet irre- Public health specialists expected that wide application of versible. After impairment is established, activity limitations MDT together with earlier diagnosis resulting from the up- can still be minimized with the help of reconstructive surgery grading of leprosy services would have a considerable effect on or appropriate assistive devices, such as orthoses, grip aids, transmission; however, by 2002, clear evidence of a reduction calipers, or prostheses. A large but unknown percentage of peo- in transmission had not been seen. ple succeed in overcoming activity limitations by themselves. Some require rehabilitation interventions, such as physical or Immunoprophylaxis and Chemoprophylaxis. Several ran- occupational therapy, reconstructive surgery, or temporary domized trials have shown that vaccination with the bacillus socioeconomic assistance. Calmette-Guérin (BCG) vaccine reduces the risk of develop- ing leprosy (Fine and Smith 1996), with the level of protection Intervention Effectiveness varying from 20 to 80 percent. Chemoprophylaxis based on For all four diseases covered in this chapter, interventions are dapsone or intramuscular acedapsone conferred overall pro- available that are effective under routine control conditions. tection against leprosy of about 60 percent (Smith and Smith The feasibility of eliminating these diseases as public health 2000). Two large trials are currently under way in Bangladesh problems from most endemic areas is therefore not in doubt; and Indonesia to investigate the efficacy of one or two doses however, questions remain about the effectiveness and Tropical Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Onchocerciasis, and Leprosy | 441 sustainability of control under specific conditions and about would have cost US$98 million. Diagnosis of megasyndromes the feasibility of eliminating the parasites and transmission. for 6 percent of infected individuals at an average cost of The vector control strategy for Chagas disease worked well in US$141 each would add US$16.9 million, and corrective sur- the Southern Cone countries, but the sylvatic reservoir of T. gery for 3 percent of the latter would add US$60 million. cruzi remains unaffected, and continued surveillance will be Cardiac pacemakers for 5 percent of infected individuals at essential. Vector control is more challenging in the Andean and US$3,000 each would add another US$30 million, so that the Central American countries, where some of the vectors are not partial direct costs for medical attention in Brazil in 1995­96 domiciliated. For LF, the number of years of MDA and the would have been US$205 million. treatment coverage required to interrupt transmission remain In Argentina, the costs for medical attention in 1992 were unknown, just as the epidemiological conditions and the num- US$435 for acute cases, US$122 per patient for asymptomatic ber of rounds of ivermectin treatment required to achieve the cases, US$336 for moderate cardiopathy, and US$1,135 for same for onchocerciasis are not yet known. For leprosy, the key severe cardiopathy. Given that Argentina had 2 million infected questions remain how much effect MDT has on transmission individuals, and assuming that 85 percent of them would have and when the incidence of new cases can be expected to decline been asymptomatic,9 percent would have had mild cardiopathy, significantly. Hence, the sustainability of control remains a 4 percent would have had moderate cardiopathy, and 2 percent critical issue. would have had severe cardiopathy, then total expenditures for The control programs for three of the diseases depend on medical attention would have amounted to US$457 million. drug donations. To date, the pharmaceutical companies donat- In Chile, aggregate treatment costs were estimated at US$37 ing the drugs and the donors supporting drug distribution million in 1991. New estimates in 1997 using the government have shown impressive commitment to the programs, but if payment schedule and an estimate of 142,000 people infected, their commitment were to lapse, the control programs would including 26,545 with myocardiopathy of which 9,652 were collapse and the diseases would return as public health prob- severe cases, resulted in an estimated cost of US$14 million to lems. Another risk is drug resistance. The control programs rely US$19 million. In Uruguay, annual costs for treatment were on just a few drugs, and even though drug resistance is not cur- estimated at US$15 million for 1996. rently apparent, if it were to emerge, the essential tools for con- These treatment costs are significantly higher than the costs trol would be lost. Hence, although elimination is in sight, the of vector control, which for 1996 were US$13 million in battle has not yet been won, and research to develop new and Argentina, US$28 million in Brazil, US$650,000 in Chile, and improved interventions and strategies for these tropical dis- US$4,000 in Uruguay. Akhavan's (1998) study in Brazil esti- eases remains important. mates a cost of US$260 per DALY prevented, and Robles's (1997) study in Bolivia indicated a cost of US$362 to prevent one year of life lost. COSTS AND COST-EFFECTIVENESS Data on the cost-effectiveness of treating those infected are OF INTERVENTIONS sparse; however, Robles's (1997) study in Bolivia estimates the costs of treating infected children under the age of five, coming The published information on cost-effectiveness of interven- up with a cost of US$3,009 per death averted or about US$100 tions for the four diseases is incomplete, and this section pro- per DALY averted. vides some new data on the cost per DALY averted that is not available in the literature. Lymphatic Filariasis The most widely used interventions for LF control are MDA, Chagas Disease vector control, and administration of diethylcarbamazine- For the Southern Cone countries, investment in the control of fortified salt. We estimate the cost-effectiveness of those strate- Chagas disease since 1991 has been about US$320 million, well gies in terms of DALYs averted from studies in India on the within the original estimates of US$190 million to US$350 mil- costs and effectiveness of control and for different scenarios for lion (Schofield and Dias 1991). Initial predictions of cost- the minimum duration of control required to achieve sus- effectiveness suggested an internal rate of return (IRR) for the tained interruption of transmission. These scenarios assume initiative of about 14 percent, but point studies during the that all three strategies are implemented in areas with similar course of the interventions suggest actual IRRs of about 30 per- levels of endemicity of Culex-transmitted Bancroftian filariasis cent for Brazil (Akhavan 1998) and more than 60 percent for and that the available cost data for India apply. the province of Salta, Argentina (Basombrio and others 1998). We consider three scenarios (table 22.2). Elim1 is an opti- Brazil had an estimated 2 million infected individuals in mistic elimination scenario under which sustained interrup- 1995. Annual follow-up of the 1.6 million asymptomatic cases tion of transmission is achieved after a relatively short period 442 | Disease Control Priorities in Developing Countries | Jan H. F. Remme, Piet Feenstra, P. R. Lever, and others Table 22.2 Costs per DALY Averted for LF Mass drug Diethylcarbamazine- administration Vector control fortified salt Scenario Elim1 Elim2 Control Elim1 Elim2 Control Elim1 Elim2 Control Duration (years) 6 10 30 10 5 30 2 4 30 Costs of control operations 0.35 0.65 2.22 2.92 6.64 22.74 0.09 0.29 3.56 per capita (US$) Cost per DALY averted (US$) 4.40 8.10 29.00 47.50 84.30 302.50 1.10 3.62 46.48 Source: Authors' calculations. of intervention (six annual rounds of MDA, 10 years of vector the previous calculations are based may be too optimistic. control, and 2 years of diethylcarbamazine-fortified salt). Elim2 However, even under much less favorable assumptions that the is a conservative elimination scenario under which sustained prevalence of hydrocele and lymphedema declines by 20 per- interruption is achieved only after a longer period of interven- cent after 7 years of MDA and by 75 percent after 30 years, the tion (10 years of MDA, 15 years of vector control, and 4 years estimated cost per DALY averted would be only 50 percent of diethylcarbamazine-fortified salt). Control is a scenario higher than those given in table 22.3, and the interventions under which transmission is brought to low levels but not would still be very cost-effective. interrupted and where control efforts will have to continue. The prevention of chronic disease also has direct economic Because of the slow dynamics of filariasis transmission and benefits (Ramaiah and Das 2004). The cost of preventing one disease, the prevalence of the chronic disease manifestations case of chronic disease through six rounds of MDA in India has (lymphedema and hydrocele) on which the DALY estimates are been estimated at US$8.41. The economic benefits include sav- based will not fully reflect the effect of control for many years. ings of 58.24 working days per year per case, yielding wages of We have therefore tried to predict the trend in chronic disease US$39.39 and treatment costs of US$1.44. On average, chronic over a 30-year period. Recent findings from a longitudinal patients lose 11 years of productive life; thus, the average eco- study (Ramaiah and others 2003) of the effect of MDA in nomic benefits total US$449.13 per chronic case averted. This Pondicherry, India, showed that the prevalence of hydrocele figure gives a benefit-cost ratio of 52.6, perhaps one of the and lymphedema had declined by 58 percent after seven annual highest for any disease control program. treatment rounds with diethylcarbamazine. We assumed that from the seventh year of intervention, any further reduction in disease prevalence was attributable exclusively to reduced inci- Onchocerciasis dence as a result of reduced transmission, and that 30 years Investment in onchocerciasis control has included about after the initiation of the intervention, the prevalence of disease US$570 million provided by donors to the OCP during would have fallen by 90 percent. We assumed that the effect of 1975­2002, US$140 million provided and earmarked for diethylcarbamazine-fortified salt was similar to that of MDA, APOC for 1996­2010, and US$10 million for OEPA for whereas for vector control we assumed that prevalence would 1991­2003. The African onchocerciasis control programs are decline with a delay of seven years. considered highly cost-effective. No cost-benefit analysis has The predicted costs per DALY averted (table 22.3) indicate yet been published for OEPA. that MDA and diethylcarbamazine-fortified salt are extremely The OCP has been highly successful. More than 40 million cost-effective. Elimination with MDA costs about US$4 to people in the program's 11 countries are now considered free US$8 per DALY averted, and even if transmission were not from infection and eye lesions, more than 1.5 million people interrupted and MDA would have to be continued for 30 years are no longer infected, and more than 200,000 cases of blind- (control scenario), the cost would be still only be around ness have been prevented. Sixteen million children born since US$29 per DALY averted. Diethylcarbamazine-fortified salt the program began are free of onchocerciasis. The socioeco- would be the cheapest intervention, but governments rarely nomic effect has also been dramatic: 25 million hectares of fer- favor it, and compliance can be difficult to ensure. Vector tile land in the river valleys were made available for resettle- control is at least 10 times more expensive in terms of DALYs ment and agriculture. A cost-benefit analysis of the OCP has averted, but it offers additional benefits in terms of malaria and estimated the net present value for the OCP over a 39-year dengue control and significant relief from mosquito nuisance. project horizon from 1974 to 2002 as US$485 million (Kim and The effect of MDA on hydrocele and lymphedema is not yet Benton 1995). This figure corresponds to an IRR of 20 percent, well established and the results of the Indian trial on which resulting mainly from increased labor because of prevention Tropical Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Onchocerciasis, and Leprosy | 443 of blindness (25 percent of benefits) and increased land use national decision makers that if investments do not continue, (75 percent of benefits). recrudescence of infection is likely. One strategy for sustaining A similar cost-benefit analysis for APOC also considered national investment is to show that ivermectin distribution sys- benefits in terms of additional labor resulting from blindness tems can be made polyfunctional. Treatment programs based prevention (Benton 1998). It did not consider land use because on MDA for intestinal parasites, schistosomiasis, and LF and on depopulation of river valleys is rarely seen in APOC countries, vitamin A distribution can be integrated with ivermectin dis- where the forest type of onchocerciasis predominates. tribution programs and thereby further improve cost-benefit Nevertheless, the estimated IRR for APOC remained almost as ratios. The use of community-directed treatment with iver- high as that for the OCP (17 percent), because the cost is lower mectin is also envisaged as a way of strengthening peripheral but the number of people served is far greater. and district health systems (Homeida and others 2002). The estimated rates of return for the OCP and APOC did not include the effects of control on onchocercal skin disease. Hence, these rates underestimate the benefits, because trouble- Leprosy some itching accounts for more than 50 percent of the DALYs Costs associated with leprosy control include case detection, attributable to onchocerciasis. The cost of ivermectin, which is treatment, prevention of disability, and rehabilitation. We cal- donated by Merck, was not included in our analyses. culate the incremental health service cost to arrive at the aver- To estimate the approximate cost per DALY averted, we con- age cost of curing a patient with leprosy. Our estimates are sidered the burden of disease and treatment with ivermectin based on the limited published cost data available, program in APOC countries. Using the latest epidemiological mapping expenditure data, and expert opinion, although costs are likely data, we estimate that, in 1995, 34.6 million people were to differ substantially by country. infected in APOC countries and that 1.86 million DALYs were As case-detection rates decrease, the average cost of detect- lost. Currently more than 44 percent of those infected are cov- ing one case increases. The previous edition of this volume ered by community-directed treatment with ivermectin, and estimated a cost of US$2 per case detected based on a expectations are that treatment will be expanded to cover most case-detection rate of about 300 per 100,000; however, case- of the remainder before the end of APOC in 2010. Information detection rates are now considerably lower in most countries from areas where ivermectin treatment has been in effect for (Dharmshaktu and others 1999; Ganapati and others 2001; more than 15 years shows that the prevalence and intensity of Smith 1999). Many leprosy control programs now rely on vol- onchocerciasis infection have fallen to low levels (Borsboom untary case finding supported by information, education, and and others 2003), and computer simulations predict that the communication activities to raise or maintain people's aware- disease could not become a public health problem again for at ness of the early signs and symptoms of leprosy. We estimate least another 10 to 20 years if treatment were halted (Remme, the cost of this approach to be about US$1 per case detected. Alley, and Plaisier 1995). We therefore estimate that 15 years of Nevertheless, if active methods are still used in areas where ivermectin treatment at 65 percent coverage will prevent at least case-detection rates are low, the cost of case detection may be 25 years of onchocercal disease. If we assume that 70 percent as high as US$108. of endemic communities will ultimately be covered by The costs of diagnosing and treating leprosy have fallen in community-directed treatment with ivermectin and that the past decade, and diagnosis by clinical examination only is 80 percent of those communities will maintain annual treat- now recommended. We therefore exclude the cost of skin ment at 65 percent coverage for at least 15 years, at least 26 mil- smears. In addition, a shortening of the treatment regimen has lion DALYs would be prevented over a 25-year period. lowered drug costs to about US$12 for a multibacillary case The predicted cost of community-directed treatment with and US$1 for a paucibacillary case. Globally, almost 40 percent ivermectin in APOC countries is US$145 million by the inter- of leprosy cases are classified as multibacillary cases, with the national donor community plus US$64 million by ministries of remaining 60 percent being paucibacillary cases. Thus, we esti- health and collaborating nongovernmental organizations, mate the average drug cost as US$5.40 per case. giving a total of US$209 million. Therefore we estimate that The cost of treatment, however, is more than the cost of the cost of community-directed treatment is approximately drugs alone. WHO guidelines recommend that a multibacillary US$7 per DALY averted. case receive supervised treatment for 12 months and that a The ultimate cost-benefit of onchocerciasis control will paucibacillary patient receive treatment for 6 months. Using depend on how long effective control programs will need to be cost data from Ethiopia and Pakistan, we estimate these maintained to keep the disease under control. National govern- treatment costs at US$20 to US$30 in low-income countries. ments and ministries of health should plan to invest in Data from studies of tuberculosis interventions show that ivermectin distribution and in surveillance activities for the community-supervised treatment may reduce costs by up to foreseeable future. Thus, a case must continually be made with 50 percent (Khan and others 2002), and this approach is being 444 | Disease Control Priorities in Developing Countries | Jan H. F. Remme, Piet Feenstra, P. R. Lever, and others advocated as part of "flexible MDT delivery" (ILA 2002) and Table 22.3 Cost-Effectiveness Estimates for the Main "accompanied MDT" (WHO 2002b). Reducing the nondrug Interventions for Each Disease costs of treating leprosy to about US$10 to US$20 per patient Cost per DALY Internal rate of may, therefore, be possible. We thus estimate the costs of treat- Intervention averted (US$) return (percent) ing a case of leprosy with MDT to be between US$15.40 and Chagas disease US$35.40 per case, depending on the strategy used. About 10 to 20 percent of new leprosy cases are likely to suf- Vector control 260 30­60 fer a reaction during or after MDT. We estimate treatment of Treatment (children under five) 100 -- those reactions to cost US$25 per patient. Of these patients, LF 1 percent may develop severe complications requiring hospital- MDA 4­29 -- ization, at an estimated cost of US$480 per patient. In addition, Diethylcarbamazine-fortified salt 1­46 -- 10 percent of new cases will develop neural or secondary Vector control 48­303 -- impairments and may require footwear and education about Onchocerciasis wound management. We estimate the lifetime cost of protec- MDA 7 17 (APOC) tive footwear at US$300 per patient (Seboka, Saunderson, and Vector control -- 20 (OCP) Currie 1998) and education at US$10 per patient. In 1 percent of cases, reconstructive surgery may be required at about Leprosy US$455 per patient. We therefore estimate the average incre- Case detection and treatment 38 -- mental cost of interventions for prevention of disability to be Prevention of disability 1­110 -- US$44.15 per new case of disability. Because about 3 percent of Source: Authors' calculations. new patients will need rehabilitation, we estimate the average MDA mass drug administration, -- not available. cost at under US$1 for each new case of leprosy detected (Jagannathan and others 1993). However, a backlog of old cases reactions and ulcers, US$75 for those needing footwear and exists. Although data in this area are weak, up to a third of the self-care education, and US$110 for those needing reconstruc- 4 million people living with leprosy globally (2 million with tive surgery. These estimates provide only a broad indication grade 1 disability and 2 million with grade 2 disability) could because data on the effectiveness of these interventions are require rehabilitation. scarce, and the application of the disability weight of 0.152 to Few data are available on the program costs associated with all interventions may overestimate their benefits. leprosy. A review of expenditure in Asia found that up to Data on the economic effect of leprosy at the national level 40 percent of the total costs could be classified as programmatic are not available. However, leprosy affects those who are eco- costs, although this amount may now be less because leprosy nomically active, with a peak in incidence at 10 to 20 years of programs have increasingly been integrated into general health age and again at 30 to 50 years of age. Studies of the impact of services. Data from Indonesia demonstrate that program costs leprosy on productivity show that deformity from leprosy can can be reduced by up to 35 percent by integrating them with reduce the probability of obtaining employment and can tuberculosis programs (Plag 1995). We therefore estimate the reduce household income and expenditure on food (Diffey and average cost of finding, treating, and preventing disabilities and others 2000; Kopparty 1995). In addition, leprosy can have a rehabilitating a new case of leprosy at US$76 to US$264. significant social impact because participation in the commu- In practice, many leprosy programs will also be providing nity may be restricted. This impact continues well beyond the disability prevention and rehabilitation interventions to a large actual treatment period because leprosy-related impairments backlog of patients, so the average cost per new case will be have a tendency to get worse over time even after the infection higher than here. Programs that face a high proportion of has been arrested. multibacillary cases and cases presenting with high levels of Summary disability are also likely to have higher costs. Assuming a cure rate of around 85 percent, we estimate the Available information indicates that interventions for the four costs of curing one patient of leprosy to be about US$93 per diseases are highly cost-effective and that the benefit-cost ratio new case. Using data from India (25 percent of those with lep- of control is high (table 22.3). rosy will self-cure, an average age of onset of 27, a disability weighting of 0.152, and a life expectancy at age 25 to 29 of RESEARCH NEEDS AND PRIORITIES 44.75), we estimate the cost per DALY of detecting and treating a new case of leprosy to be US$38. Because the diseases in this chapter are targeted for elimination In addition, assuming a 90 percent success rate, we calculate as public health problems, it is sometimes assumed that a cost per DALY of US$7 for patients needing treatment for research for these diseases is no longer necessary and that all Tropical Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Onchocerciasis, and Leprosy | 445 Table 22.4 Control Strategies, Major Challenges, and Research Needs for Each Disease Disease Principal control strategy Major problems and challenges Major research needs Chagas disease Interruption of transmission through domestic Control of nondomiciliated vectors Strategies for surveillance of nondomicil- vector control and improved blood transfusion Sustained vector control iated vectors Millions of those infected still at risk of Better drugs and diagnostics disease LF Interruption of transmission through periodic Need for high treatment coverage Strategies for high treatment coverage mass treatment Unknowns in elimination strategy Evidence base for elimination strategies Disability alleviation by local hygiene measures Limited effect of current drugs Drug that kills or sterilizes adult worms Onchocerciasis Periodic mass treatment to eliminate the Need to sustain high coverage for decades Strategies for sustained high treatment disease as a public health problem Risk of ivermectin resistance coverage Eradication not possible with current tools Feasibility of elimination with ivermectin Drug that kills or sterilizes adult worms Leprosy Case finding and multidrug treatment Incomplete MDT coverage Integration and sustainability of control Rehabilitation and prevention of disability Need to integrate and sustain control Improved diagnosis of infection Impact on transmission not known Prevention and management of nerve damage Source: Remme and others 2002. available resources should be allocated to elimination efforts. For onchocerciasis and LF, the main research priorities are However, research remains critical to address questions per- similar: implementation research to improve MDA; epidemio- taining to how to achieve elimination with currently available logical research to determine if, when, and with what treatment tools and especially to how to optimize implementation in dif- coverage the parasite reservoir can be locally eliminated for dif- ferent epidemiological, sociocultural, and health system set- ferent vector-parasite complexes; and research to develop a tings. Epidemiological questions on the required intervention macrofilaricide and improved diagnostics that would facilitate coverage, frequency, and duration need to be answered to guide elimination. elimination strategies, and research on the risk, prevention, and For leprosy, the research needs were further reviewed during control of recrudescence is crucial to ensure sustained success. a Scientific Working Group (Special Programme for Research The Special Programme for Research and Training in and Training in Tropical Diseases 2003). The meeting arrived at Tropical Diseases, a joint project of the United Nations a clear consensus of three top priorities for leprosy research: Children's Fund, United Nations Development Programme, implementation research on sustainable and integrated resid- World Bank, and WHO, recently undertook a systematic analy- ual leprosy control activities, improved diagnosis of infection, sis of research needs for each of the 10 tropical diseases in its and improved approaches for preventing and managing nerve portfolio (Remme and others 2002). This analysis involved damage. assessing the burden of disease and recent epidemiological These are the current main priorities for research in support trends, reviewing current control strategies, and identifying the of elimination. Eradication is not currently anticipated for any major problems and challenges for disease control and the of the diseases; thus, research on better tools and strategies that research needed to address these challenges. Table 22.4 summa- will allow a permanent solution for these infectious diseases is rizes the results of this analysis for the four diseases discussed also needed. Furthermore, currently available control tools may in this chapter. be lost because of factors such as resistance, and research to Chagas disease has two main research priorities. The first is develop replacement tools is essential now. the development of new vector control strategies that will allow the successful elimination campaign used in the Southern Cone countries to be extended to the Central American and CONCLUSION Andean countries, where the vectors are often not domiciliated. The second is the development of effective and affordable treat- Tropical diseases are often viewed as neglected, because the ment for the millions of people already infected and the pre- investments made to fight them appear negligible compared vention of chronic complications. with the massive amounts expended globally on the health 446 | Disease Control Priorities in Developing Countries | Jan H. F. Remme, Piet Feenstra, P. R. Lever, and others problems of developed countries. Tropical diseases are truly Argentina." Transactions of the Royal Society of Tropical Medicine and diseases of the poor, but despite the limited resources available Hygiene 92 (2): 137­43. for research and control, simple and effective interventions Bekri, W., S. Gebre, A. Mengiste, P. R. Saunderson, and S. Zewge. 1998. "Delay in Presentation and Start of Treatment in Leprosy Patients: A have been developed and delivered to populations in need for Case-Control Study of Disabled and Non-Disabled Patients in Three the four tropical diseases discussed in this chapter. Thus, expe- Different Settings in Ethiopia." International Journal of Leprosy and rience with these four diseases sends a powerful message: suc- Other Mycobacterial Diseases 66 (1): 1­9. cess is possible, even for neglected tropical diseases of poor Benton, B. 1998. "Economic Impact of Onchocerciasis Control through the African Programme for Onchocerciasis Control: An populations in developing countries. Elimination of these dis- Overview." Annals of Tropical Medicine and Parasitology 92 (Suppl. 1): eases as public health problems can be achieved, and invest- S33­39. ments in tropical disease research and control can make a sig- Borsboom, G. J. J. M., B. A. Boatin, N. J. D. Nagelkerke, H. Agoua, K. L. B. nificant contribution to poverty reduction. Akpoboua, E. W. S. Alley, and others. 2003. "Impact of Ivermectin on Onchocerciasis Transmission: Assessing the Empirical Evidence An important reason for the success was that the interven- That Repeated Ivermectin Mass Treatments May Lead to tions were extremely cost-effective. The available cost- Elimination/Eradication in West-Africa." Filaria Journal 2 (1): 8. effectiveness data, though limited, show convincingly that Boussinesq, M., S. D. Pion, Demanga-Ngangue, and J. Kamgno. 2002. intervention against these diseases is a good investment, and "Relationship between Onchocerciasis and Epilepsy: A Matched Case- Control Study in the Mbam Valley, Republic of Cameroon." the argument for investment gets better when other economic Transactions of the Royal Society of Tropical Medicine and Hygiene benefits, not reflected in DALYs, are taken into account, such as 96 (5): 537­41. increased food production when fertile land along river valleys Burkot, T., and K. Ichimori. 2002. "The PacELF Program: Will Mass Drug became available for agriculture after the control of onchocer- Administration Be Enough?" Trends in Parasitology 18 (3): 109­15. ciasis in West Africa and increased labor productivity after Dadzie, K. Y., J. Remme, A. Rolland, and B. Thylefors. 1989. "Ocular Onchocerciasis and Intensity of Infection in the Community: II. West effective filariasis control in India. African Rainforest Foci of the Vector Simulium yahense." Tropical The pharmaceutical industry also played a major role Medicine and Parasitology 40 (3): 348­54. through large drug donations, and the creation of intercountry Dharmshaktu, N. S., B. N. Barkakaty, P. K. Patnaik, and M. A. Arif. 1999. control programs provided effective mechanisms for imple- "Progress towards Elimination of Leprosy as a Public Health Problem menting interventions, technical support, and coordination. in India and Role of Modified Leprosy Elimination Campaign." Leprosy Review 70 (4): 430­39. Another reason for the success was a focused research program Dias, J. C. 1987. "Control of Chagas' Disease in Brazil." Parasitology Today that ensured the development of interventions based on simple 3 (11): 336­41. and sustainable approaches that use cheap and "appropriate" Diffey, B., M. Vaz, M. J. Soares, A. J. Jacob, and L. S. Piers. 2000. "The Effect technology and that are potentially multifunctional. of Leprosy-Induced Deformity on the Nutritional Status of Index Chagas disease, LF, onchocerciasis, and leprosy are now on Cases and Their Household Members in Rural South India: A Socioeconomic Perspective." European Journal of Clinical Nutrition target for elimination as public health problems from large 54 (8): 643­49. parts of the world. However, these diseases cannot be eradi- Dreyer, G., P. Dreyer, and J. Noroes. 2002. "Recommendations for the cated using current tools, and much remains to be done to Treatment of Bancroftian filariasis in Symptomless and Diseased expand and sustain the control efforts and undertake the nec- Patients." Revista da Sociedade Brasileira de Medicina Tropical 35 (1): 43­50. essary research to improve the control efforts as well as to Dreyer, G., J. Noroes, J. Figueredo-Silva, and W. F. Piessens. 2000. develop more definite solutions. It will be essential, therefore, "Pathogenesis of Lymphatic Disease in Bancroftian filariasis: A Clinical that donors and ministries of health not abandon these pro- Perspective." Parasitology Today 16 (12): 544­48. grams because of their success. Fine, P. E., and P. G. Smith. 1996. "Vaccination against Leprosy: The View from 1996." Leprosy Review 67 (4): 249­52. Ganapati, R., C. R. Revankar, V. V. Pai, S. Kingsley, and S. N. Prasad. 2001. REFERENCES "Can Cost of Leprosy Case Detection in Urban Areas Be Further Reduced?" International Journal of Leprosy and Other Mycobacterial Addiss, D., J. Critchley, H. Ejere, P. Garner, H. Gelband, and C. Gamble. Diseases 69 (4): 349­51. 2004. "Albendazole for Lymphatic Filariasis." Cochrane Database of Gwatkin, D. R., M. Guillot, and P. Heuveline. 1999. "The Burden of Systematic Reviews (1) CD003753. Disease among the Global Poor." Lancet 354 (9178): 586­89. Akhavan, D. 1998. Análise de custo-efetividade do programa de controle Gyapong, J. O., D. Kyelem, I. Kleinschmidt, and others. 2002. "The Use of da doença de Chagas no Brasil. Brasília: Pan American Health Spatial Analysis in Mapping the Distribution of Bancroftian filariasis Organization. in Four West African Countries." Annals of Tropical Medicine and Amazigo, U. V., M. Obono, K. Y. Dadzie, J. Remme, J. Jiya, R. Parasitology 96 (7): 695­705. Ndyomugyenyi, and others. 2002. "Monitoring Community-Directed Gyapong, J. O., V. Kumaraswami, G. Biswas, and E. A. Ottesen. 2005. Treatment Programs for Sustainability: Lessons from the African "Treatment Strategies Underpinning the Global Programme to Programme for Onchocerciasis Control (APOC)." Annals of Tropical Eliminate Lymphatic Filariasis." Expert Opinion on Pharmacotherapy Medicine and Parasitology 96 (Suppl. 1): S75­92. 6 (2): 179­200. Basombrio, M. A., C. J. Schofield, C. L. Rojas, and E. C. del Rey. 1998. "A Homeida, M., E. Braide, E. Elhassan, U. V. Amazigo, B. Liese, B. Benton, Cost-Benefit Analysis of Chagas Disease Control in North-western and others. 2002. "APOC's Strategy of Community-Directed Tropical Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Onchocerciasis, and Leprosy | 447 Treatment with Ivermectin (CDTI) and Its Potential for Providing Ramaiah, K. D., and P. K. Das. 2004. "Mass Drug Administration to Additional Health Services to the Poorest Populations: African Eliminate Lymphatic Filariasis in India." Trends in Parasitology 20 (11): Programme for Onchocerciasis Control." Annals of Tropical Medicine 499­502. and Parasitology 96 (Suppl. 1): S93­104. Ramaiah, K. D., P. K. Das, and V. Dhanda. 1994. "Estimation of Permissible ILA (International Leprosy Association). 2002. "Report of the Levels of Transmission of Bancroftian filariasis Based on Some International Leprosy Association Technical Forum, Paris, France, Entomological and Parasitological Results of a Five-Year Vector 22­28 February 2002." International Journal of Leprosy and Other Control Program." Acta Tropica 56 (1): 89­96. Mycobacterial Diseases 70 (Suppl. 1): S1­62. Ramaiah, K. D., P. K. Das, E. Michael, and H. Guyatt. 2000. "The Economic Jagannathan, S. A., V. Ramamurthy, S. J. Jeyaraj, and S. Regina. 1993. Burden of Lymphatic Filariasis in India." Parasitology Today 16 (6): "A Pilot Project on Community Based Rehabilitation in South India: 251­53. A Preliminary Report." Indian Journal of Leprosy 65 (3): 315­22. Ramaiah, K. D., P. K. Das, P. Vanamail, and S. P. Pani. 2003. "The Im- Khan, M. A., J. D. Walley, S. N. Witter, A. Imran, and N. Safdar. 2002."Costs pact of Six Rounds of Single-Dose Mass Administration of and Cost-Effectiveness of Different DOT Strategies for the Treatment Diethylcarbamazine or Ivermectin on the Transmission of Wuchereria of Tuberculosis in Pakistan: Directly Observed Treatment." Health bancrofti by Culex quinquefasciatus and Its Implications for Lymphatic Policy and Planning 17 (2): 178­86. Filariasis Elimination Programs." Tropical Medicine and International Kim, A., and B. Benton. 1995. Cost-Benefit Analysis of the Onchocerciasis Health 8 (12): 1082­92. Control Programme (OCP). Washington, DC: World Bank. Ramu, K., K. D. Ramaiah, H. Guyatt, and D. Evans. 1996. "Impact of Kopparty, S. N. 1995. "Problems, Acceptance, and Social Inequality: A Lymphatic Filariasis on the Productivity of Male Weavers in a South Study of the Deformed Leprosy Patients and their Families." Leprosy Indian Village." Transactions of the Royal Society of Tropical Medicine Review 66 (3): 239­49. and Hygiene 90 (6): 669­70. Mathers, C. D., A. Lopez, and C. J. L. Murray. 2006. "The Burden Remme, J. H. F. 1995. "The African Programme for Onchocerciasis of Disease and Mortality by Condition: Data, Methods, and Results Control: Preparing to Launch." Parasitology Today 11: 403­6. for the Year 2001." In Global Burden of Disease and Risk Factors, ------. 2004a. "The Global Burden of Onchocerciasis in 1990." In Global ed. A. Lopez, C. Mathers, M. Ezzati, D. Jamison, and C. Murray. Burden of Disease 1990. Geneva: World Health Organization. http:// New York: Oxford University Press. www3.who.int/whosis/burden/gbd2000docs/Onchocerciasis%201990. Maxwell, C. A., C. F. Curtis, H. Haji, S. Kisumku, A. I. Thalib, and S. A. pdf. Yahya. 1990. "Control of Bancroftian filariasis by Integrating Therapy ------. 2004b. "Research for Control: The Onchocerciasis Experience." with Vector Control Using Polystyrene Beads in Wet Pit Latrines." Tropical Medicine and International Health 9 (2): 243­54. Transactions of the Royal Society of Tropical Medicine and Hygiene 84 Remme, J. H. F., E. S. Alley, and A. P. Plaisier. 1995. "Estimation and (5): 709­14. Prediction in Tropical Disease Control: The Example of Meima, A., P. R. Saunderson, S. Gebre, K. Desta, G. J. van Oortmarssen, Onchocerciasis." In Epidemic Models: Their Structure and Relation to and J. D. Habbema. 1999. "Factors Associated with Impairments in Data, ed. D. Mollison, 372­92. Cambridge, U.K.: Cambridge University New Leprosy Patients: The AMFES Cohort." Leprosy Review 70 (2): Press. 189­203. Remme, J. H., F. Blas, L. Chitsulo, P. M. Desjeux, H. D. Engers, T. P. Kanyok, Molyneux, D. H. 1995. "Onchocerciasis Control in West Africa: Current and others. 2002. "Strategic Emphases for Tropical Diseases Research: Status and Future of the Onchocerciasis Control Programme." A TDR Perspective." Trends in Parasitology 18 (10): 421­26. Parasitology Today 11: 399­402. Reuben, R., R. Rajendran, I. P. Sunish, T. R. Mani, S. C. Tewari, J. Hiriyan, Murdoch, M. E., M. C. Asuzu, M. Hagan, W. H. Makunde, P. Ngoumou, and others. 2001. "Annual Single-Dose Diethylcarbamazine plus K. F. Ogbuagu, and others. 2002. "Onchocerciasis: The Clinical and Ivermectin for Control of Bancroftian filariasis: Comparative Efficacy Epidemiological Burden of Skin Disease in Africa." Annals of Tropical with and without Vector Control." Annals of Tropical Medicine and Medicine and Parasitology 96 (3): 283­96. Parasitology 95 (4): 361­78. Noma, M., B. E. Nwoke, I. Nutall, P. A. Tambala, P. Enyong, A. Namsenmo, Richards, F. O., B. Boatin, M. Sauerbrey, and A. Seketeli. 2001. "Control of and others. 2002. "Rapid Epidemiological Mapping of Onchocerciasis Onchocerciasis Today: Status and Challenges." Trends in Parasitology (REMO): Its Application by the African Programme for 17 (12): 558­63. Onchocerciasis Control (APOC)." Annals of Tropical Medicine and Richards F. O., D. Pam, A. Kal, G. Gerlong, J. Oneyka, Y. Sambo, and oth- Parasitology 96 (Suppl. 1): S29­39. ers. 2005 "Significant Decrease in the Prevalence of Wuchereria ban- Ottesen, E. A. 2000. "The Global Programme to Eliminate Lymphatic crofti Infection in Anopheline Mosquitoes Following the Addition of Filariasis." Tropical Medicine and International Health 5 (9): 591­94. Albendazole to Annual, Ivermectin-Based, Mass Treatments in Ottesen, E. A., B. O. Duke, M. Karam, and K. Behbehani. 1997. "Strategies Nigeria." Annals of Tropical Medicine and Parasitology 99 (2): 155­64. and Tools for the Control/Elimination of Lymphatic Filariasis." Robles, M. C. 1997. Analisis costo-efectividad de las intervenciones de salud Bulletin of the World Health Organization 75 (6): 491­503. en Bolivia. La Paz: Unidad de Análisis de Políticas Sociales. Peters, D. H., and T. Phillips. 2004. "Mectizan Donation Program: Salvatella, A. R., and W. Vignolo. 1996. "Una aproximación a los costos Evaluation of a Public-Private Partnership." Tropical Medicine and de internación por cardiopatia Chagasica en Uruguay." Revista da International Health 9 (4): A4­15. Sociedade Brasileira de Medicina Tropical 29 (Suppl.): 114­18. Plag, I. 1995. Guidelines for Cost Analysis in Leprosy Control Programmes. Schenone, H. 1998. "Human Infection by Trypanosoma Cruzi in Chile: Amsterdam: Royal Tropical Institute. Epidemiology Estimates and Costs of Care and Treatment of the Rajendran, R., I. P. Sunish, T. R. Mani, A. Munirathinam, S. M. Abdullah, Chagasic Patient." Boletin Chileno de Parasitologia. 53 (1­2): 23­26. D. J. Augustin, and K. Satyanarayana. 2002. "The Influence of the Mass Schmunis, G. A. 2000. "A tripanossomiase Americana e seu impacto na Administration of Diethylcarbamazine, Alone or with Albendazole, on saude publica das Americas." In Trypanosoma cruzi e doença de Chagas, the Prevalence of Filarial Antigenaemia." Annals of Tropical Medicine 2nd ed., ed. Z. Brener, Z. Andrade, and M. Barral-Neto, 1­15. Rio de and Parasitology 96 (6): 595­602. Janeiro: Guanabara-Koogan. 448 | Disease Control Priorities in Developing Countries | Jan H. F. Remme, Piet Feenstra, P. R. Lever, and others Schmunis, G. A., F. Zicker, J. R. Cruz, and P. Cuchi. 2001. "Safety of Blood van Brakel, W. H., and A. M. Anderson. 1998. "A Survey of Problems in Supply for Infectious Diseases in Latin American Countries, Activities of Daily Living among Persons Affected by Leprosy." Asia 1994­1997." American Journal of Tropical Medicine and Hygiene 65 (6): Pacific Disability and Rehabilitation Journal 9 (2): 62­67. 924­30. Vlassoff, C., M. Weiss, E. B. Ovuga, C. Eneanya, P. T. Nwel, S. S. Babalola, Schmunis, G. A., F. Zicker, F. Pinheiro, and D. Brandling-Bennett. 1998. and others. 2000. "Gender and the Stigma of Onchocercal Skin Disease "Risk of Transfusion-Transmitted Infectious Diseases in Latin in Africa." Social Science and Medicine 50 (10): 1353­68. America." Emerging Infectious Diseases 4: 5­11. WHO (World Health Organization). 1991. Control of Chagas' Disease: Schofield, C. J., and J. C. Dias. 1991. "A Cost-Benefit Analysis of Chagas' Report of a WHO Expert Committee. Technical Report 811. Geneva: Disease Control." Memórias do Instituto Oswaldo Cruz 86 (3): 285­95. WHO. Seboka, G., P. Saunderson, and H. Currie. 1998. "Footwear for Farmers ------. 1995a. Onchocerciasis and Its Control: Report of a WHO Expert Affected by Leprosy." Leprosy Review 69 (2): 182­83. Committee on Onchocerciasis Control. Technical Report 852. Geneva: Seketeli, A., G. Adeoye, A. Eyamba, E. Nnoruka, P. Drameh, U. V. Amazigo, WHO. and others. 2002. "The Achievements and Challenges of the African ------. 1995b. World Health Report 1995: Bridging the Gaps. Geneva: Programme for Onchocerciasis Control (APOC)." Annals of Tropical WHO. Medicine and Parasitology 96 (Suppl. 1): 15­28. ------. 2002a. Control of Chagas Disease: Second Report of the WHO Smith, C. M., and W. C. Smith. 2000."Chemoprophylaxis Is Effective in the Expert Committee. Technical Report 109. Geneva: WHO. Prevention of Leprosy in Endemic Countries: A Systematic Review and ------. 2002b. The Final Push Strategy to Eliminate Leprosy as a Public Meta-Analysis. MILEP2 Study Group--Mucosal Immunology of Health Problem: Questions and Answers. Geneva: WHO. Leprosy." Journal of Infection 41 (2): 137­42. ------. 2002c. Lymphatic Filariasis: The Disease and Its Control. Technical Smith, W. C. 1999. "Future Scope and Expectations: Why, When, and How Report 71. Geneva: WHO. LECs Should Continue." Leprosy Review 70 (4): 498­505. ------. 2002d. The World Health Report 2002: Reducing Risks, Promoting Special Programme for Research and Training in Tropical Diseases. 2003. Healthy Life. Technical Report 248. Geneva: WHO. Report of the Scientific Working Group Meeting on Leprosy. Geneva: ------. 2004a. World Health Organization: Leprosy Elimination Project World Health Organization. Status Report 2003. Geneva: WHO. Stolk, W. A., S. Swaminathan, G. J. van Oortmarssen, P. K. Das, and J. D. ------. 2004b. World Health Report 2004: Changing History. Geneva: Habbema. 2003."Prospects for Elimination of Bancroftian Filariasis by WHO. Mass Drug Treatment in Pondicherry, India: A Simulation Study." Journal of Infectious Diseases 188 (9): 1371­81. World Bank. 1993. World Development Report 1993: Investing in Health. New York: Oxford University Press. van Beers, S. M., M. Hatta, and P. R. Klatser. 1999. "Patient Contact Is the Major Determinant in Incident Leprosy: Implications for Future Zimmerman, P. A., K. Y. Dadzie, G. De Sole, J. Remme, E. S. Alley, and Control." International Journal of Leprosy and Other Mycobacterial T. R. Unnasch. 1992. "Onchocerca volvulus DNA Probe Classification Diseases 67 (2): 119­28. Correlates with Epidemiologic Patterns of Blindness." Journal of Infectious Diseases 165 (5): 964­68. van Brakel, W. H. 2000. "Peripheral Neuropathy in Leprosy and Its Consequences." Leprosy Review 71 (Suppl.): S146­53. Tropical Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Onchocerciasis, and Leprosy | 449 Chapter 23 Tropical Diseases Lacking Adequate Control Measures: Dengue, Leishmaniasis, and African Trypanosomiasis Pierre Cattand, Phillippe Desjeux, M. G. Guzmán, Jean Jannin, A. Kroeger, André Médici, Philip Musgrove, Mike B. Nathan, Alexandra Shaw, and C. J. Schofield Dengue, leishmaniasis, and African trypanosomiasis (sleeping present with symptoms ranging from mild, self-limiting, sickness) are serious diseases that the World Health febrile illness to severe, life-threatening disease. Two clinical Organization (WHO) characterizes as lacking effective control pictures are recognized: (a) dengue fever (DF) and (b) dengue measures. They are transmitted by insect vectors and can result hemorrhagic fever (DHF) or dengue shock syndrome (DSS). in epidemic outbreaks. Specific treatment is unavailable for The four dengue serotypes, known as dengue 1, 2, 3, and 4, dengue, although good supportive treatment can drastically constitute a complex of the flaviviridae transmitted by Aedes reduce mortality. For the leishmaniases and for sleeping sick- mosquitoes, particularly Ae. aegypti. Infection by any of the ness, treatment relies largely on antiquated drugs based on four serotypes induces lifelong immunity against reinfection by antimony and arsenic, respectively. Sustained control of the the same serotype, but only partial and transient protection insect vectors is difficult for dengue and leishmaniasis because against the others. Sequential infection by different serotypes their high reproductive potential allows the vector populations seems to be the main trigger for DHF/DSS. to recover quickly after intervention wherever adequate breed- ing conditions exist. By contrast, tsetse flies, the vectors for Disease Manifestations. The incubation period is four to six sleeping sickness, have a much lower reproductive potential days. Infants and young children usually develop fever, some- and could be eliminated over large areas, given adequate organ- times accompanied by a rash. Older children and adults may ization and surveillance. Through the African Union, African develop either a mild febrile syndrome or classic DF with fever, nations are developing a large-scale initiative for areawide headache, myalgias, arthralgia, nausea, vomiting, and rash. Skin elimination of tsetse flies, partly because of sleeping sickness, bleeding, petechiae, or ecchymosis are observed in some but also because of their importance as vectors of animal try- patients. Bleeding from the nose, gums, and gastrointestinal panosomiasis, which poses a serious constraint to livestock tract; hematuria; or hypermenorrhea can accompany the clini- development and agriculture. cal picture. Leukopenia is common and thrombocytopenia is sometimes observed. DF can be incapacitating, but the prog- DISEASE CHARACTERISTICS nosis is favorable and the case-fatality rate is low. AND TRANSMISSION By contrast, DHF/DSS can be life threatening. It is character- ized by high fever, bleeding, thrombocytopenia, and hemocon- Dengue centration (Nimmanitya 1993; PAHO 1994). Plasma leakage Dengue is a mosquitoborne viral disease with a high capacity differentiates DHF/DSS from classic DF. Severity is classified as for epidemic outbreaks. Infection can be asymptomatic or can mild (grades I and II) or severe (grades III and IV), with the 451 main difference being shock in the latter. In some epidemics, Eastern Mediterranean, Southeast Asia, and the Western hepatomegaly has been prominent. As with DF, DHF generally Pacific. According to WHO, it occurs in more than 100 coun- begins with a sudden temperature rise accompanied by facial tries and an estimated 2.5 billion people are at risk. flush and other nonspecific manifestations, such as anorexia, The increase in dengue epidemics can be attributed to rising vomiting, headache, and muscle or joint pains. The most levels of urbanization, which promote contact between humans common hemorrhagic manifestation is a positive tourniquet and Ae. aegypti; inadequate domestic water supplies; and test, although petechiae, ecchymosis, epistaxis, and gingival or increasing international travel, migration, and trade, which help gastrointestinal bleeding may also be observed. After three or disseminate vectors and the virus. Epidemiological changes in four days, when the temperature returns to normal or below, the the Americas since the 1970s illustrate this process. During the patient's condition can suddenly deteriorate with signs of circu- 1940s through the 1960s, the Ae. aegypti control program was latory disturbance. The patient may sweat, be restless, have cool successful in most of the region, with several countries declaring extremities, and show changes in pulse rate and blood pressure. complete eradication. However, after some years, reinfestation Many recover spontaneously or after brief fluid therapy, but with Ae. aegypti was apparent, and by 1995 it had returned to some proceed to shock with typical signs of circulatory failure. most of the previously infested countries. Reinvasion by the vec- Initially patients may be lethargic but become restless and rap- tor was followed by increased circulation of the virus, and the idly enter a critical stage of shock. Some patients evolve to severe region evolved from nonendemic to hypoendemic (sporadic circulatory failure (DSS), presenting a rapid and weak pulse, a epidemics caused by a single serotype) to hyperendemic (simul- narrow pulse pressure or hypotension, cold and clammy skin, taneous circulation of multiple serotypes resulting in frequent and an altered mental state. DSS is fatal in 5 to 10 percent of epidemics). DHF became a major public health problem. Before cases if fluid management is inadequate or delayed. 1981, 5 countries in the region reported only a few cases of DHF, but by 2002, 21 countries reported more than 14,000 cases and Transmission and Epidemiological Trends. The dengue virus 250 deaths (Gubler 2002; Guzmán and Kouri 2002, 2003; is transmitted from humans to humans by Aedes mosquitoes, Guzmán and others 2002). of which the most important is Ae. aegypti (Bennett and others 2002; Gubler 1979; Tardieux and others 1990). Female mosqui- toes ingest the virus while feeding on viremic individuals, and Leishmaniasis after an 8- to 12-day incubation period they can transmit the Leishmaniasis (or the leishmaniases) refers to infections caused virus to other humans during blood feeding (Watts and others by protozoan parasites of the genus Leishmania transmitted by 1987). Thereafter, the female mosquito remains infective for female sandflies (Phlebotominae). More than 20 Leishmania life. Transmission of the virus from infected females to their species are pathogenic to humans, and more than 30 species of progeny has been documented, but its epidemiological signifi- sandflies are proven vectors. The disease tends to be focal in dis- cance is not well understood (Hull and others 1984; Rosen and tribution. In anthroponotic foci, sandflies transmit parasites others 1983). from human to human, and in zoonotic foci, sandflies transmit Although believed to be of African origin, Ae. aegypti is now the parasites between mammal hosts and from them to humans. established throughout the tropics and subtropics, exploiting almost any water-filled container as larval habitat. Ae. aegypti is Disease Manifestations. The different species of Leishmania also the urban vector of yellow fever. Ae. albopictus also trans- cause illness of differing severity. Visceral leishmaniasis (VL), mits dengue and is an important secondary vector in parts of caused by species of the L. donovani complex, is usually fatal Southeast Asia and the Pacific. This species is of Asian origin if untreated. Mucocutaneous leishmaniasis, caused by the but has spread to parts of Africa, the Americas, and Europe by L. braziliensis complex, is highly disfiguring and mutilating, and depositing egg masses in used car tires, which are traded it can be fatal because of secondary complications. Cutaneous around the world. leishmaniasis (CL), caused by the L. major, L. donovani, and Major epidemics of denguelike illness were documented in L. braziliensis complexes, may be a simple, self-limiting skin the 18th and 19th centuries in Africa, the Americas, and Asia, ulcer, but it can be disabling when numerous lesions occur. and clinical descriptions of illness compatible with dengue in Diffuse cutaneous leishmaniasis, caused by the L. mexicana and China date from about 265. During 1900­50, dengue epi- L. aethiopica complexes, is longer lasting because of deficient demics occurred in Australia, China, Greece, India, Japan, immune responses. Malaysia, Thailand, and Vietnam and in the Caribbean (Gubler and Kuno 1997). DHF was first recognized in the 1950s, Epidemiological Trends. Leishmaniasis is found in 88 coun- although a similar hemorrhagic fever was reported in tries worldwide. VL occurs in 62 of those countries, with most Philadelphia in 1780, in Australia in 1897, and in Greece in of the estimated 500,000 annual cases occurring in poorer rural 1928. Dengue is now endemic in Africa, the Americas, the and suburban areas of Bangladesh, Brazil, India, Nepal, and 452 | Disease Control Priorities in Developing Countries | Pierre Cattand, Phillippe Desjeux, M. G. Guzmán, and others Sudan. Mucocutaneous leishmaniasis is mainly limited to producing a characteristic lesion or chancre. The parasites cir- South and Central America, whereas most of the estimated culate in blood and lymph, resulting in waves of parasitemia 1 million to 1.5 million annual CL cases occur in the Middle with episodes of fever, often accompanied by chills, rigor, East (Afghanistan, Algeria, the Islamic Republic of Iran, Saudi malaise, prostration, and weight loss. These symptoms may Arabia, and the Syrian Arab Republic) and in Brazil and Peru occur within days of development of the chancre and constitute (Desjeux 1996). Reliable data on incidence and prevalence are the hemolymphatic early stage. Febrile episodes become less scarce because only 33 endemic countries provide official noti- severe as the disease progresses, and after a variable period the fication of infection. parasites invade the central nervous system and cerebrospinal Leishmaniasis transmission is increasing in several areas. fluid, leading to the late stage, with meningoencephalitis typi- For example, the number of cases of CL in Kabul, Afghanistan, cally accompanied by severe and protracted headache, apathy, increased from 14,200 in 1994 to 65,000 in 2002, and the num- sleep disorders, irritability, and antisocial behavior. ber of cases of VL in northeastern Brazil increased from 1,840 The clinical features of late-stage sleeping sickness can in 1998 to 6,000 in 2002. Such increases reflect the following resemble AIDS. With T.b. rhodesiense, meningoencephalitis environmental, land-use, and behavioral changes that increase typically occurs within weeks of initial infection, whereas with exposure to the sandflies: T.b. gambiense, this syndrome occurs later, sometimes after sev- eral years. Untreated disease causes relentless deterioration in · Rural-urban migration seems to have contributed to urban- cerebral function, with patients becoming increasingly difficult izing VL in Brazil, whereas in East Africa, VL seems to be to rouse and passing into coma and death. Infection does more closely associated with migrations of seasonal workers not seem to confer immunity, so reinfection can occur after and refugees. Transborder migrations between Bangladesh, treatment. India, and Nepal are also a risk factor for VL. · New settlements in high-risk endemic areas, such as those Transmission and Vectors. Male and female tsetse flies are established by people migrating from high plateaus to trop- obligate bloodsuckers and can transmit trypanosomes, which ical plains in some Andean countries, increase their expo- undergo cyclical development in the infected flies. With T.b. sure to vectors. gambiense, the main reservoir host is people, so tsetse flies · Development in areas of zoonotic transmission, such as mainly transmit from person to person, although increasing road building, mining, oil prospecting, forestry, and eco- evidence suggests that pigs and some other animals are also tourism, and military activity increase risks for those important reservoirs. With T.b. rhodesiense, the main reservoir involved (Desjeux 2001). hosts are cattle and related animals, so transmission occurs · Deteriorating social and economic conditions in the poorer mainly from animals to humans, although transmission from suburbs of some cities may contribute to increasing trans- human to human also occurs (Okoth 1986). Mechanical (Frézil mission, especially of VL. 1983), sexual (Rochas and others 2004), and transplacental (De Raadt 1985; Libala, Wery, and Ruppol 1978; Traub and others Leishmaniasis can be an opportunistic infection in people 1978) transmission have been described but are believed to be with HIV/AIDS, and coinfections have been reported in insignificant. 34 countries (Desjeux and Alvar 2003). Malnutrition or Thirty-one species and subspecies of Glossina are recog- HIV/AIDS coinfection can also increase disease severity by nized. All probably can transmit trypanosomes, but only eight impairing the immune response. are known vectors for human infection. Animal trypanosomi- asis can be found wherever wild tsetse flies occur, but human African Trypanosomiasis trypanosomiasis is usually associated with historic foci with African trypanosomiasis is caused by parasites transmitted by strong epidemic potential. Most T.b. gambiense transmission is tsetse flies (Glossinidae). The most important are forms of attributed to G. palpalis species occupying riverine and forest Trypanosoma brucei that infect humans and livestock, and habitats in West and Central Africa, whereas T.b. rhodesiense T. congolense and T. vivax that infect only livestock. Human transmission is mainly attributed to G. morsitans species in East infection causes severe disease known as sleeping sickness, African savannas. G. fusca species, although important vectors which is acute in the case of infection with T. brucei rhodesiense of animal trypanosomiasis, are considered insignificant for but more chronic with T.b. gambiense. Both forms lead to cen- human forms. tral nervous involvement and are fatal without appropriate Tsetse flies have an unusual life cycle. An inseminated female treatment. nurtures the egg and larva in her uterus, depositing the mature larva on the ground, where it burrows and pupates. Thus, each Disease Manifestations. Parasites are transmitted by the bite of female produces only one offspring at a time. She produces up infected tsetse flies. They multiply locally in extracellular spaces, to 12 during her two- to three-month adult life span. The Tropical Diseases Lacking Adequate Control Measures: Dengue, Leishmaniasis, and African Trypanosomiasis | 453 intrinsic population growth rate is low. Even small increases in DISEASE BURDEN average daily mortality rates can cause population decline, even to extinction. All three diseases affect substantial populations. Globally, WHO estimates that 500,000 cases of DHF occur annually. Assuming Epidemiological Trends. Tsetse flies occur in parts of 37 coun- that DHF cases constitute 6 percent of all clinical dengue cases tries in Sub-Saharan Africa. Animal trypanosomiasis is wide- (that is, all other cases are classical DF) implies a total of almost spread throughout this region, but human disease is focused in 8 million new infections per year. For leishmaniasis, current areas of 20 countries. Over the entire tsetse-fly belt, WHO esti- estimates suggest an overall prevalence of 12 million people mates that 60 million people are at risk of infection, with a infected in an at-risk population of 350 million, suggesting standing prevalence of about 300,000 infections. Of these, prob- more than 2 million new infections each year. The prevalence of ably fewer than 15 percent are diagnosed and treated (Cattand, sleeping sickness is estimated at 300,000 people, with 60 million Jannin, and Lucas 2001). For T.b. rhodesiense, epidemiological people considered to be at risk. Uncertainty about the true work in Uganda estimated that for every individual correctly number of cases makes all these estimates approximate, partic- diagnosed and treated, a further 12 cases are undiagnosed and ularly because incidence is increasing. In terms of disability- unreported (Odiit 2003). The incidence of sleeping sickness has adjusted life years (DALYs) lost to these diseases (table 23.1), a been increasing steadily since the 1970s, with epidemics in sev- dengue or trypanosomiasis death accounts for 27 to 28 DALYs eral areas, particularly the Democratic Republic of Congo lost. Leishmaniasis kills less often than trypanosomiasis, but (DRC) and northern Angola. In 1998­99, some 45,000 new each death is responsible for a loss of 34 DALYs. cases were reported each year, representing a 10-fold increase since the 1960s. Most current epidemics are due to T.b. gambi- Dengue ense, although major epidemics of T.b. rhodesiense occurred in In hyperendemic areas of Southeast Asia, where multiple virus Uganda in 1978, 1980, and 1988. serotypes are circulating, DF and DHF are mainly childhood The apparent increase is largely attributable to a decline in diseases and in some countries are leading causes of pediatric tsetse and trypanosomiasis control operations, which are influ- hospitalization and death, particularly in Cambodia, Myanmar, enced both by changing political priorities and by civil unrest and Vietnam. Worldwide, 80 to 90 percent of deaths occur and war. In the DRC, some 10,000 cases were diagnosed annu- before age 15. In recent years in the Americas, DHF in adults in ally during the 1980s, but after four years with little or no con- endemic form has been reported frequently (Diaz and others trol, the number of reported cases rose to 30,000 per year. In 1988; Guzmán and others 1999; Harris and others 2000; Zagne Angola, cases rose sixfold following the interruption of control and others 1994). In some countries, the disease is more fre- operations because of war. Transmission is also occurring in quent among females, and in Cuba, significantly more severe new locations. In 1999, urban and periurban transmission was cases occur among Caucasians than among those of African reported in Kinshasa, DRC, and in Luanda, Angola, and a new descent (Kouri, Guzmán, and Bravo 1987). Dengue causes rel- focus was reported in Soroti, Uganda, where an epidemic of atively few deaths, estimated at 19,000 in 2001, corresponding T.b. rhodesiense disease followed the introduction of infected to a case-fatality rate of 3.8 percent for DHF. Nonetheless, it can cattle (Fèvre and others 2001). cause a substantial burden: in Puerto Rico during 1984­94, the Table 23.1 Number of Deaths and DALYS Caused by Dengue, Leishmaniasis, and African Trypanosomiasis, by World Bank Region, 2001 (thousands) African Dengue Leishmaniasis trypanosomiasis Total Region Deaths DALYs Deaths DALYs Deaths DALYs Deaths DALYs East Asia and the Pacific 8 217 2 48 0 0 10 265 Europe and Central Asia 0 0 0 6 0 0 0 6 Latin America and the Caribbean 2 59 0 37 0 0 2 96 Middle East and North Africa 0 8 1 48 1 22 2 78 South Asia 9 240 40 1,306 0 0 49 1,546 Sub-Saharan Africa 0 4 8 312 48 1,310 56 1,626 Total 19 528 51 1,757 49 1,332 119 3,617 454 | Disease Control Priorities in Developing Countries | Pierre Cattand, Phillippe Desjeux, M. G. Guzmán, and others DALY loss per million population was similar to that for the human-vector contact, whereas in contiguous Nuevo Laredo entire Latin America and Caribbean region from malaria, on the Mexican side, where per capita income was one-seventh tuberculosis, intestinal helminths, and the childhood disease of that in Laredo, seroprevalence was much higher despite cluster (Meltzer and others 1998). lower vector densities. In countries with weak or unprepared health services, Economic Impact. Studies in Thailand estimated the eco- epidemics of dengue can be extremely disruptive and health nomic impact of dengue as equivalent to US$12.6 million in services can be rapidly overwhelmed. Frequently, governments 1994, of which patients and their families incurred 45 percent declare states of emergency to mobilize additional resources (Okanurak, Sornmani, and Indaratna 1997). Kouri and others against dengue outbreaks, sometimes deploying the army to (1989) estimate the cost of the 1981 DHF epidemic in Cuba at eliminate or apply larvicides to larval habitats. These responses US$103 million, constituting US$41 million in medical care, are often launched at or after the peak of the epidemic, and the US$5 million in social security disability payments, US$14 decline in transmission is unjustifiably attributed to the inter- million in lost production, and US$43 million for vector vention rather than to the natural epidemic decline. control. For Southeast Asia, Shepard and others (2004) estimate individual treatment costs of US$139 for DHF and Leishmaniasis US$4.29 for DF (including health clinic visits, hospitalization, medications, travel expenses, and parents' time seeking treat- In 2001, leishmaniasis killed an estimated 51,000 people, ment for their children). This estimate implies annual costs in including 40,000 in South Asia and 8,000 in Sub-Saharan the region of US$105 million--US$69.5 million for DHF and Africa, representing 0.3 percent and less than 0.1 percent, US$35.5 million for DF. Extrapolating from current trends (on respectively, of all deaths (table 23.1). Nearly all deaths occurred the basis of cases reported to WHO since 1960), this figure will at ages 5 to 29. Males were more affected than females, espe- increase to an average of US$118 million each year through the cially in Sub-Saharan Africa, where the ratio was three to one. first decade of the 21st century. Using the Thailand study to add the value of productive work lost doubles this figure to Economic Impact. Treatment for leishmaniasis is expensive, US$236 million. Thus, during the next decade, Southeast Asian especially for VL. For many countries, the cost of treating all economies could lose a total of US$2.36 billion because of DF leishmaniasis patients would far exceed their total health and DHF. Additional economic losses are expected because of budgets. For a WHO-recommended course of pentavalent anti- the impact of dengue outbreaks on tourism. monials, current drug costs per patient are US$150 for sodium stibogluconate, US$120 for meglumine antimoniate, and Social Impact. During the transmission season, parents famil- US$30 for generic sodium stibogluconate. Cases not responding iar with DHF are anxious about their children's survival and to antimonials may require second-line, more toxic drugs, such the financial consequences of emergency medical care. In as amphotericin B at a cost of US$60 or pentamidine at a cost of Cambodia, even relatively modest out-of-pocket health expen- US$70. Less toxic amphotericin in liposomes is effective, but ditures can lead to debt and poverty (Van Damme and others costs US$1,500 per patient. The first oral drug for VL, miltefos- 2004). The psychological burden is poorly understood and ine, currently costs US$120 per patient. warrants further study. In addition to drug costs, the additional costs of drug deliv- Even though dengue affects all strata of society, it may ery can be high, especially for patients in remote areas. Patients selectively affect the poorest. Most larval habitats in dengue- often live far from a treatment center, and the expense of hos- endemic communities are artificial: water storage containers, pitalization may lead to interrupted treatment, facilitating flower vases, discarded food containers, used tires, and habitats resistance and requiring additional therapy. Without treatment, created by poor design of roof gutters and drains. Local vector severe leishmaniasis can become chronic and debilitating, inca- ecology is largely determined by community social and cultural pacitating patients and making them unable to work and vul- practices and infrastructure, and increasing urbanization typi- nerable to poverty, malnutrition, and secondary infections. cally attracts the poor to periurban settlements with deficient water supplies. Studies in the República Bolivariana de Social Impact. Even self-limiting CL can leave disfiguring Venezuela (Barrera, Avila, and Gonzalez-Teller 1993; Barrera scars, which have associated stigma and may affect marriage and others 1995) and in Thailand (Nagao and others 2003) prospects. CL can be disabling when lesions are numerous, and have shown higher Ae. aegypti infestations where water distri- the most severe form, recidivans leishmaniasis, is difficult to bution systems are deficient or unreliable. Along the border treat, long-lasting, and disfiguring. In individuals with a defec- between Mexico and the United States, Reiter and others tive cell-mediated immune response, the disseminated lesions (2003) attribute the low dengue seroprevalence reported in of diffuse cutaneous leishmaniasis resemble those of leprosy. Laredo, Texas, to factors such as air conditioning that limited They do not heal spontaneously and frequently recur after Tropical Diseases Lacking Adequate Control Measures: Dengue, Leishmaniasis, and African Trypanosomiasis | 455 treatment. Diffuse cutaneous leishmaniasis is recognized as a costs. Even though WHO now provides specific first-line special public health problem, both clinically and because of its drugs at no cost in excess of delivery and administration, severe emotional consequences. hospitalization and treatment are expensive. In the DRC, The lesions of mucocutaneous leishmaniasis can cause Gouteux and others (1987) estimate that total costs are extensive destruction and distortion of oronasal and pharyn- equivalent to at least 25 percent of a year's income from geal cavities, leading to mutilation of the face. Patients may be agriculture. shunned and, in severe cases, even incarcerated. Although mainly associated with L. braziliensis and L. guyanensis in the Social Impact. The importance of sleeping sickness lies not Americas, mucocutaneous leishmaniasis has been reported in only in the number of new cases reported, but also in its Africa, Asia, and Europe as a complication of L. donovani and potential for epidemic outbreaks causing thousands of deaths: L. major infections and, in immunosuppressed patients, of during recent epidemics in the DRC, in some villages up to L. infantum. 70 percent of the population became infected. Because of the Untreated VL is usually fatal. Even after recovery, patients severity of the disease, one case can affect all family members, may develop a chronic form of CL that usually requires pro- placing a burden on the whole community, reducing the labor longed and expensive therapy. force, interrupting agricultural activities, and jeopardizing food security. Although untreated trypanosomiasis is lethal, African Trypanosomiasis treated patients often remain incapacitated, perpetuating the cycle of poverty, malnutrition, and disease. DALYs do not take WHO estimates the number of deaths caused by sleeping sick- into account the psychosocial impact and the "minor" disabil- ness as 48,000 in 2001 (table 23.1), although current estimates ities. In adults, loss of memory and ability to concentrate is are in the range of 50,000 to 100,000 per year. Men are affected common. Such disabilities are often accompanied by reading at nearly twice the rate of women. In relation to mortality, of all and writing difficulties and occasionally by extreme incoher- parasitic diseases in Sub-Saharan Africa, trypanosomiasis ranks ence. These disabilities greatly affect everyday life, particularly behind only malaria. As concerns DALYs, the health burden is for those school-age children who, even after successful treat- similar to that of schistosomiasis. In Sudan, 33 DALYs on aver- ment, do not recover fully and cannot pursue their studies age are lost because of each premature death from sleeping (Frézil 1983). sickness (McFarland 2003). In Uganda, 23 DALYs were lost per death, but only 0.21 DALYs per successfully treated individual (Odiit and others 2004). Underreporting makes deriving esti- Burden of Animal Trypanosomiasis. Animal trypanosomi- mates for the whole continent difficult: 100,000 deaths per year asis constrains agricultural production--in particular, the use would imply more than 2 million DALYs lost, compared with of draft power. Cattle infected with T.b. brucei, T. vivax, or T. the WHO estimate of 1.31 million (830,000 for males and congolense quickly succumb to a wasting form of anemia. In 480,000 for females) (table 23.1). many areas with a high tsetse challenge, such as Central Africa, cattle are few or not present at all. Elsewhere, countries invest Economic Impact. Information on age at death indicates that an estimated total of US$30 million to US$50 million per year sleeping sickness mainly affects economically active adults. in some 35 million doses of veterinary trypanocides to prevent Data from Uganda show nearly 25 percent of cases occurring in the disease in livestock (Geerts and Holmes 1998). About those age 20 to 29 and more than 60 percent in those age 10 to 60 percent of the cattle at risk are not treated, and the disease is 39 (Odiit 2003). Thus, when people become ill, their families thought to kill about 1 million a year. Current drugs are more not only become burdened with the care of seriously ill indi- than 40 years old, and drug resistance is increasing, as are viduals but also often lose their breadwinners. Poor diagnostic problems of drug availability and accessibility, counterfeit support in many areas means that families often invest in a drugs, and drug mismanagement (Geerts and Holmes 1998). number of treatments that have no effect on the disease. In a Constraints on draft power mean that farmers can till only T.b. rhodesiense area of Uganda, Odiit and others (2004) find small plots, making subsistence farmers extremely vulnerable that some patients made up to seven visits to health facilities to food shortages. Milk yields are lower in infected cows, and before being correctly diagnosed, with just under three- animal trypanosomiasis lowers fertility and increases mortality, quarters initially being diagnosed with malaria. For the 11 of 12 thereby constraining the overall growth rate of the number of who were never diagnosed or were told that they had a differ- livestock (Swallow 2000). Kristjanson and others (1999) esti- ent fatal disease, the costs to and burdens on their families can mate annual direct losses of US$1.3 billion per year as a result only be imagined. of lowered production of milk and meat, with aggregate agri- In addition to the economic losses caused by interruption cultural losses attributable to trypanosomiasis estimated at of their work, sleeping sickness patients face direct financial US$4.5 billion per year. 456 | Disease Control Priorities in Developing Countries | Pierre Cattand, Phillippe Desjeux, M. G. Guzmán, and others MANAGEMENT AND CONTROL STRATEGIES Environmental management is generally considered the core component of dengue prevention and control, including Dengue cleanup campaigns, regular emptying and cleaning of contain- Patient Management and Treatment. Classic dengue fever is ers, installation of water supply systems, solid waste manage- generally self-limiting. No specific treatment is available, but ment, and urban planning. However, huge investments in supportive treatment must be given, including fluid replace- infrastructure are needed to increase access to safe and reliable ment when necessary. Early recognition of DHF cases-- water supplies, to provide solid waste management services, indicated by intense, continuous abdominal pain; persistent and to dispose of liquid waste. In addition to overall health vomiting; and restlessness or lethargy--and early supportive gains, such provision would have a major effect on vector ecol- treatment are of utmost importance to reduce case-fatality ogy, although the relationship is not invariably an inverse one. rates (Martinez 1992). For differential diagnosis, a wide spec- Cost-recovery mechanisms, such as the introduction of trum of viral and bacterial infections should be considered, metered water, may encourage household collection and stor- especially leptospirosis, malaria, yellow fever, chikungunya age of roof catchment rainwater that can be harvested at no virus, rubella, and influenza. cost. Although unproven, the installation of community water services in rural townships and villages may be contributing to the rural spread of dengue in Southeast Asia and elsewhere. Vector Control. Although good patient management can be At the household and community levels, where most vector effective for individual cases, currently no alternative to vector control efforts are centered, increasing attention is given to control is available for the prevention of dengue. Most endemic such activities as covering or frequently cleaning water storage countries have a vector control component in their programs; vessels, removing discarded food and beverage containers, and however, the application of vector control measures is fre- storing or disposing of used tires in such a way that they do not quently insufficient, ineffective, or both and is currently failing collect rainwater. Such tasks would seem to be simple and well to reduce the public health burden to an acceptable level. suited to engagement by communities, but with a few excep- Most Aedes control programs rely on the application of lar- tions, achievements to date have been unspectacular. vicides and adulticidal insecticide space sprays (Zaim and Nevertheless, such community-based interventions are widely Jambulingham 2004). Because Ae. aegypti characteristically seen as the most promising way of achieving sustainable con- breeds in water that does not contain high levels of organic trol through behavior change (Parks and Lloyd 2004). pollutants, control measures typically must be applied to water stored for household purposes, including drinking water. WHO currently approves five insecticides for application to Leishmaniasis potable water (FAO 1999; WHO 1991). Since the early 1970s, Leishmaniasis control is based primarily on finding and treat- the organophosphate temephos has been the most widely ing cases, combined where feasible with vector control and, in used, but increasing levels of resistance to this insecticide are some zoonotic foci, control of animal reservoirs. reducing the duration of effectiveness of treatments in some countries (Brengues and others 2003; Lima and others 2003; Diagnosis and Treatment. For VL, serological diagnosis is Rodriguez and others 2001). An additional challenge is the usually based on the enzyme-linked immunosorbent assay growing objection among householders, particularly in Latin (ELISA), indirect fluorescent antibody tests, and direct aggluti- America, to the application of chemicals to drinking water. nation tests, including a new direct agglutination test kit using Biological control agents, including larvivorous fish and lyophilized antigen, which avoids the need for refrigeration copepods, have had a demonstrable role in integrated control (Schallig and others 2001). A dipstick test based on a highly of Ae. aegypti, but operational difficulties--particularly a lack specific recombinant antigen is also available, together with a of facilities and of expertise in mass rearing and the need for latex agglutination test that can be used to detect antigens in repeated introduction of these agents into some container urine (Attar and others 2001; Sundar and others 1998). habitats--have largely precluded their widespread use. One Parasitological diagnosis relies on microscopy of aspirates of encouraging exception is Vietnam, where indigenous species of the spleen, bone marrow, and lymph nodes. predatory copepods are increasingly used to control semiper- Specific treatment includes the first-line drugs, which are manent larval habitats of Ae. aegypti (Kay and others 2002; pentavalent antimonials (sodium stibogluconate and meglu- Nam and others 2000). However, some communities have mine antimoniate), and the second-line drugs, which are strong cultural objections to the introduction of live animals amphotericin B and AmBisome (amphotericin B in lipo- into household water storage containers--for example, in somes). Miltefosine for VL was registered in India in 2002, and Thailand, where bathing with water that contains small fish or aminosidine (paromomycin) has just completed phase 3 clin- other creatures is widely regarded as unacceptable. ical trials and follow-up. For CL, parasitological diagnosis is Tropical Diseases Lacking Adequate Control Measures: Dengue, Leishmaniasis, and African Trypanosomiasis | 457 made from skin smears followed by treatment with pentava- Sleeping sickness is fatal if untreated. No vaccination lent antimonials. Treatment is given locally if lesions are few exists. Specific drugs are currently available free through and relatively small, or systemically if lesions are more numer- WHO. Pentamidine is used to treat early-stage T.b. gambiense ous. For mucocutaneous leishmaniasis, diagnosis relies on infection, and suramine is used for early-stage T.b. serology because patients generally develop a strong humoral rhodesiense. The organoarsenical compound melarsoprol response. (Arsobal) is used for the late stages of both. Eflornithine has been introduced to treat late-stage T.b. gambiense but is Vector and Reservoir Control. In foci of peridomestic or difficult to administer. Nifurtimox, although not yet regis- intradomestic transmission, vector control can be carried out tered for the treatment of sleeping sickness, has been used on by indoor residual spraying using pyrethroid insecticides. compassionate grounds to treat patients unresponsive to Individual protection using pyrethroid-impregnated bednets is melarsoprol. also used in some areas. In zoonotic foci of VL, control has also included culling stray dogs--and pet dogs if found to be Vector Control. A wide range of techniques for tsetse control infected--although this practice is often poorly accepted by is available (Maudlin, Holmes, and Miles 2004). Most current communities and is probably of limited effectiveness. Trials approaches exploit the acute susceptibility of tsetse flies to with insecticide-treated dog collars are showing some promise biodegradable pyrethroid insecticides. Spraying can be applied as an alternative way to reduce the peridomestic reservoir of from the ground to known fly resting sites or at ultra-low infection (Mazloumi Gavgani and others 2002). For zoonotic volume from the air. Spraying is carried out sequentially to CL, rodent reservoirs can be controlled using poisoned bait and kill all flies initially present and thereafter to kill each environmental management, including physical destruction of generation of newly emerging flies. The sequential aerosol rodents' burrows. technique uses extremely low levels of insecticide and has been effective in Botswana, Somalia, South Africa, and Zambia. It is also useful against epidemic outbreaks of sleeping sickness, African Trypanosomiasis where a rapid cessation in contact between humans and tsetse For human trypanosomiasis, control consists primarily of flies is needed. active and passive case finding and treatment, occasionally Tsetse flies can also be controlled using traps and targets. associated with vector control operations. Dissemination of Targets are combinations of cloth and netting baited with an sleeping sickness can be prevented by regular surveillance of odor attractant and impregnated with a pyrethroid insecticide. the population at risk, including diagnosis and treatment; con- Traps work on the same principle, but the fly is encouraged to trol of the tsetse-fly population can affect the transmission of enter a net or plastic chamber where it remains trapped. Live sleeping sickness as well as of animal trypanosomiasis. In T.b. bait techniques are also used. Cattle are treated with a veteri- rhodesiense foci, where cattle are reservoirs of the disease, treat- nary formulation of pyrethroid insecticides applied as sprays or ing cattle with trypanocides is being investigated as an addi- pour-ons, which kill both tsetse flies and ticks. This technique tional approach to controlling outbreaks. has been successfully used in Burkina Faso, Ethiopia, Kenya, Tanzania, Zambia, and Zimbabwe (Kuzoe and Schofield 2005). Case Finding and Treatment. No single clinical sign is regarded The sterile insect technique involves mass release of steril- as pathognomonic for sleeping sickness. Tests have been devel- ized male tsetse flies, which compete with local males to mate oped to detect antibodies, circulating antigens, or trypanoso- with females. Because female tsetse flies generally mate only mal DNA, but all require parasitological confirmation. For once, the result is infertile offspring and a decline of the wild mass screening, infection can be confirmed by the card agglu- tsetse population. This technique is expensive, because it tination trypanosomiasis test, which is easy to perform and rel- requires large-scale rearing of flies, and it is only recommended atively inexpensive. Parasitological confirmation is by for use once the wild tsetse population has been suppressed to microscopy of lymph node aspirates and of thin or thick blood low levels using other techniques. A combination of insecticide films. Concentration methods increase sensitivity. The most spraying and trap deployment followed by the sterile insect sensitive is the miniature anion exchange centrifugation tech- technique has been successfully used to eliminate G. austeni nique. The capillary tube centrifugation technique is less sensi- from Zanzibar (Vreysen and others 2000). tive but is commonly used in the field because of its ease and Degrees of resistance to trypanosome infection are found in rapidity of use and its low cost. the N'dama, Dwarf, and Savannah Shorthorn breeds in West Determining the stage of disease is essential, because early- Africa and, to a lesser extent, in some Orma Boran breeds and late-stage infections require different treatments. The cri- in East Africa. However, even though these cattle show toler- teria for late-stage infection are based on cerebrospinal fluid ance, can control parasitemia, and resist development of ane- analysis. mia, they can ultimately succumb to the disease. 458 | Disease Control Priorities in Developing Countries | Pierre Cattand, Phillippe Desjeux, M. G. Guzmán, and others COSTS AND COST-EFFECTIVENESS agglutination test using freeze-dried antigen, and US$1.50 for OF INTERVENTIONS the urine latex agglutination text. These tests can be used in the field. A study in Nepal (Pokhrel 1999) comparing outreach case Dengue detection using serology (the dipstick) with parasitological Few studies are available on the cost-effectiveness of vector diagnosis at health centers (bone marrow aspirate) concluded control for reducing dengue transmission. One of the difficul- that the median cost per VL case detected was US$25 in the ties is that the level of vector population control needed to outreach program, compared with US$145 at health centers (of reduce transmission is influenced by the human population's which more than 50 percent was due to absence from work). past exposure to the circulating virus serotype. A direct rela- Treatment costs increased these figures to US$131 and US$200 tionship is apparent between seroprevalence rates and levels per patient, respectively. of vector abundance needed for epidemic transmission. Thus, In India, an examination of the costs of drugs and hospital- the paradox is that, as herd immunity declines over time in ization and of the evolution of the disease under treatment response to effective vector control, progressively lower vector (cure, relapse, failure, intolerance) indicated that the final cost densities can maintain the same level of transmission. of successful treatment depends largely on the basic drug cost, Modeling of the dynamics of dengue transmission is help- which averaged US$86 per patient successfully treated with ing to improve understanding of the interrelationships miltefosine (using reduced pricing because of the large number between virus, vector, and host (Ferguson, Donnelly, and of patients), US$467 for treatment with amphotericin B, and Anderson 1999; Focks and others 1995; Newton and Reiter US$1,613 for treatment with AmBisome. Given current esti- 1992; Shepard 2001), but the absence of epidemiologically mates of about 100,000 cases of VL each year in the state of defined target levels for vector control has hindered calcula- Bihar, India, the estimated total cost of treatment using miltefo- tions of cost-effectiveness. According to Shepard and others sine as a first-line drug and amphotericin B as a second-line (2004), average annual costs for dengue vector control per drug would be about US$11 million, or approximately US$110 1,000 population were US$15 in 1998 in Indonesia, US$81 in per patient (personal communication with P. Olliaro and 1994 and US$188 in 1998 in Thailand, US$240 in 2002 in S. Sundar on treatment options for kalaazar [visceral leishman- Malaysia, and US$2,400 in 2000 in Singapore. In 1997, spend- iasis], 2003). By contrast, analysis of humanitarian relief inter- ing on dengue control in 14 Latin American countries ranged ventions by Médecins sans Frontières­Holland that combined from US$20 to US$3,560 per 1,000 population, with a median case finding with treatment after a VL epidemic in southern of US$260. For 17 Caribbean islands in 1990, the correspon- Sudan indicated total costs of US$394 per patient, or an average ding expenditures ranged from US$140 to US$8,490, with a cost of US$595 per life saved (Griekspoor, Sondorp, and Vos median of US$1,340 (Nathan 1993). By contrast, McConnell 1999). Thus, the average cost per DALY averted was US$18.40. and Gubler's (2003) study in Puerto Rico concludes that larval control programs that achieve a 50 percent reduction in dengue Vector Control. Vector control is rarely carried out as a spe- transmission and cost less than US$2.50 per person would be cific approach to leishmaniasis control, and cost-effectiveness cost-effective in that setting. From research based on analytical estimates are not available. In general, domestic and perido- models (Shepard 2001) and primary data from Singapore, we mestic sandfly vectors are more susceptible to indoor residual estimate the cost of using environmental management for con- spraying than are other domestic vectors, such as anopheline trol at US$3,139 per DALY averted and the cost of using insec- mosquitoes or triatomine bugs, so that transient suppression of ticides at US$1,992 per DALY averted. sandfly populations is seen as an additional benefit of malaria Dengue case management depends on the severity of the ill- or Chagas disease vector control in areas where these vectors ness. Despite the lack of information about cost-effective inter- coincide. However, insecticide-treated bednets, which are ventions to treat dengue cases, Shepard (2001) estimates an becoming widely deployed against malaria transmission, may average cost of US$587 per DALY averted by appropriate case also become cost-effective for reducing leishmaniasis in areas management. Were a dengue vaccine to become available, the of domestic transmission. In Yenice, Turkey, the use of impreg- Shepard model estimates that immunization would cost nated bednets reduced the incidence of CL from 1.90 percent US$3,040 per DALY averted. to 0.04 percent between 2000 and 2001 (Alten and others 2003). Leishmaniasis Case Finding and Treatment. For leishmaniasis, diagnosis African Trypanosomiasis represents a small proportion of the cost of case finding Case Finding and Treatment. WHO (1986, 1998) has ana- and treatment, with diagnostic tests becoming available at lyzed the costs of T.b. gambiense control by means of case approximately US$1.50 for the dipstick, US$3.00 for the direct finding and treatment based on practice in Côte d'Ivoire and Tropical Diseases Lacking Adequate Control Measures: Dengue, Leishmaniasis, and African Trypanosomiasis | 459 Uganda. This work and other studies indicate that, at current studied in more detail in the context of livestock disease prices, the cost of active detection using the card agglutination (Maudlin, Holmes, and Miles 2004). These costs vary accord- trypanosomiasis test with parasitological confirmation varies ing to the technique used and the environmental context, often around US$1 per person screened or slightly more for mobile ignoring overheads for organizing and planning. With that teams. However, mobile teams are more effective in screening a caveat in mind, the figures per square kilometer cited for local high proportion of the population and are also more successful tsetse-fly eradication range from about US$250 to US$550 at in ensuring that a high proportion of patients receive treat- current prices for aerial spraying (based on experience in ment. Unit costs are currently US$0.33 per person for the card Somalia, South Africa, Zambia, and Zimbabwe); US$250 to agglutination trypanosomiasis test and US$2.20 for the minia- US$400 per square kilometer for ground spraying; and US$200 ture anion exchange centrifugation technique. Less sensitive to US$400 per square kilometer for targets. However, the cost parasitological techniques, such as examination of lymph node of traps and targets falls to US$25 to US$60 per square kilome- aspirate or blood smears, cost only a few cents but may miss a ter for control or suppression operations alone. Projects third to half of patients. treating cattle with insecticides have been implemented at costs By contrast, treatment is expensive despite the availability of of US$50 to US$60 per square kilometer. Use of the sterile free drugs. Treatment of early-stage disease incurs costs of insect technique is much more expensive because it relies on more than US$100 per person, rising to more than US$250 for prior suppression of fly populations using another technique. late-stage treatment with melarsoprol and about US$700 with The overall costs of the experimental eradication of G. austeni eflornithine (WHO 1998). The long hospitalization period is a from Zanzibar using the sterile insect and other techniques major component of costs during the second stage, although were about US$3,000 per square kilometer, although the work undertaken by Burri and others (2000) on a shorter International Atomic Energy Agency (IAEA) envisages that melarsoprol regimen offers opportunities for reducing these the cost of the sterile insect technique component could costs. be reduced to less than US$800 per square kilometer as the Despite the costs and the risk of complications, treating technology is developed and applied on a sufficiently sleeping sickness patients in the second stage of the disease is large scale (Dr. Udo Feldmann, IAEA, Vienna, personal cost-effective. In Uganda, costs were less than US$10 per communication). DALY averted for melarsoprol treatment and less than US$20 per person for eflornithine (Politi and others 1995). Similarly, in southern Sudan, the cost per DALY averted ranged from PROBLEMS AND CHALLENGES US$4 to US$22 (Trowbridge and others 2001). Shaw and FOR DISEASE CONTROL Cattand (2001) considered the costs of case finding and treat- ment for T.b. gambiense infection for three delivery options Dengue and a wide range of prevalences. Given the limited informa- Potential for Vaccine Development. The occurrence of DHF tion available on DALYs gained or on the effect on transmis- in children and adults with previous dengue antibodies, sion of reducing the size of the human reservoir, they estimate acquired passively or actively, has been the most important that under different scenarios the costs per DALY averted tend challenge for the development of a dengue vaccine. Lack of a to converge. For most assumptions, the cost per DALY averted suitable animal model, insufficient knowledge of disease fell below a US$25 threshold at prevalences of 0.5 to 1.0 per- pathogenesis, and limited research funding have also had a neg- cent but rose sharply at low prevalences, which explains the ative influence. Researchers generally agree that a dengue vac- reluctance of control programs to invest in screening opera- cine must confer long-lasting protection against the four tions when prevalence is less than 0.2 percent. With better dengue serotypes. Currently, they are following different strate- quantitative understanding of the effects of screening and gies in the development of several vaccine candidates. Some removing patients from the reservoir in preventing future vaccines are currently undergoing human clinical trials--for epidemics, investigators could demonstrate that even at low example, live attenuated dengue and yellow fever chimeric vac- prevalences screening for sleeping sickness is highly cost- cines. The conventional live attenuated vaccines are entering effective. phase 3 trials, while the chimeric vaccines are presently in phase 1 and phase 2 trials. Others are in the preclinical phase of devel- Vector Control. Several countries have undertaken opment. To accelerate the development of a dengue vaccine, a community-based programs to trap tsetse flies, notably Côte new initiative--the pediatric dengue vaccine initiative--has d'Ivoire, where costs came to US$2.30 per person protected been launched, with the major objective of mobilizing per year (Laveissière and others 1994), and Uganda, which resources to accelerate the development of a safe and effective achieved a cost of US$0.50 per person protected per year pediatric dengue vaccine (Almond and others 2002; Halstead (Lancien and Obayi 1993). Vector control costs have been and Deen 2002; Pang 2003). 460 | Disease Control Priorities in Developing Countries | Pierre Cattand, Phillippe Desjeux, M. G. Guzmán, and others Vector Control. Without a vaccine, vector control remains the New Risk Factors. Environmental changes, particularly those only available strategy against dengue. Selective, integrated vec- related to climate, directly affect the incidence and prevalence tor control with community and intersectoral participation, of vectorborne diseases. However, social factors, such as active disease surveillance based on a strong health information lifestyles and population density, particularly in the case of system, emergency preparedness, capacity building and train- dengue, are also important. Using an empirical model of the ing, and vector control research constitute the major elements effect of population and climate change on the global distribu- of WHO's global strategy for dengue prevention and control. tion of dengue fever, Hales and others (2002) conclude that Since the eradication era, few examples of successful dengue predicted changes in humidity will increase the areas with a cli- prevention and control on a national scale are available. mate suitable for dengue transmission. Exceptions include Cuba and Singapore, both island states. The world is also becoming increasingly urbanized: during Cuba, with approximately 11 million inhabitants, has been able 2000­25, Asia's urban population is expected to double, and to interrupt dengue transmission. Despite being in an endemic that of Latin America and the Caribbean is expected to area, the country has maintained low vector densities and has increase by almost 50 percent. The resulting high human pop- successfully controlled epidemics in recent years (Arias 2002). ulation densities, coupled with lifestyles that increasingly Critical factors contributing to this achievement are the strong contribute to the proliferation of larval habitats and infra- dengue surveillance system, which integrates environmental, structural deficiencies in relation to water supply and sanita- entomological, epidemiological, clinical, and virological tion, are such that effective delivery of vector control on the surveillance in conjunction with the public health infrastruc- scale needed is beyond the reach of many governments. ture, and a strong vector control program, along with good The increasing global trend in international travel also facili- intersectoral coordination, active community involvement, and tates the dissemination of virus serotypes and strains between strong political commitment. vulnerable populations. A limited array of tools is available for vector control inter- Genetic variability is another element to be considered. The ventions, any one of which can control at least part of the vec- genetic diversity of the viruses is increasing, with some geno- tor population or provide personal protection. However, types associated with severe disease (Cologna and Rico-Hesse approaches are converging, at least at the policy level, toward 2003; Leitmeyer and others 1999; Rico-Hesse and others 1997). application of vector control tools through social or commu- Recombination has been demonstrated in all four serotypes, nity mobilization. Consensus is growing that community- but the implications in terms of pathogenesis are unknown. In based approaches are desirable and necessary, and many believe addition to recombination, mutations, gene flow, and other that only through such approaches can a degree of sustainabil- factors could further influence the genetic diversity and selec- ity be accomplished in relation to dengue vector control. Even tion of virulent strains (Holmes and Burch 2000). At the same though few such interventions have expanded beyond the pilot time, in addition to initial observations of the higher risk of stage, the decentralization of budgetary and operational DHF in Caucasian than in those of African descent, a few responsibilities for program delivery appears to offer opportu- reports associate some human leukocyte antigen alleles with nities for strengthening and expanding this integrated vector disease severity (Bravo, Guzmán, and Kouri 1987; LaFleur and management approach. others 2002; Loke and others 2001; Paradoa, Trujillo, and Increasing levels of resistance of Ae. aegypti to temephos Basanta 1987; Stephens and others 2002). The sequence of imply shorter intervals between treatments. This situation is infecting viruses and, more recently, the longer interval already a reality in some countries, including Brazil and several between primary and secondary infection as a risk factor for Caribbean islands (Carvalho and da Silva 1999; Rawlins 1998). DHF, add a new perspective to the problem (Guzmán and Resistance of adult mosquitoes to malathion and to pyrethroids others 2002; Nisalak and others 2003). has been reported in the Americas and in Asia (WHO 1992) and is likely to reduce the efficacy of space spraying. Given the peridomestic ecology of Ae. aegypti in most regions and the widespread use of pyrethroids for public health purposes and in Leishmaniasis household insecticide products, the rate of development of Because the primary control strategy against leishmaniasis is pyrethroid resistance is likely to accelerate.At the same time, few based on case finding and treatment, the priority for control is new insecticide products are becoming available in the public developing and implementing improved diagnostic methods health market because of the costs involved in development and better treatments that are more amenable to field use.A par- and registration compared with the returns on investment allel requirement is for the development of more cost-effective from the relatively small commercial market. The high cost of drug delivery systems, especially ones that take advantage of re-registration of existing products is also contributing to the new oral drugs, such as miltefosine, and the planned registration withdrawal of some insecticides from the market. and local production of aminosidine in India. Tropical Diseases Lacking Adequate Control Measures: Dengue, Leishmaniasis, and African Trypanosomiasis | 461 Improved Diagnostics and Treatment. Even though the Parasitological diagnosis, such as the miniature anion exchange new serological tests, such as the dipstick, lyophilized direct centrifugation test, is more expensive and complicated to use in agglutination test kit, and latex agglutination urine test, repre- field surveys despite the development of a simplified sterile kit sent major improvements, they are not yet widely used in version. Molecular diagnostics are not yet developed to a level endemic countries. Moreover, they are indirect tests that can- appropriate for widespread field use. not provide direct parasitological confirmation of infection or of cure immediately following treatment. Current parasitolog- Treatment. Despite the availability of drugs free of charge ical tests tend to be highly invasive and can be costly to per- from WHO, treatment is hampered by the length of hospital- form; therefore, simple molecular-based tests would be an ization required and by the toxicity of currently available drugs. advantage. In addition, inability to use the same drug in the early and late For leishmaniasis treatment, development of the oral VL stages of the disease complicates the treatment protocol. The drug, miltefosine, represents a substantial improvement, but it existing late-stage drug, melarsoprol, is unsafe, its secondary remains expensive and with a long treatment regimen and can- reactions are numerous, and the occurrence of a lethal not be administered to pregnant women because of the risk of encephalopathic syndrome in 5 to 10 percent of treated cases teratogenicity. Further clinical evaluation is required to estab- means that patients on melarsoprol must be hospitalized. A lish the possibility of shorter treatment regimes and the poten- new oral drug for the early stage, soon to be registered, must be tial of combination therapy to inhibit the development and introduced in the field, which will take several years. spread of drug resistance. Another oral drug, sitamaquine, is Drug resistance is well established in animal trypanosomes. currently under development and will require similar clinical For T. congolense, T. vivax, and T. evansi, resistance to all and implementation studies. Clinical trials of aminosidine available drugs has been reported, along with trypanosome (paromomycin) are proceeding, and use of this drug against VL populations with multiple drug resistance (Geerts and Holmes may become widespread. An improved understanding of dis- 1998). Much less information is available on human path- ease pathogenesis would be helpful in refining the criteria for ogenic trypanosomes. The resurgence of human African cure and in improving patients' prognosis. trypanosomiasis in recent years has been accompanied by increasing reports of treatment failure using melarsoprol. As Vaccine Development. The leishmaniases offer substantial early as 1960, T.b. rhodesiense patients in Uganda were reported opportunities for vaccine development, and a crude vaccine to have relapsed after two or more courses of melarsoprol, and against CL has been widely used in parts of the Middle East. in 1977, a 40 percent melarsoprol relapse rate was reported in Trials of a second-generation vaccine that includes three the DRC. Leishmania antigens are currently in progress. Relapses after treatment with early-stage drugs remain rare. Whether relapses after melarsoprol treatment reflect parasite Vector Control. Control of domestic and peridomestic sand- drug resistance or host factors is unknown. Furthermore, even fly vectors will probably continue as an additional benefit though increasing rates of melarsoprol failure have been accruing to programs against other insect vectors using indoor observed in several countries, the magnitude and geographic residual spraying or insecticide-treated bednets. However, in distribution of the problem have not been determined. areas where dogs are among the main reservoir hosts, increased Analyses of existing data are complicated by the lack of a use of insecticide-treated dog collars would merit further standard treatment regimen and the range of clinical and appraisal. Such collars not only would reduce the likelihood of laboratory criteria used to define a relapse. new infections in dogs, but also could reduce the risk of trans- mission from dogs to humans. Vector Control. Even though available techniques to control tsetse flies can be highly effective (Maudlin, Holmes, and Miles 2004), all are constrained by the difficulties of applying them African Trypanosomiasis on a large enough scale for long enough to achieve sustainable Improved Diagnostics. Serological diagnosis is reliable for results. Insecticide spraying is efficient but is difficult to sustain verifying infection; however, most district hospitals or periph- because of logistical constraints and high costs. Targets and eral health units have neither the facilities nor the necessary traps are effective, but their deployment is difficult to sustain expertise to perform and read serological tests. In the past, for long periods, and implementation through community serological diagnosis, based on indirect fluorescent antibody participation requires constant motivation and supervision to tests and ELISA, was restricted to central-level facilities or remain effective. To address these problems, the African Union specialized mobile teams. The card agglutination trypanosomi- has launched the Pan African Initiative (PATTEC), which asis test has substantially simplified the use of serology but focuses on identifying regions where elimination of the tsetse requires specifically equipped health units with a cold chain. fly may be feasible using currently available techniques. This 462 | Disease Control Priorities in Developing Countries | Pierre Cattand, Phillippe Desjeux, M. G. Guzmán, and others initiative is designed as part of a poverty reduction strategy that epidemic outbreaks rather than instigating emergency meas- aims at eliminating the problem of tsetseborne animal try- ures after an outbreak is in full crudescence. Moreover, because panosomiasis, but in several areas it will also reduce the risk of preemptive measures and emergency responses are competing human infections. strategies, analyses of their relative cost-effectiveness would be appropriate. SUMMARY Case finding and treatment for the leishmaniases and African trypanosomiases depend on the effectiveness of the For dengue, leishmaniasis, and African trypanosomiasis, the diagnostic and treatment packages. Such packages are available, longstanding problem is the lack of adequate specific treat- and research is required into the most cost-effective means for ment. For dengue, no specific treatment is available. For the large-scale implementation. Again, the management exercise is leishmaniases and African trypanosomiases, specific treatment to accept that some transmission will occur but to be aware that has long depended on antiquated drugs that would be consid- cases can be found and treated with minimal losses to healthy ered far too toxic for introduction under modern registration life. As with dengue, predictive surveillance will help focus systems. Even though progress is being made, especially in rela- attention on those areas where outbreaks seem most likely, and tion to the development of new oral drugs for leishmaniasis, in rapid, accurate diagnostics are crucial both to avoid the waste purely pragmatic terms what is currently available will and danger of mistreatment and to minimize delays in admin- probably represent almost the entire therapeutic arsenal for the istering the specific treatment required. But should such coming decades. Even without toxicological problems, the approaches rely on health centers, on mobile teams, or on some development and registration of a new candidate drug will, combination of the two? To what degree can the specialist diag- given current requirements, take at least a decade. nosis and treatment teams be integrated into more general Although basic research will continue (table 23.2), the cur- approaches to health care? And, most crucially, how is the epi- rent challenge is to make better use of what is already available. demiological surveillance to be organized: disease and vector Dengue can be prevented with available vector control tools notification, geographic information system mapping, analysis, and strategies designed to reduce the risk of transmission. This and prediction? method requires a sustainable surveillance system capable of For the leishmaniases, vector control seems unlikely to providing early warning and predictions based on experience become a major component of disease control except where of factors predisposing to new epidemic outbreaks. To a large sandfly distribution overlaps with that of other vectors or extent, it becomes a management exercise that accepts that where use of personal protection measures can be more widely some dengue transmission will occur but aims at preempting encouraged. For dengue, vector control is a major component, Table 23.2 Control Strategies, Major Challenges, and Research Needs of Each Disease Disease Principal control strategy Major problems and challenges Major research needs Dengue Interruption of transmission Lifestyles that provide abundant Vector control thresholds to interrupt through vector control artificial larval habitats transmission Patient management and Urban water infrastructure Behavioral changes conducive to supportive treatment and management dengue prevention Sustainable vector control Pathogenesis and disease prognosis Early diagnosis and treatment Vaccine development Vaccine development Leishmaniasis Case detection and treatment Rapid field diagnosis Molecular-based field diagnostic kits Deployment of oral drugs for treatment Effective, safe, oral drug treatments, Vaccine development including combination therapy Diagnosis and treatment strategies African Case detection and Rapid field diagnosis, including determination Molecular-based field diagnostic kits trypanosomiasis treatment of stage of infection Effective, safe, oral drug treatments Interruption of transmission Safer drug treatment regimens Vector population genetics to determine through vector control Sustainable large-scale vector control areas amenable to vector population elimination Source: Authors. Tropical Diseases Lacking Adequate Control Measures: Dengue, Leishmaniasis, and African Trypanosomiasis | 463 but unless Aedes eradication appears again on the agenda, pre- from Mexico and the United States." American Journal of Tropical dicting the levels of control required in specific situations will Medicine and Hygiene 67: 85­92. require much greater understanding of transmission dynamics. Bravo, J. R., M. G. Guzmán, and G. P. Kouri. 1987. "Why Dengue Haemorrhagic Fever in Cuba? 1: Individual Risk Factors for Dengue Significant resources have been wasted on emergency dengue Haemorrhagic Fever/Dengue Shock Syndrome (DHF/DSS)." vector control, which has subsequently been seen to have had Transactions of the Royal Society of Tropical Medicine and Hygiene 81: little more than a palliative effect, whereas sustained suppres- 816­20. sion of vector populations may require changes in urban water Brengues, C., N. J. Hawkes, F. Chandre, L. McCarroll, S. Duchon, P. Guillet, management and in human behavior that exceed the usual and others. 2003. "Pyrethroid and DDT Cross-Resistance in Aedes aegypti Is Correlated with Novel Mutations in the Voltage-Gated remit of health specialists. Sodium Channel Gene." Medical and Veterinary Entomology 17: 87­94. For African trypanosomiasis, however, the prospects for Burri, C., S. Nkunku, A. Merolle, T. Smith, J. Blum, and R. Brun. 2000. sustainable vector control are more promising. The vector's low "Efficacy of New, Concise Schedule for Melarsoprol in Treatment of reproductive rate, combined with its extreme sensitivity to Sleeping Sickness Caused by Trypanosoma brucei gambiense: A Randomised Trial." Lancet 355: 1419­25. ultra-low doses of biodegradable insecticides, put tsetse flies Carvalho, A. de F., and I. G. da Silva. 1999. "Atividade larvicida do among the most promising candidates for large-scale elimina- temephos a 1% sobre o Aedes aegypti (Lin., 1762), em diferentes cri- tion. Campaigns against tsetse flies during the past century adouros artificiais." Revista de Patologia Tropical 28: 211­32. were invariably successful until they were discontinued and the Cattand, P., J. Jannin, and P. Lucas. 2001. "Sleeping Sickness Surveillance: controlled areas became reinvaded. Thus, the operational issue An Essential Step towards Elimination." Tropical Medicine and is to design large-scale international programs that can succes- International Health 6: 348­61. sively eliminate tsetse populations and prevent reinvasion of Cologna, R., and R. Rico-Hesse. 2003. "American Genotype Structures Decrease Dengue Virus Output from Human Monocytes and controlled areas, as contemplated by the African Union's Pan Dendritic Cells." Journal of Virology 77: 3929­38. African Initiative. Cref, B. J., T. C. Jones, R. Badar, D. Sampaio, R. Teixeira, and W. D. J. In essence, all three diseases face parallel needs involving Johnson. 1987. "Malnutrition as a Risk Factor for Severe Visceral some marginal improvements to existing control techniques, Leishmaniasis." Journal of Infectious Diseases 156: 1030­33. but, most important, they require a management exercise that De Raadt, P. 1985. "Trypanosomes et leishmanioses congénitales." Archives acknowledges the long-term need for surveillance, adequate Françaises de Pédiatrie 42: 925­27. reporting, case finding, and treatment. The primary challenges Desjeux, P. 1996. "Leishmaniasis: Public Health Aspects and Control." Clinics in Dermatology 14: 417­23. seem to reside less in the domain of new tools and more in the ------. 2001. "The Increase in Risk Factors for the Leishmaniases deployment of what is already available. Worldwide." Transactions of the Royal Society of Tropical Medicine and Hygiene 95: 239­43. Desjeux, P., and J. Alvar. 2003. "Leishmania/HIV Co-infections: REFERENCES Epidemiology in Europe." Annals of Tropical Medicine and Parasitology 97 (Suppl. 1): S3­15. Almond, J., J. Clemens, H. Engers, S. B. Halstead, H. B. Khiem, A. Pablos- Diaz, A., G. Kouri, M. G. Guzmán, L. Lobaina, J. Bravo, A. Ruiz, and oth- Mendez, and others. 2002. "Accelerating the Development and ers. 1988. "Description of the Clinical Picture of Dengue Hemorrhagic Introduction of a Dengue Vaccine for Poor Children, 5­8 December Fever/Dengue Shock Syndrome (DHF/DSS) in Adults." Bulletin of the 2001, Ho Chi Minh City, Vietnam." Vaccine 20: 3043­46. Pan American Health Organization 22: 133­44. Alten, B., S. Çaglar, S. Kaynas¸, and F. M. S¸ims¸ek. 2003. "Evaluation of FAO (Food and Agriculture Organization of the United Nations). 1999. Protective Efficacy of K-OTAB Impregnated Bednets for Cutaneous Pesticide Residues in Food, 1999: Report of the Joint Meeting of the FAO Leishmaniasis Control in Southeast Anatolia, Turkey." Journal of Vector Panel of Experts on Pesticide Residues in Food and the Environment and Ecology 28 (1): 53­64. the WHO Core Assessment Group on Pesticide Residues. Paper 153. Arias, J. 2002."El Dengue en Cuba." Revista Panamericana de Salud Pública Rome: FAO Plant Production and Protection. 11: 221­22. Ferguson, N. M., C. A. Donnelly, and R. M. Anderson. 1999."Transmission Attar, A. J., M. L. Chance, S. El-Safi, J. Carney, A. Azazy, M. El-Hadi, and Dynamics and Epidemiology of Dengue: Insights from Age-Stratified others. 2001. "Latex Agglutination Test for the Detection of Urinary Sero-Prevalence Surveys." Philosophical Transactions of the Royal Antigens in Visceral Leishmaniasis." Acta Tropica 78 (1): 11­16. Society of London, Series B, Biological Sciences 354: 757­68. Barrera, R., J. Avila, and S. Gonzalez-Teller. 1993. "Unreliable Supply of Fèvre, E. M., P. G. Coleman, M. Odiit, S. C. Welburn, and M. E. J. Potable Water and Elevated Aedes aegypti Larval Indices: A Causal Woolhouse. 2001. "The Origins of a New Trypanosoma brucei rhode- Relationship?" Journal of the American Mosquito Control Association 9: siense Sleeping Sickness Outbreak in Eastern Uganda." Lancet 358: 189­95. 625­28. Barrera, R., J. C. Navarro, J. D. Mora, D. Dominguez, and J. Gonzalez. 1995. Focks, D. A., E. Daniels, D. G. Haile, and J. E. Keesling. 1995. "A Simulation "Public Service Deficiencies and Aedes aegypti Breeding Sites in Model of the Epidemiology of Urban Dengue Fever: Literature Venezuela." Bulletin of the Pan American Health Organization 29: Analysis, Model Development, Preliminary Validation, and Samples 193­205. of Simulation Results." American Journal of Tropical Medicine and Bennett, K. E., K. E. Olson, M. de Lourdes Munoz, I. Fernandez-Salas, Hygiene 53: 489­506. J. A. Farfan-Ale, S. Higgs, and others. 2002. "Variation in Vector Frézil, J. L. 1983. La trypanosomiase humaine en République Populaire du Competence for Dengue 2 Virus among 24 Collections of Aedes aegypti Congo. Travaux et documents de l'ORSTOM 155, Paris, ORSTROM. 464 | Disease Control Priorities in Developing Countries | Pierre Cattand, Phillippe Desjeux, M. G. Guzmán, and others Geerts, S., and P. H. Holmes. 1998. Drug Management and Parasite LaFleur, C., J. Granados, G. Vargas-Alarcon, J. Ruiz-Morales, C. Villareal- Resistance in Bovine Trypanosomiasis in Africa. Programme against Garza, L. Higuera, and others. 2002. "HLA-DR Antigen Frequencies in African Trypanosomiasis Technical and Scientific Series 1. Rome: Food Mexican Patients with Dengue Virus Infection: HLA-DR4 as a Possible and Agriculture Organization of the United Nations. Genetic Resistance Factor for Dengue Hemorrhagic Fever." Human Gouteux, J. P., P. Bansimba, F. Noireau, and J. L. Frezil. 1987. "Le coût du Immunology 63: 1039­44. traitement individuel de la trypanosomiase à T. b. gambiense dans le Lancien, J., and H. Obayi. 1993."La lutte contre les vecteurs de la maladie du foyer du niari (Congo)." Médecine Tropicale 47: 61­63. sommeil." Bulletin de la Société Française de Parasitiologie 11: 107­17. Griekspoor, A., E. Sondorp, and T. Vos. 1999. "Cost-Effectiveness Analysis Laveissière, C., O. Grébuat, J. J. Lemasson, A. H. Meda, D. Couret, F. Doua, of Humanitarian Relief Interventions: Visceral Leishmaniasis and others. 1994. Les communautés rurales et la lutte contre la maladie Treatment in the Sudan." Health Policy and Planning 14: 70­76. du sommeil en forêt de Côte d'Ivoire. OCCGE-WHO/TRY/94.1. Geneva: Gubler, D. J. 1979. "Variation in Susceptibility to Oral Infection with World Health Organization. Dengue Viruses among Geographic Strains of Aedes aegypti." American Leitmeyer, K. C., D. W. Vaughn, D. M. Watts, R. Salas, I. Villalobos, C. Journal of Tropical Medicine and Hygiene 28: 1045­52. Ramos, and R. Rico-Hesse. 1999. "Dengue Virus Structural Differences That Correlate with Pathogenesis." Journal of Virology 73: ------. 2002. "Epidemic Dengue/Dengue Hemorrhagic Fever as a Public 4738­47. Health, Social, and Economic Problem in the 21st Century." Trends in Microbiology 2: 100­3. Libala, K., M. Wery, and J. F. Ruppol. 1978. "Congenital Transmission of Trypanosoma gambiense." Annales de la Société Belge de Médecine Gubler, D. J., and G. Kuno, eds. 1997. Dengue and Dengue Hemorrhagic Tropicale 58: 65­66. Fever. New York: CAB International. Lima, J. B., M. P. Da-Cunha, R. C. Da Silva, A. K. Galardo, S. Soares Sda, Guzmán, M. G., M. Alvarez, R. Rodriguez, D. Rosario, S. Vazquez, I. A. Braga, and others. 2003. "Resistance of Aedes aegypti to L. Valdes, and others. 1999."Fatal Dengue Hemorrhagic Fever in Cuba, Organophosphates in Several Municipalities in the State of Rio de 1997." International Journal of Infectious Diseases 3: 130­35. Janeiro and Espírito Santo, Brazil." American Journal of Tropical Guzmán, M. G., and G. Kouri. 2002. "Dengue: An Update." Lancet Medicine and Hygiene 68: 329­33. Infectious Disease 2: 33­42. Loke, H., D. B. Betchell, C. X. T. Phuong, M. Dung, J. Schneider, N. J. ------. 2003. "Dengue and Dengue Hemorrhagic Fever in the Americas: White, and others. 2001. "Strong HLA Class I-Restricted T Cell Lessons and Challenges." Journal of Clinical Virology 27: 1­13. Responses in Dengue Hemorrhagic Fever: A Double-Edged Sword?" Guzmán, M. G., G. Kouri, L. Valdes, J. Bravo, S. Vazquez, and S. B. Halstead. Journal of Infectious Disease 184: 1369­73. 2002. "Enhanced Severity of Secondary Dengue-2 Infections: Death Martinez, E. 1992. Dengue hemorrágico en criancas. Havana: Editorial José Rates in 1981 and 1997 Cuban Outbreaks." Pan American Journal of Marti. Public Health 11: 223­27. Maudlin, I., P. H. Holmes, and M. A. Miles, eds. 2004. The Hales, S., N. de Wet, J. Maindonald, and A. Woodward. 2002. "Potential Trypanosomiases. Wallingford, U.K.: CABI Publishing. Effect of Population and Climate Changes on Global Distribution of Mazloumi Gavgani, A. S., M. H. Hodjati, H. Mohite, and C. R. Davies. Dengue Fever: An Empirical Model." Lancet 360: 830­34. 2002. "Effect of Insecticide-Impregnated Dog Collars on Incidence of Halstead, S. B., and J. Deen. 2002. "The Future of Dengue Vaccines." Lancet Zoonotic Visceral Leishmaniasis in Iranian Children: A Matched- 360: 1243­45. Cluster Randomised Trial." Lancet 360: 374­79. Harris, E., E. Videa, E. Perez, E. Sandoval, Y. Tellez, M. A. Perez, and others. McConnell, J. K., and D. J. Gubler. 2003. "Guidelines on the Cost- 2000. "Clinical, Epidemiologic, and Virologic Features of Dengue in Effectiveness of Larval Control Programs to Reduce Dengue the 1998 Epidemic in Nicaragua." American Journal of Tropical Transmission in Puerto Rico." Pan American Journal of Public Health Medicine and Hygiene 63: 5­11. 14: 9­16. Holmes, E. C., and S. S. Burch. 2000. "The Causes and Consequences of Meltzer, M. I., J. G. Rigau-Pérez, G. G. Clark, P. Reiter, and D. J. Gubler. Genetic Variation in Dengue Virus." Trends in Microbiology 8: 74­77. 1998. "Using Disability-Adjusted Life Years to Assess the Economic Hull, B., E. Tikasingh, M. de Souza, and R. Martinez. 1984. "Natural Impact of Dengue in Puerto Rico: 1984­1994." American Journal of Transovarial Transmission of Dengue 4 Virus in Aedes aegypti in Tropical Medicine and Hygiene 59: 265­71. Trinidad." American Journal of Tropical Medicine and Hygiene 33: Nagao, Y., U. Thavara, P. Chitnumsup, A. Tawatsin, C. Chansang, and D. 1248­50. Campbell-Lendrum. 2003. "Climatic and Social Risk Factors for Aedes Kay, B. H., V. S. Nam, T. V. Tien, N. T. Yen, T. V. Phong, V. T. Diep, and oth- Infestation in Rural Thailand." Tropical Medicine and International ers. 2002. "Control of Aedes Vectors of Dengue in Three Provinces of Health 8: 650­59. Vietnam by Use of Mesocyclops (Copepoda) and Community-Based Nam, V. S., N. T. Yen, M. Holynska, J. W. Reid, and B. H. Kay. 2000. Methods Validated by Entomologic, Clinical, and Serological "National Progress in Dengue Vector Control in Vietnam: Survey for Surveillance." American Journal of Tropical Medicine and Hygiene 66: Mesocyclops (Copepoda), Micronecta (Corixidae), and Fish as 40­48. Biological Control Agents." American Journal of Tropical Medicine and Kouri, G. P., M. G. Guzmán, and J. R. Bravo. 1987. "Why Dengue Hygiene 62: 5­10. Haemorrhagic Fever in Cuba? 2: An Integral Analysis." Transactions of Nathan, M. B. 1993. "Critical Review of Aedes aegypti Control Programs in the Royal Society of Tropical Medicine and Hygiene 81: 821­23. the Caribbean and Selected Neighboring Countries." Journal of the Kouri, G. P., M. G. Guzmán, J. R. Bravo, and C. Triana. 1989. "Dengue American Mosquito Control Association 9: 1­7. Haemorrhagic Fever/Dengue Shock Syndrome: Lessons from the Newton, E. A. C., and P. Reiter. 1992. "A Model of the Transmission of Cuban Epidemic." Bulletin of the World Health Organization 67: Dengue Fever with an Evaluation of the Impact of Ultra-Low Volume 375­80. (ULV) Insecticide Applications on Dengue Epidemics." American Kristjanson, P. M., B. M. Swallow, G. J. Rowlands, R. L. Kruska, and P. N. Journal of Tropical Medicine and Hygiene 47: 709­20. de Leeuw. 1999. "Measuring the Costs of African Animal Nimmanitya, S. 1993. Clinical Manifestations of Dengue/Dengue Trypanosomosis: The Potential Benefits of Control and Returns to Haemorrhagic Fever. Regional Publication, SEARO 22. New Delhi: Research." Agricultural Systems 59: 79­98. World Health Organization, Regional Office for Southeast Asia. Tropical Diseases Lacking Adequate Control Measures: Dengue, Leishmaniasis, and African Trypanosomiasis | 465 Nisalak, A., T. P. Endy, S. Nimmanitya, S. Kalayanrooj, U. Thiyakorn, R. M. Shepard, D. S. 2001."Modeling in CE Analysis." Brandeis University, Heller Scott, and others. 2003. "Serotype-Specific Dengue Virus Circulation School, Schneider Institute for Health Policy. http://www.sihp. and Dengue Disease in Bangkok, Thailand, from 1973 to 1999." brandeis.edu/Shepard/module-10-8-01.ppt. American Journal of Tropical Medicine and Hygiene 68: 191­202. Shepard, D. S., J. A. Suaya, S. Halstead, M. B. Nathan, D. J. Gubler, R. T. Odiit, M. 2003. "The Epidemiology of Trypanosoma brucei rhodesiense in Mahoney, and others. 2004. "Cost-Effectiveness of a Pediatric Dengue Eastern Uganda." Ph.D. thesis, University of Edinburgh, Scotland. Vaccine." Vaccine 22: 1275­80. Odiit, M., A. Shaw, S. C. Welburn, E. M. Fevre, P. G. Coleman, and J. J. Stephens, H. A. F., R. Klaythong, M. Sirikong, D. W. Vaughn, S. Green, S. McDermott. 2004. "Assessing the Patterns of Health-Seeking Behavior Kalayanarooj, and others. 2002."HLA-A and B Allele Associations with and Awareness among Sleeping-Sickness Patients in Eastern Uganda." Secondary Dengue Virus Infections Correlate with Disease Severity Annals of Tropical Medicine and Parasitology 98: 339­48. and the Infecting Viral Serotype in Ethnic Thais." Tissue Antigens 60: Okanurak, K., S. Sornmani, and K. Indaratna. 1997. "The Cost of Dengue 309­18. Hemorrhagic Fever in Thailand." Southeast Asian Journal of Tropical Swallow, B. M. 2000. Impacts of Trypanosomiasis on African Agriculture. Medicine and Public Health 28: 711­17. Programme against African Trypanosomiasis Technical and Scientific Okoth, J. O. 1986. "Peridomestic Breeding Sites of Glossina fuscipes fuscipes Series 2. Rome: Food and Agriculture Organization of the United Newst. in Busoga, Uganda, and Epidemiological Implications for Nations. Trypanosomiasis." Acta Tropica 43: 283­86. Tardieux, I., O. Poupel, L. Lapchin, and F. Rodhain. 1990. "Variation PAHO (Pan American Health Organization). 1994. Dengue and Dengue among Strains of Aedes aegypti in Susceptibility to Oral Infection with Hemorrhagic Fever in the Americas: Guidelines for Prevention and Dengue Virus Type 2." American Journal of Tropical Medicine and Control. PAHO Scientific Publication 548. Washington, DC: PAHO. Hygiene 43: 308­13. Pang, T. 2003. "Vaccines for the Prevention of Neglected Diseases: Dengue Traub, N., P. Hira, C. Chintu, and C. Mhango. 1978. "Congenital Fever." Current Opinions in Biotechnology 14: 332­36. Trypanosomiasis: Report of a Case Due to Trypanosoma brucei rhode- siense." East African Medical Journal 55: 477­81. Paradoa, M. L., Y. Trujillo, and P. Basanta. 1987. "Association of Dengue Hemorrhagic Fever with the HLA System." Haematologia 20: 83­87. Trowbridge, M., D. McFarland, M. Richer, M. Adeoye, and A. Moore. 2001. "Cost-Effectiveness of Programs for Sleeping Sickness Control: Parks, W., and L. Lloyd. 2004. Planning Social Mobilization and American Society of Tropical Medicine and Hygiene, 49th Annual Communication for Dengue Fever Prevention and Control: A Step-by-Step Meeting, Houston, TX, 2000, Abstract 417." American Journal of Guide. WHO/CDS/WMC/2004.2. Geneva: World Health Organization. Tropical Medicine and Hygiene 62 (Suppl. 3): 312. Pokhrel, S. 1999. "Cost-Effectiveness of Early Case Detection for Visceral Van Damme, W., L.Van Leemput, I. Por, W. Hardeman, and B. Meessen. Leishmaniasis in Nepal," M.Sc. thesis, Chulalongkorn University, 2004. "Out-of-Pocket Health Expenditure and Debt in Poor Thailand. Households: Evidence from Cambodia." Tropical Medicine and Politi, C., G. Carrin, D. Evans, F. A. S. Kuzoe, and P. D. Cattand. 1995. International Health 9: 273­80. "Cost-Effectiveness Analysis of Alternative Treatments of African Vreysen, M. J. B., K. M. Saleh, M. Y. Ali, M. A. Abdullah, Z. R. Zhu, K. G. Gambiense Trypanosomiasis in Uganda." Health Economics 4: 273­87. Juma, and others. 2000. "Glossina austeni (Diptera: Glossinidae) Rawlins, S. C. 1998. "Spatial Distribution of Insecticide Resistance in Eradicated on the Island of Unguja, Zanzibar, Using the Sterile Insect Caribbean Populations of Aedes aegypti and Its Significance." Pan Technique." Journal of Economic Entomology 93: 123­35. American Journal of Public Health 4: 243­51. Watts, D. M., D. S. Burke, B. A. Harrison, R. E. Whitmire, and A. Nisalak. Reiter, P., S. Lathrop, M. Bunning, B. Biggerstaff, D. Singer, T. Tiwari, and 1987. "Effect of Temperature on the Vector Efficiency of Aedes aegypti others. 2003. "Texas Lifestyle Limits Transmission of Dengue Virus." for Dengue 2 Virus." American Journal of Tropical Medicine and Emerging Infectious Diseases 9: 86­89. Hygiene 36: 143­52. Rico-Hesse, R., L. M. Harrison, R. A. Salas, D. Tovar, A. Nisalak, C. Ramos, WHO (World Health Organization). 1986. Epidemiology and Control of and others. 1997. "Origins of Dengue Type 2 Viruses Associated with African Trypanosomiasis. Report of a WHO Expert Committee. Increased Pathogenicity in the Americas." Virology 230: 244­51. Technical Report Series 739. Geneva: WHO. Rochas, G., A. Martins, G. Gama, F. Brandão, and J. Atougia. 2004. ------. 1991. Safe Use of Pesticides. Fourteenth Report of the WHO "Possible Cases of Sexual and Congenital Transmission of Sleeping Expert Committee on Vector Biology and Control. Technical Report Sickness." Lancet 363: 247. Series 813. Geneva: WHO. Rodriguez, M. M., J. Bisset, D. M. de Fernandez, L. Lauzan, and A. Soca. ------. 1992. Vector Resistance to Pesticides. Fifteenth Report of the WHO 2001. "Detection of Insecticide Resistance in Aedes aegypti (Diptera: Expert Committee on Vector Biology and Control. Technical Report Culicidae) from Cuba and Venezuela." Journal of Medical Entomology Series 818. Geneva: WHO. 38: 623­28. ------. 1998. African Trypanosomiasis: Control and Surveillance. Report of Rosen, L., D. A. Shroyer, R. B. Tesh, J. E. Freier, and J. C. Lien. 1983. a WHO Expert Committee. Technical Report Series 881. Geneva: "Transovarial Transmission of Dengue Viruses by Mosquitoes: Aedes WHO. albopictus and Aedes aegypti." American Journal of Tropical Medicine Zagne, S. M. O., V. G. F. Alves, R. M. R. Nogueira, M. P. Miagostovich, E. and Hygiene 32: 1108­19. Lampe, and W. Tavares. 1994. "Dengue Hemorrhagic Fever in the State Schallig, H. D. F. H., G. J. Schoone, C. C. M. Kroon, A. Hailu, F. Chappuis, of Rio de Janeiro, Brazil: A Study of 56 Confirmed Cases." Transactions and H. Veeken. 2001. "Development and Application of Simple of the Royal Society of Tropical Medicine and Hygiene 88: 677­79. Diagnostic Tools for Visceral Leishmaniasis." Medical Microbiology and Zaim, M., and P. Jambulingham. 2004. Global Insecticide Use for Vector- Immunology 190: 69­71. Borne Disease Control. 2nd ed. WHO/CDS/WHOPES/GCDPP/2004.9. Shaw, A. P. M., and P. Cattand. 2001. "Analytical Tools for Planning Cost- Geneva: World Health Organization. Effective Surveillance in Gambiense Sleeping Sickness." Médecine Tropicale 61: 412­21. 466 | Disease Control Priorities in Developing Countries | Pierre Cattand, Phillippe Desjeux, M. G. Guzmán, and others Chapter 24 Helminth Infections: Soil-Transmitted Helminth Infections and Schistosomiasis Peter J. Hotez, Donald A. P. Bundy, Kathleen Beegle, Simon Brooker, Lesley Drake, Nilanthi de Silva, Antonio Montresor, Dirk Engels, Matthew Jukes, Lester Chitsulo, Jeffrey Chow, Ramanan Laxminarayan, Catherine M. Michaud, Jeff Bethony, Rodrigo Correa-Oliveira, Xiao Shu-Hua, Alan Fenwick, and Lorenzo Savioli Helminth infections caused by soil-transmitted helminths Soil-Transmitted Helminths (STHs) and schistosomes are among the most prevalent afflic- The four most common STHs are roundworm (Ascaris lumbri- tions of humans who live in areas of poverty in the developing coides), whipworm (Trichuris trichiura), and the anthropophilic world. The morbidity caused by STHs and schistosomes is most hookworms (Necator americanus and Ancylostoma duodenale). commonly associated with infections of heavy intensity. Recent estimates suggest that A. lumbricoides infects 1.221 bil- Approximately 300 million people with heavy helminth infec- lion people, T. trichiura 795 million, and hookworms 740 million tions suffer from severe morbidity that results in more than (de Silva and others 2003) (table 24.1). The greatest numbers of 150,000 deaths annually (Crompton 1999; Montresor and oth- STH infections occur in the Americas, China and East Asia, ers 2002). In addition to their health effects, helminth infections and Sub-Saharan Africa. Strongyloides stercoralis is also a com- also impair physical and mental growth in childhood, thwart mon STH in some of these regions, although detailed informa- educational advancement, and hinder economic development. tion on the prevalence of strongyloidiasis is lacking because of Because of the geographic overlap of these afflictions and their the difficulties in diagnosing human infection. The life cycles of impact on children and adolescents, the World Health Ascaris, Trichuris, and hookworm follow a general pattern. The Organization (WHO); the World Bank; and other United adult parasite stages inhabit the gastrointestinal tract (Ascaris Nations agencies and bilaterals; and civil society are working to and hookworm in the small intestine; Trichuris in the colon), integrate STH and schistosome control through a program of reproduce sexually, and produce eggs, which are passed in periodic school-based, targeted anthelmintic drug treatments. human feces and deposited in the external environment. STH infections rarely cause death. Instead, the burden of disease is related less to mortality than to the chronic and insid- CAUSES AND CHARACTERISTICS ious effects on the hosts' health and nutritional status OF HELMINTH INFECTIONS (Stephenson, Latham, and Ottesen 2000; Stoltzfus and others 1997). Hookworms have long been recognized as an important Emphasis is placed on the four most common STH infections cause of intestinal blood loss leading to iron deficiency and pro- and the three most common schistosome infections. Together, tein malnutrition. The iron deficiency anemia that accompanies these infections account for most of the global helminth dis- moderate and heavy hookworm burdens is sometimes referred ease burden. to as hookworm disease (Hotez and others 2004). When host 467 Table 24.1 Global Prevalence and Distribution of Helminth infection with S. japonicum, which occurs in parts of China Infections and the Philippines (Ross and others 2002). Two other schisto- some species are known to cause intestinal schistosomiasis in Helminth infections Total cases Major geographic areas restricted geographical areas: S. intercalatum, found in Central STH infections 2 billion Africa, and S. mekongi, found in Cambodia and the Lao Ascariasis 1.221 billion Sub-Saharan Africa, India, People's Democratic Republic. Schistosomiasis is estimated to China and East Asia affect 187 million people worldwide (table 24.1). Trichuriasis 795 million Sub-Saharan Africa, India, A serious acute illness accompanied by fever and lym- China and East Asia phadenopathy, known as Katayama Syndrome, can result from Hookworm 740 million Sub-Saharan Africa, Americas, heavy schistosome infections. Chronic disease is mostly due to China and East Asia perforation of blood vessels and entrapment of eggs by host Schistosomiasis 187 million tissues. The host's reaction to entrapped eggs results in gran- S. haematobium 119 million Sub-Saharan Africa uloma formation. S. haematobium causes bladder wall S. mansoni 67 million Sub-Saharan Africa, Americas pathology, leading to ulcer formation, hematuria, and dysuria. S. japonicum 1 million China and East Asia Granulomatous changes and ulcers of the bladder wall and ureter can lead to bladder obstruction, dilatation, secondary Source: de Silva and others 2003. urinary tract infections and subsequent bladder calcification, renal failure, lesions of the female and male genital tracts, and hydronephrosis. S. haematobium is also associated with iron stores are depleted, the extent of iron deficiency anemia increased risk of bladder cancer. The morbidity commonly is linearly related to the intensity of hookworm infection associated with S. mansoni infection includes lesions of the (Stoltzfus and others 1997). Because of their underlying poor liver, portal vein, and spleen, leading to periportal fibrosis, iron status, children, women of reproductive age, and pregnant portal hypertension, hepatosplenomegaly, splenomegaly, and women are frequently the ones most susceptible to developing ascites. Schistosomiasis also causes chronic growth faltering hookworm anemia (Brooker, Bethony, and Hotez 2004). Iron and can contribute to anemia (Ross and others 2002). deficiency anemia during pregnancy has been linked to adverse maternal-fetal consequences, including prematurity, low birth- weight, and impaired lactation (WHO 2002). EPIDEMIOLOGY OF STH INFECTIONS Chronic STH infections resulting from Ascaris, Trichuris, AND SCHISTOSOMIASIS and hookworm can dramatically affect physical and mental development in children (WHO 2002). Studies have also The most striking epidemiological features of human helminth shown that the growth and physical fitness deficits caused by infections are aggregated distributions in human communities, chronic STH infections are sometimes reversible following predisposition of individuals to heavy (or light) infection, rapid treatment with anthelmintic drugs (Stephenson, Latham, and reinfection following chemotherapy, and age-intensity profiles Ottesen 2000). The effects on growth are most pronounced in that are typically convex (with the exception of hookworm). children with the heaviest infections, but light infections may For all the major human STH and schistosome infections also contribute to growth deficits if the nutritional status of the studied to date, worm burdens exhibit a highly aggregated community is poor (Stephenson, Latham, and Ottesen 2000). (overdispersed) distribution so that most individuals harbor just a few worms in their intestines, although a few hosts har- bor disproportionately large worm burdens (Anderson and Schistosomiasis May 1991). As a rule, 20 percent of the host population harbors Five major species of parasitic trematodes of the family approximately 80 percent of the worm population. This Schistosomatidae--Schistosoma haematobium, S. intercalatum, overdispersion has many consequences, both with regard to the S. japonicum, S. mansoni, and S. mekongi--infect humans. population biology of the helminths and the public health con- These parasites have a complex, indirect life cycle involving an sequence for the host, because heavily infected individuals are intermediate snail host. Disease is caused primarily by schisto- simultaneously at highest risk of disease and the major source some eggs, which are deposited by adult worms in the blood of environmental contamination. One feature that may help vessels surrounding the bladder or intestines. Urinary schisto- explain overdispersion is that individuals tend to be predis- somiasis, in which the bladder is affected, is caused by infection posed to heavy (or light) infections. Predisposition has been with S. haematobium, which occurs mainly in Africa. Intestinal demonstrated for all four major STHs and the schistosomes. schistosomiasis results from infection with S. mansoni, which The underlying cause of such predisposition remains poorly occurs in the Middle East, South America, and Africa, and from understood. However, a combination of heterogeneity in 468 | Disease Control Priorities in Developing Countries | Peter J. Hotez, Donald A. P. Bundy, Kathleen Beegle, and others exposure to infection or differences in susceptibility to infec- a. Prevalence tion and the ability to mount effective immunity (genetic and Percentage infected nutritional factors) is likely to be important. 100 People of all ages rapidly reacquire infection following treat- ment, but in schistosomiasis, older people reacquire infection 90 at slower rates than younger ones (Kabatereine and others 80 1999). The rate of reinfection is specific to certain species of 70 helminths and depends on the life expectancy of that species 60 (short-lived helminths reinfect more rapidly), on the intensity 50 of transmission within a given community, and on the treat- ment efficacy and coverage. The basic reproductive rate (Ro) 40 describes the transmission potential of a parasite (and thus its 30 ability to reinfect the host). It defines the average number of 20 female offspring produced during the life span of the parasite 10 that survive to reproductive maturity in the absence of density 0 dependence. Ro is determined by parasite immigration and 0 10 20 30 40 50 death rates as well as by host density (and, in schistosomiasis, Age (years) also snail density). A parasite will fail to become established unless Ro is greater than unity (Anderson and May 1991). Adult A. lumbricoides Hookworm T. trichiura S. haematobium worms usually survive between one and four years, whereas eggs can sometimes remain viable for several more years in the b. Intensity environment. Therefore, reinfection rates will remain high Mean number of worms until adults are removed with chemotherapy and until infective 110 stages, through time, become uninfective. In reality, density- dependent processes regulate parasite populations; at endemic 100 equilibrium, the effective reproductive ratio equals unity (that 90 is, each female replaces herself). Control programs rely on 80 reducing the effective reproductive ratio long enough for the 70 parasite population to be driven to local elimination. 60 Theoretically, Ro provides useful insights, and it is helpful to 50 think of control programs attempting to break the transmis- 40 sion cycle by reducing Ro to less than unity. Therefore, estimates can be made about how long and how many rounds 30 of chemotherapy are required to treat intestinal helminths. For 20 example, A. lumbricoides with an Ro of three and a life 10 expectancy of one year will need to be treated annually with 0 a drug that is 95 percent efficacious and with coverage of more 0 5 10 15 20 25 30 35 40 45 50 Age (years) than 91 percent of the population. Where Ro is five--that is, in areas where transmission is higher--treatment must be A. lumbricoides Hookworm given more frequently than once a year (Anderson and May (worms 2) (worms 7) 1991). T. trichiura S. haematobium The age-dependent patterns of infection prevalence are gen- (actual numbers) erally similar among the major helminth species, exhibiting a Source: Bundy 1995; reproduced and modified from Hotez, Arora, and others 2005. rise in childhood to a relatively stable asymptote in adulthood Figure 24.1 Age-Associated Prevalence and Intensity Profiles of (figure 24.1). Maximum prevalence of A. lumbricoides and STH and Schistosome Infections: Typical Age Profiles of Prevalence T. trichiura is usually attained before five years of age, and the and Intensity of STH Infections and Schistosomiasis maximum prevalence of hookworm and schistosome infec- tions is usually attained in adolescence or in early adulthood. linked to morbidity, the age-intensity profiles provide a clearer The nonlinear relationship between prevalence and intensity understanding of which populations are vulnerable to the has the consequence that the observed age-prevalence profiles different helminths (figure 24.1). For A. lumbricoides and provide little indication of the underlying profiles of age inten- T. trichiura infections, the age-intensity profiles are typically sity (age in relation to worm burden). Because intensity is convex in form, with the highest intensities in children 5 to Helminth Infections: Soil-Transmitted Helminth Infections and Schistosomiasis | 469 15 years of age (Bundy 1995). For schistosomiasis, a convex Climate, Water, and Season. Adequate warmth and moisture pattern is also observed, with a similar peak but with a plateau are key features for each of the STHs. Wetter areas exhibit in adolescents and young adults 15 to 29 years of age increased transmission, and in some endemic areas, both STH (Kabatereine and others 1999). In contrast, the age-intensity and schistosome infections exhibit marked seasonality profile for hookworm exhibits considerable variation, although (Brooker and Michael 2000). Recent use of geographical infor- intensity typically increases with age until adulthood and then mation systems and remote sensing has identified the distribu- plateaus (Brooker, Bethony, and Hotez 2004). In East Asia it is tional limits of STH and schistosomes on the basis of temper- also common to find the highest intensities among the elderly. ature and rainfall patterns (Brooker and Michael 2000). For However, more generally, children and young adults are at schistosomiasis, specific snail intermediate hosts prefer certain higher risk of both harboring higher levels of infection (thus types of aquatic environments. Construction of dams is known greater levels of morbidity) and becoming reinfected more to extend the range of snail habitats, thereby promoting the quickly. Both may occur at vital stages in a child's intellectual reemergence of schistosomiasis. and physical development. Risk Factors BURDEN OF THE DISEASE Both host-specific and environmental factors have been identi- fied that may affect the risk of acquiring or harboring heavy- The revised estimates in 2003 (de Silva and others 2003) use the intensity helminth infections. methodology developed by Chan and others (1994) and build on recent applications of geographical information systems to Genetics. No genes that control for human helminth infection derive updated atlases of helminth infections. To reflect recent have yet been identified. However, recent genome scans have changes in the epidemiology of infection, de Silva and others identified a locus possibly responsible for controlling S. man- used data from only 1990 onward. These data confirm that soni infection intensity on chromosome 5q31-33 and loci STH infections are the most prevalent infections of humans controlling A. lumbricoides intensity on chromosomes 1 and and that a large proportion of the population in developing 13. There is also evidence for genetic control of pathology countries is at risk. Of the 187 million cases of schistosomiasis attributable to S. mansoni, with linkage reported to a region estimated to occur worldwide, most are caused by S. haemato- containing the gene for the interferon gamma receptor 1 sub- bium in Sub-Saharan Africa (table 24.1). unit (Quinnell 2003). WHO (2002) estimates that 27,000 people die annually from STH infections and schistosomiasis (case fatality rate of Behavior, Household Clustering, and Occupation. Specific 0.0014 percent). Many investigators, however, believe that this occupations, household clustering, and behaviors influence the figure is an underestimate. Crompton (1999) estimated that prevalence and intensity of helminth infections (Bethony and 155,000 deaths annually occur from these infections (case fatal- others 2001), particularly for hookworm, in which the highest ity rate of 0.08 percent), whereas Van der Werf and others intensities occur among adults (Brooker, Bethony, and Hotez (2003), using the limited data available from Africa, estimated 2004). Engagement in agricultural pursuits, for example, the schistosomiasis mortality alone at 280,000 per year (case remains a common denominator for hookworm infection. fatality rate of 0.014 percent) because of nonfunctioning Behavioral and occupational factors, through their effect on kidneys (from S. haematobium) and hematemesis (from S. water contact, interact with environmental factors to produce mansoni). Therefore, the difference between estimates for variation in the epidemiology of schistosomiasis. helminth-associated mortality is more than 10-fold. Because it is uncommon for STHs and schistosomes to kill Poverty, Sanitation, and Urbanization. STH and schistoso- their human host, citing mortality figures provides only a small miasis depend for transmission on environments contaminated window on their health impact. Instead, measurements of dis- with egg-carrying feces. Consequently, helminths are intimately ease burden using disability-adjusted life years (DALYs) and associated with poverty, poor sanitation, and lack of clean similar tools portray a more accurate picture for helminthic water. The provision of safe water and improved sanitation are disease burden. WHO estimates the global burden of disease essential for the control of helminth infection. Although the from STH infections and schistosomiasis on the basis of the STH and schistosome infections are neglected diseases that enormous number of infected individuals, together with an occur predominantly in rural areas, the social and environ- associated low disability weight (Van der Werf and others mental conditions in many unplanned slums and squatter set- 2003). However, because an estimated 2 billion people are tlements of developing countries are ideal for the persistence of infected with STHs and schistosomes, even minor adjustments A. lumbricoides (Crompton and Savioli 1993). Schistosomiasis to the disability weights produce enormous variations in transmission can also occur in urban areas. DALYs or other measurements of disease burden. This helps to 470 | Disease Control Priorities in Developing Countries | Peter J. Hotez, Donald A. P. Bundy, Kathleen Beegle, and others explain why, for instance, in 1990 the disease burden for the malnutrition resulting from hookworm disease. Iron deficiency STH infections and schistosomiasis was almost 18 million anemia alone results in approximately 12 million DALYs lost DALYs, whereas the 2001 estimate was only 4.7 million DALYs. annually, making it the world's most important nutrition prob- In the intervening 11 years, the DALYs were as low as 2.6 mil- lem. Data on the epidemiology of iron deficiency anemia in lion. Such disparities are substantial when one considers that East Africa and elsewhere point to the important contribution the 1990 estimate ranks helminths close to major disease enti- of hookworms to this condition (Stoltzfus and others 1997). In ties such as tuberculosis, measles, and malaria, whereas the Tanzania, where hosts' iron stores are often depleted, there is a lowest estimate during the 1990s ranks helminth infections on correlation between the number of adult hookworms in the a par with gonorrhea, otitis media, and iodine deficiency. The intestine and the amount of host blood loss (Stoltzfus and Disease Control Priorities Project helminth working group others 1997). However, it is unclear whether current disability has determined that the WHO global burden of disease weights effectively incorporate the full contribution of hook- estimates are low because they do not incorporate the full clin- worm to severe iron deficiency anemia among iron-depleted ical spectrum of helminth-associated morbidity and chronic populations or whether they take host protein losses and disability, including anemia, chronic pain, diarrhea, exercise malnutrition into account. There is increasing interest in the intolerance, and undernutrition (King, Dickman, and Tisch importance of hookworm anemia in preschool children, espe- 2005). However, for this chapter, the average disability weights cially in Africa (Brooker, Bethony, and Hotez 2004), where estimated by WHO are used throughout. Some of the specific infants and preschool children are particularly vulnerable to the controversies are described below. developmental and behavioral deficits caused by iron deficien- cy anemia (Stephenson, Latham, and Ottesen 2000). Closer A. Lumbricoides and T. Trichiura infections assessment of the impact of hookworm on another important iron-deficient population--namely, women of reproductive Because the most significant physical and intellectual growth age--could also significantly increase current DALY estimates. disturbances occur as a consequence of moderate and heavy Approximately 44 million of these women harbor hookworms worm burdens, the age-associated epidemiology of A. (Bundy, Chan, and Savioli 1995). In addition, severe anemia in lumbricoides and T. trichiura infections has focused attention pregnancy is associated with neonatal prematurity, reduced on infected school-age children in developing countries birthweight, and impaired lactation (Christian, Khatry, and (Bundy 1995). In a revised estimate of the probable number of West 2004). ascariasis infections worldwide and a better categorization of the morbidity, de Silva, Chan, and Bundy (1997) indicated that 59 million of the 1.2 billion people infected (including 51 mil- Schistosomes lion children less than 15 years of age) were at risk of faltering growth, decreased physical fitness, or both as a result of infec- Scientists and public health workers disagree on the current tion. They estimated that about 1.5 million children would assessments of both morbidity and mortality attributable to never make up the deficit in growth, even if treated. In addition schistosomiasis. Several investigators have now initiated a to these chronic, insidious effects, they estimated that about process to recalculate the burden of disease attributable to 11.5 million individuals (almost all of them children) were at schistosomiasis, focusing much more on the clinical course of risk of more acute clinical illness. Their figures also indicated the different types of schistosomiasis and chronic sequelae that at least 10,500 deaths annually were directly attributable to (King, Dickman, and Tisch 2005; Michaud, Gordon, and Reich one of the serious complications of ascariasis; children account 2003). Through a comprehensive literature review combined for more than 90 percent of those deaths. The actual threshold with mathematical modeling, Van der Werf and others (2003) at which A. lumbricoides and T. trichiura worm burdens result estimate that urinary schistosomiasis in Africa results in in childhood morbidity is controversial because of the nonlin- approximately18 million cases of bladder wall pathology and ear relationship between intensity and pathogenesis and the 20 million cases of hydronephrosis, and African intestinal difficulties of measuring and attributing morbidity in under- schistosomiasis results in approximately 8.5 million cases of served populations suffering from other underlying conditions hepatomegaly. Mortality in Africa attributable to urinary and (Bundy 1995). intestinal schistosomiasis was extrapolated from these figures using a limited number of studies reporting case fatality rates for nonfunctioning kidney and hematemesis. From these Hookworm Infection extrapolations, Fenwick and others (2003) conclude that in Hookworm infection causes more DALYs lost than any other Africa the mortality attributable to urinary schistosomiasis helminthiasis with the exception of lymphatic filariasis. Even could be as high as 150,000 per year, and the number dying as these DALY measurements may still underestimate the true a result of intestinal schistosomiasis could be as high as 130,000 disease burden of iron deficiency anemia and protein per year. Helminth Infections: Soil-Transmitted Helminth Infections and Schistosomiasis | 471 COST-EFFECTIVENESS ANALYSIS tion. Sanitation is the only definitive intervention to eliminate OF INTERVENTIONS STH infections, but to be effective it should cover a high per- centage of the population. Therefore, because of the high costs Classifying Interventions involved, implementing this strategy is difficult where The three major interventions are anthelmintic drug treat- resources are limited (Asaolu and Ofoezie 2003). Moreover, ment, sanitation, and health education. when used as the primary means of control, it can take years or even decades for sanitation to be effective (Brooker, Bethony, Anthelmintic Drug Treatment. Anthelmintic drug treatment and Hotez 2004). ("deworming") is aimed at reducing morbidity by decreasing the worm burden. Repeated chemotherapy at regular intervals Health Education. Health education is aimed at reducing (periodic deworming) in high-risk groups can ensure that transmission and reinfection by encouraging healthy behav- the levels of infection are kept below those associated with iors. For STH infections and schistosomiasis, the aim is to morbidity (figure 24.2) and will frequently result in immediate reduce contamination of soil and water by promoting the use improvement in child health and development. Anthelmintic of latrines and hygienic behavior. Without a change in defeca- drug treatment can prevent the development of irreversible tion habits, periodic deworming cannot attain a stable reduc- consequences of schistosomiasis in adulthood. For ascariasis tion in transmission. Health education can be provided simply and trichuriasis, for which intensity peaks among school-age and economically and presents no contraindications or risks. children, frequent and periodic deworming may reduce Furthermore, its benefits go beyond the control of helminth transmission over time. Obstacles that diminish the effective- infections. In this perspective, it is reasonable to include this ness of periodic deworming are the low efficacy of single-dose component in all helminth control programs. mebendazole and albendazole for the treatment of hookworm and trichuriasis, respectively (Adams and others 2004; Albonico Other Control Measures. In specific epidemiological condi- and others 1994); high rates of posttreatment reinfection for tions, environmental or chemical control of snails can be useful STHs in areas of high endemicity (Albonico and others 1995); tools for reducing the transmission of schistosomiasis. and diminished efficacy with frequent and repeated use Research to develop new tools for control is in progress, includ- (Albonico and others 2003), possibly because of anthelmintic ing vaccine development programs for hookworm infection resistance (see the section "Research and Development"). and schistosomiasis (see "Research and Development"). Improved Sanitation. Improved sanitation is aimed at con- Choosing Interventions trolling transmission by reducing soil and water contamina- Periodic deworming stands out as the most cost-effective means to reduce the morbidity of STH and schistosome infections. Albendazole administration Periodic Anthelmintic Therapy. Periodic anthelmintic ther- apy, or periodic deworming, represents the main measure in Prevalence (percent) areas where infections are intensely transmitted, resources for disease control are limited, and funding for sanitation is lack- 80 ing. Drug treatment can be administered in the community High to moderate intensity using different strategies: 60 · Universal treatment. The entire community is treated, 40 irrespective of age, sex, infection status, and other Total prevalence characteristics. 20 · Targeted treatment. Treatment targets population groups, which may be defined by age, sex, or other social character- 0 istics, irrespective of the infectious status. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 · Selective treatment. Treatment targets individual-level appli- Time (months) cation of anthelmintic drugs, which is selected on the basis Source: Albonico and others, forthcoming. of either diagnosis or a suspicion of current infection. Figure 24.2 Predicted Effect on Ascaris and Trichuris Prevalence Following Frequent and Periodic Dewormings with Benzimidazole Recommended drugs for use in public health interventions Anthelmintics to control STH infection are the benzimidazole anthelmintics 472 | Disease Control Priorities in Developing Countries | Peter J. Hotez, Donald A. P. Bundy, Kathleen Beegle, and others (BZAs), albendazole (single dose: 400 mg, reduced to 200 mg for based deworming also has major externalities for untreated children between 12 and 24 months), or mebendazole (single children and the whole community. By reducing transmission dose: 500 mg), as well as levamisole or pyrantel pamoate (WHO in the community of Ascaris and Trichuris infections, deworm- 2002). Praziquantel (PZQ) (single dose: 40­60 mg/kg) is the ing substantially improves the health and school participation major drug used for the treatment of schistosomiasis. However, of both treated and untreated children, both in treatment therapy with oxamniquine has been the cornerstone for treat- schools and in neighboring schools (Bundy and others 1990; ment of S. mansoni infection in South American national Miguel and Kremer 2003). control programs over the past 20 years. The efficacy of oxam- These observations provided a basis for the adoption of res- niquine and PZQ is comparable, although that of PZQ is slightly olution 54.19 at the 2001 World Health Assembly, which urged better. The BZAs and PZQ are inexpensive; they have undergone member states to ensure access to essential drugs for STH and extensive safety testing and have been used by millions of indi- schistosome infections in endemic areas for the treatment of viduals with only a few minor side effects. Drugs that do not clinical cases and groups at high risk for morbidity (box 24.1). need dosage according to weight, such as BZAs (in school-age To achieve this goal, WHO has developed a broad partnership children), are considered easier to use for population-based that promotes the incorporation of deworming into existing interventions; however, the use of proxy indicators--for exam- institutions and programs, for both the education sectors and ple, substituting height for weight--has proved a successful the health sectors. The Partnership for Parasite Control was implementation strategy for PZQ (Hall and others 1999). launched in 2001 with the aim of mobilizing resources and promoting synergy among public and private efforts for the Distribution Strategy and Frequency of Treatment. The control of soil-transmitted helminths and schistosomiasis at selection of the distribution strategy and the frequency of global and national levels. School-based deworming has its treatment is based on epidemiological data. The recommended full effect when delivered within an integrated school health strategy for helminth control is a population-based approach, program that includes elements of the Focusing Resources on in which individuals in targeted communities are treated irre- Effective School Health (FRESH) framework. spective of their infection status (WHO 2002). This strategy is justified for several reasons, including the simplicity and safety Other At-Risk Populations. Not only school-age children can of delivering treatment. Individual diagnosis is difficult and benefit from treatment. Preschool children (one to five years expensive and offers no safety benefit. of age) are vulnerable to the developmental and behavioral The intrinsic transmission potential of the parasite species deficits caused by iron deficiency anemia, and recent analyses determines the frequency of treatment (see "Epidemiology of by Brooker, Bethony, and Hotez (2004) indicate that hook- STH Infections and Schistosomiasis," earlier in this chapter). To worm is an important contributor to anemia in that age group control morbidity in areas of intense transmission (prevalence (see "Estimating Intervention Effectiveness"). Women of greater than 70 percent and more than 10 percent of moderate- reproductive age (15 to 49 years of age) are particularly sus- and heavy-intensity infection), WHO (2002) recommends ceptible to iron deficiency anemia because of iron loss during treatment two or three times a year for STH infections. In areas menstruation and because of increased needs during with a lower intensity of transmission (prevalence between 40 pregnancy (Bundy, Chan, and Savioli 1995). In certain and 60 percent and less than 10 percent of moderate- and circumstances, male worker populations can also be at heavy-intensity infection), intervention once a year is recom- increased risk (Guyatt 2000). mended (WHO 2002). School-Age Children as a High-Risk Population. School-age Estimating Intervention Effectiveness children typically have the highest intensity of worm infection of any age group, and chronic infection negatively affects all The evidence base for the health and educational effect of aspects of children's health, nutrition, cognitive development, periodic deworming has accumulated significantly over the learning, and educational access and achievement (World Bank past decade. 2003). Regular deworming can cost-effectively reverse and prevent much of this morbidity. Furthermore, schools offer a STH Infections. All the anthelmintic drugs mentioned above readily available, extensive, and sustained infrastructure with a substantially reduce the number of adult worms in the gastroin- skilled workforce that is in close contact with the community. testinal tract. This effect is also reflected in reduced fecal egg With support from the local health system, teachers can deliver counts. In some cases, however, the efficacy of single-dose the drugs safely. Teachers need only a few hours of training to mebendazole or albendazole on hookworm and Trichuris infec- understand the rationale for deworming and to learn how to tions is low (Adams and others 2004; Albonico and others 1994). give out the pills and keep a record of their distribution. School- Moreover, pyrantel pamoate has little effect on T. trichiura. Helminth Infections: Soil-Transmitted Helminth Infections and Schistosomiasis | 473 Box 24.1 The 54th World Health Assembly The 54th World Health Assembly, which met in May 2001, countries in which such disease is endemic as an integral urged member states to ensure access to essential drugs for part of the primary health care system: (a) access to ade- schistosomiasis and STH infections in endemic areas quate diagnosis and essential anthelmintic drugs in all for the treatment of clinical cases and groups at high risk health services in all endemic areas, even at peripheral lev- for morbidity. The helminth infections of concern are the els, for the treatment of symptomatic cases and of chil- major schistosomes and STHs outlined in the text. The dren, women, and other groups at high risk of morbidity; World Health Assembly determined that simple and sus- (b) regular administration of chemotherapy to at least tainable control measures can relieve a generally underes- 75 percent of all school-age children at risk for morbidity timated and unnecessary disease burden in high- by 2010; and (c) sustained, community-based efforts to transmission areas. The following minimal targets, aimed improve sanitation, clean water supplies, and health at reducing morbidity by 80 percent, can be achieved by all education. Source: WHO 2002. Overall, however, anthelmintic treatment significantly improves Periodic Deworming. The advantage of periodic deworming physical and cognitive outcomes in the following ways: lies in its simplicity (one tablet per child) and safety. Teachers and other personnel without medical training can easily apply · Preschool children. Periodic distribution of anthelmintics the simple measures, which can be incorporated without diffi- has a positive effect on motor and language development culty in existing health and nonhealth activities that reach the and reduces malnutrition in very young children (Stoltzfus high-risk group. Several organizations, including nongovern- and others 2004). mental organizations, include an STH and schistosome · School-age children. Treating school-age children has a infection-control package within their routine activities and, considerable effect on their nutritional status (Stoltzfus and with very limited budgets, relieve the burden of helminth infec- others 2004), anemia, physical fitness, appetite, growth tions in the population covered. The costs of albendazole and (Stephenson, Latham, and Ottesen 2000), and intellectual PZQ are available through the International Drug Price development (Drake and others 2000). Indicator Guide (http://www.msh.org). Delivery systems for · Women of reproductive age. Studies of pregnant women deworming have often depended on vertical programs, in conducted by Christian, Khatry, and West (2004) in Nepal which mobile teams visit schools or communities to carry out indicate that albendazole treatment improves maternal treatment (WHO 2002). Estimated costs for this approach are hemoglobin as well as birth-weight and child survival. outlined in table 24.2. For STH infections in Tanzania, Nigeria, Schistosomiasis. As with STH infections, anthelmintic and Montserrat, the costs range from US$0.21 to US$0.51 per chemotherapy for schistosomiasis has an important effect on treatment. However, by training teachers and other school offi- child development, growth, and physical fitness (WHO 2002). cials to administer anthelmintic drugs, the system could Richter (2003) recently summarized details of the effect of PZQ achieve low-cost delivery by "piggy-backing" on existing pro- on organ pathology. In S. haematobium infections, reversal of grams in the educational sector (WHO 2002). Specific exam- urinary tract pathology can be seen six months after a cure. In ples of such programs conducted in Ghana and Tanzania are S. mansoni and S. japonicum infections, much of the intestinal summarized in the section "Implementation of Control pathology regresses after chemotherapy. However, more than Strategies: Lessons of Experience," later in this chapter. It was one PZQ treatment is usually necessary to reverse hepatic found that delivery of school-based targeted anthelmintic pathology, especially in areas of intense transmission. Early treatment could cost as little as US$0.03 per child, which may intervention with PZQ is preferable to reverse organ pathology. be as low as one-tenth of the estimated costs for vertical deliv- ery (WHO 2002). Thus, at current drug prices, the total cost Intervention Costs (drug plus delivery) of a single treatment with albendazole or Several studies have evaluated the costs of school-based peri- mebendazole may be as low as US$0.05, and that of a combined odic deworming in several different settings, whereas compara- treatment with PZQ may be as low as US$0.25 per child (WHO ble studies on other interventions are still lacking. 2002). 474 | Disease Control Priorities in Developing Countries | Peter J. Hotez, Donald A. P. Bundy, Kathleen Beegle, and others Table 24.2 Recent Examples of Delivery Costs for a Single Mass Treatment Delivery cost per treatment Strategy Drug Country US$ Percentage of total cost Mobile team Albendazole Montserrat 0.51 67 Albendazole Bangladesh -- 42 Levamisole Nigeria 0.32 81 PZQ Tanzania 0.21 24 School-based Albendazole Ghana 0.04 17 Albendazole Tanzania 0.03 13 Out-of-school children Arab Rep. of Egypt 0.16­0.21 40­47 Source: Guyatt 2003. -- not available. Integrating drug distribution through the school system qualitative judgments, and no cost-effectiveness analysis (CEA) rather than using mobile teams, along with a marked decline in estimates exist for sanitation in this context. the price of BZAs and PZQ, has resulted in a 10-fold reduction in delivery costs. However, those costs are artificially low Health Education and Communication. Measures to increase because they do not include the external costs for the coordi- the health awareness of the population are included as an nating center responsible for supporting those approaches essential component of any population-based activity aimed at (Guyatt 2003). It has been estimated, for instance, that mass controlling morbidity attributable to helminth infections. albendazole treatment of school-age children in Kenya could However, the effectiveness of those activities in reducing trans- cost more than US$3 million each year, equivalent to some mission of infection varies according to different reports. In 4 percent of current national public expenditure on health care some cases, health education can decrease costs, increase levels (Guyatt 2003). This analysis has not been evaluated against of knowledge, and decrease reinfection rates (Lansdown and actual operations, however, and current estimates from the others 2002). Health education efforts can build trust and parasite control authorities in Kenya suggest that the actual cost engage communities, aspects that are crucial to the success of is likely to be far less. Large-scale chemotherapy programs for public health initiatives. No CEA estimates exist for health helminth control continue to rely heavily on donor support, education in this context. suggesting that some affected countries may be unable to sup- port the costs of deworming. Monitoring of control programs is an important part of the Linking Costs and Effects of Interventions managerial process, and it should be carried out at minimum cost so as not to divert resources from the intervention Interventions to reduce morbidity from helminth infections (Brooker and others 2004). It is recommended that, at the fall into two categories: targeting the transmission mechanisms planning stage, approximately 5 to 10 percent of the program and treating individuals directly. The former encompasses budget be reserved for monitoring activities (Montresor and improvements in infrastructure, including water supply and others 2002). sanitation, and health education. The latter entails the periodic drug treatment of the population. Substantial improvements Improved Sanitation. When sanitation improvements are through prevention may be a long-term outcome of economic made alongside deworming, the results obtained last longer. growth in general, because wealthier households have However, the investment needed to reach the level required to improved sanitation facilities and practices, but those improve- interfere with STH transmission could be high. To correctly ments are not an option in the short term without large evaluate the advantage of such investments, one must take into investments in infrastructure. As shown in the previous sec- account the consequences for other health indicators and for tion, deworming options dominate on both effectiveness and economic development. An efficient sanitation infrastructure cost-effectiveness criteria. Costs continue to fall as drug costs removes the underlying cause of most poverty-related commu- decrease. With better data and detailed mapping of disease nicable diseases and can boost the economic development of a distribution within communities, targeting individuals at country. The resources needed to improve hygienic standards high risk becomes more feasible, thus improving the cost- can be huge and require the cooperation of several sectors of effectiveness of control programs (Michaud, Gordon, and society (Asaolu and Ofoezie 2003). Currently, these are Reich 2003). Helminth Infections: Soil-Transmitted Helminth Infections and Schistosomiasis | 475 Box 24.2 The High Cost-Effectiveness of Mass Treatment for Helminth Infection The combination of low-cost treatment and high preva- albendazole and then reinfected, the cost per DALY averted lence rates suggests that the cost per DALY averted from is estimated at US$3.41 for STH infections. That is, if treating helminth infections will be quite low. Following spending were capped at US$1 million, total DALYs would the consistent framework described in mass treatment of be reduced by nearly 300,000. The estimate of cost per school-age children for both STH infections and schistoso- DALY is higher for schistosomiasis relative to STH infec- miasis proves to be extremely cost-effective. In fact, benefit- tions because of higher drug costs and lower disability cost ratios would be even higher if the analyses incorpo- weights. Depending on whether generics or original for- rated the additional benefits associated with externalities mulations are used, the cost per DALY averted ranges from for the untreated. For a population of 1 million people in US$3.36 to US$6.92. However, in combination, treatment low- and middle-income countries, if treatment is limited with both albendazole and PZQ proves to be extremely cost to school-age children treated 1.1 times per year with effective, in the range of US$8 to US$19 per DALY averted. Source: Authors. Evidence from existing programs that narrow the interven- infections. Expanding mass treatment to children not enrolled tion to school-age children (a high-risk group) shows that the in school will result in treating populations that have higher treatment costs of chemotherapy for helminth infections are incidence and intensity, thus raising effectiveness (box 24.2). quite low--well below US$1 per school-age child. This finding is in part due to the accessibility of the target group and the cost Distributional and Equity Consequences savings of incorporating delivery into existing school and health programs. Moreover, as discussed in the following Interventions to control helminth infections can have equity sections, the economic benefits of targeting this group may be implications in several dimensions. Programs designed to tar- substantial. Still other targeted groups may also have low cost get communities with high prevalence or high intensity of per treatment when treatment is merged into existing helminth infection focus on areas with lower income, as programs. For example, interventions through prenatal care described in the sections on the causes, characteristics, and programs for pregnant women may be cost-effective. Likewise, epidemiology of such infections. Although no studies under- evidence on costs of treatment through existing integrated take benefit-incidence analysis of public spending on such management of infant and childhood illness (IMCI) programs health services, this targeting implies that state subsidies on for small children and health campaigns (such as vaccination deworming services will be of most benefit to lower-income and micronutrient programs) find low cost per case treated groups. With the increasing availability of poverty maps, (Montresor and others 2002). empirical evaluation of the equity implications of deworming Several factors can potentially alter the ranking of interven- will be feasible. tions in regard to cost-effectiveness, although there are no existing studies to evaluate this. Previous analysis may underes- timate the effectiveness and overestimate the cost-benefit AVERTED, AVERTABLE, AND ratios of mass treatment of school-age children if the NONAVERTABLE BURDEN externalities of treatment are not considered (Miguel and Kremer 2003). The cost-effectiveness of school-based deworm- In the short run, deworming can avert helminth infections. In ing programs will change as the programs are extended to cover the long run, it is assumed that as income levels grow and children who are not enrolled in school. Such program exten- infrastructure improves, the number of infections averted by sions are likely to have greater costs because they entail addi- reducing transmission will increase. However, given the slow tional staff and outreach efforts per case treated. However, the rate of poverty reduction in the 1990s for the poorest regions, effectiveness of mass treatment of school-age children (both such as Sub-Saharan Africa, waiting for economic development enrolled and not enrolled) may be greater. Children who are to lead to a reduction in infections is only, at best, a slow-paced not enrolled in school come from households with lower solution for the majority of the infected population. It is more income levels. Lower income, which leads to poorer sanitation likely that most averted infections will depend on periodic conditions, is associated with greater incidence and intensity of deworming. Thus, the question remains as to what portion of 476 | Disease Control Priorities in Developing Countries | Peter J. Hotez, Donald A. P. Bundy, Kathleen Beegle, and others existing infections is potentially avertable through recom- Studies are increasingly documenting a causal impact of mended interventions and what portion is currently averted adult health (broadly defined) on labor force participation, with existing programs. For schistosomiasis, successful pro- wages, and productivity in developing countries (Thomas and grams in several countries, including Brazil, China, and the Strauss 1997). Moreover, helminth infection is known to affect Arab Republic of Egypt, and the issues related to the sustain- some of the health conditions related to productivity--namely, ability of these successes have been described (see iron deficiency anemia and wasting. Guyatt (2000) reviews "Implementation of Control Strategies: Lessons of Experience" numerous studies relating these conditions to physical fitness later in this chapter). However, the number of averted schisto- and productivity; Haas and Brownlie (2001) review studies on somiasis infections in Sub-Saharan Africa is likely to be small, the effect of iron supplementation on work. The studies gener- because few serious attempts at widespread control have been ally show productivity gains linked to better health along the made in recent years, and not much of the burden of STH and various health dimensions studied. However, although some schistosome infections is currently averted through private evidence points to the indirect impact of STHs on income treatment. In part, the low number of averted infections may earnings, these relationships have not been adequately studied, be due to the lack of information on the part of infected indi- either directly or indirectly. viduals, the insidious nature of the condition, and the lack of More compelling links between helminth infection and eco- drugs in the public or private health delivery system. nomic well-being may exist for children. The strong association between worm burden and poor health outcomes for children suggests that infections may affect school enrollment, atten- ECONOMIC BENEFITS OF INTERVENTION dance, grade repetition, and grade attainment. In turn, the potential impact on educational outcomes has implications The characteristics of helminth infections make a compelling for the assessment of the economic benefits of intervention. case for public sector intervention if based only on the evidence Numerous studies have demonstrated the benefits of schooling, related to the intervention's effect on health. From an economic showing that the return on education is quite high. Increased perspective, the public sector has several reasons to become education is associated with, among other things, higher worker involved in improving health outcomes. First, other benefits may productivity and generally higher productivity in nonmar- be gained, in addition to the benefit for the treated individual. ket production activities, including greater farmer efficiency Second, some forms of intervention are almost pure public and productivity (Psacharopoulos and Patrinos 2002). goods; that is, no one can be excluded from using the goods or Although observational studies show that lower levels of services the interventions deliver, and the private sector is thus learning and schooling are linked to helminth infection (World unlikely to deliver them. Finally, preventive measures, such as Bank 2003), establishing a causal relationship requires ade- information on the value of washing hands,may not be delivered quately controlling for all unobserved or confounding factors. through the private sector. The lack of knowledge about infec- Miguel and Kremer (2003) note that several methodological tions and subclinical symptoms may make individuals less like- issues hamper many existing randomized treatment-control ly to seek treatment. In analyzing the gains of interventions for evaluations. First, externalities associated with interventions worm control, one should account for the burden of helminth can lead to underestimating impacts among the untreated pop- infections, which extends well beyond the health impacts and ulation. Second, sample selection and attrition issues can affect DALYs. The economic implications may be quite large. the validity of findings, although the direction of this effect The negative correlation between helminth infections and is ambiguous. Third, existing studies typically evaluate the income level is clearly demonstrated both within and between impact of deworming on cognitive skills, likely the culmination countries (de Silva and others 2003). However, causality cannot of several years of health and education investments, as be inferred from this established relationship; poverty pro- assessed by tests administered to treated and untreated chil- motes higher worm burdens, yet poor health induced by dren. Although studies find an effect on cognitive skills for helminths can lead to lower income. There may also be oppor- those with the heaviest worm burden, they do not focus on tunity costs to uninfected household members residing with other important education outcomes, which are likely to be infected persons. Few studies have been designed to evaluate, more affected in the short run by health improvements, such as either directly or indirectly, the magnitude of the effect of school enrollment and school attendance. deworming on economic productivity. The indirect evidence at The study by Miguel and Kremer (2003) in Kenya attempts the micro level suggests that helminth infection has a signifi- to address those shortcomings through improved study design cant impact on adult productivity and, subsequently, on earn- and analytical methods. In addition to providing health gains, ings. More direct evidence for children shows that helminth deworming reduced total primary school absenteeism by at infection has long-term implications for educational attain- least one-quarter in the first two years of the project. The gains ment and economic status. were largest for the youngest children, who suffered from more Helminth Infections: Soil-Transmitted Helminth Infections and Schistosomiasis | 477 Net presented value of discounted wages (US$) Beyond the current impacts on schooling and implications for cognition, helminth infection in children can have long- 125 term implications for economic outcomes in adulthood 100 through its effect on physical growth. Height has been shown to affect wage-earning capacity as well as participation in the 75 labor force for men and women (Thomas and Strauss 1997). This relationship may be strongest in settings where infection 50 rates are highest--that is, low-income areas, where physical 25 endurance yields high returns in the labor market. 0 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 Increase in school participation per year per pupil (percent) IMPLEMENTATION OF CONTROL STRATEGIES: Note: Assumptions are as follows: a 7 percent return to an additional year of school; LESSONS OF EXPERIENCE wage gains earned over 40 years in the workforce, discounted at 5 percent per year with no wage growth; and annual wage earnings of US$400 per year, which is below the estimated agricultural and nonagricultural annual wages for low-income countries in Two case studies illustrate the profound health effects of peri- the World Bank (2003). odic deworming. Figure 24.3 Returns to School Participation Case Study Number 1: Periodic Deworming in Ghana and Tanzania intense worm infections. Externalities would cause a substan- The Partnership for Child Development (PCD) undertook an tial underestimation of this effect. In terms of cost-effectiveness evaluation of the use of schools in Ghana and Tanzania for the as an educational intervention, deworming proved to be far delivery of health interventions, including research into the more effective at improving school attendance than other edu- processes, costs, and benefits (PCD 1999). The effort also cational interventions implemented in a study in Kenya. involved operations research and evaluation of programs with Deworming offers a high rate of return, increasing the net pres- regard to health and education outcomes and people's percep- ent value of discounted wages by more than US$30 per treated tions of the programs (Hall and others 1999). The results child compared with per treatment costs of under US$1. For demonstrated the following: realistic estimates of returns to schooling, these results show in general that the net present discounted value of lifetime earn- · Simple interventions, such as deworming, have the ings is high compared with the costs of treatment even for potential to improve children's health and educational small gains in school participation (figure 24.3). achievement, especially for those worst affected and most Bleakley (2003) examined the effect of a hookworm control disadvantaged. program undertaken about 1910 in the southern part of the · The delivery of school-based health services is efficient and United States. Hookworm infection was estimated to cause a cost-effective and is supported by the key stakeholders 23 percent drop in the probability of school attendance, and involved. Implementers of the school health programs in the children with greater exposure to the hookworm eradication education and health sectors and the community are posi- campaign were more likely to be literate. Moreover, the long- tive regarding the teacher's role in health provision, as long term follow-up of affected cohorts showed that hookworm as the health interventions are simple, safe, familiar, and infection in childhood led to significantly lower wages in effective and are seen as responding to local needs. adulthood. · The provision of health services through schools need not Helminth infections in preschool-age children can have con- require long and complex training, nor significantly add to sequences for subsequent schooling, such as delaying primary the workload of teachers or administrators. school enrollment and school attainment, thereby affecting · Delivery through the existing education sector could occur future labor market outcomes. Bobonis, Miguel, and Sharma effectively without any additional infrastructure, as long as (2003) conducted a study of preschool-age children, using iron the existing educational system is adequately functional. supplementation and deworming drugs administered to chil- dren two to six years of age. Preliminary results indicate that, in The results of the evaluation of these programs also high- addition to the weight gain associated with treatment, average lighted the need for deworming to be carried out in the context preschool participation rates increased sharply by 6.3 percent- of a wider framework of school health, which includes core age points among assisted children older than two, reducing activities such as effective and nondiscriminatory school health preschool absenteeism by roughly one-fifth. policies, provision of safe water and sanitation, and effective 478 | Disease Control Priorities in Developing Countries | Peter J. Hotez, Donald A. P. Bundy, Kathleen Beegle, and others health education (http://www.freshschools.org and http://www. tive performance. Plausible mechanisms by which helminths schoolsandhealth.org). suppress growth in childhood and exert negative impacts on intelligence, cognition, and school performance are largely unstudied and unknown. In addition, some reports have ques- Case Study Number 2: Schistosomiasis Control in Egypt tioned whether albendazole itself could adversely affect growth In 1937, the prevalence of schistosomiasis in rural areas was (Forrester and others 1998). Those issues require clarification about 50 percent, almost every boy had blood in his urine by as widespread deworming programs become more common. the age of 12, and bladder cancer was the commonest cancer in The impact of helminths on populations other than Egypt. Molluscicides, from copper sulfate to niclosamide, have schoolchildren, including preschool-age children, women of been used to try to kill the host snails, and drugs from antimony- childbearing age, and adult workers, appears to be substantial. based compounds, through niridazole, metrifonate, and However, those populations are understudied. Also unclear is PZQ, have been used to treat the millions of infected Egyptians. the impact of childhood STH and schistosome infections on Finally, after a 14-year control campaign using PZQ, the productivity in adulthood. The effect of chemotherapy on prevalence of schistosomiasis has been reduced to below many of the manifestations of schistosomiasis has not been 10 percent. With infection intensities now low, the serious assessed systematically. It has been postulated that PZQ treat- health consequences of schistosomiasis have disappeared. The ment of schistosomiasis promotes partial immunity by destroy- program was started in 1988, when using loans from the ing worms in the vasculature and releasing parasite antigens African Development Bank and the World Bank, Egypt invested (Colley and Secor 2004). In contrast, the frequent and periodic heavily in the purchase of PZQ, encouraging local production, treatment of STH infections (Albonico and others 1995) do not to control morbidity caused by schistosomiasis. Since the drug appear to promote natural protective immunity. was first made available in 1988, some 45 million doses have The role of helminths and coinfections also warrants fur- been dispensed. A television campaign has encouraged people ther exploration. Some studies suggest that HIV-1 infection to submit samples for diagnosis and to receive free treatment if may promote susceptibility to schistosomiasis (Secor, Karanja, their diagnoses are positive. Since 1997, a mass chemotherapy and Colley 2004), and human T-cell lymphotropic virus-1 campaign was used to target populations in high-prevalence (HTLV-1) infection may promote susceptibility to strongy- villages and children in selected governorates where prevalence loidiasis. In addition, emerging evidence indicates that STH was greater than 20 percent. In addition, molluscicides and schistosome infections may promote susceptibility to have been applied in canals around high-prevalence villages. other infectious agents, possibly including HIV/AIDS and The widespread use of PZQ has given dramatic results. malaria (Fincham, Markus, and Adams 2003). This phenome- Morbidity, including hematuria, has almost disappeared, and non, if verified in an epidemiologic study, would further bladder cancer is on the decline. increase helminth-associated DALY estimates. Anthelmintic Drug Resistance and New Drug Development RESEARCH AND DEVELOPMENT A concern about the feasibility of sustainable control with Among the important tasks to be done to control helminth BZAs is the possible emergence of drug resistance among infections are collection of better data on helminth disease human STHs. BZA resistance occurs because of the spread of burden, research on the health and economic effects (and point mutations in nematode-tubulin alleles. This phenome- safety) of periodic deworming, monitoring of the emergence of non has already resulted in widespread BZA drug resistance anthelmintic drug resistance, and development of new tools to among STHs of ruminant livestock. There is still no direct evi- supplement or complement existing control strategies. dence for BZA resistance among human STHs, although such resistance could account for an observed failure of mebenda- zole for human hookworm in southern Mali, as well as a Health and Economic Impact diminished efficacy against hookworm in Zanzibar following Overall, better estimates of disease burden are needed frequent and periodic use of mebendazole (Albonico and oth- (Michaud, Gordon, and Reich 2003), especially with respect to ers 2003). PZQ resistance must also be considered, especially as obtaining a consistent and agreed-on estimate for the DALYs it begins to be widely used in Sub-Saharan Africa (Hagan and attributed to helminth infections. In a systematic review of ran- others 2004). Should PZQ resistance develop, there will be new domized deworming trials, Dickson and others (2000) con- demands for antischistosomal drugs. Recently, the artemisins clude that, although data support the effects of deworming on have shown activity against schistosomulae and were successful weight gain, there are inconsistencies among trials and insuffi- in protecting against S. japonicum in China (Hagan and others cient evidence as to whether such interventions improve cogni- 2004). Helminth Infections: Soil-Transmitted Helminth Infections and Schistosomiasis | 479 Anthelmintic Vaccines concern that could derail global deworming efforts in much the The high rates of reinfection that can occur following treatment same way that resistance to DDT and chloroquine has affected with anthelmintic drugs and concern about emerging drug the ambitions for global malaria control. It is possible that the resistance have prompted the search for alternative control tools. specific dynamics of helminth populations will provide suffi- For most helminth infections, reduction in adult worm burden cient genetic flow to maintain susceptibility in much the same has been considered the "gold standard" for vaccine develop- way that insecticides remain the effective staple of global ment. For schistosomiasis, however, a vaccine that targets para- agribusiness. Until new technologies become available, site fecundity and egg viability, thereby reducing pathology and anthelmintic chemotherapy for school-age children remains transmission, would also represent an important breakthrough. the most practical and substantive means to control STH and A 28-kDa glutathione S-transferase (GST) has shown promise schistosome infections in the developing world. as a protective antigen for S. haematobium infection (Capron and others 2005). The S. haematobium vaccine project based on GST has successfully passed phase 1 testing; the research group, REFERENCES which is based at the Pasteur Institute, is embarking on phase 2 Adams, V. J., C. J. Lombard, M. A. Dhansay, M. B. Markus, and J. E. clinical trials in Senegal and Niger. Additional schistosomiasis Fincham. 2004. "Efficacy of Albendazole against the Whipworm vaccines are also undergoing early-stage development. Efforts Trichuris Trichiura--A Randomized, Controlled Trial." South African are also under way by the Human Hookworm Vaccine Initiative Medical Journal 94: 972­76. to develop and test a first-generation recombinant hookworm Albonico, M., Q. Bickle, M. Ramsan, A. Montresor, L. Savioli, and vaccine (Brooker and others 2005; Hotez, Bethony, and others M. Taylor. 2003. "Efficacy of Mebendazole and Levamisole Alone or in Combination against Intestinal Nematode Infections after Repeated 2005). The first vaccine manufactured under current good man- Targeted Mebendazole Treatment in Zanzibar." Bulletin of the World ufacturing practices and tested for quality control and toxicity is Health Organization 81: 343­52. the Na-ASP-2 hookworm vaccine, which was developed from Albonico, M., A. Montresor, D. W. Crompton, and L. Savioli. Forthcoming. research demonstrating human correlates of immunity and par- "Intervention for the Control of Soil-Transmitted Helminthiasis." Advances in Parasitology. tial protection data in vaccinated laboratory animals. Phase 1 Albonico, M., E. Renganathan, A. Bosman, U. M. Kisumku, K. S. Alawi, human trials for evaluating the safety and immunogenicity of and L. Savioli. 1994. "Efficacy of a Single Dose of Mebendazole on the Na-ASP-2 hookworm vaccine are in progress. Additional Prevalence and Intensity of Soil-Transmitted Nematodes in Zanzibar." research is needed to determine how an anthelmintic vaccine Tropical and Geographic Medicine 46: 142­46. can be incorporated into existing control programs, as well how Albonico, M., P. G. Smith, E. Ercole, A. Hall, H. M. Chwaya, K. S. Alawi, and L. Savioli. 1995. "Rate of Reinfection with Intestinal Nematodes after it would be used for at-risk populations not currently targeted Treatment of Children with Mebendazole or Albendazole in a Highly for periodic deworming in schools. Endemic Area." Transactions of the Royal Society of Tropical Medicine and Hygiene 89: 538­41. Anderson, R. M., and R. M. May. 1991. Infectious Diseases of Humans. Oxford, U.K.: Oxford University Press. CONCLUSIONS: PROMISES AND PITFALLS Asaolu, S. O., and I. E. Ofoezie. 2003. "The Role of Health Education and Sanitation in the Control of Helminth Infections." Acta Tropica 86: Fulfilling the mandate of World Health Assembly resolution 283­94. 54.19 will require the regular treatment of hundreds of millions Bethony, J., J. T. Williams, H. Kloos, J. Blangero, L. Alves-Fraga, G. Buck, of children over decades. The obstacles in this undertaking are and others. 2001. "Exposure to Schistosoma mansoni Infection in a formidable, and success will depend on the ability of countries Rural Area in Brazil: II. Household Risk Factors." Tropical Medicine and to identify or create reliable and sustained infrastructures for International Health 6: 136­45. this purpose. A focus on using preexisting school systems may Bleakley, H. 2003. "Disease and Development: Evidence from Hookworm Eradication in the American South." Journal of the European Economic be key to achieving this goal. The treatment of schoolchildren Association 1: 376­86. for A. lumbricoides, T. trichiura, and schistosome infections Bobonis, G., E. Miguel, C. Sharma. 2003. "Iron Deficiency Anemia and achieves large externalities that reduce infection in other School Participation." University of California­Berkeley. http://emlab. vulnerable age groups. However, the different epidemiology of berkeley.edu/users/emiguel/miguel_anemia.pdf. hookworm raises concerns about the risks to preschool chil- Brooker, S., J. Bethony, and P. J. Hotez. 2004. "Human Hookworm Infection in the 21st Century." Advances in Parasitology 58: 197­288. dren and women of reproductive age who remain untreated. Brooker, S., J. M. Bethony, L. Rodrigues, N. Alexander, S. Geiger, and Providing regular treatment to these populations appears to be P. J. Hotez. 2005. "Epidemiological, Immunological and Practical a less cost-effective option, largely because of the absence of Considerations in Developing and Evaluating a Human Hookworm a preexisting infrastructure. This situation presents a strong Vaccine." Expert Review of Vaccines 4: 35­50. argument for developing a hookworm vaccine that could be Brooker, S., and E. Michael. 2000. "The Potential of Geographical Information Systems and Remote Sensing in the Epidemiology and used to protect these vulnerable groups. It has yet to be seen Control of Human Helminth Infections." Advances in Parasitology 47: whether the emergence of BZA drug resistance is a genuine 245­87. 480 | Disease Control Priorities in Developing Countries | Peter J. Hotez, Donald A. P. Bundy, Kathleen Beegle, and others Brooker, S., S. Whawell, N. B. Kabatereine, A. Fenwick, and R. M. Haas, J. D., and T. Brownlie. 2001. "Iron Deficiency and Reduced Work Anderson. 2004. "Evaluating the Epidemiological Impact of National Capacity: A Critical Review of the Research to Determine a Causal Control Programmes for Helminths." Trends in Parasitology 11: Relationship." Journal of Nutrition 131 (Suppl.): 676S­88S. 537­45. Hagan, P., C. C. Appleton, G. C. Coles, J. R. Kusel, and L. A. Tchuem- Bundy, D. A. 1995. "Epidemiology and Transmission of Intestinal Tchuente. 2004. "Schistosomiasis Control: Keep Taking the Tablets." Helminths." In Enteric Infection 2, Intestinal Helminths, ed. M. J. G. Trends in Parasitology 20: 92­97. Farthing, G. T. Keusch, and D. Wakelin, 5­24. London: Chapman & Hall, A., C. Nokes, S. T. Wen, S. Adjei, C. Kihamia, L. Mwanri, and others. Hall Medical. 1999. "Alternatives to Bodyweight for Estimating the Dose of Bundy, D. A., M. S. Chan, and L. Savioli. 1995. "Hookworm Infection in Praziquantel Needed to Treat Schistosomiasis." Transactions of the Pregnancy." Transactions of the Royal Society of Tropical Medicine and Royal Society of Tropical Medicine and Hygiene 93: 652­58. Hygiene 89: 521­22. Hotez, P. J., S. Arora, J. Bethony, M. E. Bottazzi, A. Loukas, R. Correa- Bundy, D. A., M. S. Wong, L. L. Lewis, and J. Horton. 1990. "Control of Oliveira, and S. Brooker. 2005. "Helminth Infections of Children: Geohelminths by Delivery of Targeted Chemotherapy through Prospects for Control." In Hot Topics in Infection and Immunity in Schools." Transactions of the Royal Society of Tropical Medicine and Children, ed. A. J. Pollard and A. Finn. New York: Springer. Hygiene 84: 115­20. Hotez, P. J., J. Bethony, M. E. Bottazzi, S. Brooker, and P. Buss. 2005. Capron, A., G. Riveau, M. Capron, and F. Trottein. 2005. "Schistosomes: "Hookworm--`The Great Infection of Mankind.'" Public Library of The Road from Host-Parasite Interactions to Vaccines in Clinical Science Medicine 2: e67. Trials." Trends in Parasitology 21: 143­49. Hotez, P. J., S. Brooker, J. M. Bethony, M. E. Bottazzi, A. Loukas, and S. H. Chan, M. S., G. F. Medley, D. Jamison, and D. A. Bundy. 1994. "The Xiao. 2004. "Current Concepts: Hookworm Infection." New England Evaluation of Potential Global Morbidity Attributable to Intestinal Journal of Medicine 351: 799­807. Nematode Infections." Parasitology 109: 373­87. Kabatereine, N. B., B. J. Vennervald, J. H. Ouma, J. Kemijumbi, A. E. Christian, P., S. K. Khatry, and K. P. West. 2004. "Antenatal Anthelmintic Butterworth, D. W. Dunne, and A. J. Fulford. 1999. "Adult Resistance Treatment, Birthweight, and Infant Survival in Rural Nepal." Lancet to Schistosomiasis Mansoni: Age-Dependence to Reinfection Remains 364: 981­83. Constant in Communities with Diverse Exposure Patterns." Colley, D. G., and E. W. Secor. 2004."Immunoregulation and World Health Parasitology 118: 101­5. Assembly Resolution 54.19: Why Does Treatment Control Morbidity?" King, C. H., K. Dickman, and D. J. Tisch. 2005. "Reassessment of the Parasitology International 53: 143­50. Cost of Chronic Helmintic Infection: A Meta-analysis of Disability- Crompton, D. W. 1999. "How Much Helminthiasis Is There in the World?" Related Outcomes in Endemic Schistosomiasis." Lancet 365: Journal of Parasitology 85: 397­403. 1561­69. ------. 2001. "Ascaris and Ascariasis." Advances in Parasitology 48: Lansdown, R., A. Ledward, A. Hall, W. Issac, E. Yona, J. Matulu, and others. 285­375. 2002. "Schistosomiasis, Helminth Infection, and Health Education in Tanzania: Achieving Behaviour Change in Primary Schools." Health Crompton, D. W., and L. Savioli. 1993. "Intestinal Parasitic Infections and Education Research 17: 425­33. Urbanization." Bulletin of the World Health Organization 71: 1­7. Michaud, C. M., W. S. Gordon, and M. R. Reich. 2003. "The Global de Silva, N. R., S. Brooker, P. J. Hotez, A. Montresor, D. Engles, and L. Burden of Disease Due to Schistosomiasis." Disease Control Savioli. 2003. "Soil-Transmitted Helminth Infections: Updating the Priorities Project Working Paper 19. http://www.fic.nih.gov/dcpp/ Global Picture." Trends in Parasitology 19: 547­51. wps/wp19.pdf. de Silva, N. R., M. S. Chan, and D. A. P. Bundy. 1997. "Morbidity and Miguel, E. A., and M. Kremer. 2003. "Worms: Identifying Impacts on Mortality Due to Ascariasis: Re-estimation and Sensitivity Analysis of Education and Health in the Presence of Treatment Externalities." Global Numbers at Risk." Tropical Medicine and International Health Econometrica 72 (1): 159­217. 2: 519­28. Montresor, A., D. W. T. Crompton, T. W. Gyorkos, and L. Savioli. 2002. Dickson, R., S. Awasthi, P. Williamson, C. Demellweek, and P. Garner. Helminth Control in School-Age Children: A Guide for Managers of 2000. "Effects of Treatment for Intestinal Helminth Infection on Control Programmes. Geneva: World Health Organization. Growth and Cognitive Performance in Children: Systematic Review of Randomised Trials." British Medical Journal 320: 1697­701. PCD (Partnership for Child Development). 1999. "The Cost of Large- Scale School Health Programmes Which Deliver Anthelmintics to Drake, L. J., M. C. H. Jukes, R. J. Sternberg, and D. A. P. Bundy. 2000. Children in Ghana and Tanzania." Acta Tropica 73 (2): 183­204. "Geohelminth Infections (Ascariasis, Trichuriasis, and Hookworm): Cognitive and Developmental Impacts." Seminars in Pediatric Psacharopoulos, G., and H. Patrinos. 2002. "Returns to Investment in Infectious Diseases 11: 245­51. Education: A Further Update." Working Paper 2881, World Bank, Washington, DC. Fenwick, A., L. Savioli, D. Engels, N. R. Bergquist, and M. H. Todd. 2003. "Drugs for the Control of Parasitic Diseases: Current Status and Quinnell, R. J. 2003. "Genetics of Susceptibility to Human Helminth Development in Schistosomiasis." Trends in Parasitology 19: 509­15. Infection." International Journal of Parasitology 33: 1219­31. Fincham, J. E., M. B. Markus, and V. J. Adams. 2003. "Could Control of Richter, J. 2003. "The Impact of Chemotherapy on Morbidity Due to Soil-Transmitted Helminthic Infection Influence the HIV/AIDS Schistosomiasis." Acta Tropica 86: 161­83. Pandemic?" Acta Tropica 86: 315­33. Ross, A. G., P. B. Bartley, A. C. Sleigh, G. R. Olds, Y. Li, G. M. Williams, and Forrester, J. E., J. C. Bailar III, S. A. Esrey, M. V. Jose, B. T. Castillejos, and D. P. McManus. 2002. "Schistosomiasis." New England Journal of G. Ocamp. 1998. "Randomised Trial of Albendazole and Pyrantel in Medicine 346: 1212­20. Symptomless Trichuriasis in Children." Lancet 353: 1103­8. Secor, W. E., D. M. Karanja, and D. G. Colley. 2004. "Interactions Guyatt, H. L. 2000. "Do Intestinal Nematodes Affect Productivity in between Schistosomiasis and Human Immunodeficiency Virus in Adulthood?" Parasitology Today 16: 153­58. Western Kenya." Memorias do Instituto Oswaldo Cruz 99 (5 Suppl. 1): 93­95. ------. 2003. "The Cost of Delivering and Sustaining a Control Programme for Schistosomiasis and Soil-Transmitted Helminthiasis." Stephenson, L. S., M. C. Latham, and E. A. Ottesen. 2000. "Malnutrition Acta Tropica 86: 267­74. and Parasitic Helminth Infections." Parasitology 121 (Suppl.): S23­28. Helminth Infections: Soil-Transmitted Helminth Infections and Schistosomiasis | 481 Stoltzfus, R. J., H. M. Chwaya, A. Montresor, J. M. Tielsch, J. K. Jape, M. Van der Werf, M. J., S. J. de Vlas, S. Brooker, C. W. Looman, N. J. Albonico, and L. Savioli. 2004. "Low Dose Daily Iron Supplementation Nagelkerke, J. D. Habbema, and D. Engels. 2003. "Quantification of Improves Iron Status and Appetite but Not Anemia, Whereas Quarterly Clinical Morbidity Associated with Schistosome Infection in Sub- Anthelmintic Treatment Improves Growth, Appetite, and Anemia in Saharan Africa." Acta Tropica 86: 125­39. Zanzibari Preschool Children." Journal of Nutrition 134: 348­56. WHO (World Health Organization). 2002. Prevention and Control of Stoltzfus, R. J., M. L. Dreyfuss, H. M. Chwaya, and M. Albonico. 1997. Schistosomiasis and Soil-Transmitted Helminthiasis. WHO Technical "Hookworm Control as a Strategy to Prevent Iron Deficiency Anemia." Series Report 912. Geneva: WHO. Nutrition Reviews 55: 223­32. World Bank. 2003. School Deworming at a Glance. Public Health at a Thomas, D., and J. Strauss. 1997. "Health and Wages: Evidence on Men Glance Series. http://www.worldbank.org/hnp. and Women in Urban Brazil." Journal of Econometrics 77: 159­85. 482 | Disease Control Priorities in Developing Countries | Peter J. Hotez, Donald A. P. Bundy, Kathleen Beegle, and others Chapter 25 Acute Respiratory Infections in Children Eric A. F. Simoes, Thomas Cherian, Jeffrey Chow, Sonbol Shahid- Salles, Ramanan Laxminarayan, and T. Jacob John Acute respiratory infections (ARIs) are classified as upper res- CAUSES OF ARIS AND THE BURDEN OF DISEASE piratory tract infections (URIs) or lower respiratory tract infec- tions (LRIs). The upper respiratory tract consists of the airways ARIs in children take a heavy toll on life, especially where med- from the nostrils to the vocal cords in the larynx, including the ical care is not available or is not sought. paranasal sinuses and the middle ear. The lower respiratory tract covers the continuation of the airways from the trachea Upper Respiratory Tract Infections and bronchi to the bronchioles and the alveoli. ARIs are not confined to the respiratory tract and have systemic effects URIs are the most common infectious diseases. They include because of possible extension of infection or microbial toxins, rhinitis (common cold), sinusitis, ear infections, acute pharyn- inflammation, and reduced lung function. Diphtheria, per- gitis or tonsillopharyngitis, epiglottitis, and laryngitis--of tussis (whooping cough), and measles are vaccine-preventable which ear infections and pharyngitis cause the more severe diseases that may have a respiratory tract component but also complications (deafness and acute rheumatic fever, respec- affect other systems; they are discussed in chapter 20. tively). The vast majority of URIs have a viral etiology. Except during the neonatal period, ARIs are the most com- Rhinoviruses account for 25 to 30 percent of URIs; respiratory mon causes of both illness and mortality in children under five, syncytial viruses (RSVs), parainfluenza and influenza viruses, who average three to six episodes of ARIs annually regardless of human metapneumovirus, and adenoviruses for 25 to 35 per- where they live or what their economic situation is (Kamath cent; corona viruses for 10 percent; and unidentified viruses for and others 1969; Monto and Ullman 1974). However, the pro- the remainder (Denny 1995). Because most URIs are self-limit- portion of mild to severe disease varies between high- and low- ing, their complications are more important than the infections. income countries, and because of differences in specific etiolo- Acute viral infections predispose children to bacterial infections gies and risk factors, the severity of LRIs in children under five of the sinuses and middle ear (Berman 1995a), and aspiration is worse in developing countries, resulting in a higher case- of infected secretions and cells can result in LRIs. fatality rate. Although medical care can to some extent mitigate both severity and fatality, many severe LRIs do not respond to Acute Pharyngitis. Acute pharyngitis is caused by viruses in therapy, largely because of the lack of highly effective antiviral more than 70 percent of cases in young children. Mild pharyn- drugs. Some 10.8 million children die each year (Black, Morris, geal redness and swelling and tonsil enlargement are typical. and Bryce 2003). Estimates indicate that in 2000, 1.9 million of Streptococcal infection is rare in children under five and more them died because of ARIs, 70 percent of them in Africa and common in older children. In countries with crowded living Southeast Asia (Williams and others 2002). The World Health conditions and populations that may have a genetic predispo- Organization (WHO) estimates that 2 million children under sition, poststreptococcal sequelae such as acute rheumatic fever five die of pneumonia each year (Bryce and others 2005). and carditis are common in school-age children but may also 483 occur in those under five. Acute pharyngitis in conjunction pharyngeal cultures do not always reveal the pathogen that is with the development of a membrane on the throat is nearly the cause of the LRI. Bacterial cultures of lung aspirate speci- always caused by Corynebacterium diphtheriae in developing mens are often considered the gold standard, but they are not countries. However, with the almost universal vaccination of practical for field application. Vuori-Holopainen and Peltola's infants with the DTP (diphtheria-tetanus-pertussis) vaccine, (2001) review of several studies indicates that S. pneumoniae diphtheria is rare. and Hib account for 13 to 34 percent and 1.4 to 42.0 percent of bacterial pneumonia, respectively, whereas studies by Adegbola Acute Ear Infection. Acute ear infection occurs with up to and others (1994), Shann, Gratten, and others (1984), and Wall 30 percent of URIs. In developing countries with inadequate and others (1986) suggest that Hib accounts for 5 to 11 percent medical care, it may lead to perforated eardrums and chronic of pneumonia cases. ear discharge in later childhood and ultimately to hearing Reduced levels of clinical or radiological pneumonia in clin- impairment or deafness (Berman 1995b). Chronic ear infection ical trials of a nine-valent pneumococcal conjugate vaccine following repeated episodes of acute ear infection is common in provide an estimate of the vaccine-preventable disease burden developing countries, affecting 2 to 6 percent of school-age chil- (valency indicates the number of serotypes against which the dren.The associated hearing loss may be disabling and may affect vaccine provides protection; conjugate refers to conjugation of learning. Repeated ear infections may lead to mastoiditis, which polysaccharides to a protein backbone). In a study in The in turn may spread infection to the meninges. Mastoiditis and Gambia, 37 percent of radiological pneumonia was prevented, other complications of URIs account for nearly 5 percent of all reflecting the amount of disease caused by S. pneumoniae, and ARI deaths worldwide (Williams and others 2002). mortality was reduced by 16 percent (Cutts and others 2005). Upper respiratory tract colonization with potentially patho- genic organisms and aspiration of the contaminated secretions Lower Respiratory Tract Infections have been implicated in the pathogenesis of bacterial pneumo- The common LRIs in children are pneumonia and bronchi- nia in young children. Infection of the upper respiratory tract olitis. The respiratory rate is a valuable clinical sign for diag- with influenza virus or RSVs has been shown to increase the nosing acute LRI in children who are coughing and breathing binding of both H. influenzae (Jiang and others 1999) and rapidly. The presence of lower chest wall indrawing identifies S. pneumoniae (Hament and others 2004; McCullers and more severe disease (E. Mulholland and others 1992; Shann, Bartmess 2003) to lining cells in the nasopharynx. This finding Hart, and Thomas 1984). may explain why increased rates of pneumococcal pneumonia Currently, the most common causes of viral LRIs are RSVs. parallel influenza and RSV epidemics. A study in South Africa They tend to be highly seasonal, unlike parainfluenza viruses, showed that vaccination with a nine-valent pneumococcal con- the next most common cause of viral LRIs. The epidemiology jugate vaccine reduced the incidence of virus-associated pneu- of influenza viruses in children in developing countries monia causing hospitalization by 31 percent, suggesting that deserves urgent investigation because safe and effective vac- pneumococcus plays an important role in the pathogenesis of cines are available. Before the effective use of measles vaccine, virus-associated pneumonia (Madhi, Petersen, Madhi, Wasas, the measles virus was the most important viral cause of respi- and others 2000). ratory tract­related morbidity and mortality in children in Entry of bacteria from the gut with spread through the developing countries. bloodstream to the lungs has also been proposed for the patho- genesis of Gram-negative organisms (Fiddian-Green and Baker Pneumonia. Both bacteria and viruses can cause pneumonia. 1991), but such bacteria are uncommon etiological agents of Bacterial pneumonia is often caused by Streptococcus pneumo- pneumonia in immune-competent children. However, in niae (pneumococcus) or Haemophilus influenzae, mostly type b neonates and young infants, Gram-negative pneumonia is not (Hib), and occasionally by Staphylococcus aureus or other strep- uncommon (Quiambao forthcoming). tococci. Just 8 to 12 of the many types of pneumococcus cause Viruses are responsible for 40 to 50 percent of infection in most cases of bacterial pneumonia, although the specific types infants and children hospitalized for pneumonia in developing may vary between adults and children and between geographic countries (Hortal and others 1990; John and others 1991; locations. Other pathogens, such as Mycoplasma pneumoniae Tupasi and others 1990). Measles virus, RSVs, parainfluenza and Chlamydia pneumoniae, cause atypical pneumonias. Their viruses, influenza type A virus, and adenoviruses are the most role as a cause of severe disease in children under five in devel- important causes of viral pneumonia. Differentiating between oping countries is unclear. viral and bacterial pneumonias radiographically is difficult, The burden of LRIs caused by Hib or S. pneumoniae is partly because the lesions look similar and partly because bac- difficult to determine because current techniques to establish terial superinfection occurs with influenza, measles, and RSV bacterial etiology lack sensitivity and specificity. The results of infections (Ghafoor and others 1990). 484 | Disease Control Priorities in Developing Countries | Eric A. F. Simoes, Thomas Cherian, Jeffrey Chow, and others In developing countries, the case-fatality rate in children 34 percent, much higher than the 5 to 10 percent for children with viral pneumonia ranges from 1.0 to 7.3 percent (John and not infected with HIV (Bobat and others 1999; Madhi, others 1991; Stensballe, Devasundaram, and Simoes 2003), Petersen, Madhi, Khoosal, and others 2000; Nathoo and others with bacterial pneumonia from 10 to 14 percent and with 1993; Zwi, Pettifior, and Soderlund 1999). Pneumocystis jirove- mixed viral and bacterial infections from 16 to 18 percent ci and cytomegalovirus are important opportunistic infections (Ghafoor and others 1990; Shann 1986). in more than 50 percent of HIV-infected infants (Jeena, Coovadia, and Chrystal 1996; Lucas and others 1996). Gram- Bronchiolitis. Bronchiolitis occurs predominantly in the first negative bacteria are also important in more than 70 percent of year of life and with decreasing frequency in the second and HIV-infected malnourished children (Ikeogu, Wolf, and Mathe third years. The clinical features are rapid breathing and lower 1997). Patient studies have confirmed the frequent association chest wall indrawing, fever in one-third of cases, and wheezing of these bacteria but added S. pneumoniae and S. aureus as (Cherian and others 1990). Inflammatory obstruction of the important pathogens (Gilks 1993; Goel and others 1999). The small airways, which leads to hyperinflation of the lungs, and first South African report on the overall burden of invasive collapse of segments of the lung occur. Because the signs and pneumococcal disease reported a 41.7-fold increase in HIV- symptoms are also characteristic of pneumonia, health workers infected children compared with uninfected children (Farley may find differentiating between bronchiolitis and pneumonia and others 1994). difficult. Two features that may help are a definition of the sea- sonality of RSVs in the locality and the skill to detect wheezing. RSVs are the main cause of bronchiolitis worldwide and can INTERVENTIONS cause up to 70 or 80 percent of LRIs during high season (Simoes 1999; Stensballe, Devasundaram, and Simoes 2003). The Interventions to control ARIs can be divided into four basic recently discovered human metapneumovirus also causes bron- categories: immunization against specific pathogens, early chiolitis (Van den Hoogen and others 2001) that is indistin- diagnosis and treatment of disease, improvements in nutrition, guishable from RSV disease. Other viruses that cause bronchi- and safer environments (John 1994). The first two fall within olitis include parainfluenza virus type 3 and influenza viruses. the purview of the health system, whereas the last two fall under public health and require multisectoral involvement. Influenza. Even though influenza viruses usually cause URIs in adults, they are increasingly being recognized as an impor- tant cause of LRIs in children and perhaps the second most Vaccinations important cause after RSVs of hospitalization of children with Widespread use of vaccines against measles, diphtheria, per- an ARI (Neuzil and others 2002). Although influenza is consid- tussis, Hib, pneumococcus, and influenza has the potential to ered infrequent in developing countries, its epidemiology substantially reduce the incidence of ARIs in children in devel- remains to be investigated thoroughly. The potential burden of oping countries. The effects of measles, diphtheria, and pertus- influenza as a cause of death in children is unknown. Influenza sis vaccines are discussed in chapter 20. The limited data on virus type A may cause seasonal outbreaks, and type B may influenza in developing countries do not permit detailed cause sporadic infection. Recently, avian influenza virus has analysis of the potential benefits of that vaccine. This chapter, caused infection, disease, and death in small numbers of indi- therefore, focuses on the potential effects of Hib and pneumo- viduals, including children, in a few Asian countries. Its poten- coccal vaccines on LRIs. tial for emergence in human outbreaks or a pandemic is unknown, but it could have devastating consequences in devel- Hib Vaccine. Currently three Hib conjugate vaccines are avail- oping countries (Peiris and others 2004) and could pose a able for use in infants and young children. The efficacy of Hib threat to health worldwide. New strains of type A viruses will vaccine in preventing invasive disease (mainly meningitis, but almost certainly arise through mutation, as occurred in the case also pneumonia),has been well documented in several studies in of the Asian and Hong Kong pandemics in the 1950s and industrialized countries (Black and others 1992; Booy and oth- 1960s. ers 1994; Eskola and others 1990; Fritzell and Plotkin 1992; Heath 1998; Lagos and others 1996; Santosham and others 1991) and in one study in The Gambia (K. Mulholland and HIV Infection and Pediatric LRIs others 1997). All studies showed protective efficacy greater than Worldwide, 3.2 million children are living with HIV/AIDS, 90 percent against laboratory-confirmed invasive disease, 85 percent of them in Sub-Saharan Africa (UNAIDS 2002). In irrespective of the choice of vaccine. Consequently, all industri- southern Africa, HIV-related LRIs account for 30 to 40 percent alized countries include Hib vaccine in their national immu- of pediatric admissions and have a case-fatality rate of 15 to nization programs, resulting in the virtual elimination of Acute Respiratory Infections in Children | 485 invasive Hib disease because of immunity in those vaccinated immunogenic conjugate vaccines (Mulholland 1998; Obaro and a herd effect in those not vaccinated. Available data from a 1998; Temple 1991). few developing countries show a similar herd effect (Adegbola The 7-PCV and 9-PCV have been evaluated for efficacy and others 1999; Wenger and others 1999). against invasive pneumococcal disease in four trials, which The initial promise and consequent general perception was demonstrated a vaccine efficiency ranging from 71.0 to 97.4 per- that Hib vaccine was to protect against meningitis, but in devel- cent (58 to 65 percent for HIV-positive children, among whom oping countries the vaccine is likely to have a greater effect on rates of pneumococcal disease are 40 times higher than in HIV- preventing LRIs. The easily measured effect is on invasive dis- negative children) (Black and others 2000; Cutts and others ease, including bacteraemic pneumonia. The vaccine probably 2005; Klugman and others 2003; O'Brien and others 2003). has an effect on nonbacteremic pneumonia, but this effect is In the United States, the 7-PCV was included in routine vac- difficult to quantify because of the lack of an adequate method cinations of infants and children under two in 2000. By 2001 the for establishing bacterial etiology. In Bangladesh, Brazil, Chile, incidence of all invasive pneumococcal disease in this age group and The Gambia, Hib vaccine has been associated with a reduc- had declined by 69 percent and disease caused by the serotypes tion of 20 to 30 percent in those hospitalized with radiograph- included in the vaccine and related serotypes had declined by ically confirmed pneumonia (de Andrade and others 2004; 78 percent (Whitney and others 2003). Similar reductions were Levine and others 1999; K. Mulholland and others 1997; WHO confirmed in a study in northern California (Black and others 2004a). However, results of a large study in Lombok, Indonesia, 2001). A slight increase in rates of invasive disease caused by were inconclusive with regard to the effect of Hib vaccine on serotypes of pneumococcus not included in the vaccine was pneumonia (Gessner and others 2005). observed, but it was not large enough to offset the substantial reduction in disease brought about by the vaccine. The studies Pneumococcal Vaccines. Two kinds of vaccines are currently also found a significant reduction in invasive pneumococcal available against pneumococci: a 23-valent polysaccharide vac- disease in unvaccinated older age groups, especially adults age cine (23-PSV), which is more appropriate for adults than chil- 20 to 39 and age 65 and older, suggesting that giving the vaccine dren, and a 7-valent protein-conjugated polysaccharide vaccine to young children exerted a considerable herd effect in the com- (7-PCV). A 9-valent vaccine (9-PCV) has undergone clinical munity. Such an advantage is likely to occur even where the trials in The Gambia and South Africa, and an 11-valent vac- prevalence of adult HIV disease is high and pneumococcal dis- cine (11-PCV) is being tried in the Philippines. ease may be recurrent and life threatening. Studies of the efficacy of the polysaccharide vaccine in The effect of the vaccine on pneumococcal pneumonia as preventing ARIs or ear infection in children in industrialized such is difficult to define given the problems of establishing the countries have shown conflicting results. Whereas some studies bacterial etiology of pneumonia. Three studies have evaluated of this vaccine show no significant efficacy (Douglas and Miles the effect of the vaccine on radiographic pneumonia (irres- 1984; Sloyer, Ploussard, and Howie 1981), studies from Finland pective of the etiological agent) and have shown a 20.5 to show a generally protective effect against the serotypes 37.0 percent reduction in radiographically confirmed pneumo- contained in a 14-PSV (Douglas and Miles 1984; Karma and nia (9.0 percent for HIV-positive individuals) (Black and others others 1980; Makela and others 1980). The efficacy was more 2000; Cutts and others 2005; Klugman and others 2003). marked in children over two years of age than in younger chil- Several field trials have evaluated the efficacy of PCV against dren. The only studies evaluating the effect of the polysaccha- ear infection. Even though the vaccine resulted in a significant ride vaccine in children in developing countries are a series of reduction in culture-confirmed pneumococcal otitis, no net three trials conducted in Papua New Guinea (Douglas and reduction of ear infection was apparent among vaccinated chil- Miles 1984; Lehmann and others 1991; Riley and others 1981; dren, probably because of an increase in the rates of otitis Riley, Lehmann, and Alpers 1991). The analysis of the pooled caused by types of pneumococci not covered by the vaccine, data from these trials showed a 59 percent reduction in LRI H. influenzae and Moraxella catarrhalis (Eskola and others mortality in children under five at the time of the vaccination 2001; Kilpi and others 2003). However, a trial in northern and a 50 percent reduction in children under two. On the basis California showed that the vaccine had a protective effect of these and other studies, the investigators concluded that the against frequent ear infection and reduced the need for tympa- vaccine had an effect on severe pneumonia. The greater-than- nostomy tube placement (Fireman and others 2003). Thus, a expected efficacy in these trials was attributed to the greater vaccine for ear infection may be beneficial in developing coun- contribution of the more immunogenic adult serotypes in tries with high rates of chronic otitis and conductive hearing pneumonia in Papua New Guinea (Douglas and Miles 1984; loss and should be evaluated by means of clinical trials. Riley, Lehmann, and Alpers 1991). On account of the poor The most striking public health benefit of a vaccine in immunogenicity of the antigens in the 23-PSV against preva- developing countries would be a demonstrable reduction in lent pediatric serotypes, attention is now directed at more mortality. Although the primary outcome in The Gambia trial 486 | Disease Control Priorities in Developing Countries | Eric A. F. Simoes, Thomas Cherian, Jeffrey Chow, and others was initially child mortality, it was changed to radiological others 1992). On the basis of these and other data (Campbell, pneumonia. Nevertheless, the trial showed a 16 percent Byass, and others 1989; Kolstad and others 1997; Perkins and (95 percent confidence level, 3 to 38) reduction in mortality. others 1997; Redd 1994; Simoes and others 1997; Weber and This trial was conducted in a rural area in eastern Gambia others 1997), WHO recommends a respiratory rate cutoff of where access to round-the-clock curative care, including case 50 breaths per minute for infants age 2 through 11 months and management, is difficult to provide. This trial demonstrates 40 breaths per minute for children age 12 months to 5 years. that immunization delivered through outreach programs will Rapid breathing, as defined by WHO, detects about 85 per- have substantial health and economic benefits in such popula- cent of children with pneumonia, and more than 80 percent of tions. One additional study evaluating the effect of an 11-PCV children with potentially fatal pneumonia are probably suc- on radiological pneumonia is ongoing in the Philippines; cessfully identified and treated using the WHO diagnostic cri- results are expected in the second half of 2005. teria. Antibiotic treatment of children with rapid breathing has been shown to reduce mortality (Sazawal and Black 2003). The problem of the low specificity of the rapid breathing criterion Case Management is that some 70 to 80 percent of children who may not need The simplification and systematization of case management for antibiotics will receive them. Nevertheless, for primary care early diagnosis and treatment of ARIs have enabled significant workers for whom diagnostic simplicity is essential, rapid reductions in mortality in developing countries, where access to breathing is clearly the most useful clinical sign. pediatricians is limited. WHO clinical guidelines for ARI case management (WHO 1991) use two key clinical signs: respiratory Pneumonia Diagnosis Based on Chest Wall Indrawing. rate, to distinguish children with pneumonia from those with- Children are admitted to hospital with severe pneumonia when out, and lower chest wall indrawing, to identify severe pneumo- health workers believe that oxygen or parenteral antibiotics nia requiring referral and hospital admission. Children with (antibiotics administered by other than oral means) are needed audible stridor when calm and at rest or such danger signs of or when they lack confidence in mothers' ability to cope. The severe disease as inability to feed also require referral. Children rationale of parenteral antibiotics is to achieve higher levels of without these signs are classified as having an ARI but not pneu- antibiotics and to overcome concerns about the absorption of monia. Children showing only rapid breathing are treated for oral drugs in ill children. pneumonia with outpatient antibiotic therapy. Children who The Papua New Guinea study (Shann, Hart, and Thomas have a cough for more than 30 days are referred for further 1984) used chest wall indrawing as the main indicator of sever- assessment of tuberculosis and other chronic infections. ity, but studies from different parts of the world show large dif- ferences in the rates of indrawing because of variable defini- Pneumonia Diagnosis Based on Rapid Breathing. The initial tions. Restriction of the term to lower chest wall indrawing, guidelines for detecting pneumonia based on rapid breathing defined as inward movement of the bony structures of the chest were developed in Papua New Guinea during the 1970s. In a wall with inspiration, has provided a better indicator of the study of 200 consecutive pediatric outpatients and 50 consecu- severity of pneumonia and one that can be taught to health tive admissions (Shann, Hart, and Thomas 1984), 72 percent of workers. It is more specific than intercostal indrawing, which children with audible crackles in the lungs had a respiratory frequently occurs in bronchiolitis. rate of 50 or more breaths per minute, whereas only 19 percent In a study in The Gambia (Campbell, Byass, and others of children without crackles breathed at such a rapid rate. 1989), a cohort of 500 children from birth to four years old Therefore, the initial WHO guidelines used a threshold of was visited at home weekly for one year. During this time, 50 breaths per minute, at or above which a child with a cough 222 episodes of LRI (rapid breathing, any chest wall indrawing, was regarded as having pneumonia. nasal flaring, wheezing, stridor, or danger signs) were referred The major concern was the relatively low sensitivity of this to the clinic. Chest indrawing was present in 62 percent of these approach, which could miss 25 to 40 percent of cases of pneu- cases, many with intercostal indrawing. If all children with any monia. A study in Vellore, India, found that sensitivity could chest indrawing were hospitalized, the numbers would over- be improved by lowering the threshold to 40 for children age 1 whelm pediatric inpatient facilities. to 4, while keeping the 50 breaths per minute cutoff for infants Studies in the Philippines and Swaziland (E. Mulholland age 2 months through 11 months (Cherian and others 1988). and others 1992) found that lower chest wall indrawing was Subsequent studies showed that when these thresholds were more specific than intercostal indrawing for a diagnosis of used, sensitivity improved from 62 to 79 percent in the severe pneumonia requiring hospital admission. In the Vellore Philippines and from 65 to 77 percent in Swaziland, but at study (Cherian and others 1988), lower chest wall indrawing the same time, the specificity fell from 92 to 77 percent in the correctly predicted 79 percent of children with an LRI who Philippines and 92 to 80 percent in Swaziland (Mulholland and were hospitalized by a pediatrician. Acute Respiratory Infections in Children | 487 Antimicrobial Options for Oral Treatment of Pneumonia. chloramphenicol are common, this complication is extremely The choice of an antimicrobial drug for treatment is based on rare in young children. There is no evidence that intramuscu- the well-established finding that most childhood bacterial lar chloramphenicol succinate is more likely to produce side pneumonias are caused by S. pneumoniae or H. influenzae. A effects than other forms and routes of chloramphenicol. single injection of benzathine penicillin, although long last- ing, does not provide adequate penicillin levels to eliminate Hypoxemia Diagnosis Based on WHO Criteria. The ARI H. influenzae. WHO has technical documents to help assess the case-management and integrated management of infant and relevant factors in selecting first- and second-line antimicro- childhood illness (IMCI) strategies depend on accurate referral bials and comparisons of different antimicrobials in relation to of sick children to a hospital and correct inpatient manage- their antibacterial activity, treatment efficacy, and toxicity ment of LRI with oxygen or antibiotics. Hypoxemia (deficiency (WHO 1990). of oxygen in the blood) in children with LRI is a good predic- The emergence of antimicrobial resistance in S. pneumoniae tor of mortality, the case-fatality rate being 1.2 to 4.6 times and H. influenzae is a serious concern. In some settings, in vitro higher in hypoxemic LRI than nonhypoxemic LRI (Duke, tests show that more than 50 percent of respiratory isolates of Mgone, and Frank 2001; Onyango and others 1993), and oxy- both bacteria are resistant to co-trimoxazole, and penicillin gen reduces mortality. Thus, it is important to detect hypox- resistance to S. pneumoniae is gradually becoming a problem emia as early as possible in children with LRI to avert death. worldwide. Although diagnoses of acute LRIs are achieved very easily by In pneumonia, unlike in meningitis, in vitro resistance of recognizing tachypnoea, and although severe LRI is associated the pathogen does not always translate into treatment failure. with chest wall indrawing, the clinical recognition of hypox- Reports from Spain and South Africa suggest that pneumonia emia is more problematic. Different sets of clinical rules have caused by penicillin-resistant S. pneumoniae can be successfully been studied to predict the presence of hypoxemia in children treated with sufficiently high doses of penicillin. Amoxicillin is with LRI (Cherian and others 1988; Onyango and others 1993; concentrated in tissues and in macrophages, and drug levels are Usen and others 1999). Although some clinical tools have a directly correlated with oral dosages. Therefore, higher doses high sensitivity for detecting hypoxemia, a good number of than in the past--given twice a day--are now being used to hypoxemic children would still be missed using these criteria. successfully treat ear infections caused by penicillin-resistant Pulse oximetry is the best tool to quickly detect hypoxemia in S. pneumoniae. Amoxicillin is clearly better than penicillin for sick children. However, pulse oximeters are expensive and have such infections. The situation with co-trimoxazole is less clear recurring costs for replacing probes, for which reasons they are (Strauss and others 1998), and in the face of high rates of co- not available in most district or even referral hospitals in devel- trimoxazole resistance, amoxicillin may be superior for chil- oping countries. dren with severe pneumonia. Treatment Guidelines. Current recommendations are for co- Intramuscular Antibiotics for Treatment of Severe trimoxazole twice a day for five days for pneumonia and intra- Pneumonia. Even though chloramphenicol is active against muscular penicillin or chloramphenicol for children with severe both S. pneumoniae and H. influenzae, its oral absorption is pneumonia. The problems of increasing resistance to co- erratic in extremely sick children. Thus, the WHO guidelines trimoxazole and unnecessary referrals of children with any chest recommend giving intramuscular chloramphenicol at half the wall indrawing have led to studies exploring alternatives to the daily dose before urgent referral of severe pneumonia cases. An antibiotics currently used in ARI case management. One study additional rationale is that extremely sick children may have indicated that amoxicillin and co-trimoxazole are equally effec- sepsis or meningitis that are difficult to rule out and must be tive for nonsevere pneumonia (Catchup Study Group 2002), treated immediately. Although intravenous chloramphenicol is though amoxicillin costs twice as much as co-trimoxazole. With superior to intramuscular chloramphenicol, the procedure can respect to the duration of antibiotic treatment, studies in delay urgently needed treatment and adds to its cost. Bangladesh, India, and Indonesia indicate that three days of oral Investigators have questioned the adequacy and safety of co-trimoxazole or amoxicillin are as effective as five days of intramuscular chloramphenicol. Although early studies sug- either drug in children with nonsevere pneumonia (Agarwal gested that adult blood levels after intramuscular administra- and others 2004; Kartasasmita 2003). In a multicenter study of tion were significantly less than those achieved after intra- intramuscular penicillin versus oral amoxicillin in children with venous administration, the intramuscular route gained severe pneumonia, Addo-Yobo and others (2004) find similar wide acceptance following clinical reports that confirmed its cure rates. Because patients were treated with oxygen when efficacy. Local complications of intramuscular chlorampheni- needed for hypoxemia and were switched to other antibiotics if col succinate are rare, unlike the earlier intramuscular prepara- the treatment failed, this regimen is not appropriate for treating tions. Although concerns about aplastic anemia following severe pneumonia in an outpatient setting. 488 | Disease Control Priorities in Developing Countries | Eric A. F. Simoes, Thomas Cherian, Jeffrey Chow, and others WHO recommends administering oxygen, if there is ample of pneumococcal vaccines, because global and regional esti- supply, to children with signs and symptoms of severe pneu- mates of the pneumococcal pneumonia burden are currently monia and, where supply is limited, to children with any of being developed and will not be available until later in 2005. In the following signs: inability to feed and drink, cyanosis, res- addition, current vaccine prices are relatively stable in devel- piratory rate greater than or equal to 70 breaths per minute, oped countries, but the prices for low- and middle-income or severe chest wall retractions (WHO 1993). Oxygen should countries are expected to be substantially lower when vaccines be administered at a rate of 0.5 liter per minute for children are purchased through a global tender. younger than 2 months and 1 liter per minute for older chil- We evaluate case-management intervention strategies for dren. Because oxygen is expensive and supply is scarce, espe- LRIs in children under five. Health workers who implement cially in remote rural areas in developing countries, WHO rec- case management diagnose LRIs on the basis of fast breathing, ommends simple clinical signs to detect and treat hypoxemia. lower chest wall indrawing, or selected danger signs in children Despite those recommendations, a study of 21 first-level facil- with respiratory symptoms. Because this method does not dis- ities and district hospitals in seven developing countries found tinguish between pneumonia and bronchiolitis, nor between that more than 50 percent of hospitalized children with LRI bacterial and viral pneumonia, we group these conditions into were inappropriately treated with antibiotics or oxygen the general category of "clinical pneumonia" (Rudan and oth- (Nolan and others 2001)--and in several, oxygen was in short ers 2004). This approach assumes that a high proportion of supply. Clearly, providing oxygen to hypoxemic babies is clinical pneumonia is of bacterial origin and that health work- lifesaving, though no randomized trials have been done to ers can considerably reduce case fatality through breathing rate prove it. diagnosis and timely administration of antibiotics (Sazawal and Black 2003). We calculated treatment costs by World Bank Prevention and Treatment of Pneumonia in HIV-Positive region using standardized input costs provided by the volume Children. Current recommendations of a WHO panel for editors and costs published in the International Drug Price managing pneumonia in HIV-positive children and for pro- Indicator Guide (Management Sciences for Health 2005) and phylaxis of Pneumocystis jiroveci are as follows (WHO 2003): other literature (table 25.1). The analysis addresses four cate- gories of case management, which are distinguished by the · Nonsevere pneumonia up to age 5 years. Oral co-trimoxazole severity of the infection and the point of treatment: should remain the first-line antibiotic, but oral amoxicillin should be used if it is affordable or if the child has been on · nonsevere pneumonia treated by a community health co-trimoxazole prophylaxis. worker · Severe or very severe pneumonia. Normal WHO case- · nonsevere pneumonia treated at a health facility management guidelines should be used for children up to · severe pneumonia treated at a hospital 2 months old. For children from 2 to 11 months, injectable · very severe pneumonia treated at a hospital. antibiotics and therapy for Pneumocystis jiroveci pneu- monia are recommended, as is starting Pneumocystis jirove- Information about these categories of case management and ci pneumonia prophylaxis on recovery. For children age 12 their outcomes is drawn from a report on the methodology to 59 months, the treatment consists of injectable antibi- and assumptions used to estimate the costs of scaling up select- otics and therapy for Pneumocystis jiroveci pneumonia. ed health interventions aimed at children (WHO and Child Pneumocystis jiroveci pneumonia prophylaxis should be Adolescent Health forthcoming). We assumed a total of three given for 15 months to children born to HIV-infected follow-up visits for each patient treated by a community health mothers; however, this recommendation has seldom been worker rather than the twice-daily follow-ups for 10 days rec- implemented. ommended by the report. We also assumed that all severe pneumonia patients receive an x-ray examination, rather than COST-EFFECTIVENESS OF INTERVENTIONS just 20 percent as suggested by the report. Moreover, we assumed a five-hour workday for a community health worker, Pneumonia is responsible for about a fifth of the estimated the minimum workday required for community health work- 10.6 million deaths per year of children under five. Where pri- ers under the Child Health and Survival initiative of the U.S. mary health care is weak, reducing mortality through public Agency for International Development (Bhattacharyya and health measures is a high priority. As noted earlier, the available others 2001). interventions are primary prevention by vaccination and sec- Table 25.2 presents region-specific estimates of average ondary prevention by early case detection and management. treatment costs per episode for the four case-management The cost-effectiveness of Hib vaccines is discussed in chap- strategies. Because we considered the prices of tradable com- ter 20. We did not attempt an analysis of the cost-effectiveness modities such as drugs and oxygen to be constant across Acute Respiratory Infections in Children | 489 Table 25.1 Inputs for Case Management of Pneumonia in Low- and Middle-Income Countries Condition and intervention Cost per unit (2001 US$) Quantity Percentage of patients Nonsevere pneumonia at the community level Oral amoxicillin (15 mg/kg) 0.03/dose 3 doses/day for 3 days 100 Acetaminophen (100-mg tablet) 0.001/dose 6 doses 100 Community health worker houra 1.83/hour 1 initial 1-hour visit and 3 follow-up visits 100 Nonsevere pneumonia at the facility level Oral amoxicillin (15 mg/kg) 0.03/dose 3 doses/day for 3 days 100 Acetaminophen (100-mg tablet) 0.001/dose 6 doses 100 Oral salbutamol (2-mg tablet) 0.003/dose 3 doses/day for 4 days 10 Outpatient health facility visita 1.72/visit 1 visit 100 Severe pneumonia at the hospital level Oral amoxicillin (15 mg/kg) 0.03/dose 3 doses/day for 5 days 100 Nebulized salbutamol (2.5 mg) 0.13/dose 6 doses/day for 4 days 50 Injectable ampicillin (50 mg/kg) 0.21/dose 4 doses/day for 3 days 100 X-raya 9.21/test 1 test 100 Oxygen (1 liter/minute)b 20/day 3.5 days 50 Inpatient hospital carea 10.8/day 3 days 100 Very severe pneumonia at the hospital level Oral amoxicillin (15 mg/kg) 0.03/dose 3 doses/day for 5 days 100 Nebulized salbutamol (2.5 mg) 0.13/dose 6 doses/day for 4 days 50 Injectable ampicillin (50 mg/kg) 0.21/dose 4 doses/day for 5 days 100 Injectable gentamicin (2.5 mg/kg) 0.14/dose 1 dose/day for 10 days 100 Oral prednisolone (1 mg/kg) 0.02/dose 1 dose/day for 3 days 5 X-raya 9.21/test 1 test 100 Oxygen (1 liter/minute)b 20/day 5 days 100 Inpatient hospital carea 10.8/day 5 days 100 Source: Management Sciences for Health 2005. Note: We assumed that the average patient weighs 12.5 kilograms. a. Provided by the volume editors. Input costs vary by region. b. Median costs obtained from Dobson 1991; Pederson and Nyrop 1991; Schneider 2001; WHO 1993. Table 25.2 Average per Episode Treatment Costs of Case-Management Interventions for Acute Lower Respiratory Infection (2001 US$) Region Nonsevere, community level Nonsevere, facility level Severe, hospital level Very severe, hospital level Low- and middle-income countries 8 2 82 172 East Asia and the Pacific 6 2 75 160 Latin America and the Caribbean 13 4 134 256 Middle East and North Africa 22 3 113 223 South Asia 5 2 66 148 Sub-Saharan Africa 7 2 64 145 Source: Authors' calculations. regions, regional variations were due to differences in hospital We calculated region-specific cost-effectiveness ratios and health worker costs. Latin America and the Caribbean and (CERs) for a model population of 1 million in each region, fol- the Middle East and North Africa had the highest treatment lowing the standardized guidelines for economic analyses (see costs. chapter 15 for details). Input variables included the treatment 490 | Disease Control Priorities in Developing Countries | Eric A. F. Simoes, Thomas Cherian, Jeffrey Chow, and others costs detailed in tables 25.1 and 25.2, region-specific LRI than at the community level, and of all four case-management morbidity rates, adapted from Rudan and others (2004), categories, treatment of very severe clinical pneumonia at the region-specific mortality rates and age structures provided by hospital level was the least cost-effective. Treatment of non- the volume editors, and region-specific urban to rural popula- severe clinical pneumonia at the facility level was more cost- tion ratios (World Bank 2002). The Europe and Central Asia effective than treatment by a community health worker region was excluded from this analysis because of a lack of because of the lower cost of a single visit to a health facility than incidence information. In the absence of region-specific infor- of multiple visits by a health worker. The CER of providing all mation, we assumed uniform intervention effectiveness rates. levels of treatment to all low- and middle-income countries Disability-adjusted life years are averted through reduced was estimated at US$398 per disability-adjusted life year. duration of illness and decreased mortality with treatment. We Because we assumed that effectiveness rates were constant, assumed an average illness duration of 8.5 days for those not regional variations in the CER for each case-management cat- treated and of 6.0 days for those treated. We used a case-fatality egory were due only to variations in the intervention costs, and reduction of 36.0 percent on account of treatment (Sazawal the relative cost-effectiveness rankings for the strategies was the and Black 2003) and a diagnosis specificity of 78.5 percent for same for all the regions. Variation in the CERs for providing all patients diagnosed based on breath rate alone. The disability categories of care was also due to region-specific urban to rural weight cotemporaneous with infection was 0.28. We did not population ratios. We assumed that all patients in urban areas consider disabilities caused by chronic sequelae of LRIs because seek treatment at the facility level or higher, whereas 80 percent it is unclear whether childhood LRI causes long-term impaired of nonseverely ill patients in rural areas receive treatment at lung function or whether children who develop impaired lung the community level and the remainder seek treatment at the function are more prone to infection (von Mutius 2001). facility level. Because a single year of these interventions yields only cotemporaneous benefits--because effectively treated individ- IMPLEMENTATION OF ARI CONTROL STRATEGIES: uals do not necessarily live to life expectancy given that they are likely to be infected again the following year--we calculated the LESSONS OF EXPERIENCE cost-effectiveness of a five-year intervention. This time period The lessons of ARI prevention and control strategies that have enabled us to consider the case in which an entire cohort of been implemented by national programs include the vaccina- newborns to four-year-olds avoids early childhood clinical tion and case-management strategies discussed below. pneumonia mortality because of the intervention and receives the benefit of living to life expectancy. Finally, this analysis con- sidered only long-run marginal costs, which vary with the Vaccine Strategies number of individuals treated, and did not include the fixed Hib vaccine was introduced into the routine infant immuniza- costs of initiating a program where none currently exists. tion schedule in North America and Western Europe in the Table 25.3 presents the region-specific CERs of the four early 1990s. With the establishment of the Global Alliance for case-management categories as well as the CER for providing Vaccines and Immunization (GAVI) and the Vaccine Fund, all four categories to a population of 1 million people. Among progress is being made in introducing it in developing coun- all low- and middle-income countries, treatment of nonsevere tries, although major hurdles remain. By 2002, only 84 of the clinical pneumonia was more cost-effective at the facility level 193 WHO member nations had introduced Hib vaccine. Five Table 25.3 CERs of Case-Management Interventions for Pneumonia (2001US$/disability-adjusted life year) Nonsevere, Nonsevere, Severe, Very severe, Provision of all Region community level facility level hospital level hospital level four interventions Low- and middle-income countries 208 50 2,916 6,144 398 East Asia and the Pacific 439 91 6,511 13,945 900 Latin America and the Caribbean 547 424 14,719 28,106 1,941 Middle East and North Africa 733 180 6,810 13,438 1,060 South Asia 140 28 1,931 4,318 264 Sub-Saharan Africa 139 24 1,486 3,376 218 Source: Authors' calculations. Acute Respiratory Infections in Children | 491 countries have since been approved for support from GAVI for specific mortality was reduced by 42, 36, and 36 percent, Hib vaccine introduction in 2004­5. respectively. These data clearly show that relatively simplified, The United States added 7-PCV to the infant immunization but standardized, ARI case management can have a significant program in 2000. Several other industrialized countries have effect on mortality, not only from pneumonia, but also from plans to introduce the vaccine into their national immuniza- other causes in children from birth to age four. Currently, the tion programs in 2005, whereas others recommend the use of ARI case-management strategy has been incorporated into the vaccine only in selected high-risk groups. In some of these the IMCI strategy, which is now implemented in more than last countries, the definition of high risk is quite broad and 80 countries (see chapter 63). includes a sizable proportion of all infants. The currently Despite the huge loss of life to pneumonia each year, the licensed 7-PCV lacks certain serotypes important in developing promise inherent in simplified case management has not been countries,but the 9-PCV and 11-PCV would cover almost 80 per- successfully realized globally.One main reason is the underuse of cent of serotypes that cause serious disease worldwide. health facilities in countries or communities in which many chil- Despite the success of Hib vaccine in industrial countries dren die from ARIs. In Bangladesh, for example, 92 percent of and the generally appreciated importance of LRIs as a cause of sick children are not taken to appropriate health facilities (WHO childhood mortality, as a result of a number of interlinked fac- 2002). In Bolivia, 62 percent of children who died had not been tors, uptake in developing countries has been slow. Sustained taken to a health care provider when ill (Aguilar and others use of the vaccine is threatened in a few of the countries that 1998). In Guinea, 61 percent of sick children who died had not have introduced the vaccine. First, the magnitude of disease been taken to a health care provider (Schumacher and others and death caused by Hib is not recognized in these countries, 2002). Schellenberg and others' (2003) study in Tanzania shows partly because of their underuse of bacteriological diagnosis that children of poorer families are less likely to receive antibi- (a result of the lack of facilities and resources). Second, because otics for pneumonia than children of better-off families and that the coverage achieved with traditional Expanded Program on only 41 percent of sick children are taken to a health facility. Immunization vaccines remains low in many countries, adding Thus, studies consistently confirm that sick children, especially more vaccines has not been identified as a priority. Third, from poor families, do not attend health facilities. developing countries did not initiate efforts to establish the A number of countries have established large-scale, sustain- utility of the vaccine until after the vaccine had been licensed able programs for treatment at the community level: and used routinely for several years in industrialized countries. Consequently, Hib vaccination has been perceived as an inter- · The Gambia has a national program for community-level vention for rich countries. As a result of all these factors, actual management of pneumonia (WHO 2004b). demand for the vaccine has remained low, even when support · In the Siaya district of Kenya, a nongovernmental organiza- has been available through GAVI and the Vaccine Fund. tion efficiently provides treatment by community health In 2004, the GAVI board commissioned a Hib task force to workers for pneumonia and other childhood diseases explore how best to support national efforts to make evidence- (WHO 2004b). based decisions about introducing the Hib vaccine. On the · In Honduras, ARI management has been incorporated in basis of the task force's recommendations, the GAVI board the National Integrated Community Child Care Program, approved establishment of the Hib Initiative to support those whereby community volunteers conduct growth monitor- countries wishing either to sustain established Hib vaccination ing, provide health education, and treat pneumonia and or to explore whether introducing Hib vaccine should be a diarrhea in more than 1,800 communities (WHO 2004b). priority for their health systems. A consortium consisting of · In Bangladesh, the Bangladesh Rural Advancement the Johns Hopkins Bloomberg School of Public Health, the Committee and the government introduced an ARI control London School of Hygiene and Tropical Medicine, the Centers program covering 10 subdistricts, using volunteer commu- for Disease Control and Prevention, and the WHO has been nity health workers. Each worker is responsible for treating selected to lead this effort. childhood pneumonia in some 100 to 120 households after a three-day training program. · In Nepal during 1986­89, a community-based program for Case-Management Strategies management of ARIs and diarrheal disease was tested in Sazawal and Black's (2003) meta-analysis of community-based two districts and showed substantial reductions in LRI trials of the ARI case-management strategy includes 10 studies mortality (Pandey and others 1989, 1991). As a result, the that assessed its effects on mortality, 7 with a concurrent program was integrated into Nepal's health services and is control group. The meta-analysis found an all-cause mortality being implemented in 17 of the country's 75 districts by reduction of 27 percent among neonates, 20 percent among female community health volunteers trained to detect and infants, and 24 percent among children age one to four. LRI- treat pneumonia. 492 | Disease Control Priorities in Developing Countries | Eric A. F. Simoes, Thomas Cherian, Jeffrey Chow, and others · In Pakistan, the Lady Health Worker Program employs established, delivery systems will have to be established, and approximately 70,000 women, who work in communities countries will need financial support so that the vaccines can be providing education and management of childhood pneu- introduced into their immunization programs. These activities monia to more than 30 million people (WHO 2004b). are being initiated before the launch of vaccine formulations designed for use in developing countries, so as to inform capac- RESEARCH AND DEVELOPMENT AGENDA ity planning, product availability, and pricing. The research and development agenda outlined below summa- rizes the priorities that have been established by advisory Case-Management Strategies groups to the Initiative for Vaccine Research (vaccine interven- In 2003, WHO's Division of Child and Adolescent Health con- tion strategies) and the WHO Division of the Child and vened a meeting to review data and evidence from recent ARI Adolescent Health (case-management strategies). case-management studies and to suggest the following revisions to case-management guidelines and future research priorities: Vaccine Intervention Strategies The GAVI task force on Hib immunization made a number of · Nonsevere pneumonia: recommendations that vary depending on the country. -- Improve the specificity of clinical diagnostic criteria. Countries that have introduced Hib vaccine should focus on -- Reassess WHO's current recommended criteria for documenting its effect and should use the data to inform detecting and managing treatment failure, given the national authorities, development partners, and other agencies high rates of therapy failure. involved in public health to ensure sustained support to such -- Reanalyze data from short-course therapy studies to vaccination programs. Countries eligible for GAVI support that better identify determinants of treatment failure. have not yet introduced Hib vaccines are often hindered by a -- Carry out placebo-controlled trials among children pre- lack of local data and a lack of awareness of regional data. They senting with wheezing and pneumonia in selected set- can address these issues through subregional meetings at which tings that have a high prevalence of wheezing to deter- country experts can pool data and review information from mine whether such children need antibiotics. other countries. In addition, most of the countries need to · Severe pneumonia: In a randomized clinical trial in a con- carry out economic analyses that are based on a standardized trolled environment, Addo-Yobo and others (2004) showed instrument. Finally, all countries that face a high Hib disease that oral amoxicillin is as effective as parenteral penicillin or burden need to develop laboratory facilities so that they can ampicillin; however, the following actions need to be under- establish the incidence of Hib meningitis at selected sites. taken before it can be recommended on a general basis: Countries in which the disease burden remains unclear may -- Analyze data on exclusions from the trial. have limited capacity to document the occurrence of Hib dis- -- Identify predictors that may help distinguish children ease using protocols that are based on surveillance for menin- who require hospitalization and who subsequently gitis invasive disease. They will need to explore the possibilities deteriorate. of using alternative methods for measuring disease burden, -- Reassess WHO's current recommended treatment including the use of vaccine-probe studies. failure criteria for severe pneumonia, given the overall On the basis of experience with introducing Hib and hepa- high rates of therapy failure. titis B vaccines, GAVI took a proactive approach and in 2003 -- Conduct descriptive studies in a public health setting in established an initiative based at the Johns Hopkins School of several centers worldwide, to evaluate the clinical out- Public Health in Baltimore to implement an accelerated devel- comes of oral amoxicillin in children age 2 to 59 months opment and introduction program for pneumococcal vaccines who present with lower chest wall indrawing. (the PneumoADIP; see http://www.preventpneumonia.org). -- Document the effectiveness of WHO's treatment guide- The program's intent is to establish and communicate the value lines for managing children with pneumonia and HIV of pneumococcal vaccines and to support their delivery. infection. Establishing the value of the vaccine involves developing local · LRI deaths: evidence about the burden of disease and the vaccines' poten- -- To help develop more effective interventions to reduce tial effect on public health. This effort can be accomplished LRI mortality, study the epidemiology of LRI deaths in through enhanced disease surveillance and relevant clinical tri- various regions in detail, using routine and advanced als in a selected number of lead countries. Once established, the laboratory techniques. evidence base will be communicated to decision makers and · Oxygen therapy: key opinion leaders to ensure that data-driven decisions are -- Carry out studies to show the effectiveness of oxygen for made. Once the cost-effectiveness of routine vaccination is managing severe respiratory infections. Acute Respiratory Infections in Children | 493 -- Collect baseline information about the availability and vaccine may fall with the entry of more manufacturers into the delivery of oxygen and its use in hospital settings in low- market in the next few years. Nevertheless, convincing evidence income countries. of the vaccines' cost-effectiveness is required to facilitate -- Explore the utility of pulse oximetry for optimizing oxy- national decisions on introducing the vaccine and using it sus- gen therapy in various clinical settings. tainedly. In low-income countries, positive cost-benefit and -- Undertake studies to improve the specificity of clinical cost-effectiveness ratios alone appear to be insufficient to signs in the overlapping signs and symptoms of malaria enable the introduction of these vaccines into national immu- and pneumonia. nization programs. -- Study rapid diagnostic tests for malaria to assess their effectiveness in differentiating between malaria and pneumonia. REFERENCES -- Examine the effect of widespread use of co- Addo-Yobo, E., N. Chisaka, M. Hassan, P. Hibberd, J. M. Lozano, P. Jeena, trimoxazole on sulfadoxine-pyrimethamine resistance and others. 2004. "Oral Amoxicillin versus Injectable Penicillin for to Plasmodium falciparum. Severe Pneumonia in Children Aged 3 to 59 Months: A Randomised Multicentre Equivalency Study." Lancet 364 (9440): 1141­48. · Etiology: Data on the etiology of pneumonia in children are Adegbola, R. A., A. G. Falade, B. E. Sam, M. Aidoo, I. Baldeh, D. Hazlett, somewhat out of date, and new etiological studies are need- and others. 1994. "The Etiology of Pneumonia in Malnourished and ed that use modern technology to identify pathogens. Well-Nourished Gambian Children." Pediatric Infectious Disease Journal 13 (11): 975­82. Adegbola, R. A., S. O. Usen, M. Weber, N. Lloyd-Evans, K. Jobe, CONCLUSIONS: PROMISES AND PITFALLS K. Mulholland, and others. 1999. "Haemophilus influenzae Type B Meningitis in The Gambia after Introduction of a Conjugate Vaccine." The evidence clearly shows that the WHO case-management Lancet 354 (9184): 1091­92. approach and the wider use of available vaccines will reduce Agarwal, G., S. Awasthi, S. K. Kabra, A. Kaul, S. Singhi, and S. D. Walter ARI mortality among young children by half to two-thirds. The (ISCAP Study Group). 2004. "Three-Day versus Five-Day Treatment systematic application of simplified case management alone, with Amoxicillin for Non-Severe Pneumonia in Young Children: A Multicentre Randomised Controlled Trial." British Medical Journal the cost of which is low enough to be affordable by almost any 328: 791. http://bmj.bmjjournals.com/cgi/content/full/328/7443/791. developing country, will reduce ARI mortality by at least one- Aguilar, A. M., R. Alvarado, D. Cordero, P. Kelly, A. Zamora, and R. Salgado. third. The urgent need is to translate this information into 1998. Mortality Survey in Bolivia: The Final Report--Investigating and actual implementation. Identifying the Causes of Death for Children under Five.Vienna,VA: Basic Support for Institutionalizing Child Survival Project. The case-management strategy has to be applied and Berman, S. 1995a. "Otitis Media in Children" New England Journal of prospectively evaluated so that emerging problems of antimi- Medicine 332 (23): 1560­65. crobial resistance, reduced efficacy of current treatment with . 1995b. "Otitis Media in Developing Countries." Pediatrics 96 the recommended antimicrobials, or emergence of unexpected (1, part 1): 126­31. pathogens can be detected early and remedial steps can be taken Bhattacharyya, K., P. Winch, K. LeBan, and M. Tien. 2001. "Community rapidly. If community-level action by health workers is supple- Health Worker Incentives and Disincentives: How They Affect Motivation, Retention, and Sustainability." Basic Support for mented by the introduction of the IMCI strategy at all levels of Institutionalizing Child Survival Project for the U.S. Agency for primary care, then both applying and evaluating this strategy International Development, Arlington, VA. will be easier. Such synergy may also help in gathering infor- Black, R. E., S. S. Morris, and J. Bryce. 2003. "Where and Why Are mation that will help further fine-tune clinical signs, so that 10 Million Children Dying Every Year?" Lancet 361 (9376): 2226­34. even village health workers can better distinguish bronchiolitis Black, S. B., H. R. Shinefield, B. Fireman, E. Lewis, P. Ray, J. R. Hansen, and and wheezing from bacterial pneumonia. The criticism that the others (Northern California Kaiser Permanente Vaccine Study Center Group). 2000. "Efficacy, Safety, and Immunogenicity of Heptavalent case-management steps may result in overuse of antimicrobials Pneumococcal Conjugate Vaccine in Children." Pediatric Infectious should be countered by documenting their current overuse and Disease Journal 19 (3): 187­95. incorrect use by doctors and other health workers. Although Black, S. B., H. R. Shinefield, B. Fireman, and R. Hiatt. 1992. "Safety, there is a resurgent interest in basing interventions at the com- Immunogenicity, and Efficacy in Infancy of Oligosaccharide Conjugate Haemophilus influenzae Type B Vaccine in a United States munity level, our analysis suggests that doing so may not be Population: Possible Implications for Optimal Use." Journal of cost-effective. Indeed, ARI case management at the first-level Infectious Diseases 165 (Suppl. 43): S139­43. facility may still be the most cost-effective when coupled with Black, S. B., H. R. Shinefield, J. Hansen, L. Elvin, D. Laufer, and better care-seeking behavior interventions. F. Malinoski. 2001. "Postlicensure Evaluation of the Effectiveness of The international medical community is only beginning to Seven-Valent Pneumococcal Conjugate Vaccine." Pediatric Infectious Disease Journal 20 (12): 1105­7. appreciate the potential benefits of Hib and pneumococcal vac- Bobat, R., H. M. Coovadia, D. Moodley, and A. Coutsoudis. 1999. cines. They are currently expensive compared with Expanded "Mortality in a Cohort of Children Born to HIV­1 InfectedWomen from Program on Immunization vaccines, but the price of Hib Durban, South Africa." South African Medical Journal 89 (6): 646­48. 494 | Disease Control Priorities in Developing Countries | Eric A. F. Simoes, Thomas Cherian, Jeffrey Chow, and others Booy, R., S. Hodgson, L. Carpenter, R. T. Mayon-White, M. P. Slack, J. A. Fritzell, B., and S. Plotkin. 1992. "Efficacy and Safety of a Haemophilus Macfarlane, and others. 1994. "Efficacy of Haemophilus influenzae influenzae Type B Capsular Polysaccharide-Tetanus Protein Conjugate Type B Conjugate Vaccine PRP-T." Lancet 344 (8919): 362­66. Vaccine." Journal of Pediatrics 121 (3): 355­62. Bryce, J., C. Boschi-Pinto, K. Shibuya, R. E. Black, and the WHO Child Gessner B. D., A. Sutanto, M. Linehan, I. G. Djelantik, T. Fletcher, I. K. Health Epidemiology Reference Group. 2005. "WHO Estimates of the Gerudug, and others. 2005. "Incidences of Vaccine-Preventable Causes of Death in Children." Lancet 365: 1147­52. Haemophilus Influenzae Type B Pneumonia and Meningitis in Indonesian Children: Hamlet-Randomised Vaccine-Probe Trial." Campbell, H., J. R. M. Armstrong, and P. Byass. 1989."Indoor Air Pollution Lancet 365 (9453): 43­52. in Developing Countries and Acute Respiratory Infection in Children." Lancet 1 (8645): 1012. Ghafoor, A., N. K. Nomani, Z. Ishaq, S. Z. Zaidi, F. Anwar, M. I. Burney, and others. 1990. "Diagnoses of Acute Lower Respiratory Tract Infections Campbell, H., P. Byass, A. C. Lamont, I. M. Forgie, K. P. O'Neill, N. Lloyd- in Children in Rawalpindi and Islamabad, Pakistan." Reviews of Eans, and B. M. Greenwood. 1989. "Assessment of Clinical Criteria for Infectious Diseases 12 (Suppl. 8): S907­14. Identification of Severe Acute Lower Respiratory Tract Infections in Children." Lancet 1 (8633): 297­99. Gilks, C. F. 1993. "Pneumococcal Disease and HIV Infection." Annals of Internal Medicine 118: 393­94. Catchup Study Group. 2002. "Clinical Efficacy of Co-Trimoxazole versus Goel, A., L. Bamford, D. Hanslo, and G. Hussey. 1999. "Primary Amoxicillin Twice Daily for Treatment of Pneumonia: A Randomized Staphylococcal Pneumonia in Young Children: A Review of 100 Cases." Controlled Clinical Trial in Pakistan." Archives of Disease in Childhood Journal of Tropical Pediatrics 45 (4): 233­36. 86: 113­18. Hament, J. M., P. C. Aerts, A. Fleer, H. Van Dijk, T. Harmsen, J. L. Kimpen, Cherian, T., T. J. John, E. A. Simoes, M. C. Steinhoff, and M. John. 1988. and T. F. Wolfs. 2004. "Enhanced Adherence of Streptococcus pneumo- "Evaluation of Simple Clinical Signs for the Diagnosis of Acute Lower niae to Human Epithelial Cells Infected with Respiratory Syncytial Respiratory Tract Infection." Lancet 2: 125­28. Virus." Pediatric Research 55 (6): 972­78. Cherian, T., E. A. Simoes, M. C. Steinhoff, K. Chitra, M. John, Heath, P. T. 1998. "Haemophilus influenzae Type B Conjugate Vaccines: P. Raghupathy, and others. 1990. "Bronchiolitis in Tropical South A Review of Efficacy Data." Pediatric Infectious Disease Journal 17 India." American Journal of Diseases of Children 144 (9): 1026­30. (9 Suppl.): S117­22. Cutts, F. T., S. M. A. Zaman, G. Enwere, S. Jaffar, O. S. Levine, J. B. Okoko, Hortal, M., C. Mogdasy, J. C. Russi, C. Deleon, and A. Suarez. 1990. and others. 2005. "Efficacy of Nine-Valent Pneumococcal Conjugate "Microbial Agents Associated with Pneumonia in Children from Vaccine against Pneumonia and Invasive Pneumococcal Disease in The Uruguay." Reviews of Infectious Diseases 12 (Suppl. 8): S915­22. Gambia: Randomised, Double-Blind, Placebo-Controlled Trial." Ikeogu, M. O., B. Wolf, and S. Mathe. 1997. "Pulmonary Manifestations in Lancet 365 (9465): 1139­46. HIV Seropositive and Malnourished Children in Zimbabwe." Archives de Andrade, A. L., J. G. de Andrade, C. M. Martelli, S. A. Silva, R. M. de of Disease in Childhood 76: 124­28. Oliveira, M. S. Costa, and others. 2004. "Effectiveness of Haemophilus Jeena, P. M., H. M. Coovadia, and V. Chrystal. 1996. "Pneumocystis carinii influenzae B Conjugate Vaccine on Childhood Pneumonia: A Case- and cytomegalo Virus Infections in Severely Ill HIV-Infected African Control Study in Brazil." International Journal of Epidemiology 33 (1): Infants." Annals of Tropical Paediatrics 16: 361­68. 173­81. Jiang, Z., N. Nagata, E. Molina, L. O. Bakaletz, H. Hawkins, and J. A. Patel. Denny, F. W. Jr. 1995. "The Clinical Impact of Human Respiratory Virus 1999. "Fimbria-Mediated Enhanced Attachment of Nontypeable Infections." American Journal of Respiratory and Critical Care Medicine Haemophilus influenzae to Respiratory Syncytial Virus-Infected 152 (4, part 2): S4­12. Respiratory Epithelial Cells." Infection and Immunity 67: 187­92. Dobson,M.1991."Oxygen Concentrators Offer Cost Savings for Developing John, T. J. 1994. "Who Determines National Health Policies?" In Countries: A Study Based on New Guinea." Anaesthesia 146: 217­19. Vaccination and World Health: Fourth Annual Public Health Forum of Douglas, R. M., and H. B. Miles. 1984. "Vaccination against Streptococcus the London School of Hygiene and Tropical Medicine, ed. F. T. Cutts and pneumoniae in Childhood: Lack of Demonstrable Benefit in Young P. G. Smith, 205­11. Chichester, U.K.: John Wiley. Australian Children." Journal of Infectious Diseases 149 (6): 861­69. John, T. J., T. Cherian, M. C. Steinhoff, E. A. Simoes, and M. John. 1991. Duke T, J. Mgone, and D. Frank. 2001. "Hypoxaemia in children with "Etiology of Acute Respiratory Infections in Children in Tropical severe pneumonia in Papua New Guinea." International Journal of Southern India." Reviews of Infectious Diseases 13 (Suppl. 6): S463­69. Tuberculosis and Lung Disease 5: 511­19. Kamath, K. R., R. A. Feldman, P. S. S. Rao, and J. K. Webb. 1969. "Infection Eskola, J., H. Kayhty, A. K. Takala, H. Peltola, P. R. Ronnberg, E. Kela, and and Disease in a Group of South Indian Families." American Journal of others. 1990. "A Randomized, Prospective Field Trial of a Conjugate Epidemiology 89: 375­83. Vaccine in the Protection of Infants and Young Children against Karma, P., J. Luotonen, M. Timonen, S. Pontynen, J. Pukander, E. Herva, Invasive Haemophilus influenzae Type B Disease." New England Journal and others. 1980. "Efficacy of Pneumococcal Vaccination against of Medicine 323 (20): 1381­87. Recurrent Otitis Media: Preliminary Results of a Field Trial in Finland." Annals of Otology, Rhinology, and Laryngology Suppl. 89 Eskola, J., T. Kilpi, A. Palmu, J. Jokinen, J. Haapakoski, E. Herva, and oth- (3, part 2): 357­62. ers. 2001. "Efficacy of a Pneumococcal Conjugate Vaccine against Acute Otitis Media." New England Journal of Medicine 344 (6): 403­9. Kartasasmita, C. 2003. "Three versus Five Days Oral Cotrimoxazole for Nonsevere Pneumonia." Paper presented at the World Health Farley, J. J., J. C. King, P. Nair, S. E. Hines, R. L. Tressier, and P. E. Vink. Organization Consultative Meeting on Reviewing Current Research and 1994. "Invasive Pneumococcal Disease among Infected and Un- Management of Acute Respiratory Infections, Geneva, September 29­ infected Children of Mothers with Human Immunodeficiency Virus October 1. Infection." Journal of Pediatrics 124: 853­58. Kilpi, T., H. Ahman, J. Jokinen, K. S. Lankinen, A. Palmu, H. Savolainen, Fiddian-Green, R. G., and S. Baker. 1991. "Nosocomial Pneumonia in the and others. 2003. "Protective Efficacy of a Second Pneumococcal Critically Ill: Product of Aspiration or Translocation?" Critical Care Conjugate Vaccine against Pneumococcal Acute Otitis Media in Infants Medicine 19: 763­69. and Children: Randomized, Controlled Trial of a Seven-Valent Fireman, B., S. B. Black, H. R. Shinefield, J. Lee, E. Lewis, and P. Ray. Pneumococcal Polysaccharide-Meningococcal Outer Membrane 2003. "Impact of the Pneumococcal Conjugate Vaccine on Otitis Protein Complex Conjugate Vaccine in 1,666 Children." Clinical Media." Pediatric Infectious Disease Journal 22 (1): 10­16. Infectious Diseases 37 (9): 1155­64. Acute Respiratory Infections in Children | 495 Klugman, K. P., S. A. Madhi, R. E. Huebner, R. Kohberger, N. Mbelle, Neuzil, K. M., Y. Zhu, M. R. Griffin, K. M. Edwards, J. M. Thompson, S. J. N. Pierce, and others. 2003. "A Trial of a 9-Valent Pneumococcal Tollefson, and P. F. Wright. 2002. "Burden of Interpandemic Influenza Conjugate Vaccine in Children with and Those without HIV in Children Younger Than 5 Years: A 25-Year Prospective Study." Infection." New England Journal of Medicine 349 (14): 1341­48. Journal of Infectious Diseases 185: 147­52. Kolstad, P. R., G. Burnham, H. D. Kalter, N. Kenya-Mugisha, and R. E. Nolan T., P. Angos, A. J. Cunha, L. Muhe, S. Qazi, E. A. Simoes, and others. Black. 1997. "The Integrated Management of Childhood Illness in 2001. "Quality of Hospital Care for Seriously Ill Children in Less- Western Uganda." Bulletin of the World Health Organization 75 Developed Countries." Lancet 357 (9250): 106­10. (Suppl. 1): 77­85. Obaro, S., A. Leach, and K. W. McAdam. 1998. "Use of Pneumococcal Lagos, R., I. Horwitz, J. Toro, O. San Martin, P. Abrego, C. Bustamante, and Polysaccharide Vaccine in Children," Lancet 352 (9127): 575. others. 1996. "Large Scale, Postlicensure, Selective Vaccination of O'Brien, K. L., L. H. Moulton, R. Reid, R. Weatherholtz, J. Oski, L. Brown, Chilean Infants with PRP-T Conjugate Vaccine: Practicality and and others. 2003. "Efficacy and Safety of Seven-Valent Conjugate Effectiveness in Preventing Invasive Haemophilus influenzae Type B Pneumococcal Vaccine in American Indian Children: Group Infections." Pediatric Infectious Disease Journal 15: 216­22. Randomised Trial." Lancet 362 (9381): 355­61. Lehmann, D., T. F. Marshall, I. D. Riley, and M. P. Alpers. 1991. "Effect of Onyango F. E., M. C. Steinhoff, E. M. Wafula, S. Wariua, J. Musia, and Pneumococcal Vaccine on Morbidity from Acute Lower Respiratory J. Kitonyi. 1993. "Hypoxaemia in Young Kenyan Children with Acute Tract Infections in Papua New Guinean Children." Annals of Tropical Lower Respiratory Infection."British Medical Journal 306 (6878):612­15. Paediatrics 11 (3): 247­57. Pandey, M. R., N. M. Daulaire, E. S. Starbuck, R. M. Houston, and Levine, O. S., R. Lagos, A. Munoz, J. Villaroel, A. M. Alvarez, P. Abrego, and K. McPherson. 1991. "Reduction in Total Under-Five Mortality in others. 1999. "Defining the Burden of Pneumonia in Children Western Nepal through Community-Based Antimicrobial Treatment Preventable by Vaccination against Haemophilus influenzae Type B." of Pneumonia." Lancet 338 (8773): 993­97. Pediatric Infectious Disease Journal 18 (12): 1060­64. Pandey, M. R., P. R. Sharma, B. B. Gubhaju, G. M. Shakya, R. P. Neupane, Lucas, S. B., C. S. Peacock, A. Hounnou, K. Brattegaard, K. Koffi, A. Gautam, and others. 1989. "Impact of a Pilot Acute Respiratory M. Honde, and others. 1996."Disease in Children Infected with HIV in Infection (ARI) Control Programme in a Rural Community of the Hill Abidjan, Côte d'Ivoire." British Medical Journal 312: 335­38. Region of Nepal." Annals of Tropical Paediatrics 9 (4): 212­20. Madhi, S. A., K. P. Klugman, and the Vaccine Trialist Group. 2004. "A Role Pederson, J., and M. Nyrop. 1991. "Anaesthetic Equipment for a for Streptococcus pneumoniae in Virus-Associated Pneumonia." Nature Developing Country." British Journal of Anaesthesia 66: 264­70. Medicine 10 (8): 811­13. Peiris, J. S., W. C. Yu, C. W. Leung, C. Y. Cheung, W. F. Ng, J. M. Nicholls, Madhi, S. A., K. Petersen, A. Madhi, M. Khoosal, and K. P. Klugman. 2000. and others. 2004. "Re-emergence of Fatal Human Influenza A Subtype "Increased Disease Burden and Antibiotic Resistance of Bacteria H5N1 Disease." Lancet 363 (9409): 617­19. Causing Severe Community Acquired Lower Respiratory Tract Infections in Human Immunodeficiency Virus 1 Infected Children." Perkins, B. A., J. R. Zucker, J. Otineo, H. S. Jafari, L. Paxton, S. C. Redd, and Clinical Infectious Diseases 31: 170­76. others. 1997. "Evaluation of an Algorithm for Integrated Management of Childhood Illness in an Area of Kenya with High Malaria Madhi, S. A., K. Petersen, A. Madhi, A. Wasas, and K. P. Klugman. 2000. Transmission." Bulletin of the World Health Organization 75 (Suppl. 1): "Impact of Human Immunodeficiency Virus Type 1 on the Disease 33­42. Spectrum of Streptococcus pneumoniae in South African Children." Pediatric Infectious Disease Journal 19 (12): 1141­47. Quiambao, B. P., E. A. Simoes, E. Abucejo-Ladesma, L. S. Gozum, S. P. Lupisan, L. T. Sombrero, and P. J. Ruutu (ARIVAC Consortium). Makela, P. H., M. Sibakov, E. Herva, J. Henrichsen, J. Luotonen, Forthcoming. "Serious Community Acquired Pediatric Infections in M. Timonen, and others. 1980. "Pneumococcal Vaccine and Otitis Rural Asia (Bohol Island, Philippines)." Pediatric Infectious Disease Media." Lancet 2 (8194): 547­51. Journal. Management Sciences for Health. 2005. International Drug Price Indicator Redd, S. 1994. "Diagnosis and Management of Acute Respiratory Guide. Cambridge, MA: Management Sciences for Health. Infections in Lesotho." Health Policy and Management 5: 255­60. McCullers, J.A., and K. C. Bartmess. 2003."Role of Neuraminidase in Lethal Riley, I. D., F. A. Everingham, D. E. Smith, and R. M. Douglas. 1981. Synergism between Influenza Virus and Streptococcus pneumoniae." "Immunization with a Polyvalent Pneumococcal Vaccine: Effect of Journal of Infectious Diseases 187: 1000­9. Respiratory Mortality in Children Living in the New Guinea Monto, A. S., and B. M. Ullman. 1974. "Acute Respiratory Illness in an Highlands." Archives of Disease in Childhood 56 (5): 354­57. American Community: The Tecumseh Study." Journal of the American Riley, I. D., D. Lehmann, and M. P. Alpers. 1991. "Pneumococcal Vaccine Medical Association 227 (2): 164­69. Trials in Papua New Guinea: Relationships between Epidemiology of Mulholland, E. K., E. A. Simoes, M. O. Castales, E. J. McGrath, E. M. Pneumococcal Infection and Efficacy of Vaccine." Reviews of Infectious Manalac, and S. Gove. 1992. "Standardized Diagnosis of Pneumonia Diseases 13 (Suppl. 6): S535­41. in Developing Countries."Pediatric Infectious Disease Journal 11: 77­81. Rudan, I., L. Tomaskovic, C. Boschi-Pinto, and H. Campbell (WHO Mulholland, K., S. Hilton, R. Adegbola, S. Usen, A. Oparaugo, C. Child Health Epidemiology Reference Group). 2004. "Global Estimate Omosigho, and others. 1997. "Randomised Trial of Haemophilus of the Incidence of Clinical Pneumonia among Children under Five influenzae Type-B Tetanus Protein Conjugate Vaccine for Prevention of Years of Age." Bulletin of the World Health Organization 82 (12): Pneumonia and Meningitis in Gambian Infants." Lancet 349 (9060): 895­903. 1191­97. Santosham, M., M. Wolff, R. Reid, M. Hohenboken, M. Bateman, J. Goepp, Mulholland, K., S. Usen, R. Adegbola, and M. Weber. 1998. "Use of and others. 1991. "The Efficacy in Navajo Infants of a Conjugate Pneumococcal Polysaccharide Vaccine in Children." Lancet 352 (9127): Vaccine Consisting of Haemophilus influenzae Type B Polysaccharide 575­76. and Neisseria meningitidis Outer-Membrane Protein Complex." New England Journal of Medicine 324 (25): 1767­72. Nathoo, K. J., F. K. Nkrumah, D. Ndlovu, D. Nhembe, J. Pirie, and H. Kowo. 1993. "Acute Lower Respiratory Tract Infection in Hospitalized Sazawal, S., and R. E. Black. 2003. "Pneumonia Case Management Trials Children in Zimbabwe." Annals of Tropical Paediatrics 13: 253­61. Group: Effect of Pneumonia Case Management on Mortality in 496 | Disease Control Priorities in Developing Countries | Eric A. F. Simoes, Thomas Cherian, Jeffrey Chow, and others Neonates, Infants, and Preschool Children--A Meta-analysis of Wall, R. A., P. T. Corrah, D. C. Mabey, and B. M. Greenwood. 1986. "The Community-Based Trials." Lancet Infectious Diseases 3: 547­56. Etiology of Lobar Pneumonia in The Gambia." Bulletin of the World Schellenberg, J. A., C. G. Victora, A. Mushi, D. de Savigny, D. Schellenberg, Health Organization 64 (4): 553­58. H. Mshinda, and others. 2003. "Inequities among the Very Poor: Weber, M. W., E. K. Mulholland, S. Jaffar, H. Troedsson, S. Gove, and B. M. Health Care for Children in Rural Southern Tanzania." Lancet 361 Greenwood. 1997. "Evaluation of an Algorithm for the Integrated (9357): 561­66. Management of Childhood Illness in an Area with Seasonal Malaria in Schneider,G.2001."Oxygen Supply in Rural Africa:A Personal Experience." The Gambia." Bulletin of the World Health Organization 75 (Suppl. 1): International Journal of Tuberculosis and Lung Disease 5 (6): 524­26. 25­32. Schumacher, R., E. Swedberg, M. O. Diallo, D. R. Keita, H. Kalter, and Wenger, J. D., J. DiFabio, J. M. Landaverde, O. S. Levine, and T. Gaafar. O. Pasha. 2002. Mortality Study in Guinea: Investigating the Causes of 1999. "Introduction of Hib Conjugate Vaccines in the Non- Death in Children under Five. Arlington, VA: Save the Children and the industrialized World: Experience in Four `Newly Adopting' Countries." Basic Support for Institutionalizing Child Survival Project. Vaccine 18 (7­8): 736­42. Shann, F. 1986. "Etiology of Severe Pneumonia in Children in Developing Wenger, J. D., and M. M. Levine, eds. 1997. Epidemiological Impact of Countries." Pediatric Infectious Disease 5 (2): 247­52. Conjugate Vaccine on Invasive Disease Caused by Haemophilus influen- zae Type B. New York: Marcel Dekker. Shann, F., M. Gratten, S. Germer, V. Linnemann, D. Hazlett, and R. Payne. 1984. "Aetiology of Pneumonia in Children in Goroka Hospital, Papua Whitney, C. G., M. M. Farley, J. Hadler, L. H. Harrison, N. M. Bennett, New Guinea." Lancet 2 (8402): 537­41. R. Lynfield, and others. 2003. "Decline in Invasive Pneumococcal Disease after the Introduction of Protein-Polysaccharide Conjugate Shann, F., K. Hart, and D. Thomas. 1984. "Acute Lower Respiratory Tract Vaccine." New England Journal of Medicine 348 (18): 1737­46. Infections in Children: Possible Criteria for Selection of Patients for Antibiotic Therapy and Hospital Admission." Bulletin of the World WHO (World Health Organization). 1990. "Antibiotics in the Treatment Health Organization 62: 749­51. of Acute Respiratory Infections in Young Children." Unpublished doc- ument WHO/ARI/90.10, available on request from the Division of Simoes, E. A. 1999. "Respiratory Syncytial Virus Infection." Lancet 354 Child Health and Development, formerly the Division of Diarrhoeal (9181): 847­52. and Acute Respiratory Disease Control, WHO, Geneva. Simoes, E. A., T. Desta, T. Tessema, T. Gerbresellassie, M. Dagnew, and . 1991. "Management of the Young Child with an Acute S. Gove. 1997. "Performance of Health Workers after Training in Respiratory Infection. Supervisory Skills Training Module." Un- Integrated Management of Childhood Illness in Gondar, Ethiopia." published document, available on request from the WHO Division of Bulletin of the World Health Organization 75 (Suppl. 1): 43­53. Child Health and Development, formerly the Division of Diarrhoeal Sloyer, J. L. J., J. H. Ploussard, and V. M. Howie. 1981. "Efficacy of and Acute Respiratory Disease Control, WHO, Geneva. Pneumococcal Polysaccharide Vaccine in Preventing Acute Otitis . 1993."Oxygen Therapy for Acute Respiratory Infections in Young Media in Infants in Huntsville, Alabama." Reviews of Infectious Diseases Children in Developing Countries." Geneva, WHO. http://www. 3 (Suppl.): S119­23. who.int/child-adolescent-health/New_Publications/CHILD_ Stensballe, L. G., J. K. Devasundaram, and E. A. Simoes. 2003. "Respiratory HEALTH/WHO_ARI_93.28.htm. Syncytial Virus Epidemics: The Ups and Downs of a Seasonal Virus." . 2002. The Multicountry Evaluation of IMCI Effectiveness, Cost and Pediatric Infectious Disease Journal 22 (2 Suppl.): S21­32. Impact (MCE): Progress Report, May 2001­April 2002. Strauss, W. L., S. A. Qazi, Z. Kundi, N. K. Nomani, and B. Schwartz WHO/FCH/CAH/02.16. Geneva: Division of Child and Adolescent (Co-trimoxazole Study Group). 1998. "Antimicrobial Resistance and Health and Development, WHO. Clinical Effectiveness of Co-trimoxazole versus Amoxycillin for . 2003. Consultative Meeting on Management of Children with Pneumonia among Children in Pakistan: Randomised Controlled Pneumonia and HIV Infection, 30­31 Jan. 2003, Harare, Zimbabwe. Trial." Lancet 352: 270­74. WHO/FCH/CAH/03.4. Geneva: WHO. Temple, K., B. Greenwood, H. Inskip, A. Hall, M. Koskela, and M. . 2004a. "Review Panel on Haemophilus influenzae Type B (Hib) Leinonen. 1991. "Antibody Response to Pneumococcal Capsular Disease Burden in Bangladesh, Indonesia, and Other Asian Countries, Polysaccharide Vaccine in African Children." Pediatric Infectious Bangkok, 28­29 January 2004." Weekly Epidemiological Record 79 (18): Disease Journal 10 (5): 386­90. 173­75. Tupasi, T. E., M. G. Lucero, D. M. Magdangal, N. V. Mangubat, M. E. . 2004b. "WHO/UNICEF Joint Statement: Management of Sunico, C. U. Torres, and others. 1990. "Etiology of Acute Lower Pneumonia in Community Settings." WHO/FCH/CAG/04.06. WHO, Respiratory Tract Infection in Children from Alabang, Metro Manila." Geneva. Reviews of Infectious Diseases 12 (Suppl. 8): S929­39. WHO (World Health Organization) and Child Adolescent Health. UNAIDS (Joint United Nations Programme on HIV/AIDS). 2002. AIDS Forthcoming. Report on the Methodology and Assumptions Used to Epidemic Update. Geneva: UNAIDS. Estimate Costs of Scaling Up Selected Child Health Interventions to 95% Usen S., M. Weber, K. Mulholland, S. Jaffar, A. Oparaugo, C. Omosigho, in Order to Reduce Under-Five Mortality. Geneva: WHO. and others. 1999. "Clinical Predictors of Hypoxaemia in Gambian Williams, B. G., E. Gouws, C. Boschi-Pinto, J. Bryce, and C. Dye. 2002. Children with Acute Lower Respiratory Tract Infection: Prospective "Estimates of Worldwide Distribution of Child Deaths from Acute Cohort Study." British Medical Journal 318 (7176): 86­91. Respiratory Infections." Lancet Infectious Diseases 2: 25­32. Van den Hoogen, B. G., J. C. de Jong, J. Groen, T. Kuiken, R. de Groot, R. A. World Bank. 2002. World Development Indicators. CD-ROM. Washington, Fouchier, and A. D. Osterhaus. 2001. "A Newly Discovered DC: World Bank. Human Pneumovirus Isolated from Young Children with Respiratory Tract Disease." Nature Medicine 7: 719­24. Zwi, K. J., J. M. Pettifior, and N. Soderlund. 1999. "Paediatric Hospital Admissions at a South African Urban Regional Hospital: The Impact von Mutius, E. 2001. "Pediatric Origins of Adult Lung Disease." Thorax of HIV, 1992­1997." Annals of Tropical Paediatrics 19: 135­42. 56: 153­57. Vuori-Holopainen, E., and H. Peltola. 2001. "Reappraisal of Lung Tap: Review of an Old Method for Better Etiologic Diagnosis of Childhood Pneumonia." Clinical Infectious Diseases 32 (5): 715­26. Acute Respiratory Infections in Children | 497 Chapter 26 Maternal and Perinatal Conditions Wendy J. Graham, John Cairns, Sohinee Bhattacharya, Colin H. W. Bullough, Zahidul Quayyum, and Khama Rogo The Millennium Declaration includes two goals directly rele- these interventions require a functioning health system to have vant to maternal and perinatal conditions: reducing child mor- an effect at the population scale. Levels of maternal and peri- tality and improving maternal health. The fact that two out of natal mortality are thus regarded as sensitive indicators of the the eight Millennium Development Goals (MDGs) are exclu- entire health system (Goodburn and Campbell 2001), and they sively targeted at mothers and children is testament to the sig- can therefore be used to monitor progress in health gains more nificant proportion of the global burden of disease they suffer generally. What is also clear is that maternal mortality and the and to the huge inequities within and between countries in the neonatal component of child mortality continue to represent magnitude of their burden. Achieving these goals is inextrica- two of the most serious challenges to the attainment of the bly linked at the biological, intervention, and service delivery MDGs, particularly in South Asia and Sub-Saharan Africa. levels (Bale and others 2003). An estimated 210 million women become pregnant each Maternal and child health services have long been seen as year, and close to 60 million of these pregnancies end with the inseparable partners, although over the past 20 years the rela- death of the mother ( 500,000) or the baby or as abortions. tive emphasis within each, particularly at a policy level, has var- This chapter focuses on the adverse events of pregnancy and ied (De Brouwere and Van Lerberghe 2001). The launch of the childbirth and on the intervention strategies to eliminate and Safe Motherhood Initiative in the late 1980s, for example, ameliorate this burden. brought heightened attention to maternal mortality, whereas the International Conference on Population and Development (ICPD) broadened the focus to reproductive health and, more EPIDEMIOLOGY OF MATERNAL recently, to reproductive rights (Germain 2000). Those shifts AND PERINATAL CONDITIONS can be linked with international programmatic responses and terminology--with the preventive emphasis of, for instance, Much has been written about the lack of reliable data on prenatal care being lowered as a priority relative to the treat- maternal and perinatal conditions in developing countries ment focus of emergency obstetric care. For the child, inte- (AbouZahr 2003; Graham 2002; Save the Children 2001). Weak grated management of childhood illnesses has brought routine information systems, inadequate vital registration, and renewed emphasis to maintaining a balance between preventive reliance on periodic household surveys as the main source of and curative care. The particular needs of the newborn, how- population-based data are all familiar obstacles to improving ever, have only started to receive significant attention in the public health in poor countries (Godlee and others 2004). past three or four years (Foege 2001). Recognizing the implications of these obstacles for prioritizing Although health experts agree that the single clinical inter- health needs and interventions is important and is now ventions needed to avert much of the burden of maternal and endorsed by a global movement toward evidence-based deci- perinatal death and disability are known, they also accept that sion making for policy and practice (Evans and Stansfield 2003). 499 However, there has been much less appreciation of the conse- perinatal period can happen at any age, although it tends to quences for evaluations of effectiveness--and thus cost- take place during the neonatal period (up to 28 days of life). By effectiveness--of the weaknesses in current outcomes measure- contrast, perinatal deaths include both stillborn babies and ment and in routine data collection.Those weaknesses also affect those who are born alive but die before the end of the seventh the monitoring of progress toward the MDGs. Initiatives for day. Early neonatal deaths only include live births. improved health surveillance are thus urgently needed (CMH 2002). For the vast majority of the world's population, the mag- nitude of adverse maternal and perinatal outcomes is not known Nature and Characteristics reliably. It is impossible to determine whether many of the pat- Pregnancy and childbirth are not inherently pathological. terns apparently observed, especially at a cause-specific level, Maintaining an effective balance, however, between preserving are real or are artifacts of the measurement process. normality and ensuring a state of readiness to deal with abnor- mality represents a fundamental challenge to health systems and a tension in safe motherhood programming. Although this Definitions balance between prevention and treatment is not peculiar to The terms maternal and perinatal encompass a continuum of maternal and perinatal conditions (or complications), the fol- health states--from the most positive (complete physical, men- lowing additional characteristics are relevant to assessing the tal, and social well-being) to the most negative--and a huge burden as well as the effectiveness of interventions: number of clinical conditions. This chapter focuses on eight major conditions, hereafter referred to as the focus conditions, · The principle of "first, do no harm" has particular signifi- which are estimated to account for about 75 percent of mater- cance in this area, because many preventive practices related nal deaths and more than 60 percent of perinatal deaths. For to pregnancy and childbirth can readily become harmful in the mother, these conditions are hemorrhage, sepsis, hyperten- unskilled hands--for example, inappropriately early induc- sive disorders of pregnancy, obstructed labor, and unsafe abor- tion of labor or poor forceps technique. The iatrogenic bur- tion. For the baby, they are low birthweight, birth asphyxia, and den of maternal and perinatal conditions is rarely factored infection (table 26.1). into assessments of intervention effectiveness. We define maternal conditions as encompassing events · The lives of two individuals, mother and baby, are poten- occurring from conception to 42 days postpartum (WHO tially at stake (Stoll and Measham 2001); however, interven- 1992a). The chapters on women's health, family planning, ado- tions will not necessarily benefit both equally, and indeed, lescent health, and surgery address the longer-term sequelae of some will be in direct conflict. pregnancy and childbirth; the preconception period; preg- · A large number of maternal and perinatal conditions pre- nancy at an early age; and specific interventions, such as repair sent clinically not as single entities but as complexes, such of obstetric fistulas. Within the period from conception to 42 as hemorrhage and sepsis or preterm delivery and birth days postpartum, two broad categories of conditions can be asphyxia. For the mother, the situation may be further com- distinguished: those arising specifically from pregnancy and plicated by the role of underlying conditions, such as parturition (direct obstetric conditions), and those aggravated by HIV/AIDS underlying puerperal sepsis. or aggravating to pregnancy (indirect obstetric conditions). · The most extreme negative outcome, death of both the Because the latter conditions, such as malaria, HIV/AIDS, or mother and the baby, is highly concentrated around the anemia, are not exclusive to pregnant or parturient women, time of delivery, from the onset of labor or abortion to they are not dealt with here but in the relevant disease-specific 48 hours postpartum or postabortion. Estimates indicate chapters. that about two-thirds of maternal deaths occur within this Regarding perinatal conditions, we focus on those for which time window (AbouZahr 1998), and the proportion for interventions can be directed to the baby through the mother perinatal deaths appears to be even higher (Bale and others during pregnancy or delivery. Our discussion is complemented 2003). For the mother, however, a growing number of stud- by the discussion in chapter 27, which concentrates on the ies highlight the contribution of direct and indirect causes neonate, including special care of the small baby and emer- of deaths, including violence, when a one-year postpartum gency care of the sick newborn. reference period is used (Etard, Kodio, and Traore 1999; Hoj Formal definitions of perinatal conditions tend to vary by and others 2003). data source. Taken literally, they refer to conditions that arise in · The initial clinical presentation of some conditions can be the perinatal period (Murray and Lopez 1998), which are not severe, with rapid escalation to a life-threatening state, and the same as events that occur in the perinatal period--that is, these conditions often require surgical intervention. from 28 weeks of gestation to the end of the seventh day of life. · A distinct clinical feature of some maternal conditions is For example, death resulting from conditions that arise in the their unpredictability (AbouZahr 1998). This fact has had a 500 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others Table 26.1 Maternal and Perinatal Focus Conditions and Risk Factors for These Conditions Case Average duration Risk factors for condition Risk factors for death from condition Definition or complications fatality ratea until death if Timing of Distal or Direct, Distal or Direct, Condition and sequelae (percent) condition fatal presentation proximate physiological proximate physiological Maternal Hemorrhage Definition Not 12 hours 28 weeks of gestation Primigravidity Placental abnormalities Remote location Lack of blood Antepartum hemorrhage: available up to delivery Grand multiparity (including placenta previa; Anemia transfusion bleeding from the genital tract (greater than 4) abruption; placenta Coagulopathies Badly managed third during the last 3 months of accreta, percreta, increta; stage of labor pregnancy Fibroids other adhesions) Delay or absence of Primary postpartum hemorrhage: 1.0b 2 hours Delivery to 24 hours Anemia Polyhydramnios oxytocic treatment excessive bleeding (more than after delivery Multiple gestation 500 milliliters) from the genital Previous third-stage tract following delivery complication Previous cesarean section Preeclampsia, eclampsia Intrauterine death Hepatitis Induced labor Prolonged labor Precipitate labor Forceps delivery Cesarean section Chorioamnionitis Disseminated intravascular coagulation Sepsis Definition 1.3 6 days Delivery to 6 weeks Immunosuppression Prolonged labor Delivery by Misdiagnosis Infection of the genital tract or postpartum Anemia Obstructed labor untrained personnel Inappropriate use of extragenital infections following Immunosuppression antibiotics Maternal Sexually transmitted Premature rupture of childbirth infections membranes Anemia Lack of access to Inadequate prenatal Frequent pelvic Lack of knowledge intravenous and care examinations about warning signs antibiotics Perinatal Intrauterine death Lack of postnatal Foreign body insertion care Conditions (for example, herbs) Cultural practices Episiotomy | Instrumental delivery 501 (Continues on the following page.) 502 Table 26.1 Continued | Disease Case Average duration Risk factors for condition Risk factors for death from condition Definition or complications fatality ratea until death if Timing of Distal or Direct, Distal or Direct, Control Condition and sequelae (percent) condition fatal presentation proximate physiological proximate physiological Cesarean section Priorities Unhygienic delivery conditions in Developing Retained products of conception Hypertensive Definition 1.7 2 days (eclampsia) 28 weeks of gestation Extremes of maternal Multiple gestations Cultural practices Appearance of Countries disorders of Raised blood pressure with to 2 days postpartum age Molar pregnancy Lack of knowledge complications, such as pregnancy proteinuria Primigravidity cardiovascular and Previous history of Lack of prenatal cerebral complications, | Genetic predisposition pregnancy-induced care W hemolysis, elevated endy Racial or geographical hypertension or chronic liver enzyme, low predisposition hypertension J. platelets syndrome Graham, Diabetes and chronic Disseminated hypertension intravascular John Lack of prenatal care coagulation Cairns, Eclampsia Obstructed labor Definition 0.7 3 days During labor Malnutrition Cephalopelvic Lack of access to Uterine rupture Sohinee Labor in which progress is Rickets in childhood disproportion cesarean delivery Hemorrhage arrested by mechanical factors Bony deformity of Malpresentation, position Lack of access to Sepsis Bhattacharya, pelvis instrumental Exhaustion, delivery and Achondroplasia dehydration symphysiotomy Short stature and Scarred uterus Primigravidity others Inappropriate use Grand multiparity of oxytocin Adolescent pregnancy Unsafe abortion Definition 0.3 6 hours to 6 days After first missed Unwanted pregnancy Absence of aseptic Sociocultural Perforated uterus Procedure for terminating an period to 22 weeks of Adolescence technique factors Poisoning from unintended pregnancy carried gestation or fetal Unmarried status Foreign body insertion Lack of access to abortifacients out by people lacking the weight of less than Poisoning from safe termination Peritonitis necessary skills or in an 500 grams Absence of legal abortifacients services environment that does not abortion services Septic shock Lack of access to conform to minimal medical Lack of access to Acute renal failure postabortion care standards or both contraception Hepatorenal failure Lack of access to safe Bowel injury, abortion services perforation Sexually transmitted Hemorrhagic shock infections Peritonitis Perinatal c Low birthweight (less Complications or sequelae 50 5 days Less than 24 hours Extremes of maternal Multiple pregnancy Lack of adequate Birth asphyxia than 2,500 grams)d Respiratory insufficiency in age Short interpregnancy neonatal care Intraventricular preterm infants with lung Race, ethnicity interval facility hemorrhage immaturity presenting as Low socioeconomic First or second trimester Lack of knowledge Central nervous system respiratory distress syndrome status bleeding and understanding injury because of surfactant deficiency Unmarried status Placenta previa Respiratory infection Neonatal cerebral injury caused 80 3 days 1­4 days by periventricular hemorrhage Lack of education Preeclampsia Respiratory distress mediated by perinatal stress Parity (0 or greater Anemia syndrome such as hypotension or trauma than 4) Hyperemesis Necrotizing Severe physiological jaundice of 50 1­5 days 2­5 days Smoking, alcohol use enterocolitis Isoimmunization preterm infant Maternal malnutrition Cholestatic liver Fetal abnormalities Difficulties in establishing 20 1­14 days First day disease Maternal diabetes or Cervical incompetence spontaneous feeding and hypertension Other infections inability to tolerate feeds Oligohydramnios or Genetic factors Sudden infant death resulting from prematurity polyhydramnios syndrome Rubella, other viral Failure of closure of the ductus 70 3 days to months 3­14 days infection Neonatal coagulopathy arteriosus, frequently seen in preterm babies with lung disease Poor obstetric history Hypoglycemia and other 2 7 days Birth Diethylstilboestrol, metabolic disorders related to other toxic exposure prematurity High altitude Absent or inadequate prenatal care Birth asphyxia Complications or sequelae 20 20 minutes Birth (5 minutes) Drugs taken during Prolonged or obstructed Badly conducted Central nervous system (excluding birth Absent or depressed breathing labor, including labor labor injury trauma) at birth 3 days to life Birth to first 12 hours anesthesia Abruptio placentae Lack of fetal Neonatal Neonatal encephalopathy: 30 Maternal diabetes Placental infarct, monitoring encephalopathy clinically evident disturbance in Maternal hypertension insufficiency Lack of partograph (seizures and recurrent neurological behavior, commonly apnea) Preeclampsia Postmaturity Lack of neonatal with early neonatal seizures in resuscitation Maternal term babies, resulting from an Any other severe Prematurity or low facilities event causing hypoxia during illness birthweight delivery Multiple pregnancies and Perinatal Placenta previa or separation Cord prolapse Conditions (Continues on the following page.) | 503 504 Table 26.1 Continued | Disease Case Average duration Risk factors for condition Risk factors for death from condition Definition or complications fatality ratea until death if Timing of Distal or Direct, Distal or Direct, Control Condition and sequelae (percent) condition fatal presentation proximate physiological proximate physiological Priorities Infection Complications or sequelae 30­40 5 days First 3 days Lack of adequate Premature rupture of Congenital HIV Preterm delivery Neonatal sepsis of early onset prenatal care membranes infection Septic shock resulting from intrauterine or in Maternal infection Preterm delivery Lack of adequate Respiratory failure Developing intrapartum infection Lack of maternal Birth asphyxia neonatal care Hepatorenal failure Neonatal sepsis of late onset 15 5 days After 3 days immunization Unhygienic delivery and resulting from nosocomial Coagulopathies Unhygienic cultural cord care Countries infection or lack of immunity to practices commensal bacteria Tetanus neonatorum, commonly 80 3­7 days 3­14 days | W resulting from unhygienic cutting endy of the cord or care of the cord J. stump Graham, Congenital syphilis resulting 30 5 days Birth onward from transplacental infection John with Treponema pallidum after 18 weeks gestation Cairns, HIV infection transmitted either Direct effects mainly Sohinee intrapartum or postpartum after neonatal period Source: Maternal conditions: Chamberlain 1995; case fatality rates: AbouZahr 2003; mechanical factors of obstructed labor: WHO 1994; unsafe abortion: WHO 1992b; low birthweight: Bale and others 2003, Robertson 1993, Yasmin and others 2001; birth asphyxia: Bale and Bhattacharya, others 2003, p. 324, Robertson 1993; infections: Robertson 1993; risk factors: Calder and Dunlop 1992, Murray and Lopez 1998. a. Case fatality rates assume that no intensive care is available, because this is the norm in South Asia and Sub-Saharan Africa. b. Case fatality for severe pph (blood loss 1000 ml). c. Excludes stillbirths. d. Includes preterm deliveries and small for gestational age. and others profound effect on the prioritization of interventions in safe developing country (1 in 6) and the lowest estimate for a devel- motherhood, and it is an area in urgent need of further oped country (1 in 29,800) (WHO 2004b). This differential is research. The situation is confused by the alternative end- often cited as the largest discrepancy between the developing points, such as death or disability, and by the extent to which and developed world of all public health statistics, reflecting there are clear and predictable risk factors. Table 26.1 sum- major differences both in obstetric risk, as measured by the marizes some of these key characteristics as they relate to the maternal mortality ratio, and in levels of fertility, as reflected in eight focus conditions. the total fertility rate. In terms of medical causes of maternal mortality, even greater caution is needed regarding the reliability of any pat- Causes and Conceptual Frameworks terns observed, because of their dependence on whether the One of the most frequently quoted figures in safe motherhood data are health service based or population based and on cod- is that 88 to 98 percent of maternal deaths are avoidable with ing conventions. Figure 26.1a shows the percentage distribu- moderate levels of health care (WHO 1986). This advocacy tion among direct causes at a crude global level. Direct causes statement simplifies the multiple pathways leading to death account for about 80 percent of all maternal deaths, with indi- and, thus, the multiple opportunities for primary and second- rect causes responsible for the remainder. Of the direct causes, ary prevention. In part, this simplicity is a further reflection of hemorrhage is generally regarded as the most common and the grouping together of clinical conditions that in reality are may be underestimated, because health facilities are unaware of distinctly different in terms of prevalence, case fatality, and many such deaths, given the short interval between onset and scope for intervention, such as eclampsia and puerperal sepsis death (see table 26.1). In terms of indirect causes, the pattern or congenital anomalies and birth asphyxia. The multiple end- varies enormously between different parts of the world, points and conditions, for both the mother and the fetus or primarily according to the prevalence of HIV/AIDS, malaria, newborn, have implications for what is regarded as an and tuberculosis. antecedent (a cause, a determinant, or a risk factor)1 and what The published data on severe maternal morbidity are is regarded as a consequence (an outcome or a sequela). weaker still. A recent World Health Organization (WHO) sys- A large number of conceptual frameworks depict path- tematic review indicates how prevalence figures vary hugely ways to adverse maternal and perinatal outcomes (Bale and according to the criteria used to identify cases (Say, Pattinson, others 2003; McCarthy and Maine 1992). Several identify three and Gulmezoglu 2004). Using disease-specific criteria, WHO levels of contributory factors, which are also found in causal found that prevalence ranged from 0.80 to 8.23 percent. Using models for general health outcomes (WHO 2002): (a) distal, organ system criteria, WHO found that the range was 0.38 to (b) proximal or intermediate, and (c) physiological or direct. 1.09 percent. Finally, using management-based criteria, WHO Table 26.1 highlights the risk factors for the focus maternal and found that the range was 0.01 to 2.99 percent. Estimates sug- perinatal conditions. The distal determinants emphasize that gest that for every maternal death, at least 16 or 17 other maternal and perinatal well-being is not just a medical issue. women suffer a life-threatening complication during preg- Improvements throughout the health sector must be comple- nancy or childbirth (Gay and others 2003) and at least 30 mented by attention to wider social, economic, and cultural women are left with long-term disabilities, such as an obstet- factors as well as to reproductive rights (CMH 2002). Many ric fistula (UNFPA 2003). These estimates must be regarded as conceptual frameworks also differentiate between the timing of crude approximations, most originating from small-scale interventions: before pregnancy, during pregnancy, during studies and most in urgent need of updating and verification. labor and delivery, or during the postpartum period. Similarly, Given the varying case fatality rates shown in table 26.1, the a further distinction can be made in terms of the timing of the fact that the distributional pattern for morbidity (fig- outcome, although from a programmatic perspective, such a ure 26.1b) does not completely mirror the one for mortality is temporal focus may lead to fragmented care for women and not surprising. their babies. As concerns mortality in babies, an estimated 5.7 million perinatal deaths occur each year, 47 percent as stillbirths and 53 percent in the first week of life (J. Zupan, personal commu- Levels, Trends, and Differentials nication, August 25, 2004). Many of those deaths are linked The latest regional estimates of maternal mortality are for directly with complications experienced by the mothers, and 2000­1 (table 26.2), with most of the figures for the developing several studies have shown that the survival prospects for a world produced by modeling (WHO 2004b). More than 99 per- baby whose mother dies are generally poor--less than 1 per- cent of annual maternal deaths occur in the developing world. cent in one study in Bangladesh (Koenig, Fauveau, and At a national level, the magnitude of the differential in terms of Wojtyniak 1991). In 2004, neonatal deaths represented 36 per- lifetime risk is almost 500-fold between the highest figure for a cent of all deaths of children under five in developing Maternal and Perinatal Conditions | 505 Table 26.2 Estimates of Maternal Mortality by Region, 2000­1 Range of uncertainty of Maternal mortality Number of Estimated Lifetime risk maternal mortality ratio (maternal maternal deaths number of of maternal ratio estimates Total deaths per 100,000 as modeled maternal death (1 in Lower Upper fertility Region live births), 2000 by WHO, 2000 deaths, 2001 number shown) estimate estimate rate Central and Eastern 64 3,400 3,000 770 29 100 1.6 Europe, Commonwealth of Independent States, Baltic states, Europe, and Central Asia East Asia and the 110 37,000 37,000 360 44 210 2.0 Pacific Eastern and 980 123,000 -- 15 490 1,500 5.5 Southern Africa Latin America and 190 22,000 16,000 160 110 280 2.6 the Caribbean Middle East and 220 21,000 15,000 100 85 380 3.7 North Africa South Asia 560 205,000 199,000 43 370 760 3.5 Sub-Saharan Africa 940 240,000 237,000 16 400 1,500 5.7 Western and 900 118,000 -- 16 310 1,600 5.9 Central Africa High-income countries 13 1,300 1,000 4,000 8 17 1.6 Low- and middle- 440 527,000 507,000 61 230 680 3.0 income countries Low-income countries 890 236,000 -- 17 410 1,400 5.4 World 400 529,000 508,000 74 210 620 2.7 Source: WHO 2004b, 2004d; UN 2002. -- not available. Note: The regions are those used by the United Nations Children's Fund. countries, with about 1 million of these 3.94 million neonatal or special inquiries (see, for example, Pattinson 2002). Recent deaths occurring in the first week of life (Jamison and others WHO (2004c) statistics on unsafe abortion show an apparent 2004). Table 26.3 presents modeled estimates for early neonatal decrease in incidence in all world regions, although the risk of deaths in 2001. The data on the magnitude and patterns of still- death remains high at 50 per 100,000 live births, and in parts of births remain particularly poor. Sub-Saharan Africa the risk is as high as 140 per 100,000 live Given weak sources of information, the dearth of reliable births (Rogo, Bohmer, and Ombaka 1999). These adverse trends data is hardly surprising. At a global level, a major diffi- events, however, are often also the most seriously under- culty arises from the need to use models to estimate maternal reported, as elaborated further in chapter 57. mortality. As the basic methodology for the models has changed The availability of reliable trends data for perinatal mortali- over time, the data are not appropriate for trend assessment. ty is even more problematic. A demand for population-based AbouZahr and Wardlaw (2001) provide patchy support for estimates for newborn mortality is comparatively recent; thus, downward trends in some parts of the world, mostly on the there has been insufficient time to accumulate multiple data basis of civil registration data and mostly restricted to countries points. Demographic and health surveys (DHSs) are a key with maternal mortality ratios of less than 100 per 100,000 live source for tracking trends in infant and child mortality. Several births--thus notably excluding South Asia and Sub-Saharan DHSs now have data that can be disaggregated to show neona- Africa. Even where declines appear to have occurred, they did so tal deaths, but only a few have information on stillbirths, and prior to 1990. Countries with sustained falls since then, such as the quality of that information is still being assessed. Argentina and China, cannot be regarded as representative of all Information from WHO suggests that early neonatal death rates developing countries. Cause-specific trend data are extremely fell slightly, from 28 per 1,000 live births around 1980 to about rare, often gathered through small-scale hospital-based studies 25 per 1,000 in 2000, for low- and middle-income countries, 506 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others a. Maternal mortality and the equivalent trend for stillbirths is suggested to be a drop from 36 per 1,000 deliveries to 22 per 1,000 deliveries (J. Zupan, Other maternal personal communication, August 25, 2004). 22% Hemorrhage Two types of differentials are particularly relevant: geo- 28% graphic (or regional) and socioeconomic. Table 26.2 indicates the wide variation in the magnitude of maternal mortality across regions, and a similar difference can be seen between Unsafe abortion countries. In terms of absolute numbers of deaths, just 13 13% countries account for 70 percent of the global total (WHO Sepsis 2004b).2 Caution is again needed, because the poorest coun- Obstructed 15% tries also have the weakest information systems and, therefore, labor Hypertensive 8% have estimates derived solely from modeling. One regression disorders model (WHO 2004b), for example, uses independent variables, 14% such as the percentage of deliveries with health professionals b. Maternal morbidity present and the proportion of deaths of women of reproduc- Hemorrhage tive age that are maternal deaths. Those variables are them- Other maternal 18% 22% selves subject to error and likely to be least reliable where information systems are weakest. Geographic differences in maternal mortality within countries are poorly documented, although remote populations are often assumed to suffer the Sepsis 16% highest levels because of poor access to emergency obstetric care. Although this assumption seems logical, few reliable data are available to confirm or refute it, and the possibility of high Unsafe abortion 26% Hypertensive levels of mortality in urban areas linked to unsafe abortion disorders (Thonneau and others 2002) makes the topic of geographic Obstructed 9% differentials a priority for research. labor 9% Until recently, socioeconomic differentials in mortality have Source: Mortality: WHO 2004d; Morbidity: Murray and Lopez 1998. tended to be inferred from utilization patterns for prenatal care Note: Nonobstetric (indirect) causes of death and morbidity, such as tuberculosis and health professionals at delivery. The DHSs continue to and malaria, have been excluded. provide the main data sources in this regard, for both interna- tional and national analyses, and they demonstrate huge differ- Figure 26.1 Medical Causes of Direct Maternal Mortality and Morbidity (percentage distribution) ences between wealth quintiles. A relevant recent development, however, is the familial technique, which can be used to examine socioeconomic differences in maternal mortality using existing survey data (Graham and others 2004). Because Table 26.3 Early Neonatal Deaths by Gender and Cause, 2001 (thousands) Worlda South Asia Sub-Saharan Africaa Cause All Male Female All Male Female All Male Female Perinatal conditionsb 2,522 1,400 1,123 1,086 596 489 573 332 241 Low birthweightc 1,301 710 591 757 406 351 243 141 102 Birth asphyxia 739 432 307 192 122 70 240 139 101 (including birth trauma) Other perinatal conditionsd 482 258 225 137 68 68 90 52 38 Source: WHO 2004d. a. Excludes the island of Mayotte. b. Excludes stillbirths, congenital malformations, neonatal tetanus, congenital syphilis, acquired infections (respiratory and sepsis), and diarrhea. c. Includes preterm deliveries and small for gestational age. d. Includes all conditions originating in the perinatal period (P00­P96 codes in perinatal chapter of WHO 1992a), apart from low birthweight and asphyxia. Maternal and Perinatal Conditions | 507 Table 26.4 DALYs for Perinatal and Maternal Conditions by Gender, Selected Regions, 2001 (thousands) Worlda South Asia Sub-Saharan Africaa Condition All Male Female All Male Female All Male Female Maternal 26,789 n.a. 26,789 10,069 n.a. 10,069 9,743 n.a. 9,743 Hemorrhage 3,928 n.a. 3,928 1,718 n.a. 1,718 1,643 n.a. 1,643 Sepsis 5,348 n.a. 5,348 1,857 n.a. 1,857 1,843 n.a. 1,843 Hypertensive disorders of 1,895 n.a. 1,895 742 n.a. 742 842 n.a. 842 pregnancy Obstructed labor 2,506 n.a. 2,506 1,185 n.a. 1,185 919 n.a. 919 Unsafe abortion 3,507 n.a. 3,507 1,467 n.a. 1,467 1,557 n.a. 1,557 Perinatalb 90,505 49,384 41,117 37,721 20,442 17,279 20,046 11,351 8,697 Low birthweightc 43,073 23,241 19,832 25,015 13,292 11,723 7,891 4,501 3,391 Birth asphyxia (including 31,972 17,945 14,025 8,283 4,957 3,326 9,256 5,195 4,062 birth trauma) Other perinatal conditionsd 15,460 8,198 7,260 4,423 2,193 2,230 2,899 1,655 1,244 Source: WHO 2004d. n.a. not applicable. a. Excludes the island of Mayotte. b. Excludes stillbirths, congenital malformations, neonatal tetanus, congenital syphilis, acquired infections (respiratory and sepsis), and diarrhea. c. Includes preterm deliveries and small for gestational age. d. Includes all conditions originating in the perinatal period (P00­P96 codes in perinatal chapter of WHO 1992a) apart from low birthweight and asphyxia. maternal health and health care are clearly associated with still- The former focused on avoidable mortality resulting primarily births and early neonatal deaths, the same differentiating fac- from direct obstetric conditions, whereas the latter considered tors are likely to apply to perinatal outcomes. Indeed, data from population risk assessments and highlighted the contribution many DHSs show large gaps between rich and poor in relation of indirect obstetric problems--especially micronutrient to neonatal mortality, with the greatest average disparity being deficiencies--and the role for preventive strategies. Clearly, the found in Latin American and the Caribbean (http://www. choice between different measures of burden has a crucial worldbank.org/poverty/health/). influence both on the strategic approach to achieving health gains and on the prioritization of interventions. Attributable Burden The estimation of maternal and perinatal conditions as part of international assessments of the burden of disease has long INTERVENTIONS been controversial, and much has been written about the prob- Given the scope and nature of the burden of maternal and peri- lems and potential distortions of priorities (AbouZahr 1998; natal conditions, no quick fix is available and, thus, no single Sadana 2001). Some of those criticisms relate to methods of intervention warrants exclusive attention. Rather, clusters or valuation based on disability-adjusted life years (DALYs), espe- packages of interventions need to be considered, and this cially in relation to discounting and the omission of stillbirths, understanding has long been reflected in maternity services and others to the inaccuracies and selectivity of the base data throughout the world (Milne and others 2004). Even though on the incidence of complications, on case fatalities, and on these clusters can be characterized or differentiated solely on disabilities. Table 26.4 presents DALYs for South Asia and Sub- the basis of content--namely, the component interventions-- Saharan Africa for the focus conditions for 2001. Those two in practice, the health system or implementation context is also regions together account for 74 percent of the global burden of a defining factor. maternal conditions and 64 percent of the global burden of perinatal conditions. The significance of the burden of maternal and perinatal conditions is clear from two recent global assessments (CMH Levels and Types of Interventions 2002; WHO 2002). The approaches the two initiatives adopted Box 26.1 presents one example of a comprehensive strategy for have led to different conclusions about public health priorities. safe motherhood. It illustrates the range of programmatic 508 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others Box 26.1 Components of a Comprehensive Safe Motherhood Strategy The following are part of a comprehensive safe mother- -- screening and treatment for syphilis hood strategy: -- antiretrovirals, where voluntary counseling and testing undertaken, and breastfeeding advice · community education on safe motherhood and new- -- tetanus toxoid immunization born care -- treatment of urinary tract infections · evidence-based prenatal care and counseling · skilled assistance at delivery -- nutritional advice · care of obstetric complications and emergencies -- iron and folate supplements (multivitamins and · postpartum care micronutrients) · safe abortion and postabortion services -- iodization of edible oils and salt and vitamin A in · family-planning information and services areas of endemic deficiency · adolescent reproductive health education and services -- blood pressure screening Source: Dayaratna and others 2000. issues raised by maternal and perinatal conditions: maternal death and disability may be avoided by effective, timely, and appropriate clinical interventions, often referred to · the scope for both primary and secondary prevention as emergency obstetric care. · the difference between the individual receiving specific inter- Given this complexity and the multiple approaches used to ventions (here, the mother) and the beneficiary (the baby) address maternal and perinatal conditions, no perfect frame- · the multiple effects of single (component) interventions on work for categorizing interventions exists. We, therefore, clus- different outcomes ter the alternative intervention pathways on the basis of the · the multiple benefits to the same outcome of different following three parameters: interventions · the short- and long-term time frames for interventions and · level of care--home, primary, and secondary outcomes · time period--pregnancy, labor and delivery, and · the balance between supply-side and demand-side postpartum interventions · strategic approach--population-based versus personal · the role for interventions outside the health sector. interventions. Three main pathways are available for averting adverse out- comes: preventing pregnancy, preventing complications, and Quality of Evidence preventing death or disability from complications. The first Pregnancy and childbirth have been the subjects of medical pathway is the only truly primary preventive strategy. It investigation for centuries and, indeed, are among the oldest requires intervention to avert the occurrence or mistiming of clinical specialties.As a consequence, a substantial body of opin- pregnancy by means of effective family-planning methods, as ion exists on the signs, symptoms, etiology, prognosis, natural discussed in chapter 57. This preventive approach is relevant history, and management and treatment options for many for those women who are able to and wish to avoid or delay maternal and perinatal complications, particularly in developed pregnancy, but it has a limited role for those not in this posi- countries. Much of it can be regarded as conventional wisdom tion, estimated at between 15 and 57 percent of women age 15 acquired through practice. In contrast, a comparatively small to 29 (WHO 2002). As concerns the primary prevention of proportion of interventions can be regarded as based on evi- complications, comparatively limited reliable evidence is avail- dence, by contemporary scientific standards, and arrived at able on the true size of the avoidable fraction for many condi- through the conduct of robust research. Thus, in specification of tions at a population level. The emphasis in this preventive the content of intervention clusters, a built-in tension exists pathway is on maintaining normality and on managing mild between using the best available knowledge and using only evi- complications--and thus on good quality of care. Finally, dence that passes minimum quality criteria. Equally important Maternal and Perinatal Conditions | 509 is recognizing the fundamental distinction between knowing primary sources are available, but there are a variety of mod- what is effective at an individual case-management level, for eled estimates, such as Prata and others (2004), Walsh and oth- which an evidence base exists for maternal and perinatal condi- ers (1993), and Winikoff and Sullivan (1987). Model estimates tions, and demonstrating effectiveness at the aggregate levels of vary enormously in terms of the size of the effect, depending composite strategies and entire countries or regions, for which primarily on assumptions about the proportion of maternal robust evidence is extremely limited (Graham 2002). deaths caused by unsafe abortion. Investigators estimate the potential gain from the avoidance of unintended or mistimed pregnancies to be a 20 percent decrease in maternal deaths in Population-Based Interventions developing countries (Donnay 2000; Kurjak and Bekavac 2001; The primary aim of population-based interventions is to UNICEF 1999). reduce the risks leading to adverse outcomes at the population level rather than at the individual level (WHO 2002). Nutritional Interventions. Maternal undernutrition encom- Population-based interventions are essentially preventive and passes two main dimensions: underweight and micronutrient seek to promote healthy behaviors, thereby reducing incidence deficiencies (principally iron and vitamin A). Unlike many of in the entire population. In the case of maternal and perinatal the direct maternal complications, which are acute at onset and conditions, such an approach could be adopted for two major of relatively short duration, these nutritional problems are risk factors: lack of contraception and maternal undernutri- chronic and long term and, indeed, are intergenerational tion. The grade of evidence for these population-based inter- (Tomkins 2001). The physiological mechanisms by which ventions is primarily level C for the former, but a mixture of A undernutrition exerts an influence on outcomes in the mother and B for the latter.3 and baby are not entirely understood, but a large body of epi- demiological evidence supports associations with, for example, Fertility Behavior Change. Fertility behavior is ultimately the fetal growth or length of pregnancy (Villar and others 2002). primary exposure factor for both maternal and perinatal con- Those findings have originated mostly from populations with ditions. Investigators have shown that the frequency (number either severe levels of undernutrition or significant cofactors, and spacing), the timing with regard to age, and the desirabil- such as malaria and other infections. ity of pregnancy are associated with increased risks, although Considerable uncertainty surrounds the issue of timing some dispute remains about the effect of birth intervals. potential interventions, with conflicting opinions about mak- Researchers have also investigated the influence of those factors ing targeted interventions during pregnancy; addressing on perinatal conditions, finding clear associations with old or undernutrition among girl children or adolescents, and apply- young maternal age, short interpregnancy intervals, and high ing strategies for women of reproductive age, including peri- or first birth order, with many of those variables also being conceptual women (Gay and others 2003; Rush 2000). Further interrelated (Bale and others 2003). debate relates to the use of supplements versus food fortifica- Lack of effective use of contraception may result in tion. A systematic review by Villar and others (2002) of ran- unwanted or mistimed pregnancies. Unintended pregnancies domized controlled trials to prevent or treat adverse maternal are known to be associated with adverse maternal outcomes, outcomes and preterm delivery concludes that limited evidence including unsafe abortion. Contraceptive behavior is clearly supports large-scale interventions with multivitamins, miner- determined by a host of socioeconomic, cultural, religious, and als, or protein-energy supplementation, but that iron and folic medical factors (Hussain, Fikree, and Berendes 2000; Marston acid are effective against anemia. Rouse (2003) emphasizes the and Cleland 2003; Mwageni, Ankomah, and Powell 2001), potential cost-effectiveness of vitamin A or beta-carotene sup- which also have a bearing on intervention options. Most of the plementation in reducing maternal mortality if the findings of options on the demand side focus on information, education, West and others (1999) from Nepal are replicable elsewhere. and communication; those on the supply side focus on client- friendly services. At a macro level, those intervention options have been credited with the substantial increase in contracep- Personal Interventions tive use in developing countries over the past 40 years, which, When we consider interventions directed at individuals rather in turn, is seen as a contributor to the overall fall in the total than whole populations, the need for a continuum of care for fertility rate from 6 to 3 (Cleland and Ali 2004). Nevertheless, a mother and baby in terms of time (before and after delivery), significant unmet need for contraception persists in many place (linking home and health services through an effective developing countries, with high levels of unsafe abortion as a referral chain), and person (the provider of care) is important. proxy indicator of that need. A variety of conceptual frameworks emphasize this continuum As regards evidence of the effectiveness of family planning and the dangers of fragmentation. Care to prevent or treat in explicitly reducing maternal mortality or disability, no the vast majority of maternal and perinatal conditions can be 510 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others provided at home, at the primary level (clinic or health center), Primary-Level Care. Primary-level care is widely regarded as and at the secondary level (district hospital),4 with the district the crucial entry point to maternity services--and also to care or equivalent regarded as the essential planning unit for service before and after pregnancy. The focus here is essentially pre- delivery (WHO 1994). This system is comparable to the ventive, but with the capacity to detect problems, to manage "close-to-client" health system that the Commission on mild complications appropriately, and to stabilize and then Macroeconomics and Health (CMH 2002) has proposed, refer cases that require higher-level care. Although the name whereby trained staff members other than doctors provide used for primary care facilities varies from country to country, much of the care, with an emphasis on primary prevention and we employ the commonly used term health center. In terms of management of acute conditions. functionality in relation to maternal and perinatal care, the health center should provide prenatal, delivery (including Home-Based Care. Two topical interventions that fall into the management of complicated abortion), and postpartum care category of home-based care are (a) information, education, (including family planning and postabortion counseling), as and communication and birth preparedness and (b) male well as care of the newborn. involvement (for home-based newborn care, see chapter 27). The management of complicated cases is usually discussed Evidence in this cluster of interventions falls predominantly at two levels: basic emergency obstetric care (BEmOC) and into the level C category. comprehensive emergency obstetric care (CEmOC), the dis- tinction being made on the basis of the number of signal or Birth Preparedness Many descriptive studies indicate that essential clinical functions performed.5 This distinction forms women, relatives, and other members of the community the basis of a set of process indicators that the United Nations frequently do not recognize danger signs in pregnancy, child- (UN) has endorsed for program monitoring (UNFPA 2003). birth, or the puerperium, and that lack of recognition can have The capacity of health centers to provide BEmOC depends on serious consequences for mother and baby (Gay and others the availability of supplies, drugs, infrastructure, and skilled 2003). Health education interventions at prenatal clinics providers. Some of the signal functions may not always be per- appear to be less successful at raising awareness and increasing formed by midwives or nurses, sometimes because of the regu- the use of emergency obstetric care than the use of pictorial lation of roles by the government or professional bodies. For cards (Khanum and others 2000) or community education this reason, a further distinction can be made between full (Bailey, Szaszdi, and Schieber 1995). BEmOC, which comprises six functions, some of which may Birth preparedness includes planning for the place and the require a doctor, and obstetric first aid, which includes two sig- attendant at delivery, as well as arranging for rapid transfer to a nal functions universally performed by midwives and nurses: health center or hospital, when needed, and sometimes identi- the administration of antibiotics or oxytocics, intravenously or fying a compatible blood donor in the case of hemorrhage intramuscularly. (Portela and Santarelli 2003). Initiatives to promote birth pre- paredness can clearly be home or community based, but stud- Routine Prenatal Care The literature available on routine pre- ies have emphasized the importance of linkages with prenatal natal care is extensive, and there is a long history of assessing the care so as to include appropriate recommendations for intra- component interventions (Hall, MacIntyre, and Porter 1985; partum care (Shehu, Ikeh, and Kuna 1997). In circumstances in Rooney 1992). In safe motherhood programs, prenatal care which prenatal services are of poor quality or are underused, provides one of the rare examples of robust assessment of an traditional birth attendants or relatives are often the only intervention package (Villar and others 2001). As Bale and oth- source of information; thus, initiatives need to reach those ers (2003) note, even though many of the component clinical individuals too. interventions are effective in terms of perinatal outcomes (Bergsjo and Villar 1997), reliable evidence of an effect on Male Involvement Many studies have observed positive bene- maternal mortality in developing countries is not available fits from the involvement of male partners in care-seeking (McDonagh 1996). However, where early detection is followed behavior related to pregnancy and delivery (Gay and others by appropriate treatment, prenatal care does seem to reduce 2003). That involvement is now advocated as an essential adverse outcomes from specific maternal conditions, including element of WHO's Making Pregnancy Safer Initiative (WHO hypertensive disorders of pregnancy, urinary tract infections, 2003). Models and mechanisms for achieving this involvement and breech presentations (Carroli, Rooney, and Villar 2001; have not been robustly evaluated, and considerable controversy Villar and Bergsjo 1997). Conversely, the limited effectiveness of concerns those that are based on behavioral and social prenatal risk screening at a population level is now widely cognitive theories that presume lack of knowledge as the acknowledged (Graham 1998). The poor predictive value of root problem (Portela and Santarelli 2003; Raju and Leonard many screening tools for maternal complications reinforces the 2000). importance of access to emergency obstetric care for all women Maternal and Perinatal Conditions | 511 who develop a need for it and underlies calls for skilled atten- evidence supports a number of clinical interventions, such as dance at all deliveries. Many health experts, however, do accept active management of the third stage of labor, as well as essen- screening and treatment for syphilis and immunization with tial newborn care. tetanus toxoid as important prenatal interventions (Bale and Once again, the principal sources of data on levels and others 2003). Similarly, the prevention and treatment of anemia trends in coverage of skilled attendants at delivery are the and of malaria, with prophylaxis or bednets, are widely regarded DHSs. The data, however, are based on women's self-reports of as essential elements of routine prenatal care. Nutritional sup- who attended their deliveries, include only live births, and have plementation, however, remains more controversial. major definitional uncertainties. Some countries, for example, Prenatal care has been assessed not only in terms of content, use terms such as supervised deliveries and include as attendants but also in relation to alternative models of the number and both auxiliaries and trained traditional birth attendants (see timing of visits (Munjanja, Lindmark, and Nystrom 1996). Bell, Curtis, and Alayon 2003 for a critique of these data). A Strong evidence exists on the cost-effectiveness of a targeted, global analysis of trends in deliveries by skilled attendants four-visit schedule (Villar and others 2001) that includes an showed wide variations in progress across different regions, educational element on the recognition of danger signs and the with the latest figures for Sub-Saharan Africa, Asia, and Latin use of skilled attendance at delivery. America and the Caribbean for 1990­2003 being 48, 59, and The principal sources of international data on levels, trends, 82 percent, respectively (AbouZahr and Wardlaw 2001; and differentials in prenatal care coverage are the DHSs. The WHO 2004a). The proportion of deliveries with health profes- latest statistics show comparatively high coverage levels sionals present (doctors, midwives, nurses) is one of the proxy when measured in terms of one or more visits--levels average indicators for the MDG on maternal health (Graham and 71 percent for Sub-Saharan Africa--but comparatively little Hussein 2004). It demonstrates not only major differentials improvement between 1990 and 2000. Within countries, wide between countries, but also wide variation in uptake across socioeconomic differentials in uptake are apparent. socioeconomic groups within countries (De Brouwere and Van Lerberghe 2001). Although skilled attendants do not necessar- Delivery Care As indicated earlier, the risks of adverse out- ily operate only in fixed health facilities such as health centers, comes in mother and baby are usually highest during the intra- the DHS data show low levels of professional attendance in the partum period. Even though health experts have long appreci- community. Promoting skilled attendance is thus essentially ated this fact, prioritization of this element of safe motherhood advocating for institutionalizing deliveries. is comparatively recent. Much has been written both on this shift in emphasis and on the underlying rationale, as well as Postpartum Care Primary care services continue to neglect on what skilled attendance at delivery should comprise (De the postpartum period despite significant morbidity among Brouwere and Van Lerberghe 2001). Investigators have sug- mothers and babies during this time. Routine performance of gested a variety of conceptual models for defining content, with postnatal checks is not widespread, and most contacts with varying degrees of emphasis on the attendant and on the services after delivery tend to focus on educational messages enabling environment (Bell and others 2003). All these models on, for example, danger signs, breastfeeding, nutrition, and recognize that skilled attendance encompasses both normal lifestyle. and complicated deliveries, with the focus on the former and on the management of mild complications at the primary level, Postabortion Care One significant area of service delivery that as is consistent with BEmOC, and with referral to CEmOC at does not fit well with descriptive frameworks based on prena- the secondary level when necessary. tal, intrapartum, and postpartum care is the management of Key unresolved issues at the primary level relate to the complicated abortions. Unsafe abortion accounts for a signifi- skills and scope of work of the attendant, especially in relation cant proportion of the burden of maternal conditions, but it is to being a multipurpose health worker, and to the potential still treated as the poor relation in the debate on intervention role of nonprofessionals, such as auxiliaries and trained strategies (De Brouwere and Van Lerberghe 2001). In particu- traditional birth attendants (Buttiens, Marchal, and De lar, with the prioritization in recent years of skilled attendance Brouwere 2004). Work by Koblinsky and Campbell (2003) has at delivery, both the service base for and the provider of helped to inform this debate by proposing four basic models postabortion care have become less well defined (Dayaratna of delivery care that vary according to configurations of place and others 2000). This crucial element of obstetric care falls of delivery and attendant. Evidence on the effectiveness of the into BEmOC in the case of mild complications and CEmOC alternative models at a population level is lacking, and sup- for more serious cases, but whether it is regarded as part of port for skilled attendance at delivery is, thus, based primari- prenatal, delivery, or postnatal services appears to vary from ly on historical and contemporary ecological analysis (De setting to setting. Moreover, postabortion care illustrates the Brouwere and Van Lerberghe 2001). Conversely, high-grade dangers of the fragmentation of broader reproductive health 512 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others care, because primary prevention and counseling after treat- signatory countries. However, as observed at the ICPD 10 ment for complications tend to fall within the remit of family- Conference, many promised changes remain at the level of pol- planning services, whereas emergency care at the primary level icy pronouncement and have not yet been implemented. The is usually provided as part of maternity services and at the sec- stagnation is most notable in relation to maternal mortality ondary level may fall within obstetrics or gynecology services. and the HIV pandemic, especially in Sub-Saharan Africa. The failure to fully implement the ICPD consensus can be attrib- Secondary-Level Care. Secondary-level care is hospital-based uted to lack of political will, inadequate funding for programs care, generally at the district level, including CEmOC. As a cen- to further reproductive health, and weak health systems. It is ter for referral, this level of care needs to be linked to the pri- too early to judge the effect of the MDG proclamation mary level through an effective chain of communications (Johansson and Stewart 2002), although it could well suffer the (Murray and others 2001). The focus at the district hospital is same fate unless special attention is given to maternal and child on secondary prevention, with the ability to manage the prin- health in the context of sectorwide approaches and Poverty cipal maternal and perinatal conditions discussed earlier; thus, Reduction Strategy Papers (UNFPA 2003). Some suspect that district hospitals must be able to provide surgical interventions both these modalities may not give reproductive health the and the requisite backup, such as blood banks (Kusiako, focus and attention it requires, because competing needs may Ronsmans, and Van der Paul 2000). In many countries, crowd it out. Others argue, however, that sectorwide approaches however, the district hospital is also the local provider of pre- can be a boon for maternal health because they offer a more ventive services, including prenatal and normal delivery care; as effective platform for addressing ailing health systems such, it is responsible for attending to a wide mix of uncompli- (Goodburn and Campbell 2001). cated and complicated cases. Whether at the national or international level, advocacy for Although no high-grade evidence of the effectiveness of maternal and perinatal health should focus on the following CEmOC is available, many health experts agree that maternal seven key message areas: mortality cannot be significantly reduced in the absence of such care (Bale and others 2003). The issue thus becomes one · magnitude of the problem of the cost-effectiveness of other strategies, given the presence · factors influencing maternal and perinatal outcomes of CEmOC. The UN agencies have endorsed the threshold of · functions of maternal health programs and which interven- one CEmOC facility per 500,000 people. Data indicating the tions work attainment of this ratio--and, indeed, the percentage of met · consequences of not addressing maternal and perinatal need for CEmOC--are not widely available. Similarly, reliable health information on geographic or socioeconomic differentials in · costs of improving maternal and perinatal health access to CEmOC is extremely limited. · responsibilities at each level of the health system and beyond · policy and legal impediments to implementing comprehen- sive safe motherhood and newborn health programs. Policy Considerations and Approaches The health of mothers and babies is a human right and needs Major advocacy networks, such as the Partnership for Safe to be underpinned by policies and laws that increase access to Motherhood and Newborn Health, the White Ribbon Alliance, information and good-quality, affordable health services and the Healthy Newborn Partnership, seek to promote mater- (Germain 2000). A positive policy environment is crucial for nal and newborn health at the global level. Their purpose is to promoting maternal health and reducing the burden of mater- create awareness by changing the language of discourse, build- nal and perinatal conditions. Such policy considerations need ing international political commitment, developing global to go beyond the health sector to include related issues, such as guidelines, and improving access to technical information for transportation, nutrition, girls' access to education, and gender providers and program managers. biases in the control of economic resources. Through a human rights­based approach, programs can be fashioned to ensure that every woman has the right to make informed decisions COST-EFFECTIVENESS OF SELECTED about her own health and has access to quality services before, INTERVENTION PACKAGES during, and after childbirth (Freedman 2001). The ICPD marked a dramatic shift not only by putting the Cost-effectiveness analysis (CEA) faces several major chal- concepts of rights and choice center stage, but also by intro- lenges with respect to evaluating the prevention and treatment ducing the reproductive health paradigm. The first decade of of maternal and perinatal conditions. First is the sheer range of the ICPD plan of action was marked by major improvements conditions and potential interventions. The breadth of the in policies related to maternal health in most of the 179 clinical area implies the need to make tough choices with Maternal and Perinatal Conditions | 513 respect to which packages of interventions to compare. A sec- others 2003), we evaluated intervention packages with respect ond and related challenge is the lack both of reliable data on the to a counterfactual (base scenario), varying the content and burden of conditions and of high-grade evidence on the effec- coverage. We also performed sensitivity analyses to examine the tiveness and costs of packages. As a result, we can assess only effects of changing the values of key variables for costs, effec- the relative cost-effectiveness of different packages of interven- tiveness, or both. Each intervention package scenario specifies tions by means of modeling. Thus, the third set of challenges different dimensions of prenatal and intrapartum care provid- is associated with modeling, which makes the analysis vulner- ed at primary and secondary care facilities. As regards the able to all the usual criticisms of the modeling of cost- assumed pathways through which women with normal or effectiveness--in particular, uncertainty about the direction of complicated pregnancies may or may not access care, the cru- any bias introduced and the difficulty of establishing the valid- cial entry point in our model is prenatal care. That choice influ- ity of the model (Sheldon 1996). Finally, there are the related ences the detection and treatment of mild and severe compli- issues of the appropriateness to maternal and perinatal condi- cations during the antepartum period at both the primary and tions of standard outcome measures used in the model--in the secondary levels, as well as the proportion of women deliv- particular, DALYs, which exclude stillbirths and indirect mater- ering with a health professional present and with improved nal conditions (AbouZahr 1999; De Brouwere and Van access to emergency care for intrapartum or abortion-related Lerberghe 2001). complications. In our CEA model, these effects are achieved primarily through two types of interventions: Selected Intervention Packages · improvements in the quality of care, incorporating the tech- For some of the reasons mentioned in the previous subsection, nical content or the proportion of women in receipt of the researchers have made few attempts to model packages of inter- care needed (that is, met need) ventions for maternal and perinatal conditions, and many of · increases in the coverage of care--namely, the proportion of those attempts do not specify content in sufficient detail to women accessing care. replicate the package. Our approach is to define content by beginning with a literature search of best practices in prevent- Routine prenatal care can be characterized in terms of ing and managing the focus maternal and perinatal conditions, whether it is a basic or an enhanced package--in other words, acknowledging that, by excluding conditions that impose a its technical content (table 26.5)--and by the percentage of lesser burden, we ignore interventions that might be highly women accessing the package--in other words, its coverage. effective and cost-effective. We then grouped those interven- Delivery at a primary-level health center is viewed as having a tions that are considered effective and that are either being or single quality dimension in terms of content--namely, whether likely to be implemented on a substantial scale into packages of BEmOC is available for women who develop mild complica- care, bearing in mind previous CEA work, such as the WHO tions, including complicated abortion (table 26.5). BEmOC is mother-baby package (WHO 1994). Expert panels then assumed to require the presence of a doctor at the health cen- reviewed the component interventions and the packages and ter; otherwise, only obstetric first aid is presumed to be avail- assisted with identifying resource requirements. Given the able, covering just the two signal functions described earlier. complementary CEA elsewhere in this volume on interventions A percentage of women with severe complications who relevant to maternal and perinatal conditions such as family access primary care will go on to secondary care. This percent- planning, we focus on care during pregnancy, postpregnancy age is assumed to be 20 or 50 percent of complicated cases care, and care immediately postdelivery--in other words, on attending primary care. Our model makes no provision for clusters or packages of interventions typically referred to as pre- women who access secondary care directly in the event of a natal care, delivery or intrapartum care, and emergency obstetric serious complication, although it does allow for those who care. Table 26.5 outlines the content of those packages. were attending the hospital as their local provider of primary When one considers the intervention packages, contextual care. Of those women who access the secondary care facility factors are clearly crucial. Given the particularly high burden in from the primary level, a proportion will receive the CEmOC South Asia and Sub-Saharan Africa, we chose those two regions that they need (assumed to vary between 50 and 90 percent of as the specific health system scenarios for this chapter. Those complicated cases that reach secondary care). This figure regions are also characterized by high levels of poverty and reflects such issues as staff skills and motivation and the avail- encompass some of the most heavily indebted countries in the ability of drugs and equipment. For the other quality-of-care world. element--namely, the technical content of CEmOC--we consider two levels: with (enhanced package) and without Comparison of Alternative Intervention Package Scenarios. (base package) selected interventions for high-risk babies Following the approach of generalized CEA (Hutubessy and (table 26.5). 514 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others Table 26.5 Care Packages at the Primary and Secondary Levels Level of care Base Enhanced and condition Content package package Routine prenatal care Clinical examination, including for severe anemia, height and weight, blood pressure at the primary levela Obstetric examination for gestational age estimation and uterine height, fetal heart, detection of malpresentation and position, and referral Gynecological examination Urine test (multiple dipstick) Laboratory tests: hemoglobin, blood type and rhesus status, syphilis and other symptomatic testing for sexually transmitted diseases Advice on emergencies, delivery, lactation, and contraception Education about clean delivery, warning signs, and premature rupture of membranes Iron and folic acid supplementation Multivitamin supplementation -- Tetanus toxoid immunization HIV voluntary testing and counseling -- Antimalarial chemoprophylaxis in endemic areas -- Screening and treatment for syphilis Balanced protein-energy supplementation for all women -- Delivery care at the Clean delivery technique, clean cord cutting, clean delivery of baby and placenta primary levelb Active management of the third stage of labor, including oxytocics Episiotomy in appropriate cases Recognition and first-line management of delivery complications (for example, obstructed labor, early detection of cephalopelvic disproportion, malposition and malpresentation, previous cesarean delivery, postpartum hemorrhage, and preeclampsia or eclampsia) and referral Intravenous fluid Intravenous uterotonics, if bleeding occurs Partograph Essential newborn care Intravenous antibiotics Magnesium sulfate -- Forceps or vacuum extraction -- Manual removal of placenta -- Removal of retained products of conception -- Corticosteroids for preterm labor -- Antiretrovirals for prevention of mother-to-child transmission of HIV -- Antibiotics for premature rupture of membranes -- CEmOC package at the secondary level c Postpartum hemorrhage Recognition of high-risk cases and arrangements for delivery in a facility Grouping of blood Iron and folate supplementation Blood transfusion Uterotonic drugs, oxytocics Bimanual compression of uterus Manual removal of placenta Uterine packing or balloon tamponade Fluid replacement Hysterectomy Removal of products of conception Secondary postpartum hemorrhage management (antibiotics, uterotonics, removal of products of conception, and fluid and blood replacement) (Continues on the following page.) Maternal and Perinatal Conditions | 515 Table 26.5 Continued Level of care Base Enhanced and condition Content package package Antepartum hemorrhage Early detection of major placenta previa and abruption Grouping and saving blood Iron and folate supplementation Cesarean section for major-degree placenta previa, abruption with a live baby Blood and fluid replacement Oxytocics Sepsis Antibiotics for premature rupture of membranes, cesarean section Fluid and blood transfusion Intravenous antibiotics Evacuation of products of conception Drainage of abscess Treatment of shock with fluids or blood, nitroglycerine Pregnancy-induced Early detection and management of preeclampsia hypertension Calcium supplementation in high-risk cases Aspirin to prevent preeclampsia Antioxidants to prevent preeclampsia Intravenous magnesium sulfate Antihypertensive drugs to reduce blood pressure Immediate delivery if more than 36 weeks Magnesium sulfate and antihypertensives for postpartum eclampsia Obstructed labor Partograph Cesarean section Symphysiotomy Destructive operation Antibiotics Fluid and blood transfusion Hysterectomy Abortion Evacuation of retained products of conception Intravenous antibiotics Fluid or blood transfusion Postabortion contraceptive advice Ectopic pregnancy Proof puncture (culdocentesis) -- -- Laparotomy and salpingectomy -- -- Blood transfusion (autotransfusion) High-risk infant Forceps or vacuum extractiond Corticosteroids for preterm labor -- Antiretrovirals for prevention of mother-to-child transmission of HIV -- Antibiotics for premature rupture of membranes -- Source: Authors. -- not available. a. The base package includes the four-visit schedule recommended by WHO (Villar and others 2001). b. The base package includes the provision of obstetric first aid (intravenous or intramuscular antibiotics and oxytocics). The enhanced package includes the availability of a doctor, and thus the full range of BEmOC (UNFPA 2003). In some settings, experienced midwives or clinical officers may perform all six BEmOC functions. c. At the hospital level, prenatal or delivery care will also be provided for normal, uncomplicated cases and, thus, also includes all care listed in the first two panels of the table. d. Forceps or vacuum delivery can also be used for several other conditions, such as prolonged labor (not obstructed), fetal distress, preterm birth, aftercoming head of breech, and preeclampsia to speed up delivery. 516 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others The base case for our CEA model assumes the following: resources. We have assumed that increases in care can be achieved without major capital investments and that human · basic technical content for the prenatal care package resources are not in short supply; therefore, more could be used · prenatal care coverage for 50 percent of pregnancies (with given wage rates) as required for increased activity and · only obstetric first aid (two signal functions) available in enhanced coverage. health centers Table 26.7 summarizes the findings of the CEA in terms of · 20 percent of women with severe complications accessing incremental cost-effectiveness ratios (ICERs) for the six pri- secondary care mary comparisons between the base scenario and alternative · 50 percent of those severe cases receiving the CEmOC that intervention packages for a population of 1 million. Table 26.8 they need. gives details of total costs, deaths averted, life years saved, and DALYs averted. Table 26.9 shows the findings of the sensitivity The different assumptions regarding quality of care and analysis in terms of how the ICERs change when different coverage can be combined in many different ways, yielding a assumptions (see annex 26.A) are made with respect to effec- large number of potential packages and a larger number of tiveness, met need, and inpatient costs. potential comparisons between those and the base package. In interpreting the results, note that they are point esti- However, not all possible scenarios are meaningful. For exam- mates. Even though they are based on the best information ple, because the base prenatal care package does not screen for currently available, all the inputs into the model are subject to HIV, matching that package with enhanced delivery care that some degree of uncertainty. Without access to robust data on provides antiretrovirals to reduce vertical transmission would individual costs and effects or without specifying distributions be inappropriate. We identified six packages for comparison for each variable, it is impossible to identify confidence limits with the base case, representing a range of safe motherhood for the estimated ICERs. Thus, we do not know, for example, strategies and focusing on prenatal and delivery care. Table 26.6 whether the difference in the incremental cost per DALY summarizes these alternatives and indicates their essential averted for Sub-Saharan Africa between increased coverage at characteristics from a safe motherhood perspective. the primary level (US$92) and improved quality of CEmOC (US$151) reflects a genuine difference in cost-effectiveness or whether there are overlapping confidence intervals Resource Use and Costs (table 26.7). We adopted an ingredients approach (Creese and Parker 1994) With those important caveats in mind, at first sight the to identify resource use. For this type of bottom-up costing, we results for South Asia and Sub-Saharan Africa appear quite dif- prepared lists for primary- and secondary-level care facilities of ferent. For each intervention package, regardless of the specific types of personnel, drugs, supplies (medical and nonmedical), assumptions made, the cost per DALY averted is always lower medical and surgical equipment relevant for the interventions, in Sub-Saharan Africa. The higher costs of care in Sub-Saharan and capital items (vehicles, buildings, building space). For most Africa (see annex 26.A) are thus more than compensated for by of the scenarios, our identification of resources was based on the higher effectiveness, which is a result of the region's greater the WHO mother-baby package costing tool (WHO 1999), with burden. However, some important similarities are apparent necessary modifications because of the content of care packages between South Asia and Sub-Saharan Africa. Leaving aside indicated in table 26.5. We estimated the costs for clinical per- options 3b and 5b (the options without nutritional supple- sonnel on the basis of salaries for different grades according to ments), the results for both regions show a consistent pattern. the guidelines provided by the volume editors for the two Improvements in the overall quality of care, especially at the selected regions. The time required by different staff members primary level through the provision of BEmOC (option 3a), for each care intervention and the changes in time and person- together with increased overall coverage (option 5a), are the nel because of varying content and coverage of packages were most cost-effective intervention packages--and both include informed by expert panel reviews, and we then calculated the nutritional supplements. They are followed by increased cover- costs. We valued the other nontraded inputs using information age at the primary level (option 2). Improved quality of primarily provided by WHO-CHOICE (2004). CEmOC (option 4) is the least cost-effective option. Removing nutritional supplements from the packages makes relatively lit- tle difference in Sub-Saharan Africa, slightly increasing cost- Cost-Effectiveness Ratios effectiveness, but in South Asia, options 3b and 5b become less The CEA involves a number of fixed and variable assumptions cost-effective with the nutritional supplements removed. The (see annex 26.A). The most important assumptions concern explanation lies in the ICERs of nutritional supplements as the reducible burden of these conditions, the effectiveness of such, which are US$48 or US$45 in South Asia and US$118 or the interventions, and the availability of appropriate human US$110 in Sub-Saharan Africa, depending on whether the Maternal and Perinatal Conditions | 517 518 Table 26.6 Comparisons Undertaken for CEA | Disease Primary level Secondary level Control Abbreviated Quality of care: Quality of care: Quality of care: Safe description Option technical Percentage receiving technical motherhood Resource Priorities of package number Coverage content Coveragea care needed content Interpretation strategy implications Routine maternity Base 50 percent of See first two 20 percent of 50 percent of those See table 26.5c Basic package of Content of package Costs typical of in care pregnant women panels of complicated cases reaching the secondary prenatal and essentially the same WHO mother- Developing attend prenatal table 26.5 at the primary level receive the CEmOC delivery care as WHO mother-baby baby package care; 50 percent of level referred to needed package, plus pregnant women the secondary magnesium sulfate Countries have professional level and active manage- intrapartum careb ment of labor | Increased primary- 2 70 percent prenatal No change from No change from No change from base No change from Benefit from Information, educa- Costs of informa- W endy level coverage care; 70 percent base base base increasing coverage tion, and communica- tion, education, delivery care tion for increasing and communica- J. Graham, uptake of prenatal tion; increased and delivery care personnel; drugs John Improved overall 3a No change from Enhanced prenatal No change from 70 percent Enhanced CEmOC Benefit from enhanc- Provision of BEmOC Costs of doctors quality of care base and delivery care base (adds interventions ing quality (content at the primary level and equipment at Cairns, with nutritional (BEmOC) for high-risk babies) and receipt of care the primary level, supplements needed) at the primary training for Sohinee and secondary levels BEmOC and CEmOC, costs of Bhattacharya, BPS Improved overall 3b No change from Enhanced prenatal No change from 70 percent Enhanced CEmOC As for 3a without BPS As for 3a As for 3a without quality of care base and delivery care base (adds interventions costs of BPS and without nutritional (BEmOC) without for high-risk babies) others supplements BPS Improved quality 4 No change from No change No change 80 percent No change Benefit from increased Improved quality Cost of additional of CEmOC base from base from base from base percentage of women of CEmOC personnel time with severe complica- and drugs tions receiving the CEmOC needed Improved overall 5a 70 percent prenatal Enhanced prenatal 50 percent 90 percent Enhanced CEmOC Benefit from improved Comprehensive Costs of quality of care and care; 70 percent and delivery care (adds interventions quality (technical con- package: improved providing and coverage with delivery care (BEmOC) for high-risk babies) tent and percentage coverage and content running ambu- nutritional receiving care needed) with BPS lances, costs supplements and coverage at the of additional primary and secondary personnel and levels drugs, training for BEmOC and CEmOC, costs of BPS Improved overall 5b 70 percent prenatal Enhanced prenatal 50 percent 90 percent Enhanced CEmOC Benefit from improved Improved coverage As for 5a without quality of care and care; 70 percent and delivery care (adds interventions quality and coverage and content without the costs of BPS coverage without delivery care (BEmOC) without for high-risk babies) at the primary and BPS nutritional BPS secondary levels supplements without BPS Source: Authors. BPS balanced protein-energy supplementation. a. Defined in terms of the percentage of complicated cases at the primary level referred to and reaching the secondary level. b. Includes obstetric first aid for complicated cases, including abortion and postpartum complications. c. The secondary level will also provide some prenatal and delivery care for normal cases, as defined in the first two panels of table 26.5 for the base package at the primary level. Maternal and Perinatal Conditions | 519 Table 26.7 ICERs per Million Population, South Asia and Sub-Saharan Africa (U.S. dollars) Incremental cost per Incremental cost per Incremental cost per death averted life-year saved DALY averted Option Alternative compared with South Sub-Saharan South Sub-Saharan South Sub-Saharan number the base package Asia Africa Asia Africa Asia Africa 2 Increased primary-level coverage 6,129 3,337 217 119 148 92 3a Improved overall quality of care with 5,017 2,729 165 90 142 83 nutritional supplements 3b Improved overall quality of care 8,975 2,538 296 84 240 77 without nutritional supplements 4 Improved quality of CEmOC 10,532 5,089 372 195 255 151 5a Improved overall quality of care and 5,297 2,915 177 98 144 86 coverage with nutritional supplements 5b Improved overall quality of care and 7,944 2,865 269 96 203 84 coverage without nutritional supplements Source: Authors' calculations. Table 26.8 Costs and Effectiveness of Intervention Packages per Million Population, South Asia and Sub-Saharan Africa Percentage Number Number of Number of of DALYs Option Total costs of deaths life years DALYs averted that number Intervention package (US$) averted saved averted are maternal South Asia 1 Routine maternity care 408,976 79 2,240 3,273 50 2 Increased primary-level coverage 603,071 111 3,136 4,582 50 3a Improved overall quality of care with nutritional 829,505 163 4,793 6,225 26 supplements 3b Improved overall quality of care without nutritional 757,433 118 3,415 4,727 35 supplements 4 Improved quality of CEmOC 420,918 80 2,272 3,320 50 5a Improved overall quality of care and coverage with 1,287,354 245 7,201 9,354 26 nutritional supplements 5b Improved overall quality of care and coverage without 1,186,123 177 5,131 7,103 35 nutritional supplements Sub-Saharan Africa 1 Routine maternity care 602,646 192 5,406 6,969 47 2 Increased primary-level coverage 859,027 269 7,568 9,757 47 3a Improved overall quality of care with nutritional 1,164,833 398 11,652 13,753 24 supplements 3b Improved overall quality of care without nutritional 1,049,209 368 10,733 12,770 26 supplements 4 Improved quality of CEmOC 617,724 195 5,483 7,069 47 5a Improved overall quality of care and coverage with 1,785,971 597 17,508 20,664 24 nutritional supplements 5b Improved overall quality of care and coverage 1,633,956 552 16,127 19,188 26 without nutritional supplements Source: Authors' calculations. 520 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others Table 26.9 Sensitivity Analysis Results, South Asia and Sub-Saharan Africa (incremental cost per DALY averted, US$) Effectiveness Met need Inpatient cost assumption assumption assumption Best Option number Alternative compared with base package estimate High Low High Low High Low South Asia 2 Increased primary-level coverage 148 113 163 147 150 213 109 3a Improved overall quality of care with nutritional supplements 142 100 163 143 144 142 143 3b Improved overall quality of care without nutritional supplements 240 180 326 241 242 240 240 4 Improved quality of CEmOC 255 193 311 373 260 446 204 5a Improved overall quality of care and coverage with nutritional 144 104 164 144 149 152 136 supplements 5b Improved overall quality of care and coverage without nutritional 203 153 250 203 210 227 189 supplements Sub-Saharan Africa 2 Increased primary-level coverage 92 70 104 91 93 191 84 3a Improved overall quality of care with nutritional supplements 83 64 90 83 84 83 83 3b Improved overall quality of care without nutritional supplements 77 61 85 77 78 77 77 4 Improved quality of CEmOC 151 114 166 228 151 326 130 5a Improved overall quality of care and coverage with nutritional 86 66 94 86 89 123 82 supplements 5b Improved overall quality of care and coverage without nutritional 84 66 93 84 87 123 79 supplements Source: Authors' calculations. comparison is with or without increased coverage (options 5a cent) lead to larger numbers of women also benefiting from the and 3a, respectively). This difference reflects the high burden rest of the care package in terms of obstetric first aid and from low birthweight in South Asia and, thus, the gain from CEmOC. nutritional supplements. This issue is important for safe motherhood and newborn Comparing the content of the three most cost-effective health, because the role of prenatal care has been subject intervention packages (3a, 5a, and 2) suggests that much can be to intense debate about its benefits relative to resource use achieved through improvements at the primary care level. (De Brouwere and Van Lerberghe 2001; Maine and Rosenfield Improved quality in relation to managing complications--in 1999). Much of this discussion has focused on the lack of evi- other words, the provision of BEmOC--and increases in cov- dence on the direct contribution of prenatal care to reducing erage (a combination of options 3a and 2) at the primary level maternal mortality (McDonagh 1996; Rooney 1992), which, in are likely to have even lower ICERs than those shown in turn, is explained partly by the poor performance of at-risk table 26.7. This finding is consistent with the Commission on screening tools. However, differentiating the contribution to Macroeconomics and Health's emphasis on close-to-client the prevention of maternal deaths of the prenatal care compo- services (CMH 2002), and it is highlighted further in chap- nent alone is difficult. Ultimately, life-saving interventions ter 53. As noted earlier, given the importance of prompt depend on the functioning of the entire health system, includ- intervention in the event of obstetric complications, the effec- ing an effective referral network. tiveness of intervention packages that may reduce delays by Our model also made assumptions about women's willing- bringing services closer to communities is hardly surprising. ness and capacity to respond to referral to higher levels of care The benefits from option 2 were achieved essentially by in case of complications. This willingness and capacity depend increasing prenatal care coverage from 50 to 70 percent, on many factors and are undoubtedly also driven by commu- because our model assumes that those women taking advan- nities' perceptions of quality of care. As noted earlier, coverage tage of professional delivery are those who have also had pre- rates of prenatal care are already high in many Sub-Saharan natal contact. Prenatal care is, thus, a crucial entry point to the African countries, but significant socioeconomic differentials health system. Small changes in prenatal care coverage (20 per- are apparent within countries. Our model does not address this Maternal and Perinatal Conditions | 521 equity dimension but, given the recent work showing higher have identified even more cost-effective intervention packages, risks of maternal death among the poorest groups (Graham such as a combination of options 3a and 2. and others 2004), targeting disadvantaged women for improve- ments in uptake might be worth considering (Gwatkin and Deveshwar-Bahl 2002; De Brouwere and Van Lerberghe 2001). ECONOMIC BENEFITS OF INTERVENTION Whereas option 2, increased primary-level coverage, relates predominantly to the demand side of the health system A narrow definition of the economic benefits of safe mother- (Williams 1987), the most cost-effective packages (3a and 5a) hood interventions would focus primarily on the impact of focus on the supply side, particularly at the health center level. maternal mortality and morbidity on household investment The latter packages are particularly relevant to the baby, includ- and consumption. Investment in this context refers not so ing screening of the HIV status of the mother and treatment much to financial investment as to investment in improving with antiretrovirals at the time of delivery to reduce the risk of housing conditions, agricultural productivity, education, and mother-to-child transmission, as well as provision of anti- so on. The key elements to capture include the loss of produc- malarials. As a consequence, these options have a particularly tivity and the disruption of planned investment and consump- marked effect on the burden from perinatal conditions, tion. In addition to the loss of a woman's own productivity, accounting for two-thirds to three-fourths of the total DALYs consequent effects are likely on the productivity of other averted (table 26.8). Note that these cost-effective options household members--effects that may be particularly long include a doctor at the health center level to provide all six lived in the case of young children whose health and education BEmOC functions. In some situations, highly skilled midwives suffer because of their mother's death. The household will also will be able to act in this capacity, which would reduce costs be worse off because it will have diverted resources from pre- and further increase cost-effectiveness. ferred consumption and investment activities in response to The most comprehensive packages in our model provide for the health crisis. Thus, recognizing the dynamic consequences improved quality of care and coverage at both the primary and of maternal death and disability and selecting an appropriate the secondary levels (options 5a and 5b). Costing US$1.79 and time horizon for the analysis are important. US$1.63 per capita, respectively, in Sub-Saharan Africa (as The potential benefits to individual households arising from calculated from the total costs of these packages shown in investments in safe motherhood are relatively clear, although table 26.8, and divided by the base of 1 million people), these challenges in quantifying and valuing them remain. The bene- are also the most expensive packages. Not surprisingly, there- fits may, however, be more widely spread in that improvements fore, these two options avert much higher numbers of DALYs, in safe motherhood may reduce poverty, which in turn may with the package that includes nutritional supplementation stimulate economic development. Increased economic develop- averting nearly three times as many DALYs as the base package ment may then feed back into further improvements in mater- (table 26.8). In CEA, generally the most comprehensive pack- nal health, generating a virtuous cycle. The mechanisms where- ages--that is, those that result in the greatest gain in quality by changes in maternal health affect other parts of the economy and coverage and, thus, cost the most--are often not cost- may be identified by a close examination of the influence of effective, and yet our analysis found otherwise. This finding maternal health on productivity and educational attainment. may partly be explained by the linear assumptions about effec- A number of links may exist between safe motherhood and tiveness in the model and the assumption that the marginal the performance of the health care system; therefore, strategies cost of care is constant. Such a finding also stresses both the to improve safe motherhood may be a means of achieving importance of a well-functioning health system (rather than an wider health service improvements (Goodburn and Campbell excessive focus on one element) and the absence of any quick 2001). Jowett (2000, 213) notes that "to improve a facility's fix. Moreover, we did not model these more comprehensive capacity to respond to obstetric emergencies, it is necessary to options as perfect but unrealistic scenarios. We also still have the skills and supplies to deal with trauma, give blood allowed for 30 percent of pregnant women not attending pre- transfusions and anesthesia, and have a functional operating natal care, 50 percent of severe complications at a primary level theatre." Thus, initiatives in safe motherhood could be an entry not reaching CEmOC, and 10 percent of those reaching sec- point for wider health sector reform and improvement. ondary care not receiving the emergency treatment they need. Finally, a note of caution is warranted on the interpretation of the CEA results. First, our model has necessarily used a num- LESSONS FOR IMPLEMENTATION ber of assumptions for which data are extremely limited, and it remains fairly crude, having been subject to only a limited sen- The findings from the CEA indicate potential health gains and sitivity analysis. Second, many comparisons are possible from the reduced burden that may be achieved by implementing our model, but we have selected only six. Thus, we may not selected packages of interventions. Such implementation 522 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others assumes, first, that decision makers accept the evidence and are to hospitals and financial barriers to access on the part of the willing and able to act and, second, that an enabling health sys- poor. tem environment exists within which the requisite scale and The financing of prenatal and delivery care services at an quality of care can be effectively delivered. These factors are not adequate and sustainable level is a subject of much debate peculiar to safe motherhood, but they undoubtedly help and uncertainty, given the difficulty of distinguishing these explain the significant gap between evidence and action that elements from broader health expenditure categories (De many argue is one of the main obstacles to progress (Godlee Brouwere and Van Lerberghe 2001). Given the low level of and others 2004; Villar and others 2001). The gains from bridg- overall per capita expenditure on health in developing ing this gap would be significant: the MDGs for child survival countries--estimated at US$13 in 2002 for the poorest 49 and maternal health might become more than mere rhetoric countries (Bale and others 2003)--attaining our base interven- for poor regions if intervention packages of the scope and tion package (costing approximately US$0.41 per capita in nature described here were implemented. The most cost- South Asia and US$0.60 in Sub-Saharan Africa) does not effective of the packages averted nearly 50 percent more direct sound unrealistic at current resource levels (see table 26.8, and maternal deaths than the base package. This gain would be divide by base population of 1 million people). encouraging, but the prospects for achieving it by 2015 are The effects of health sector reforms, particularly decentral- weak (Johansson and Stewart 2002). ization of management and budget holding, appear to be mixed At the macro level, a supportive policy environment clearly in terms of increasing resource flows into maternity services, is crucial, as noted earlier. At the micro level, an enabling health with both apparent positive benefits, as in Bolivia (De Brouwere system implies a reduction in the disequilibrium between the and Van Lerberghe 2001), and negative effects through demand and supply sides (Williams 1987), with particular the exacerbation of inequities (Russell and Gilson 1997). attention to three interrelated issues: access, quality, and Effective management decisions on finance, access, and quality finance. The CEA reported in this chapter emphasizes the require information, an essential ingredient for stimulating potential benefits to mother and baby of improved access to action. To allocate scarce resources where they are likely to care, particularly the importance of entry to the health system achieve the greatest gain, countries need information to assess through primary-level services. The increases in coverage could the burden of ill health, evaluate the performance of current be achieved by a variety of mechanisms but clearly require both intervention strategies, identify the scope for improvement and demand- and supply-side interventions. implement changes, and close the loop by evaluating effects and On the supply side, this chapter has shown that improved cost-effectiveness (Lawn, McCarthy, and Ross 2001). quality of care at both the primary and the secondary levels Even though the challenges that the poorest countries face encompassing technical, infrastructural, and human resource today clearly differ in many respects from those that developed dimensions (Pittrof, Campbell, and Filippi 2002) is a particu- or transition countries experienced in the past, six historical larly cost-effective option. The widespread call for all women to lessons provide particularly relevant insights. First, examples deliver with skilled attendance immediately raises major ques- abound of supportive policy contexts and individual champions tions about quality of care and capacity, because much of the of progress in addressing maternal and newborn health, such developing world faces an acute shortage, as well as an unequal as those reported by De Brouwere and Van Lerberghe (2001). geographic distribution, of health professionals. Second, historical data on the uptake of prenatal care demon- Our CEA assumes that redistributing human resources strate that community-based providers and advocates played a within countries will accommodate the increased uptake of crucial role. Third, the role of various professionals and profes- care by women, although the most effective mechanisms for sional bodies has not always been positive, particularly as achieving this goal, such as incentives, use of nonphysicians, regards the "war" between advocates for home and institutional and increased private sector involvement, have not yet been deliveries (Koblinsky and Campbell 2003). Moreover, good his- established (De Brouwere and Van Lerberghe 2001). What is torical evidence indicates that excessive rates of forceps deliver- clear, however, is the importance of the interplay between sup- ies and other interventions were significant contributors to ply and demand, with the supply of quality care stimulating maternal mortality in countries such as the United Kingdom demand for care and vice versa. Quality care includes an effec- and the United States (Buekens 2001). Fourth, primary-level tive referral system (Murray and others 2001) to ensure the care depends on an effective referral system being in place to required match between the various levels of care different maintain the confidence of both women and providers (Loudon women and their babies need at different times (De Brouwere 1997). Fifth, to reduce the burden of maternal and perinatal and Van Lerberghe 2001). Such systems require not only conditions, the system of health care financing must facilitate financial resources to support transportation, communica- access for the poorest groups and guarantee service quality tions, and feedback mechanisms, but also structured fee and (De Brouwere and Van Lerberghe 2001). Finally, the role of pop- exemption strategies to reduce both inappropriate self-referral ulation-based information on births and maternal deaths was Maternal and Perinatal Conditions | 523 crucial in ensuring that actions were locally relevant (Sorenson change, such as a particular increase in the uptake of prena- and others 1998), in demonstrating progress, and thus in stim- tal care, may not be known. Thus, the ICERs may be too low, ulating further action. This crucial role is particularly apparent in that they do not fully capture the costs of the intervention. in the literature on several European countries in the past cen- · Estimating cost-effectiveness. More sophisticated economic tury (Graham 2002; De Brouwere and Van Lerberghe 2001). models need to be developed to facilitate the evaluation of a wider range of safe motherhood strategies, particularly as RESEARCH AND DEVELOPMENT NEEDS better primary evidence becomes available from other stud- ies and initiatives using a variety of outcome measures The priorities for research and development arising from this (Cairns, McNamee, and Hernandez 2003). Similarly, proba- chapter need to be put in the context of wider requirements for bilistic sensitivity analysis would be a valuable development safe motherhood and newborn health that have been well artic- that would permit fuller exploration of the uncertainties ulated elsewhere (see, for example, Bale and others 2003). The regarding the model's parameters. general heading under which the specific needs emerging from this chapter can be grouped is evidence-based decision making, which has five crucial requirements: CONCLUSIONS · recognizing the weakness of current approaches to allocat- In 2001, maternal and perinatal conditions represented the sin- ing scarce health care resources in poor countries gle largest contributor to the global burden of disease, at near- · making efforts to improve the scope and quality of data on ly 6 percent of total DALYs (Mathers and others 2004). the burden from maternal and perinatal conditions Reducing that burden is widely stated as a priority at both · carrying out robust evaluation of the costs and effectiveness national and international levels, but the track record of trans- of intervention strategies lating the rhetoric into action on a sufficiently large and equi- · using reliable evidence to inform the decision-making table scale to make a difference at the population level remains process disappointing. The literature abounds with examples of this · implementing prioritized strategies and robust, continuous disappointment (see, for example, Maine and Rosenfield 1999; assessment of their performance. Weil and Fernandez 1999). Many reasons account for the lim- ited progress, especially in the poorest regions of the world, and Within those major areas, specific topics relevant to the CEA researchers offer many interpretations of the bottlenecks. Lack undertaken here include the following: of evidence on the size of the burden and on the effectiveness of alternative intervention strategies figures prominently in · Ascertaining the burden of maternal and perinatal conditions. these interpretations. Greater clarity and consensus are needed on the scope of this The modeling in this chapter is, therefore, based on imper- important burden category and the implications of signifi- fect knowledge and needs to be supplemented with data from cant current exclusions, such as indirect maternal conditions primary evaluations. The findings do, however, provide some and stillbirths. Practical assessment tools are needed to tentative insights into programmatic options that may repre- enable meta-analysis and other modeling approaches to sys- sent the optimal use of resources in South Asia and Sub- tematically factor in data constraints. Huge gaps in knowl- Saharan Africa. In this context, three issues deserve emphasis. edge exist with regard to the levels and consequences of First, for intervention packages to achieve the degree of cost- maternal morbidity (Say, Pattinson, and Gulmezoglu 2004), effectiveness shown here, improvements are needed across the contribution of iatrogenic factors, the unpredictability of health systems, and both the supply and the demand sides need maternal complications, and the levels of mortality. Most to be addressed. Second, crucial entry points to this system can of those gaps require significant developments in relation to be achieved at the primary level, particularly through prenatal available measurement tools and in poor countries' capacity care. The effect of increasing the volume of women in contact to use them as part of routine health surveillance. These with these services is likely to manifest itself in an increased improvements not only are needed to inform future CEA but proportion of deliveries with skilled attendance and of deliveries also have wider implications for global health monitoring. in which women obtain access to emergency obstetric care. · Implementing change. In addition to evidence on the content Finally, the quality of these services is crucial, and even with of intervention strategies, assessments of how to implement only 50 percent uptake of care, benefits can still be achieved in changes are urgently needed. A limitation of our analysis is terms of overall DALYs averted and of reduced maternal and that, even though the model may be a reasonable represen- perinatal mortality. tation of the resource and health consequences of different Initiatives to improve the quality of care, particularly at a intervention packages, the way to achieve the required primary level, thus appear to be cost-effective options for the 524 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others poorest regions of the world. Overall those findings appear to perinatal complications. About 2 percent of mothers are lend support to a safe motherhood and newborn health assumed to require treatment for preterm delivery, and 1 per- strategy that is close to the client and boosts community confi- cent for premature rupture of membranes. dence in health systems. In practice, the proportion of women with serious complica- tions receiving comprehensive emergency obstetric care varies widely, from 3 percent in Cameroon to 75 percent in Sri Lanka ANNEX 26.A: CEA MODEL ASSUMPTIONS (Averting Maternal Death and Disability Working Group on Indicators 2003). The scenarios considered in this chapter We assumed that there are four primary-level health facilities assume that either 20 or 50 percent of women with serious com- (health centers) and one secondary-level care facility (district plications reach secondary care, and that 50, 70, 80, or 90 percent hospital) for every 500,000 people. We estimated the numbers of those women receive the elements of comprehensive emer- of pregnancies and births from the crude birth rate for each gency obstetric care that they need, depending on which inter- region. We assumed that pregnant mothers attending for rou- vention package is being considered. For the sensitivity analysis, tine prenatal care are equally distributed between the five facil- we used low values of 30, 50, 60, and 70 percent and high values ities and that each facility provides similar routine prenatal and of 70, 80, 90, and 95 percent. We assumed that ambulances are delivery care. Routine prenatal care is assumed to comprise available, so that when the proportion of mothers with severe four visits--except for mothers with complications, who make complications reaching secondary care is increased, the addi- six visits. Mothers with complications are referred to the dis- tional costs are only the additional driver time and the increased trict hospital after their first visit if they cannot be treated at the costs of running and maintaining the vehicle. health center. We assumed that complications such as anemia The prevalence and incidence of different maternal condi- and sexually transmitted diseases are treated without referral to tions are taken from the WHO mother-baby package (WHO secondary care, as are preeclampsia and incomplete abortion, if 1994). World Health Organization estimates of the burden of a doctor is present at the facility. The average number of bed different maternal and perinatal conditions (WHO 2004d) days is assumed to be three days for normal deliveries and six have been applied to a population of 1 million, with a particu- days for cesarean section and other complications. Table 26.A1 lar crude birth rate to generate an estimate of the potential shows the U.S. dollar costs per inpatient bed day used in the number of deaths that could be avoided, the years of life that main analysis and in the sensitivity analysis. could be saved, and the DALYs that could be averted. The We assumed the existence of excess capacity, so that an assumptions regarding the effectiveness of the interventions increase in prenatal care coverage from 50 to 70 percent would with respect to maternal and perinatal conditions were based not require an increase in the number or capacity of existing primarily on the WHO's mother-baby package and a review of health care facilities, and the increased costs would mostly be the literature; they are shown in table 26.A2. We assumed that increases in variable costs. For increased coverage of prenatal each intervention has the same effect on the number of deaths, care, we assumed a need for increased expenditure on educa- years of life saved, and DALYs. The effectiveness of interven- tion, information, and communication. Enhanced prenatal tions is assumed to be additive. care and comprehensive emergency obstetric care are assumed to require additional expenditures on training, assumed to be 10 percent of total personnel costs. We assumed that the addi- ACKNOWLEDGMENTS tional costs of basic emergency obstetric care compared with obstetric first aid are largely due to providing doctors at each We would like to thank the many individuals who helped pre- health center. We also assumed that 8 percent of mothers pare this chapter. In particular, we acknowledge the expert require cesarean section as a result of either maternal or input regarding perinatal conditions from Joy Lawn and Jelka Table 26.A1 Costs Per Inpatient Bed Day, South Asia and Sub-Saharan Africa (U.S. dollars) South Asia Sub-Saharan Africa Cost of inpatient bed day Primary level Secondary level Primary level Secondary level Best estimate 6.51 8.50 6.17 8.05 Low 2.64 3.45 1.92 2.51 High 14.52 18.94 41.79 54.52 Source: DCPP2: Guidelines for Authors. Maternal and Perinatal Conditions | 525 Table 26.A2 Assumed Effectiveness of Interventions analyses; and level C is assigned to case series, case studies, or expert (percentage of DALYs, deaths, and years of life lost averted) opinion. 4. This chapter does not deal with tertiary and specialist levels of care Best or with rehabilitative care or care for chronic conditions. 5. The six functions of BEmOC are (a) administering antibiotics intra- Condition estimate Low High venously or intramuscularly, (b) administering oxytocics intravenously or Maternal intramuscularly, (c) manually removing the placenta, (d) administering anticonvulsants intravenously or intramuscularly, (e) carrying out instru- Hemorrhage 85 80 90 mental delivery, and (f) removing retained products of conception. The two Sepsis 75 70 90 additional functions in CEmOC are blood transfusion and cesarean sec- Hypertensive disorders of pregnancy 76 71 95 tion. For a facility to be regarded as a BEmOC or CEmOC site, respectively, it must perform all six or all eight functions regularly and must be assessed (including eclampsia) every three to six months (UNFPA 2003). Obstructed labor 80 75 95 Unsafe abortion 75 70 90 Perinatal REFERENCES Low birthweight AbouZahr, C. 1998. "Maternal Mortality Overview." In Health Dimensions In context without nutritional supplementsa 8 3 14 of Sex and Reproduction, ed. C. J. Murray and A. D. Lopez, 111­64. In context with nutritional supplementsa 28 23 44 Geneva: World Health Organization. Birth asphyxia (including birth trauma) ------. 1999. "Disability-Adjusted Life Years and Reproductive Health: A Critical Analysis." Reproductive Health Matters 7 (14): 118­29. In context without enhanced delivery care packagea 40 35 60 ------. 2003. "Global Burden of Maternal Death and Disability." British In context with enhanced delivery care packagea 70 65 90 Medical Bulletin 67: 1­11. Infections, including tetanus 60 55 80 AbouZahr, C., and T. Wardlaw. 2001. "Maternal Mortality at the End of a Sepsis (newborn) 40 35 60 Decade: Signs of Progress?" Bulletin of the World Health Organization 79 (6): 561­68. HIV/AIDS 60 55 80 Averting Maternal Death and Disability Working Group on Indicators. Source: expert panels; WHO 1994, 2004d; Steketee and others 2001; Prendiville, Elbourne, and 2003. "Program Note: Using UN Process Indicators to Assess Needs in Chalmers 1998; Eclampsia Trial Collaborative Group 1995. Emergency Obstetric Services: Morocco, Nicaragua, and Sri Lanka." Note: Two extra interventions added to WHO mother-baby package: active management of the International Journal of Gynaecology and Obstetrics 80: 222­30. third stage of labor and magnesium sulfate for hypertensive disorders of pregnancy. Bailey, P., J. Szaszdi, and B. Schieber. 1995. "Analysis of the Vital Events a. See table 26.6. Reporting System of the Maternal and Neonatal Health Project: Quetzaltenango, Guatemala." MotherCare Technical Working Paper 3, John Snow, Arlington, VA. Zupan. Thanks are also given to our colleagues at the University of Aberdeen, particularly Joyce Boor, Julia Hussein, Bale, J., B. Stoll, A. Mack, and A. Lucas, eds. 2003. Improving Birth Outcomes: Meeting the Challenges in the Developing World. Washington, Emma Pitchforth, Nara Tagiyeva-Milne, and Karen Witten. We DC: National Academy of Sciences and Institute of Medicine. acknowledge and thank our colleagues Paul McNamee and Bell, J., S. L. Curtis, and S. Alayon. 2003. "Trends in Delivery Care in Six Rodolpho Hernandez from the Health Economics Research Countries." Department of Homeland Security Analytical Studies 7, Unit at the University of Aberdeen for their thorough review of Opinion Research Corporation and Macro International Research Partnership for Skilled Attendance for Everyone (SAFE), Calverton, the cost-effectiveness analysis. Thanks also to our expert panel MD. members: Deanna Ashley, Gary Darmstadt, Catherine Bell, J., J. Hussein, B. Jentsch, G. Scotland, C. Bullough, and W. J. Graham. Hauptfleisch, Jilly Ireland, Joy Lawn, Cecil Klufio, Elizabeth 2003."Improving Skilled Attendance at Delivery: A Preliminary Report Molyneux, Ashalata Shetty, Sribala Sripad, Vijay Kumar Tandle, of the SAFE Strategy Development Tool." Birth 30 (4): 227­34. Sumesh Thomas, and Jelka Zupan. Bergsjo, P., and J. Villar. 1997. "Scientific Basis for the Content of Routine Antenatal Care: II. Power to Eliminate or Alleviate Adverse Newborn Outcomes; Some Special Conditions and Examinations." Acta Obstetrica et Gynecologica Scandinavica 76 (1): 15­25. NOTES Buekens, P. 2001."Is Estimating Maternal Mortality Useful?" Bulletin of the World Health Organization 79 (3): 179. 1. Antecedent is here defined as a factor that changes the probability of Buttiens, H., B. Marchal, and V. De Brouwere. 2004. "Skilled Attendance at an adverse outcome or sequela, either positively (protecting) or, more usu- Childbirth: Let Us Go beyond the Rhetorics." Tropical Medicine and ally, negatively (aggravating). A risk factor may be a leading contributor to International Health 9 (6): 653­54. the global burden because of high prevalence in the population or because of a large increase in the probability of adverse outcomes (WHO 2002). Cairns, J., P. McNamee, and R. Hernandez. 2003. "Measurement and 2. Afghanistan, Angola, Bangladesh, China, the Democratic Republic Valuation of Economic Outcomes." Economic Outcomes Work of Congo, Ethiopia, India, Indonesia, Kenya, Nigeria, Pakistan, Tanzania, Program, Draft Concept Paper, IMMPACT (Initiative for Maternal and Uganda. Mortality Programme Assessment), Dugald Baird Centre, University of 3. We use a simple three-way distinction for levels of evidence. Level A Aberdeen, Scotland. refers to evidence from randomized clinical trials or systematic overviews Calder, A. A., and W. Dunlop, eds. 1992. High Risk Pregnancy. Oxford, of trials; level B relates to nonrandomized studies, often with multivariate U.K.: Butterworth-Heinemann. 526 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others Carroli, G., C. Rooney, and J. Villar. 2001. "How Effective Is Antenatal Care Gwatkin, D., and G. Deveshwar-Bahl. 2002. "Socioeconomic Inequalities in Preventing Maternal Mortality and Serious Morbidity? An Overview in Use of Safe Motherhood Services in Developing Countries." Paper of the Evidence." Paediatric and Perinatal Epidemiology 15 (Suppl. 1): presented at the Inter-Agency Safe Motherhood Meeting, London, 1­42. February 6. Chamberlain, G., ed. 1995. Turnbull's Obstetrics. 2nd ed. London: Hall, M., S. MacIntyre, and M. Porter. 1985. Antenatal Care Assessed. Churchill Livingstone. Aberdeen, Scotland: Aberdeen University Press. Cleland, J., and M. Ali. 2004. "Reproductive Consequences of Hoj, L., D. da Silva, K. Hedegaard, A. Sandstrom, and P. Aaby. 2003. Contraceptive Failure in 19 Developing Countries." Obstetrics and "Maternal Mortality: Only 42 Days?" British Journal of Obstetrics and Gynecology 104 (2): 314­20. Gynaecology 110 (11): 995­1000. CMH (Commission on Macroeconomics and Health). 2002. Improving the Hussain, R., F. F. Fikree, and H. W. Berendes. 2000. "The Role of Son Health Outcomes of the Poor. Report of Working Group 5 of the Preference in Reproductive Behaviour in Pakistan." Bulletin of the Commission on Macroeconomics and Health. Geneva: World Health World Health Organization 78 (3): 379­88. Organization. Hutubessy, R. C., R. M. Baltussen, T. Tan-Torres Edejer, and D. Evans. 2003. Creese, A., and D. Parker. 1994. Cost Analysis in Primary Care: A Training "Generalised Cost-Effectiveness Analysis: An Aid to Decision Making Manual for Programme Managers. Geneva: World Health in Health." In Making Choices in Health: WHO Guide to Cost- Organization. Effectiveness Analysis, ed. T. Tan-Torres Edejer, R. M. Baltussen, T. Dayaratna,V.,W.Winfrey,W. McGreevey, K. Hardee, K. Smith, E. Mumford, Adam, R. Hutubessy, A. Acharya, D. B. Evan, and C. J. L. Murray, and others. 2000. Reproductive Health Interventions: Which Ones Work 277­88. Geneva: World Health Organization. and What Do They Cost? Washington, DC: Policy Project. Jamison, D. T., J. S. Jamison, J. Lawn, S. Shahid-Salles, and J. Zupan. 2004. De Brouwere, V., and W. Van Lerberghe, eds. 2001. Safe Motherhood "Incorporating Deaths Near the Time of Birth into Estimates of the Strategies: A Review of the Evidence. Vol. 17 of Studies in Health Global Burden of Disease." Working Paper 26, Disease Control Services Organisations and Policy. Antwerp, Belgium: ITG Press. Priorities Project, Bethesda, MD. Donnay, F. 2000. "Maternal Survival in Developing Countries: What Has Johansson, C., and D. Stewart. 2002. "The Millennium Development Been Done; What Can Be Achieved in the Next Decade." International Goals: Commitments and Prospects." Working Paper 1, Human Journal of Gynecology and Obstetrics 70 (1): 89­97. Development Report Office Working Papers and Notes, United Eclampsia Trial Collaborative Group. 1995. "Which Anticonvulsant for Nations Development Programme, New York. Women with Eclampsia? Evidence from the Collaborative Eclampsia Jowett, M. 2000. "Safe Motherhood Interventions in Low-Income Trial." Lancet 345 (8963): 1455­63. Countries: An Economic Justification and Evidence of Cost Etard, J. F., B. Kodio, and S. Traore. 1999. "Assessment of Maternal Effectiveness." Health Policy 53 (3): 201­28. Mortality and Late Maternal Mortality among a Cohort of Pregnant Khanum, P. A., M. A. Quaiyum, A. Islam, and S. Ahmed. 2000. Women in Bamako, Mali." British Journal of Obstetrics and Gynaecology "Complication of Pregnancy and Childbirth: Knowledge and Practices 106 (1): 60­65. of Women in Rural Bangladesh." Working Paper 131, International Evans, T., and S. Stansfield. 2003. "Health Information in the New Centre for Diarrhoeal Disease Research, Dhaka. Millennium: A Gathering Storm?" Bulletin of the World Health Koblinsky, M. A., and O. Campbell, eds. 2003. Reducing Maternal Organization 81 (12): 856. Mortality: Learning from Bolivia, China, Egypt, Honduras, Indonesia, Foege, W. 2001. "Managing Newborn Health in the Global Community." Jamaica, and Zimbabwe. Washington, DC: World Bank. American Journal of Public Health 91 (10): 1563­64. Koenig, M. A., V. Fauveau, and B. Wojtyniak. 1991. "Mortality Reductions Freedman, L. P. 2001. "Using Human Rights in Maternal Mortality from Health Interventions: The Case of Immunization in Bangladesh." Programs: From Analysis to Strategy." International Journal of Population and Development Review 17 (1): 87­104. Gynecology and Obstetrics 75 (1): 51­60. Kurjak, A., and I. Bekavac. 2001. "Perinatal Problems in Developing Gay, J., K. Hardee, N. Judice, K. Agarwal, K. Flemming, A. Hairston, Countries: Lessons Learned and Future Challenges." Journal of and others. 2003. What Works: A Policy and Program Guide to the Perinatal Medicine 29 (3): 179­87. Evidence on Family Planning, Safe Motherhood, and STI/HIV/AIDS Kusiako, T., C. Ronsmans, and L. Van der Paul. 2000. "Perinatal Mortality Interventions. Safe Motherhood Module 1. Washington, DC: Policy Attributable to Complications of Childbirth in Matlab, Bangladesh." Project. Bulletin of the World Health Organization 78 (5): 621­27. Germain, A. 2000. "Population and Reproductive Health: Where Do We Lawn, J. E., B. J. McCarthy, and S. R. Ross. 2001. The Healthy Newborn: A Go Next?" American Journal of Public Health 90 (12): 1845­47. Reference Manual for Program Managers. Atlanta, GA: Centers for Godlee, F., N. Pakenham-Walsh, D. Ncayiyana, B. Cohen, and A. Packer. Disease Control and Prevention and CARE. http://www.cdc. 2004. "Can We Achieve Health Information for All by 2015?" Lancet gov/reproductivehealth/health_newborn.htm. 364 (9430): 295­300. Loudon, I. 1997. "Midwives and the Quality of Maternal Care." In Goodburn, E., and O. Campbell. 2001. "Reducing Maternal Mortality in Midwives, Society and Childbirth: Debates and Controversies in the the Developing World: Sector-Wide Approaches May Be the Key." Modern Period, ed. H. Marland and A. M. Rafferty, 180­200. London British Medical Journal 322 (7291): 917­20. and New York: Routledge. Graham, W. J. 1998. "Every Pregnancy Faces Risks." Planned Parenthood Maine, D., and A. Rosenfield. 1999. "The Safe Motherhood Initiative: Why Challenges 1: 13­14. Has It Stalled?" American Journal of Public Health 89 (4): 480­82. ------. 2002. "Now or Never: The Case for Measuring Maternal Marston, C., and J. Cleland. 2003. "Do Unintended Pregnancies Carried to Mortality." Lancet 359 (9307): 701­04. Term Lead to Adverse Outcomes for Mother and Child? An Assessment Graham, W. J., A. E. Fitzmaurice, J. S. Bell, and J. A. Cairns. 2004. "The in Five Developing Countries." Population Studies 57 (1): 77­94. Familial Technique for Linking Maternal Death and Poverty." Lancet Mathers, C. D., A. Lopez, C. Stein, D. Ma Fat, C. Rao, M. Inoue, and others. 363 (9402): 23­27. 2004. "Deaths and Disease Burden by Cause: Global Burden of Disease Graham, W. J., and J. Hussein. 2004. "The Right to Count." Lancet 363 Estimates for 2001 by World Bank Country Groups." Working Paper (9402): 67­68. 18, Disease Control Priorities Project, Bethesda, MD. Maternal and Perinatal Conditions | 527 McCarthy, J., and D. Maine. 1992. "A Framework for Analyzing the Save the Children. 2001. State of the World's Newborns: Saving Newborn Determinants of Maternal Mortality." Studies in Family Planning Lives. Washington, DC: Save the Children. 23 (1): 23­33. Say, L., R. Pattinson, and M. Gulmezoglu. 2004. "WHO Systematic Review McDonagh, M. 1996. "Is Antenatal Care Effective in Reducing Maternal of Maternal Morbidity and Mortality: The Prevalence of Severe Morbidity and Mortality?" Health Policy and Planning 11 (1): 1­15. Acute Maternal Morbidity (Near Miss)." Reproductive Health 1: 3. Milne, L., G. Scotland, N. Tagiyeva-Milne, and J. Hussein. 2004. "Safe http://www.reproductive-health-journal.com/content/1/1/3. Motherhood Program Evaluation: Theory and Practice." Journal of Shehu, D., A. T. Ikeh, and M. J. Kuna. 1997. "Mobilizing Transport for Midwifery and Women's Health 49 (4): 338­44. Obstetric Emergencies in Northwestern Nigeria." International Journal Munjanja, S. P., G. Lindmark, and L. Nystrom. 1996. "Randomised of Gynecology and Obstetrics 59 (2): 173­80. Controlled Trial of a Reduced-Visits Programme of Antenatal Care in Sheldon, T. A. 1996. "Problems of Using Modeling in the Economic Harare, Zimbabwe." Lancet (North American ed.) 348 (9024): 364­69. Evaluation of Health Care." Health Economics 5: 1­11. Murray, C. J. L., and A. D. Lopez, eds. 1998. Health Dimensions of Sex and Sorensen, G., K. Emmons, H. K. Hunt, and D. Johnston. 1998. Reproduction. In Vol. 3 of Global Burden of Disease and Injury. "Implications of the Results of Community Intervention Trials." Cambridge, MA: Harvard University Press. Annual Review of Public Health 19: 379­416. Murray, S. F., S. Davies, P. Kumwenda, and A. Yusuf. 2001. "Tools for Steketee, R. W., B. L. Nahlen, M. E. Parise, and C. Mendez. 2001. "The Monitoring the Effectiveness of District Maternity Referral Systems." Burden of Malaria in Pregnancy in Malaria-Endemic Area." American Health Policy and Planning 16 (4): 353­61. Journal of Tropical Medicine and Hygiene 64 (1.2 Suppl.): 28­35. Mwageni, E., A. Ankomah, and R. Powell. 2001. "Sex Preference Stoll, B. J., and A. R. Measham. 2001. "Children Can't Wait: Improving the and Contraceptive Behaviour among Men in Mbeya Region, Future for the World's Poorest Infants." Journal of Pediatrics 139 (5): Tanzania." Journal of Family Planning and Reproductive Health Care 729­33. 27 (2): 85­89. Thonneau, P., N. Goyaux, S. Goufodji, and J. Sundby. 2002. "Abortion and Pattinson, B. 2002. Saving Mothers: Second Report on Confidential Enquiries Maternal Mortality in Africa." New England Journal of Medicine 347 into Maternal Deaths in South Africa, 1999­2001. Pretoria: Department (24): 1984­85. of Health. Tomkins, A. 2001. "Nutrition and Maternal Morbidity and Mortality." Pittrof, R., O. Campbell, and V. Filippi. 2002."What Is Quality in Maternity British Journal of Nutrition 85 (2): 93­99. Care? An International Perspective." Acta Obstetrica et Gynecologica Scandinavica 81 (4): 277­83. UN (United Nations). 2002. UN Population Division Population Estimates and Projections, 2000 rev. Portela, A., and C. Santarelli. 2003. "Empowerment of Women, Men, Families, and Communities: True Partners for Improving Maternal UNFPA (United Nations Population Fund). 2003. Maternal Mortality and Newborn Health." British Medical Bulletin 67 (1): 59­72. Update 2002: A Focus on Emergency Obstetric Care. New York: UNFPA. Prata, N., F. Greig, J. Walsh, and M. Potts. 2004. "Setting Priorities for UNICEF (United Nations Children's Fund). 1999. "World Summit for Safe Motherhood Interventions in Resource Scarce Settings." Paper Children Goals: End of Decade Indicators for Monitoring Progress." submitted to the Population Association of America, 2004 Annual Executive Directive EXD/1999-03, New York. Meeting, School of Public Health, University of California­Berkeley, Villar, J., H. Ba'aqeel, G. Piaggio, P. Lumbiganon, J. M. Belizan, U. Farnot, April 1­3. and others. 2001. "WHO Antenatal Care Randomized Trial for the Prendiville, W. J., D. Elbourne, and I. Chalmers. 1988. "The Effects of Evaluation of a New Model of Routine Antenatal Care." Lancet (North Routine Oxytocics Administration in the Management of the Third American ed.) 357 (9268): 1551­64. Stage of Labour: An Overview of the Evidence from Controlled Trials." Villar, J., and P. Bergsjo. 1997. "Scientific Basis for the Content of Routine British Journal of Obstetrics and Gynaecology 95 (1): 3­16. Antenatal Care I." Acta Obstetrica et Gynecologica Scandinavica 76 (1): Raju, S., and A. Leonard. 2000. Men as Supportive Partners in Reproductive 1­14. Health: Moving from Rhetoric to Reality. New Delhi: Population Villar, J., M. Merialdi, A. M. Gulmezoglu, E. Abalos, G. Carroli, R. Kulier, Council South and East Asia Regional Office. and M. de Oni. 2002. "Nutritional Interventions during Pregnancy for Robertson, N. R. C. 1993. A Manual of Normal Neonatal Care. London: the Prevention or Treatment of Maternal Morbidity and Preterm Arnold. Delivery: An Overview of Randomized Controlled Trials." Journal of Rogo, K., L. Bohmer, and C. Ombaka. 1999. "Developing Community- Nutrition 133 (5): 1606S­25S. Based Strategies to Decrease Maternal Morbidity and Mortality Walsh, J., A. Fiefer, A. Measham, and P. Gertler. 1993. "Maternal and Due to Unsafe Abortion: Pre-intervention Report." East African Perinatal Health." In Disease Control Priorities in Developing Countries, Medical Journal 76 (11 Suppl.): S1­71. ed. D. Jamison, H. Mosley, A. Measham, and J. Bobadilla, 363­90. Rooney, C. 1992. Antenatal Care and Maternal Health: How Effective Is It? Oxford, U.K.: Oxford Publications. A Review of the Evidence. WHO/MSM/92.4. Geneva: World Health Weil, O., and H. Fernandez. 1999. "Is Safe Motherhood an Orphan Organization. Initiative?" Lancet 354 (9182): 940­43. Rouse, D. J. 2003. "Potential Cost-Effectiveness of Nutrition Interventions West, K. P. Jr., J. Katz, S. K. Khatry, S. C. LeClerq, E. K. Pradhan, S. R. to Prevent Adverse Pregnancy Outcomes in the Developing World." Shrestha, and others. 1999."Double Blind, Cluster Randomised Trial of Journal of Nutrition 133 (Suppl.): 1640S­44S. Low Dose Supplementation with Vitamin A or Beta Carotene on Rush, D. 2000. "Nutrition and Maternal Mortality in the Developing Mortality Related to Pregnancy in Nepal." British Medical Journal World." American Journal of Clinical Nutrition 72 (Suppl.): 212­40. 318 (7183): 570­75. Russell, S., and L. Gilson. 1997. "User Fee Policies to Promote Health WHO (World Health Organization). 1986. "Maternal Mortality: Helping Service Access for the Poor: A Wolf in Sheep's Clothing." International Women off the Road to Death." World Health Organization Chronicle Journal of Health Services 27 (2): 359­79. 40: 175­183. Sadana, R. 2001. "Quantifying Reproductive Health and Illness." Ph.D. ------. 1992a. International Classification of Diseases and Related Health dissertation, Harvard School of Public Health, Boston, MA. Problems, 10th rev. Geneva: WHO. 528 | Disease Control Priorities in Developing Countries | Wendy J. Graham, John Cairns, Sohinee Bhattacharya, and others ------. 1992b. The Prevention and Management of Unsafe Abortion. ------. 2004d. Global Burden of Disease for the Year 2001 by World Bank WHO/MSM/92.5. Report of a Technical Working Group. Geneva: Region, for Use in Disease Control Priorities in Developing Countries, WHO. 2nd ed. Bethesda, MD: National Institutes of Health. http://www. ------. 1994. Mother-Baby Package. WHO/FHE/MSM/94.11. Geneva: fic.nih.gov/dcpp/gbd.html. WHO. WHO-CHOICE (Choosing Interventions That Are Cost Effective). 2004. ------. 1999. Mother-Baby Package Costing Spreadsheet. WHO/FCH/ "Choosing Interventions That Are Cost Effective." Geneva. http:// RHR/99.17. Geneva: WHO. www3.who.int/whosis/menu.cfm?path=whosis,cea&language= english. ------. 2002. Reducing Risks, Promoting Healthy Life: The World Health Report 2002. Geneva: WHO. Williams, A. 1987. "Health Economics: The Cheerful Face of the Dismal Science?" In Health and Economics, ed. A. Williams. 1­11. Basingstoke ------. 2003. Making Pregnancy Safer: Global Action for Skilled Attendants and London: Macmillan. for Pregnant Women. WHO/RHR/02.17. Geneva: WHO. ------. 2004a. "Global Monitoring and Evaluation, Proportion of Births Winikoff, B., and M. Sullivan. 1987. "Assessing the Role of Family Plan- Attended by Skilled Health Personnel: Global, Regional, and ning in Reducing Maternal Mortality." Studies in Family Planning Subregional Estimates." Geneva, WHO, Department of Reproductive 18 (3): 128­43. Health and Research. http://www.who.int/reproductive-health/ Yasmin, S., D. Osrin, E. Paul, and A. Costello. 2001. "Neonatal Mortality of global_monitoring/data_regions.html. Low Birth-Weight Infants in Bangladesh." Bulletin of the World Health ------. 2004b. Maternal Mortality in 2000: Estimates Developed by WHO, Organization 79 (7): 608­14. UNICEF, UNFPA. Geneva: WHO. ------. 2004c. Unsafe Abortion: Global and Regional Estimates of the Incidence of Unsafe Abortion and Associated Mortality in 2000. 4th ed. Geneva: WHO. Maternal and Perinatal Conditions | 529 Chapter 27 Newborn Survival Joy E. Lawn, Jelka Zupan, Geneviève Begkoyian, and Rudolf Knippenberg The second half of the 20th century witnessed impressive This chapter provides an overview of neonatal deaths, pre- reductions in the risk of under-five child mortality, which was senting the epidemiology as a basis for program priorities and halved between 1960 and 1990. The greatest reduction was for summarizing the evidence for interventions within a health children after the first month of life, with relatively little systems framework, providing cost and impact estimates for decrease in the neonatal period (the first 28 days of life). packages that are feasible for universal scale-up. The focus of Neonatal deaths, estimated at approximately 4 million annually, the chapter is restricted to interventions during the neonatal now account for 38 percent of the world's deaths of children period. The priority interventions identified here are largely under five. The fourth Millennium Development Goal (MDG) well known, yet global coverage is extremely low. The chapter aspires to a global target, by 2015, of reducing the under-five concludes with a discussion of implementation in country mortality rate by two-thirds, which implies approximately 30 programs with examples of scaling up, highlighting gaps in deaths per 1,000 live births for children under five. Currently, knowledge. there are an estimated 30 deaths per 1,000 live births in the neonatal period alone. Thus, the fourth MDG cannot be achieved without substantial reduction in neonatal deaths NEONATAL DEATHS (Lawn, Cousens, and Zupan 2005). Addressing neonatal mortality requires links within the One reason neonatal survival has received little attention rela- continuum of care from maternal health through pregnancy, tive to the huge number of deaths is the invisibility of those childbirth, and early neonatal care, and into child health pro- deaths. Most deaths during the neonatal period occur at home grams. Such services can be delivered through a combination and are often unregistered even in transition countries of care at the family-community level, outreach, and clinical (Lumbiganon and others 1990). Social invisibility is linked to care (figure 27.1). Yet neither child survival nor safe mother- an expectation of high mortality; many traditional societies do hood programs have adequately addressed newborn deaths. not name newborns for up to six weeks. Data presented here are The first week of life, when 75 percent of neonatal and 50 per- derived from full-coverage vital registration for 72 countries, cent of maternal deaths occur, is associated with low health which cover less than 4 percent of all neonatal deaths; demo- care coverage, particularly in poor communities. Investing in graphic and health surveys, which cover 75 percent of global maternal, neonatal, and child health (MNCH) services will neonatal deaths; and statistical modeling, for the 20 percent of improve the survival of newborns and reduce stillbirths and neonatal deaths in countries without data (WHO forthcom- maternal and child deaths. The first weeks of life are also a ing). Population-based data on neonatal morbidity or long- time of behavioral transition, representing an opportunity to term disability in low- and middle-income countries (LMICs) promote healthy behaviors that have benefit beyond the are scarce. The World Health Organization (WHO) has neonatal period. estimated that three conditions (birth asphyxia, prematurity, 531 Skilled obstetric and immediate newborn care Emergency newborn care for illness, including resuscitation especially sepsis management care Emergency obstetric care to manage complications Extra care of very low birth weight strategy Clinical such as obstructed labor and hemorrhage babies including kangaroo mother care delivery Family Prenatal care package Postnatal care to support healthy planning practices service Early detection and referral of Outreach services by complications care Prepregnancy Counseling and preparation for Clean delivery Healthy home care including breastfeeding of health and newborn care and breastfeeding Simple early promotion, hygienic cord and skin care, thermal nutrition Emergency preparedness newborn care care, promotion of demand for quality care Family- Interventions such as extra care of small Packages community babies and case management of pneumonia depending on local situation Prepregnancy Pregnancy Postnatal and Infant and Birth health care newborn care wks child care Neonatal period Early fetal Late fetal period period Early Late terms Perinatal I (22 weeks gestation to 7 days after birth) Perinatal II Epidemiological (28 weeks gestation to 7 days after birth) Note: International Classification of Diseases version 10 recommends perinatal I for national data collection. The World Health Organization recommends perinatal II for international comparisons of data. Figure 27.1 The Continuum of Care for Mothers, Newborns, and Children, Showing Epidemiological Terms around the Time of Birth and Packages of Care Relevant to Newborn Health, According to Service Delivery Level and "other perinatal causes"), collectively termed perinatal (60 per 1,000 live births), Mozambique (55 per 1,000 live causes, contribute to 6.3 percent of global disability-adjusted births), and Ethiopia (52 per 1,000 live births). life years (WHO 2003a). Although these causes represent only part of the neonatal burden, the WHO estimate is more than When Do Newborns Die? triple that of HIV, yet receives remarkably little attention. Each year 3 million newborns die during their first seven days of life, accounting for 75 percent of all neonatal deaths. At least Where Do Newborns Die? 1 million babies die during their first 24 hours of life (Lawn, Only 1 percent of neonatal deaths occur in high-income coun- Cousens, and Zupan 2005). If mortality rates during the first tries. These countries have average neonatal mortality rate five years of life are adjusted to rates per week, the risk in the (NMR) of 4 per 1,000 live births, whereas in LMICs the aver- first week of life is massively higher than during any other time age NMR is 33 per 1,000 live births, with a range of 2 to 70 of life: 24 per 1,000 in the first week compared with 3 per week (table 27.1). The highest number of neonatal deaths occur in for the rest of the first month and only 0.12 per week after the South Asia because of the large populations of this region. The first year of life. Yet the first week is the very period in the con- six countries with the highest numbers of neonatal deaths in tinuum of care when services are most likely to be lacking, par- 2000 include the populous nations of India (1.09 million ticularly in poor communities, where most deaths occur. neonatal deaths annually), China (416,000), Pakistan (298,000), Nigeria (247,000), Bangladesh (153,000), and Time Trends in Neonatal Mortality Ethiopia (147,000). Of the 20 countries with the highest As shown in figure 27.2, the disparity in NMRs between LMICs NMRs, 80 percent are in Sub-Saharan Africa. The highest rates and high-income countries is increasing over time, especially occur in countries there that have experienced recent civil during the early neonatal period, which saw an almost 60 per- unrest, such as Liberia (65 per 1,000 live births), Sierra Leone cent reduction in high-income countries between 1983 and 532 | Disease Control Priorities in Developing Countries | Joy E. Lawn, Jelka Zupan, Geneviève Begkoyian, and others Table 27.1 NMRs and Neonatal Deaths by Region for 2000, and Variation in NMR by Income Quintile and by Region Percentage of Median NMR by income NMR per 1,000 deaths during the quintile by region live births Number (percentage) neonatal period (range across Poorest Richest of neonatal deaths among children Region countries) quintile quintile (thousands) under five World 30 -- -- 3,998 (100) 38 High-income countries 4 (1­11) -- -- 42 (1) 63 Low- and middle-income countries 33 (2­70) -- -- 3,956 (99) 38 Region Africa 44 ( 9­70) 48 34 1,128 (28) 24 Americas 12 ( 4­34) 35 18 195 (5) 48 Eastern Mediterranean 40 ( 4­63) 38 28 603 (15) 40 Europe 11 ( 2­38) -- -- 116 (3) 49 South Asia 38 (11­43) 50 28 1,442 (36) 50 Western Pacific 19 ( 1­40) 28 17 514 (13) 56 Source: Authors' calculations, based on NMRs and under-five mortality, WHO and UNICEF estimates; NMR by income quintile based on analysis of demographic and health survey data for 50 countries, 1995­2002. -- not available. NMR per 1,000 live births Historical data also show more rapid reductions in 100 postneonatal mortality, steady reductions in late neonatal Postneonatal period mortality, and slower reductions in early neonatal deaths. In Late neonatal period 80 Early neonatal period England, the NMR fell from more than 30 per 1,000 live births in 1940 to 10 per 1,000 in 1975. This fall occurred before inten- 44 60 sive care, which was introduced only when the NMR had fallen 35 below 15 per 1,000. The greatest reduction of NMR coincided 29 40 with the introduction of free prenatal care, high coverage of 16 12 8 skilled childbirth care, and the availability of antibiotics. 20 Although the number of postneonatal and late neonatal deaths 28 5 24 25 1 1 is amenable to public health interventions (such as immuniza- 2 1 1 9 6 4 0 tion and improved hygiene and nutrition), larger reduction of 1980 1995 2000 1980 1995 2000 early neonatal deaths and of maternal deaths requires a system LMICs High-income countries that provides effective clinical care--particularly during child- Source: Authors' calculations, based on UNICEF, various years; WHO 1998c; and WHO birth, which is more challenging. forthcoming. Figure 27.2 Trends in Early and Late Neonatal and Postneonatal Direct Causes of Neonatal Death Mortality, by Country Income Levels Fewer than 3 percent of the world's neonatal deaths occur in countries that have vital registration data that are reliable 2000, compared with about a 15 percent reduction in LMICs. enough to use in cause-of-death analysis. Population-based There has been no measurable decline in the regional average information in high-mortality settings often depends on NMR for sub-Saharan Africa. However some regions have verbal autopsy tools of variable quality. The Child Health made significant progress in reducing NMRs, particularly Latin Epidemiology Reference Group undertook an extensive exercise America and the Western Pacific. Some low-income countries to derive global estimates for program-relevant causes of neona- such as Bangladesh, Indonesia, and Sri Lanka have achieved tal death, including preterm birth, asphyxia, sepsis/pneumonia, NMR reductions of 40 to 50 percent. In South Asia and Sub- neonatal tetanus, diarrhea, and other causes, with the latter Saharan Africa, the decline in late neonatal deaths was influ- including specific but less prevalent causes such as jaundice. For enced by the halving of neonatal tetanus deaths that occurred low-mortality countries, vital registration data from 45 coun- during the 1990s as a result of increased tetanus toxoid protec- tries with full vital registration coverage (cumulative sample size tion and clean delivery practices. By 2000, two-thirds of LMICs of 96,797) were included. For high-mortality countries, studies had eliminated neonatal tetanus and an additional 22 countries were identified through extensive systematic searches, and a were nearing this goal. meta-analysis was performed after applying inclusion criteria Newborn Survival | 533 and using standard case definitions (56 studies, cumulative risk of neonatal death than factors identified during pregnancy, sample size of 13,685). Models were developed to simultane- which are in turn associated with greater increases in risk ously estimate the seven selected causes of death by country than prepregnancy factors (Lawn, Cousens, and Zupan 2005). (Lawn, Cousens, and Wilczynska forthcoming). Obstructed labor and malpresentation present the highest risk Three causes of death--infections (pneumonia, diarrhea, and require skilled intervention. The mother's death substan- and tetanus) (36 percent), preterm birth (28 percent), and tially increases the risk of death for her child. Greenwood and asphyxia (23 percent)--account for the majority of neonatal others (1987) report that of mothers who died in labor (N 8), deaths. Causes of death vary between the early and late neona- all the babies died within one year. tal periods, with deaths caused by preterm birth, asphyxia, and Delays in access to care for severely ill young infants are com- congenital defects occurring predominantly during the first mon. Peterson and others (2004), in a study in Uganda, find that week of life and infection being the major cause of neonatal almost 80 percent of the caregivers of severely ill young infants deaths thereafter. Neonatal tetanus, a totally preventable condi- did not comply with recommended referrals to a health facility. tion, still accounts for more than a quarter of a million deaths, The reason given in 90 percent of the cases was lack of money, even after the second global elimination deadline has passed. underscoring the need for pro-poor financing mechanisms and Most neonatal tetanus deaths occur in 20 countries in South promotion of community demand for care. This recalls the Asia and Sub-Saharan Africa, all of which have very high "three delays" model for maternal deaths, which outlines delays NMRs. Variation in causes of neonatal death is seen between in recognition of illness and in access to care and provision of and within countries, closely associated with the NMR level. care once at a health facility (Thaddeus and Maine 1994). Where the NMR is high (more than 45 per 1,000 live births), Poverty is the root cause of many maternal and neonatal more than half of neonatal deaths are due to infections; where deaths, either because it increases the prevalence of risk factors the NMR is low, prematurity and congenital abnormalities are such as maternal infections or because it reduces access to care. the major causes of death (Lawn, Cousens, and Zupan 2005). An analysis of 50 demographic and health surveys between Hence, information regarding the local epidemiology is impor- 1995 and 2002 reveals that, within regions, the poorest quin- tant in prioritizing interventions. tiles have an NMR that is, on average, 20 to 50 percent higher than that for the highest income quintile (table 27.1). Indirect Causes of Neonatal Death Deliberate programmatic focus is required to ensure that care reaches poor families. An estimated 20 million low birthweight (LBW) infants (that is, weighing less than 2,500 grams), are born each year-- 25 percent of them in South Asia (Blanc and Wardlow 2005). Applying Lessons from Epidemiology to Programs Although globally only 16 percent of newborns have LBW, 60 to There are almost 4 million neonatal deaths annually. Given that 80 percent of neonatal deaths occur in LBW infants (Lawn, the proportion of child deaths that occur during the neonatal Cousens, and Wilczynska forthcoming). LBW is due to short period (currently 38 percent) will increase over time, the MDG gestation (preterm birth), intrauterine growth restriction for child survival cannot be met without a significant reduction (IUGR), or both. Globally, almost one-third of neonatal deaths in the NMR. Most neonatal deaths are in Sub-Saharan Africa are directly attributable to preterm birth. In contrast, an analy- and South Asia and are due to preventable causes. Historical sis of vital registration data for 45 countries and of five data demonstrate that the NMR can be reduced to 15 per 1,000 population-based studies suggests that a maximum of 1 to without intensive care. 2 percent of neonatal deaths are directly attributable to IUGR Priority should be given to two main gaps in the provision in full-term neonates (Lawn, Cousens, and Wilczynska forth- of care. The first is the continuum of care by time. The period coming). Prematurity and full-term IUGR are also indirect through pregnancy and childbirth into infancy contains a gap causes or risk factors for neonatal deaths, particularly deaths at childbirth and during the first week of life, when most resulting from infection. The relative risk among preterm neonatal deaths--and also most maternal deaths--occur. infants is much higher than for full-term IUGR infants (Yasmin Addressing this gap will involve strengthening safe mother- and others 2001). Complex technology is not necessary to avoid hood and child survival services and institutionalizing links at most deaths in moderately preterm newborns. Extra attention the subnational, national, and global levels. The second gap to warmth and feeding and to prevention or early treatment of is between levels of care (figure 27.1)--particularly with the infections is crucial (Lawn, McCarthy, and Ross 2001). family-community level, since poor and rural communities Maternal health and health care are important determinants account for the majority of neonatal and maternal deaths. of neonatal survival. Neonatal outcomes are affected by female Approaches are needed to better link homes and health care, health throughout the life cycle, from child, through adoles- supplying care closer to such communities, increasing demand cence, and into pregnancy (Pojda and Kelley 2000). In general, for skilled care, and empowering communities, including poor intrapartum risk factors are associated with greater increases in communities, to make healthful decisions. 534 | Disease Control Priorities in Developing Countries | Joy E. Lawn, Jelka Zupan, Geneviève Begkoyian, and others INTERVENTIONS and afterward (WHO 2003c). Cleanliness (for example, cord care and hand washing), warmth provision, and exclusive We undertook literature searches and categorized interventions breastfeeding reduce neonatal illnesses, especially infection. by time period (during pregnancy, and intrapartum and post- Implementation of this package will depend on the setting, the natal or neonatal periods) (table 27.2). We focus on interven- coverage of facility delivery, and the availability of community tions delivered during the neonatal period that are likely to workers or other channels but is feasible even in poorly reduce neonatal deaths, as opposed to those delivered during developed health systems (Knippenberg and others 2005). The the neonatal period that yield later benefits (for example, role and value of the mother are central. prevention of mother to child transmission of HIV). Although Although much has been written to describe traditional rigorous evaluation of evidence is vital, evidence is not available newborn care practices, few studies assess behavior change. An for some well-established interventions. In the case of neonatal exception is the study by Meegan and others (2001) of the resuscitation, for example, a randomized controlled trial is Masai in Kenya, where behavioral messages about cord care impossible for ethical reasons, yet the intervention is a corner- practices were associated with the virtual elimination of neona- stone of neonatal care in high-income settings. Some impor- tal tetanus--with no increase in tetanus toxoid immunizations. tant practices, such as cleanliness, have undergone little rigor- The Warmi Project in rural Bolivia demonstrated that raising ous evaluation but are obviously beneficial (Bhutta and others community awareness of maternal, fetal, and neonatal health 2005). On the basis of level of evidence and feasibility of imple- issues through women's community groups increased family- mentation, we grouped interventions into three categories: planning coverage, attendance at prenatal and postnatal services, and the presence of trained traditional birth assistants · Universally applicable interventions are selected on the basis at childbirth, resulting in a 62 percent reduction of perinatal of mortality impact, cost, and feasibility. Some of these mortality (O'Rourke, Howard-Grabman, and Seoane 1998). A interventions are feasible only after skilled care is available. cluster randomized trial in rural Nepal, where 90 percent of Other interventions are feasible immediately, even in the women deliver at home, also used female facilitators working absence of skilled care. A particular example is improved with women's groups. Comparing the 12 intervention villages family care practices. Interventions may apply to different with their paired villages showed a 30 percent reduction in the newborns as follows: NMR (mainly late NMRs) mediated through increased health -- essential newborn care for all babies at all levels seeking and improved home behaviors (such as doubling the -- extra newborn care for babies with specific risk factors, rates of practices such as hand washing and use of clean deliv- such as LBW ery kits) and strengthening of the health system (Manandhar -- emergency newborn care for babies who are ill, particu- and others 2004). larly those with infections. A family-community package promoting good home care · Additional interventions should apply where neonatal mor- of the newborn--particularly cleanliness, warmth, and exclu- tality is lower and capacity is greater. These interventions are sive breastfeeding--would have an expected reduction in the more complex, requiring more skilled staff members and NMR of 10 to 40 percent, varying with the baseline NMR and additional commodities, and therefore cost more for less the potential for accessing care. The effect might be greater if reduction in mortality. Universal scaling up cannot be rec- the package successfully addressed harmful local practices. The ommended at present, but these interventions become effect of early care seeking for illness will depend on the capac- important for further reduction of mortality and disabilities ity of the primary and referral health care levels to manage after universal care packages are in place. neonatal illness. Thus, community-level interventions with no · Situational interventions are necessary because of locally supply-side strengthening will have only a limited effect. Many prevalent risk factors, such as HIV or malaria. questions remain about how best to work with families and communities, given widely differing cultures and behaviors The packages of newborn care selected for universal scale-up and the varying capacities of existing community health work- are summarized in table 27.3 and discussed in the following sub- ers (Darmstadt and others 2005), and about the wider applica- sections, starting with family-community interventions and fol- tion of demand subsidies. lowed by essential, extra, and emergency newborn care packages. Essential Newborn Care at the Time of Birth Family-Community Care of the Newborn WHO (2003d) defines essential newborn care as the care of the Family care of the newborn is important for all newborns. It newborn at birth, including cleaning, drying, and warming the includes promoting positive behaviors such as breastfeeding infant; initiating exclusive breastfeeding early; and caring for and demand for health care throughout the neonatal period the cord. Essential care of the newborn is necessary for all Newborn Survival | 535 Table 27.2 Interventions to Reduce Fetal and Neonatal Mortality by Timing of Intervention and by Scalability of Intervention Additional interventions (where the health care system has additional capacity and the NMR Interventions for universal coverage (priority is lower; for example, transition Situational interventions (where Period interventions for high-mortality settings) countries) specific conditions are prevalent) Prepregnancy Family planning [B]: Rubella immunization either of girls HIV prevalent: · delay age of first pregnancy to after 18 only or of all population if regular · primary prevention strategies [B] coverage can be maintained at more · space births by two to three years · voluntary counseling and testing than 80 percent of the population [A] · provide opportunity for women to reduce births and option of antiretroviral Periconceptual or preconceptual to their desired number and to avoid pregnancy therapy [A] provision of folate [A] after age 45 High prevalence of recessive Information counseling and support for Prevention, identification, and management of conditions (such as sickle cell disease) sexually transmitted diseases [A] · smoking [RF A] or high rates of consanguineous marriages: offer genetics counseling Micronutrient deficiency prevention strategies · alcohol and drug abuse [RF A] [RF A] · iodination of salt [B] · women experiencing violence [RF A] During pregnancy Four-visit prenatal care package, including Identification and treatment of HIV prevalent: · Essential for all · two tetanus immunizations [A] bacteriuria [A] · primary prevention strategies [B] pregnancies · iron and folate supplements [B] Information counseling and support for · voluntary counseling and testing · syphilis screening and treatment [A] · smoking cessation [RF A] and option of antiretroviral therapy [A] · identification and referral of multiple · alcohol and drug abuse [RF A] pregnancy, abnormal lie, preeclampsia [B] · healthy diet and avoidance of · birth planning and emergency preparedness [C] unhelpful dietary taboos [C] · prenatal counseling and preparation for · women experiencing violence [RF A] breastfeeding [C] · Extra care for Extra prenatal care (more frequent visits, more External cephalic version for breech Malaria endemic: those at risk of skilled caregiver) if presentation at 36 weeks [A] · intermittent presumptive treatment complications · multiple pregnancy or abnormal lie (breech or Fetal growth monitoring [A] monthly after 20 weeks [A] transverse) [RF A] · insecticide-treated bednets [B based · pregnancy-induced hypertension or on effect on LBW, not on NMR] preeclampsia [RF A] Hookworm infestation prevalent: · diabetes [RF A] · presumptive treatment with · severe anemia [RF A] mebendazole [B] · previous fetal or neonatal death [RF A] · Emergency for Management of emergencies, including In utero transfer of high-risk Iodine deficiency prevalent: those with · preeclampsia or eclampsia [A] pregnancies [B] · iodine supplementation [B] complications (first · bleeding in pregnancy [A*] Famine: referral level and above) · uterine infection [RF A] · targeted food supplementation [B] Group B streptococcus prevalent: · screening and treatment [A] Birth Skilled care in labor, including Supportive companion in labor [A] Mother HIV positive: · Essential · monitoring progress of labor (partograph), · antiretroviral therapy [A] maternal and fetal well-being [A] · infection control [A*] Newborn resuscitation if required [A*] 536 | Disease Control Priorities in Developing Countries | Joy E. Lawn, Jelka Zupan, Geneviève Begkoyian, and others Table 27.2 Continued Additional interventions (where the health care system has additional capacity and the NMR Interventions for universal coverage (priority is lower; for example, transition Situational interventions (where Period interventions for high-mortality settings) countries) specific conditions are prevalent) · Extra care Extra care if Tocolytics in preterm labor and transfer Maternity waiting home if limited · preterm ( 37 weeks) or prolonged ( 18 hours) to higher-level care if available [A] access to emergency obstetric care, rupture of membranes or evidence of chorioam- If preterm labor, then give prenatal high-risk condition identified, and nionitis; give antibiotics to woman [A] steroid injection to mother [A] culturally acceptable [B] · failure to progress in labor including instrumen- tal vaginal delivery (vacuum) if required [RF A] Newborn resuscitation if required [A*] · Emergency Emergency obstetric care for acute intrapartum emergencies: [A*] · obstructed labor and fetal distress · bleeding, infections, or eclampsia Neonatal resuscitation if required [A*] Postnatal and Essential newborn care for all newborns, Trained breastfeeding counselors Hepatitis B prevalent: newborn including undertaking home visits [A] · give hepatitis B immunization · Essential · early and exclusive breastfeeding [B] Vitamin K (cost-effective as early [A] · warmth provision and avoidance of bathing prophylaxis for all babies in transition Mother HIV positive: during first 24 hours [C] countries) [B] · provide counseling and support for · infection control, including cord care and Routine newborn screening programs feeding choices [C] hygiene [B] for sickle cell disease, glucose 6 phosphate dehydrogenase · postpartum vitamin A provided to mother [B] deficiency [B] · eye antimicrobial provided to prevent ophthalmia [A] · information and counseling for home care and emergency preparedness [C] · Extra care Extra care for small babies (preterm or term Provide special or intensive care for Mother with tuberculosis: IUGR) and multiple births, severe congenital preterm babies [A] · keep baby with mother and give abnormalities: izoniazid prophylaxis · extra attention to warmth, feeding support, and Mother with syphilis: early identification and management of complications [B] · treat the baby even if asymptomatic [A*] · kangaroo mother care [A: morbidity not mortality data] · vitamin K injection [B] · Emergency Emergency care providing specific and supportive Provide special care for sick and small care according to evidence-based guidelines for babies using skilled nurses and a the following: higher nurse-to-patient ratio [B] · severe infections [A] · neonatal encephalopathy (following acute intrapartum insult) · severe jaundice or bleeding [A*] · neonatal tetanus Source: Authors, based on extensive literature review. References detailed on http://www.fic.nih.gov/dcpp/. Note: A rigorous meta-analysis or at least one good randomized controlled trial exists, RF A evidence regarding risk is strong, B well-conducted clinical studies exist but no randomized controlled trial done, C some descriptive evidence and expert committee consensus exists, A* unethical to test rigorously and widely practiced as standard (for example, blood transfusion, neonatal resuscita- tion). Bold text signifies priority packages or interventions considered in detail in this chapter. Newborn Survival | 537 Table 27.3 Packages for Universal Scale-up of Newborn Care Estimated current Reduction coverage (percent) Number of in all- target popula- Sub- cause Intervention tion per year South Saharan NMR package Contents (millions) Implementation strategy Asia Africa (percent) Comments on evidence Family- Healthy home care practices All newborn Women's groups and 36 28 10­40 Mortality reduction based community (exclusive breastfeeding, infants: community health workers on studies in high NMR care of the warmth protection, clean World 130 doing postnatal visits, with settings with weak health newborn at cord care, care seeking for links to the formal health care systems. Extra care of LBW South Asia and home after emergencies); if birth out- system, including support for infants and community Sub-Saharan birth side a facility, then clean referral. If appropriate, extra management of acute Africa 63 delivery kit. care of moderately small respiratory infections not babies at home and included in range shown. community-based manage- ment of acute respiratory infections. Essential Immediate drying, warmth, All newborn Skilled attendant, or if no 11 14 20­30 Based on conservative newborn care early breastfeeding, hygiene infants: skilled attendant available, combining of single at the time of maintenance, and infection World 130 some simple postnatal interventions (for example, birth prevention practices are feasible at breastfeeding) in the South Asia and home with other cadres package. Sub-Saharan of workers. Africa 63 Neonatal Resuscitation after birth if Newborns not Skilled attendant. 3 3 10­25 Limited studies, mainly from resuscitation required breathing at lower NMR settings with birth: high percentage of asphyxia World 6.5 deaths, so range from studies was reduced. South Asia and Sub-Saharan Africa 3.2 Extra care of Extra support for warmth LBW neonates: Facility-based care for 10 10 20­40 Most studies are nonran- small new- (kangaroo mother care), World 20.0 severely preterm babies. domized controlled trials at borns feeding, and illness Community-based care is the community level in South Asia and identification and effective for moderately settings with extremely high Sub-Saharan management preterm babies. LBW rates. Effect depends Africa 10.7 on baseline NMR and LBW rates. Emergency Management of ill infants, Neonates with Facility-based care with 20 20 20­50 Meta-analysis of effect on care of ill especially those with illnesses: antibiotics and supportive the NMR of oral antibiotic newborns neonatal infections World 13.0 care. Community-based management of acute management with oral respiratory infections in the South Asia and antibiotics for acute community in high-mortality Sub-Saharan respiratory infections. settings. Africa 6.3 Neonatal Neonatal packages as All newborn Supply of care throughout 5 5 -- No study data identified. packages above, in addition to family infants: pregnancy, childbirth, and Marginal budgeting for plus MCH planning, prenatal care, and World 130 postnatal period with bottlenecks tool suggests package comprehensive obstetric increased demand and 58 percent in South Asia South Asia and care packages improved referral systems. and 71 percent in Sub-Saharan Sub-Saharan Africa. Africa 63 Source: Local data or Darmstadt and others 2005; Knippenberg and others 2005; Lawn, Cousens, and Zupan 2005. Note: The range of reduction of all-cause NMRs given for each package is independent of the others; hence, the total is greater than 100 percent. 538 | Disease Control Priorities in Developing Countries | Joy E. Lawn, Jelka Zupan, Geneviève Begkoyian, and others infants and is ideally provided by a skilled attendant, but in the Nevertheless, only 11 percent of babies in South Asia and absence of skilled care, many of the tasks can be carried out at 14 percent in Sub-Saharan Africa are exclusively breastfed to home by alternative cadres of workers. WHO's essential care three months. The Bellagio group estimated a 15 percent package includes resuscitation, which we consider separately reduction in the NMR through 99 percent coverage of exclusive because the skill level required is more complex. breastfeeding and an 11 percent impact reduction through Clean care of the umbilical cord (clean blade and tie) is clean delivery (Jones and others 2003). Conservatively, an es- important in reducing the incidence of neonatal tetanus and sential newborn care package may result in a 10 to 25 percent umbilical sepsis, but evidence for topical treatment of the cord reduction in the NMR, but field trials of a combined package remains unclear (Zupan and Garner 2000). Hand washing is are still required. No economic assessments were identified. important at all levels of care. Hypothermia is an important and preventable contributor to morbidity and mortality, espe- cially in preterm babies. The so-called warm chain involves Newborn Resuscitation ensuring that childbirth takes place in a warmed room, drying Approximately 5 to 10 percent of newborns do not breathe the newborn, encouraging skin-to-skin contact between the spontaneously and require stimulation. About half of those newborn and the mother, and avoiding bathing for at least have difficulty initiating breathing, requiring resuscitation 12 hours (Lawn, McCarthy, and Ross 2001). In hospitals in (WHO 1998a). The major reasons for failure to breathe include LMICs, many newborns are hypothermic, and staff knowledge preterm birth and acute intrapartum events resulting in hyp- and practices could be improved (Dragovich and others 1997). oxic brain injury. Basic resuscitation using a self-inflating bag The effects of exclusive breastfeeding have been intensively and air is effective for the majority of these newborns, although studied, and the positive effect on infant mortality is unequiv- some may be too premature or have already experienced severe ocal, although studies often do not specify the effect on neona- hypoxic brain injury and die despite resuscitation. tal mortality and morbidity. The WHO collaborative trial Monitoring labor and providing effective obstetric care can found the risk of mortality in nonbreastfed neonates to be reduce the need for resuscitation (Dujardin, Sene, and Ndiaye 2.5 to 7.0 times greater than for breastfed neonates (WHO 1992), but resuscitation may be required even with good Collaborative Group 2000). The practice of keeping well babies obstetric care. Therefore, every skilled attendant should be close to their mothers and allowing feeding on demand competent in newborn resuscitation (box 27.1). increases breastfeeding rates, reducing both hypothermia and For most babies who do not breathe at birth, ventilation with nosocomial infections (WHO 1998b). Unfortunately, as exem- a self-inflating bag and mask is lifesaving, and the time to first plified by the low proportion of hospitals that are certified as breath differs little between use of a self-inflating bag and mask baby friendly, this practice is poorly implemented. Supportive and use of endotracheal intubation. Evidence is growing that policy, such as the International Code of Marketing of most newborns can be successfully resuscitated without the use Breastmilk Substitutes, is also important at the national level. of oxygen (Saugstad 2001), although a small proportion of The effect of essential newborn care has not been formally infants require such advanced resuscitation techniques as endo- tested as a package, although exclusive breastfeeding, cleanli- tracheal intubation, oxygen, chest compression, or drugs. Such ness, infection control measures, and hypothermia avoidance advanced resuscitation is appropriate only in institutions that all individually reduce neonatal mortality and morbidity. provide ventilation. In the 1980s, the high cost of a self-inflating Box 27.1 Institutionalizing a Neonatal Resuscitation Program in a Chinese Province A hospital-based study from China reports baseline sur- neonatal death and the second leading cause of infant veillance of 1,722 newborns followed by a two-year death nationally. They also recognized that child survival prospective assessment of 4,751 newborns, while institut- goals could not be met unless asphyxia was addressed. ing standardized resuscitation guidelines. Previous tradi- They developed and implemented an evidence-based tional resuscitation involved infusing central stimulants neonatal resuscitation program, training staff in using the plus vitamin C and 50 percent glucose; wiping the baby new guidelines. The early NMR fell significantly--by with alcohol; and pressing the philtrum. Health profes- 66 percent, to 3.4 per 1,000. sionals recognized that asphyxia was the leading cause of Source: Zhu and others 1997. Newborn Survival | 539 bag and mask led to the development of a prototype mouth- infants, and all LBW babies were targeted for increased home to-mask device operated by blowing expired air. A study by visits. Special sleeping bags were provided for warmth, and Massawe and others (1996) in two teaching hospitals, one in support was given for breastfeeding and early treatment of pos- Tanzania and the other in India, found that resuscitators using sible infections. The NMR fell by 87 percent in the moderately this device could maintain a maximum of only 20 breaths per preterm group (35 to 37 weeks) (Bang and others 1999). minute, one-third of the recommended rate. Low-cost (less So-called kangaroo mother care involves continuous skin- than US$5) versions of the bag and mask are now available, and to-skin contact between mother and baby to provide thermal it is the recommended device for resuscitation. stability and promote exclusive breastfeeding for clinically Although small-scale studies show that nonprofessional stable preterm infants. The published evidence relates to facili- cadres can learn the technique of resuscitation (Bang and oth- ty-based care with or without kangaroo mother care after ers 1999; Kumar 1995), a significant effect on mortality has not discharge. Mortality impact data for kangaroo mother care are been demonstrated, and the feasibility of maintaining compe- lacking, but a review by Conde-Agudelo, Diaz-Rossello, and tency and the cost-effectiveness of training such cadres have yet Belizan (2000) that included three randomized controlled tri- to be ascertained. If traditional birth assistants attend, say, als found that serious morbidity was reduced by about 60 per- 20 deliveries a year, they would encounter a baby requiring cent at the six-month follow-up visit. Although cost has not resuscitation an average of only once a year, so the effect would been formally evaluated, it must be considerably less than for be lower, and the cost per life saved higher, compared with a incubator care. The lack of assessments of kangaroo mother facility-based midwife who does 200 or more deliveries a year. care at community level is a research gap. Thus, more research is required before home resuscitation by A few studies in health facilities in LMICs have reported traditional birth assistants can become a widespread policy. In increased survival of LBW infants with improved care. One the meantime, it should be ensured that where skilled atten- from Papua New Guinea demonstrated a 56 percent reduction dants exist, they have the skills and equipment to perform in the NMR with the introduction of standards for care and of neonatal resuscitation. basic technology (Duke, Willie, and Mgone 2000). Data from Because a randomized controlled trial would be considered Ghana showed a 28 percent reduction in mortality for LBW unethical, the studies identified apply a before-and-after com- infants with support for breastfeeding, attention to warmth, parison. No cost assessments were identified. Achieving wider and early management of infections and jaundice using stan- coverage of resuscitation is a challenge, especially for the dard protocols (Lawn, McCarthy, and Ross 2001). 47 percent of the world's babies born at home. The reported effect for extra care of LBW babies in the com- munity varies between 20 and 40 percent, excluding Bang and Extra Care for Small Babies others' (1999) study because additional interventions were Because 60 to 80 percent of neonatal deaths occur in LBW involved. Given the high LBW prevalence in these studies, the babies, targeting this group for additional preventive and early effect may be less in other settings with a lower LBW preva- curative care is a logical approach to mortality reduction. lence. Data from facilities that do not offer intensive care Addressing deaths among severely preterm infants (fewer than suggest a similar or slightly larger effect. Cost-effectiveness 32 weeks of gestation) is more complex, but most preterm assessments were not identified. infants are moderately preterm (33.0 to 36.9 weeks). Excess mortality from acquired infections and other complications Emergency Care for Ill Newborns can largely be prevented or managed without intensive care. For many of the world's 4 million neonatal deaths, the immedi- A number of community-based studies have undertaken ate cause is a neonatal illness presenting as an emergency either simplified identification of small babies and provided extra care soon after birth (such as complications of preterm birth and at home, especially feeding (including the use of a dropper or asphyxia) or later (because of neonatal tetanus or community- cup feeding if required), warmth promotion, and cord cleanli- acquired infections). Other important but less prevalent ness. The reported NMR reductions range from 25 percent conditions include jaundice and hemorrhagic disease of the (Pratinidhi and others 1986) to 42 percent (Daga and others newborn. Long-term disability follows many neonatal condi- 1988). Datta (1985) applied a comprehensive approach, includ- tions, but it is poorly documented. Many serious neonatal prob- ing weighing all babies and providing extra home support to lems present with similar signs: inability to feed, breathing dif- LBW infants through feeding counseling and early recognition ficulty, and temperature instability. All those conditions have of and referral for illness, alongside strengthening of local high fatality rates, particularly neonatal tetanus (Institute of health systems. Compared with a control area, the NMR was Medicine 2003) and neonatal encephalopathy (Ellis and others reduced by more than 30 percent, with the greatest reduction 1999), and preventive interventions may be the most realistic among the group of 1,500- to 2,500-gram babies. In Bang and option in those conditions. Early phototherapy for jaundice others' (1999) study, 90 percent of neonatal deaths were in LBW reduces both mortality and chronic disability subsequent to 540 | Disease Control Priorities in Developing Countries | Joy E. Lawn, Jelka Zupan, Geneviève Begkoyian, and others kernicteris and is feasible in facilities (WHO 2003b). We focus neonatal estimates by country (Lawn, Cousens, and on the clinical neonatal management of infection, which is the Wilczynska forthcoming). most prevalent neonatal illness and the most feasible to scale up. · Baseline coverage estimates for the neonatal packages pre- A meta-analysis of community-based trials of case manage- sented in table 27.3 are taken from local data, if available ment of pneumonia in Africa and Asia yields a summary (for example, exclusive breastfeeding prevalence), or drawn estimate for NMR reduction of 27 percent (Sazawal and from coverage estimates in the Lancet newborn series Black 2003). The antibiotic regime used was mainly oral co- (Darmstadt and others 2005; Knippenberg and others 2005; trimoxazole, although two studies included injectable peni- Lawn, Cousens, and Zupan 2005). cillins. Bang and others' (1999) study in rural India reports a · Impact estimates for neonatal mortality are from the litera- 62 percent reduction in the NMR with a home-based package for ture, as presented in this chapter. The range uses the 95 per- neonatal sepsis that included injectable gentamicin, although cent confidence interval, rounded to the nearest 5 percent this reduction may be related to a number of simultaneously where available (table 27.3). If the data were from an introduced interventions in addition to the gentamicin. efficacy trial or a before-and-after trial, the range in the lit- The effect of emergency care on neonatal sepsis can be erature was reduced to reflect the expected effectiveness, assumed to be similar to the range in Sazawal and Black's (2003) based on expert opinion. Cause-specific mortality was used meta-analysis: 20 to 60 percent. Published cost data were not to allow combinations of effects across packages, and the identified apart from the Bang and others (1999) study, which assumptions applied were aligned with those used in the indicated a cost of US$5.30 per neonate treated. This cost esti- Lancet neonatal series (Darmstadt and others 2005)-- mate includes the time of community health workers and the although the packages here differ, because this chapter is cost of equipment and drugs, but not associated supervision or restricted to the neonatal period. The assumptions for system costs. cause-specific impact are detailed at http://www.fic.nih. gov/dcpp. The effect for outcomes other than neonatal ones MARGINAL IMPACT AND COST OF SCALING UP was based on data in the marginal budgeting for bottlenecks tool, primarily from the Lancet Bellagio series (Jones and UNIVERSAL NEONATAL PACKAGES others 2003) and Cochrane reviews. Effects are combined in Because newborn health depends on services in the continuum a residual manner; for example, deaths averted by preventive of care for mother, newborn, and child, a vertical program strategies are removed from the pool before curative would be duplicative, expensive, and inappropriate (Tinker and approaches are applied, and hence the total effect is less than others 2005). Hence costing and impact estimates will be based the sum of the effects. Years of life lost were calculated using on marginal additions of neonatal-specific packages to existing local average life expectancy discounted at 3 percent per maternal and child health (MCH) services (table 27.4). This year. This measure equates to the fatal outcome component scenario reflects the reality in many South Asian and Sub- of disability-adjusted life years, as described in chapter 15. Saharan African contexts, where MCH services exist but do not · Specific costs of adding the intervention packages are calcu- yet include newborn interventions. We will cost packages, lated on the basis of the cadre of worker, additional person- because packages are more cost-effective than single interven- nel time, in-service training, supervision, performance tions, and the emphasis is on the packages described for uni- incentives, travel and subsistence costs for referral care, versal scale-up (table 27.3). The benefits take into account only drugs, and equipment. Demand promotion and community neonatal deaths averted, whereas many of the interventions will mobilization are included. The costs of time, training, and also reduce maternal deaths, stillbirths, and childhood morbid- incentives are based on national salary levels, using real ity and disability--and therefore the benefits underestimate country data or World Bank databases. The costs of com- gains for both the fourth and the fifth MDGs. modities are based on the UNICEF supply system Costing and impact simulations are provided using the (http://www.supply.unicef.dk/Catalogue/). The cost of "marginal budgeting for bottlenecks" tool, a prioritization tool strengthening health systems, including improving manage- developed by the United Nations Children's Fund (UNICEF), ment and logistics, constructing new facilities, and deploy- the World Bank, and WHO. The inputs for the analysis pre- ing and training new cadres of workers, is included in the sented here are as follows: comprehensive MCH package. · Baseline epidemiology uses NMRs from the latest demo- Table 27.4 presents the estimated NMR effects and per graphic and health surveys by country or state and recent capita costs, in selected Indian states and Sub-Saharan African local relevant demographic data, such as crude birth rates. countries, of strengthening health systems to increase coverage · Cause-specific neonatal mortality estimates by country are with existing MCH packages (without neonatal care after from the Child Health Epidemiology Reference Group's birth). It then presents the additional specific costs of including Newborn Survival | 541 542 | Disease Control Priorities in Developing Table 27.4 Estimated Marginal Effect and Cost of Adding Neonatal Packages to Existing MCH Packages for Three Scenarios in Selected Indian States and Sub-Saharan African Countries Countries Additional cost per neonatal NMR reduction 2004­15 YLL averted for 20 percent (range of lower and upper efficacy, percent) Cost per capita (US$) increase in coverage for lower | and upper efficacy ranges (US$) Joy Indiaa Sub-Saharan Africab Indiaa Sub-Saharan Africab E. Package Scenario 1 Scenario 2 Scenario 1 Scenario 2 Scenario 1 Scenario 2 Scenario 1 Scenario 2 Indiaa Sub-Saharan Africab Lawn, MCH package (no 12­12 27­27 11­11 24­24 2.00 4.80 2.40 5.10 480 506 Jelka neonatal care after birth)c Marginal impact or cost Zupan, of adding neonatal packages Geneviève to the MCH package Family-community packaged 3­8 5­15 2­6 6­14 0.30 0.60 0.23 0.37 100­257 100­270 Clinical packagese 0­9 0­22 1­9 3­22 0.04 0.10 0.11 0.23 11­265 25­360 Begkoyian, Total impact or cost with 13­26 27­58 11­23 24­46 2.40 5.50 2.80 5.70 244­516 282­583 combined MCH and (average 380) (average 432) neonatal-specific packages and others Source: Estimates by authors, using the "marginal budgeting for bottlenecks" tool as detailed in the text. YLL years of life lost. YLL includes neonatal fatal outcomes only and is based on local life expectancy discounted at 3 percent per year. Scenario 1: increasing coverage by 20 percent. Scenario 2: meeting fourth MDG, necessitating about 45­60 percent NMR reduction, depending on the percentage of under-five mortality that is neonatal. Note: No specific neonatal outreach package is shown, because this is in prenatal care as part of the MCH package or home postnatal visits in the family-community package a. Five states in India are represented: Gujarat, Madhya Pradesh, Orissa, Rajasthan, and West Bengal. b. Five countries in Sub-Saharan Africa are represented: Benin, Ethiopia, Madagascar, Mali, and Rwanda. c. The MCH package consists of family planning, prenatal and obstetric care, and child health services (comprehensive integrated management of infant and childhood illness including prevention and community activities) and includes system strengthening costs. d. Includes interventions listed in table 27.3 under family-community package plus extra care of moderately small babies at home and community-based management of acute respiratory infections. e. Includes clinical care packages listed in table 27.3 (essential newborn care, neonatal resuscitation, extra care of small newborns, and emergency care of ill newborns). neonatal packages at the family-community level and in clini- the package is delivered may reduce the cost of the package, cal services and, finally, the combined costs for comprehensive but it also necessitates extra supervision and attention to links MNCH. Results are shown for two coverage scenarios: with the formal health system. Box 27.2 describes the projected effect and cost of various packages in Ethiopia for a 12-year · Scenario 1: increasing coverage of the interventions by program to improve maternal and child survival targeted at 20 percent from the baseline achieving the fourth MDG by 2015. Outreach services such as · Scenario 2: increasing coverage to the level required to meet prenatal care alone have an effect of about 10 percent on the fourth MDG, necessitating about a 45 to 60 percent NMRs, but when they are combined with a family package reduction in NMR, depending on the baseline percentage of using community health promoters, an additional 30 percent under-five mortality that is neonatal. reduction in the NMR is projected in Ethiopia. Outreach and family care options are more feasible initially. Table 27.4 shows that the addition of neonatal packages will Yet if commitment toward moving to strengthen the clinical reduce neonatal deaths at an average cost of about US$0.50 per care system is lacking, the potential reduction in NMRs capita per year for up to a 15 percent reduction in NMR at the over time from those options is limited, and the cost per death family-community level and about US$0.20 per capita for a 22 averted is higher. Although the estimated cost (averaged over percent NMR reduction at the clinical care level. Although the 12 years, with gradually increasing amounts) is low, the input is cost per capita is low for clinical care, the cost per case treated is higher than the current government and donor health expendi- higher, and the lag time to scale up is longer. The family- ture of the countries examined. Thus, spending in India would community neonatal package in India is estimated to cost have to be doubled, and in some African countries probably US$100 to US$257 per year of life saved (table 27.4), which cor- tripled. Considerable new funding is required at the national responds to about US$2,800 to US$7,800 per death averted.That and international levels, as well as more efficient allocation and is similar to the results of US$3,442 per neonatal death averted absorption of existing funds (Martines and others 2005). or US$111 per life year saved in a community participatory package in Nepal (US$4,397 and US$142, respectively, with health system strengthening) (Manandhar and others 2004). IMPLEMENTATION The comprehensive MNCH package (the MCH package plus integrated neonatal packages) is more expensive than the Effective interventions exist and are low cost, especially when neonatal packages alone: US$2.40 to US$2.80 per capita and added to existing programs, but current coverage is low, espe- per year for a 20 percent increase in coverage, and US$5.50 to cially for the poor, who have the highest mortality risk. US$5.70 to achieve the mortality reduction necessary to meet Approximately 53 percent of women worldwide deliver with a the fourth MDG (including the health system strengthening skilled attendant: fewer than 30 percent in the poorest coun- and demand-side approaches required). However, the effect of tries and more than 98 percent in the richest countries. In the MNCH packages on the NMR is more than double that of Sub-Saharan Africa, average coverage with skilled care has the neonatal packages alone--for example, a reduction of up increased at only 0.2 percent per year in the past decade; to 58 percent in NMRs in Africa, compared with up to 22 per- without faster progress, coverage of skilled attendance will still cent using interventions in the neonatal period only. This find- be less than 50 percent in 2015. Analysis in 50 low-income ing emphasizes the advantages of a comprehensive approach countries showed that the richest 20 percent of women were, across the continuum of care. Hence, the average cost per year on average, almost five times as likely to use a skilled attendant of life saved is still low at US$380 (India) and US$432 (Sub- as the poorest 20 percent (Knippenberg and others 2005). Saharan Africa) for a 20 percent increase in coverage, includ- Hence, coverage is low, progress is slow, and inequity is high. ing costs of system strengthening. If the coverage of the MCH Each country or decision-making unit starts with a different plus neonatal packages were to reach 90 percent, those pack- epidemiology and varying coverage and capacity in its health ages would avert up to 71 percent of neonatal deaths in the system. No single recipe for strengthening newborn care in African countries and up to 76 percent in the Indian states. health systems is available. Scaling up MNCH care will involve In settings where the current coverage of skilled care is low, systematic steps to assess local situations and opportunities, opportunities exist to start with family care and extra care of improve care within current constraints, and overcome supply LBW babies while building toward more challenging clinical and demand constraints--especially for the poor. No country packages. Some clinical care packages--such as simple extra or program can achieve multiple new interventions at once, care of the small baby or the provision of oral antibiotics for and scaling up human resources takes time. Therefore, phasing pneumonia later in the neonatal period--can be adapted for approaches is essential not only to allow faster approaches to delivery through community health systems. Varying the reach the poor soon, but also to allow consistent strengthening cadres of worker involved or the level of health system at which of the health system (Knippenberg and others 2005). Newborn Survival | 543 Box 27.2 Steps to Increase Coverage of Key MNCH Interventions in Ethiopia Ethiopia is one of the poorest countries in the world, with school graduate per 50 families will be trained to promote gross national income of US$100 (in 2000), less than healthy family behaviors. half the average for Sub-Saharan Africa. Neonatal deaths Estimates based on the marginal budgeting for bottle- of some 135,000 a year account for 29 percent of child necks tool suggest that, during the first eight years, progres- deaths. According to a 2000 demographics and health sur- sive scaling up of the health extension and health promoters vey, coverage of care is extremely low: only 6 percent of packages, together with some upgrading of clinical services, women deliver with a skilled attendant present and only will cost an additional US$4 per person per year. That effort 8 percent receive postnatal care within 48 hours of deliv- could result in a 30 percent reduction in the NMR, attribut- ery. The poor and those in rural areas have even lower cov- able mostly to improved behaviors, such as clean delivery erage. Health professionals are in short supply. At the same and exclusive breastfeeding, and to increased demand for time, obstetric services may be unused even when accessi- care. Increased coverage with family planning and tetanus ble because of issues of affordability and acceptability toxoid vaccination through the health extension package (most health workers are male). accounts for about 10 percent of the NMR reduction. By the In 2004, the government and major stakeholders held a end of the 12-year period, an additional 30 percent reduc- national partnership conference to develop a national plan tion in NMR is expected from strengthening clinical servic- for scaling up child survival interventions. The govern- es.A comprehensive package of family-based, outreach, and ment decided on a health extension package that would clinical services is projected to reduce the NMR by nearly deploy two female health extension workers to each kebele 50 percent, associated with a 25 percent reduction in the (commune of 5,000 inhabitants). Those workers are maternal mortality ratio--as compared with a less than mainly responsible for MNCH interventions, such as 5 percent reduction in the maternal mortality ratio with immunization, micronutrient supplementation, and fam- family and outreach care alone. The incremental annual ily planning, but they also have other public health and cost of almost US$10 per person is more than three times some clinical responsibilities. In addition, one primary current public spending on health of US$2.70. Source: Knippenberg and others 2005. Step 1: Assess the Situation and Advocate for Action midwives. National champions can be effective in promoting for Newborn Health progress. Global partnerships may also play a role in facilitating Careful examination of local data is required (Lawn, McCarthy, broad national plans and promoting donor convergence in and Ross 2001). Newborn health should be included in general implementation (Tinker and others 2005). health sector and public sector planning--for instance, The government of Nepal recently held a series of stake- for education and transportation. When governments set holder meetings and developed a plan for a national newborn mortality reduction targets for children under five, they should health strategy. Representatives from such diverse backgrounds consider setting simultaneous targets for reducing NMRs as neonatology, safe motherhood programs, and community (Martines and others 2005). The level of participation-- mobilization efforts met over a five-month period to create an involving multiple stakeholders, including women and com- operational plan for newborn care through 2017 (Khadka, munities--and the political will to implement and finance such Moore, and Vikery 2003). plans are also crucial to success. Reaching every pregnant woman and every newborn with effective care involves every- Step 2: Achieve Optimal Newborn Care within the one: the family and community provide home care and advo- Constraints of the Current Health System cate for access to preventive and curative care; the health system Because situations vary even within countries, data-driven pri- supplies care during normal pregnancy, childbirth, and postna- oritization and good leadership are crucial to using resources tal care, along with emergency obstetric and young infant care well (Lawn, McCarthy, and Ross 2001). Program areas related services if required; and the government and global policy to newborn health include safe motherhood, child survival, makers provide supportive policy and resources, in particular immunization, family planning, and nutrition, along with to ensure that there are enough health care providers, such as management of sexually transmitted diseases, prevention of 544 | Disease Control Priorities in Developing Countries | Joy E. Lawn, Jelka Zupan, Geneviève Begkoyian, and others Box 27.3 Adding Newborns to IMCI in India An estimated 1.1 million neonatal deaths occur annually weeks if the infant is LBW, to provide essential new- in India--approximately 28 percent of the world's total. born care, extra care of the LBW infant, and early Between 1960 and 1990, India achieved a 50 percent identification and referral for sepsis reduction in infant mortality, but in the 1990s, the decline -- improved coordination between auxiliary nurse- in the infant mortality rate slowed, partially because of the midwives and community health workers to assist increasing proportion of infant deaths during the neona- with the integration of health and nutrition services tal period. The government looked for ways to add to at household levels. existing programs and to increase coverage of services, The marginal cost of adding N (for neonatal) into IMCI especially given that most neonatal deaths occur in the in relation to clinical care is estimated at less than US$0.10 first few days of life in home settings. per person, given the existence of traditional IMCI pro- Two major adaptations have been made to the standard grams. Training the health and nutrition workers (2 per IMCI approach: 1,000 population) and providing home visits is estimated · Integrated management of neonatal illness was intro- to cost US$0.22 per person. In 2002, the government began duced into the global generic guidelines for IMCI, which to test the integrated management package in 50 districts do not cover illness in the first week of life. of United Nations Children's Fund areas of programming. · Focus on outreach services and family care, taking the This initiative has prompted policy makers to scale up program into communities to achieve higher coverage, is implementation throughout the country during the being promoted through a variety of strategies, namely: 2005­10 phase of its Reproductive and Child Health -- three home visits in 10 days for normal weight Program. babies, with a further three in the subsequent three Source: Adapted from K. Suresh, M. Babille, and V. K. Paul, personal communication, April 2004. maternal-child transmission of HIV, and prevention of malaria cause of neonatal deaths and launched the Neonatal during pregnancy. The reality is that such interventions have Resuscitation Program, developing a course with standard not reached most women and children and that existing serv- guidelines and certification of competency (Deorari and others ices fail to coordinate along the continuum of care. This situa- 2001). Between 1990 and 1992, more than 12,000 physicians tion results in gaps in service and missed opportunities. In and nurses were trained. The effect of the program was evalu- Africa, for example, the regional average for prenatal care cov- ated in 14 teaching hospitals in India. Changes in resuscitation erage is 64 percent, yet coverage of tetanus toxoid immuniza- practices were noted, and asphyxia-related mortality fell signif- tion is 42 percent (Knippenberg and others 2005). Syphilis icantly. The prevalence of survivors with disabilities was not treatment is another opportunity that frequently is missed dur- assessed. ing prenatal care (Gloyd, Chai, and Mercer 2001). Including the An alternative model of skill strengthening has been tested newborn in transport and funding programs that currently in South Africa, where significant improvements in knowledge address only maternal emergencies may be of little marginal and skills have been documented as a result of the Perinatal cost for significant benefit. In India, where integrated manage- Education Programme, a distance-run self-taught course ment of infant and childhood illness (IMCI) is being scaled up, (Woods and Theron 1995). More than 30,000 midwives in the marginal cost of adding selected neonatal conditions to the South Africa have passed the examinations, and the program's clinical care component of IMCI is low, estimated at less than manuals are used in many undergraduate medical and nursing US$0.10 per capita (box 27.3). schools. In many settings in South Asia and Sub-Saharan Africa, even Numerous publications have detailed suboptimal hospital where midwives are in place they do not have the skills required management of women in labor or newborns, variously for newborn care. Competency-based training in neonatal reported as contributing to 10 to 75 percent of all perinatal resuscitation is a rarity and must be incorporated into preser- deaths (Lawn and Darmstadt forthcoming). Thus, there is vice as well as in-service training (box 27.1). India's National scope for improving outcomes and client satisfaction in virtu- Neonatology Forum identified birth asphyxia as a leading ally all settings. For example, in much of Sub-Saharan Africa, a Newborn Survival | 545 Box 27.4 South Africa and the Perinatal Problem Identification Programme: Locally Owned Data for Decision Making Care for pregnant women and newborns in South Africa further 14 percent of avoidable factors were administra- ranges from unattended childbirth in rural mud huts to tive and, in particular, were related to transportation and advanced obstetric and intensive neonatal care. National lack of staff members. About 25 percent of the avoidable perinatal mortality is estimated at 40 per 1,000 live births, factors involving health workers pertained to intrapartum with regional and racial disparities. During the 1990s, care, especially poor monitoring (not using the parto- growing awareness of the importance and preventability graph) and inadequate response to problems identified of newborn deaths resulted in the development of the during labor. Half of the cesarean sections were delayed by Perinatal Problem Identification Programme. Under the an hour or more. program, basic data are entered into a computer program The program identified the following national priori- that calculates perinatal mortality, supporting the identifi- ties to reduce perinatal deaths: cation of avoidable factors to aid the prioritization of · reducing intrapartum asphyxia, especially in rural actions to address key problems. More than 44 sites across areas--for instance, using maternity waiting homes and the country use the Perinatal Problem Identification addressing transport delays Programme, covering almost 80,000 births annually, or · improving intrapartum management by means of proto- approximately 10 percent of deliveries, with 3,045 perina- cols (partograph and effective monitoring), competency- tal deaths (2000). Avoidable factors were identified in based training, and ongoing audit 83 percent of deaths, with half of these being patient · implementing syphilis screening and treatment more related, such as a delayed response to complications. A effectively. Source: Authors, based on data from Pattinson 2002. significant proportion of women deliver in facilities that collect reach those women now. Feasible strategies to reduce NMRs data that could be used to identify achievable improvements in exist (for example, efforts to improve family behaviors, tetanus care (box 27.4). toxoid immunization campaigns, and community-based man- agement of acute respiratory infections) and have been demon- strated in poorly developed health care systems. Interim strate- Step 3: Phase the Systematic Scaling-Up of Newborn Care gies are available, such as linking a group of traditional birth Although some resource-poor countries have succeeded in attendants with skilled attendants (Koblinsky, Campbell, and building functional systems (box 27.5), the process, especially Heichelheim 1999) or medical assistants to perform cesarean for clinical care, takes time. Professional care during childbirth sections. Policy conflicts between skilled and community and childhood illnesses is the ideal, but significant costs are approaches are not helpful. Both approaches are required. With involved in increasing the numbers of professionals and phased program planning, community services can be used retaining them, especially in rural posts. Even maintaining cur- now while professional care is being strengthened. The com- rent staff presents challenges, given low pay and high frustra- munity services can then promote demand for skilled care tion. To markedly increase coverage requires new commitment (Knippenberg and others 2005). now to a massive expansion in the number of midwives and to innovative approaches to retain staff, especially in hard-to- serve areas. Supply constraints must be systematically identi- Step 4: Monitor Coverage and Measure Effect and Cost fied and targeted--notably, human resources, accessibility to In most high-mortality countries, NMRs are measured only facilities, financial barriers, and supply of commodities and intermittently (typically every five years through demographic drugs (Knippenberg and others 2005). Demand-side strategies and health surveys). Tracking of coverage indicators, and espe- are also important, including consideration of subsidies for cially equity of coverage, is important for managing program preventive care or transport for emergency care. decision making. Information is lacking, and the information In the meantime, most neonatal deaths continue to occur in that is available is often not used to improve care. Governments underserved and poor communities that will wait the longest must be encouraged to report funding, coverage, and outcomes for access to skilled care. Each year, 60 million women deliver related to national plans for maternal, neonatal, and child sur- without skilled care present. There is a moral imperative to vival. Donors should also be accountable for reporting funding 546 | Disease Control Priorities in Developing Countries | Joy E. Lawn, Jelka Zupan, Geneviève Begkoyian, and others Box 27.5 Reducing Newborn Deaths Is Possible in Low-Income Countries Sri Lanka achieved neonatal mortality of 11 per 1,000 live nities and without user charges. The period 1980­2000 saw births in 2000 despite a low gross national product per a further 50 percent reduction in the NMR without the use capita of US$800 and less than US$1.50 per capita per year of intensive care, apart from one unit in the capital. of health spending on maternal and neonatal health. In Malaysia also followed a policy of rapid scale-up of the 1959, maternal and neonatal mortality were high, with an coverage of skilled care at birth. It trained large numbers NMR of 50 per 1,000 live births, and gross national prod- of midwives and encouraged collaboration with tradition- uct per capita was US$290. Maternal and infant mortality al birth attendants to promote a gradual transition to were halved by 1980 because skilled childbirth care was skilled care over several decades. The NMR is now 6 per scaled up and because prenatal, childbirth, and postnatal 1,000 live births, and 95 percent of women deliver with a and newborn care services were provided close to commu- skilled attendant. Source: Adapted from Koblinsky 2003. flows and ensuring that commitments are kept (Martines and Understanding is lacking of the effects of maternal infec- others 2005). tions, particularly of synergies between HIV, malaria, and sex- ually transmitted diseases (Ticconi and others 2003) as well as RESEARCH PRIORITIES the potential synergy between maternal infections and appar- ent asphyxial injury to neonates (Peebles and Wyatt 2002). The overwhelming priority in newborn health research Incidence and intervention data regarding neonatal mor- remains how to reach underserved populations. This effort bidity and disability at the population level are entirely lacking involves demonstrating the effect, cost, and scaling up process in LMICs. Improved tools for assessing cause-specific mortali- for packages of interventions. The processes of adapting effec- ty and morbidity outcomes are required to advance answers to tive packages to different settings using various cadres of health many of these questions (Lawn, Cousens, and Wilczynska workers and of identifying indicators of successful implemen- forthcoming). tation that are replicable are all basic to scaling up yet have been little studied.Costing of newborn health interventions is a major gap. Virtually no published examinations of the marginal CONCLUSIONS benefits and costs of adding neonatal interventions to exist- ing programs aimed at safe motherhood, IMCI, HIV/AIDS, Reductions in neonatal mortality are necessary to meet the malaria, and sexually transmitted diseases are available. A fourth MDG. High-impact, low-cost, feasible interventions are demonstration of such synergies will help influence policy available. They could avert approximately 70 percent of the makers to incorporate neonatal issues into these and other pro- world's 4 million neonatal deaths, according to analysis pre- grams. Testing innovative approaches to protect poor families sented here, an estimate similar to the estimates in the Lancet from user costs is also important. neonatal series (Darmstadt and others 2005). Large gains in Given that preterm birth accounts for almost 30 percent of neonatal survival are linked to other health gains, such as neonatal deaths and contributes indirectly to many more reduced childhood morbidity and disability, prevention of still- deaths, reducing the incidence of preterm birth and decreasing births, and improved maternal survival, thus contributing also deaths among preterm infants are important areas for study. to the achievement of the fifth MDG. Low-tech extra care of small babies has the potential to reduce The success of some low-income countries is encouraging, deaths significantly, but the effectiveness of various home and but in South Asia and Sub-Saharan Africa, coverage is generally facility packages, including the potential of emollients for pre- low, progress is slow, and inequity is high. While countries con- venting infections, needs to be tested. A large industry is devel- tinue to move toward a more comprehensive health care sys- oping high-tech devices for newborn care to address the 2 per- tem, simpler approaches at family-community level and cent of neonatal deaths in rich countries. Yet there is little through outreach services can save many lives now, even in the investment in the development and testing of low-cost, simple, poorest settings. Well-known interventions, such as neonatal robust devices in the settings where most fetal and neonatal resuscitation and case management of infections, can be added deaths occur. to other programs, particularly safe motherhood and IMCI Newborn Survival | 547 programs, at low marginal cost. However, to reach the MDGs, Deorari, A. K., V. K. Paul, M. Singh, and D. Vidyasagar. 2001. "Impact of skilled care is required. Scaling up coverage to ensure profes- Education and Training on Neonatal Resuscitation Practices in 14 Teaching Hospitals in India."Annals of Tropical Paediatrics 21 (1): 29­33. sional midwives reach those in underserved areas will require Dragovich, D., G. Tamburlini, A. Alisjahbana, R. Kambarami, major new investment to generate and retain more skilled staff J. Karagulova, O. Lincetto, and others. 1997. "Thermal Control of the members, along with the necessary supportive infrastructure. Newborn: Knowledge and Practice of Health Professionals in Seven This investment will involve increased spending, which--as Countries." Acta Paediatrica 86 (6): 645­50. shown here--may double current national health expenditures Dujardin, B., H. Sene, and F. Ndiaye. 1992. "Value of the Alert and Action Lines on the Partogram." Lancet 339 (8805): 1336­38. per capita in Asia and triple them in many African countries. Duke, T., L. Willie, and J. M. Mgone. 2000. "The Effect of Introduction of Even if poor countries spend more and spend better, outside Minimal Standards of Neonatal Care on In-Hospital Mortality." Papua funding will be required. New Guinea Medical Journal 43 (1­2): 127­36. Current investment in MNCH by most national govern- Ellis, M., N. Manandhar, P. S. Shrestha, L. Shrestha, D. S. Manandhar, and ments and international donors is utterly inadequate compared A. M. Costello. 1999. "Outcome at One Year of Neonatal with investment in conditions that have higher profiles yet Encephalopathy in Kathmandu, Nepal." Developmental Medicine and Child Neurology 41 (10): 689­95. lower mortality rates. The deaths of 10,000 newborns each day Gloyd, S., S. Chai, and M. A. Mercer. 2001. "Antenatal Syphilis in Sub- are unconscionable when most could be saved now at relatively Saharan Africa: Missed Opportunities for Mortality Reduction." Health low cost if the political will to do so existed. Policy and Planning 16 (1): 29­34. Greenwood, A. M., B. M. Greenwood, A. K. Bradley, K. Williams, F. C. Shenton, S. Tulloch, and others. 1987. "A Prospective Survey of the ACKNOWLEDGMENTS Outcome of Pregnancy in a Rural Area of The Gambia." Bulletin of the World Health Organization 65 (5): 635­43. The following individuals are gratefully acknowledged for Institute of Medicine. 2003. Improving Birth Outcomes: Meeting the Challenge of the Developing World. Washington, DC: National Institutes reviewing this chapter: Gary Darmstadt, Affette Mccaw-Binns, of Science. Barbara Stoll, and Anne Tinker. We thank Saving Newborn Jamison, D. T., S. Shahid-Salles, J. S. Jamison, J. Lawn, and J. Zupan. 2006. Lives, especially Julia Ruben, for editing assistance. "Incorporating Deaths Near the Time of Birth into Estimates of the Joy E. Lawn was supported by the Bill & Melinda Gates Global Burden of Disease." In Global Burden of Disease and Risk Factors, eds. Alan Lopez, Colin Mathers, Majid Ezzati, Dean Jamison, Foundation through a grant to Save the Children/USA for the and Christopher Murray. New York: Oxford University Press. Saving Newborn Lives initiative. Jones, G., R. W. Steketee, R. E. Black, Z. A Bhutta, and S. S. Morris. 2003. "How Many Child Deaths Can We Prevent this Year?" Lancet 362 (9377): 65­71. REFERENCES Khadka, N., J. Moore, and C. Vikery. 2003. "Nepal's Neonatal Health Strategy: A Policy Framework for Program Development." In Shaping Bang, A. T., R. A. Bang, S. B. Baitule, M. H. Reddy, and M. D. Deshmukh. Policy for Maternal and Neonatal Health: A Compendium of Case 1999. "Effect of Home-Based Neonatal Care and Management of Studies, ed. S. Crump, 47­52. Washington, DC: JHPIEGO. Sepsis on Neonatal Mortality: Field Trial in Rural India." Lancet 354 http://www.mnh.jhpiego.org/resources/shapepolicy.asp. (9194): 1955­61. Knippenberg, R., J. E. Lawn, G. L. Darmstadt, G. Bekyorian, H. Fogstadt, Bhutta, Z., G. L. Darmstadt, B. Hasan, and R. Haws. 2005. "Community- N. Waleign, and V. Paul. 2005. "Systematically Scaling Up Newborn Based Interventions for Improving Perinatal and Neonatal Outcomes Care in Countries." Neonatal Series Paper 3. Lancet 365: 1087­98. in Developing Countries: A Review of the Evidence." Pediatrics 115 (2): Koblinsky, M. A., ed. 2003. Reducing Maternal Mortality: Learning from 520­603. Bolivia, China, Egypt, Honduras, Indonesia, Jamaica, and Zimbabwe. Blanc, A., and T. Wardlow. 2005. "Monitoring Low Birth Weight: An Washington, DC: World Bank. http://www-wds.worldbank.org/ Evaluation of International Estimates and an Updated Estimation servlet/WDSContentServer/WDSP/IB/2003/06/06/000094946_030528 Procedure." Bulletin of the World Health Organization 83 (3): 178­85. 0402518/Rendered/PDF/multi0page.pdf. Conde-Agudelo, A., J. L. Diaz-Rossello, and J. M. Belizan. 2000. "Kangaroo Koblinsky, M. A., O. Campbell, and J. Heichelheim. 1999. "Organizing Mother Care to Reduce Morbidity and Mortality in Low Birthweight Delivery Care: What Works for Safe Motherhood?" Bulletin of the Infants." Cochrane Database of Systematic Reviews (2) CD00277 World Health Organization 77 (5): 399­406. [PMID:]. Kumar, R. 1995. "Birth Asphyxia in a Rural Community of North India." Daga, S. R., A. S. Daga, S. Patole, S. Kadam, and Y. Mukadam. 1988. "Foot Journal of Tropical Pediatrics 41 (1): 5­7. Length Measurement from Foot Print for Identifying a Newborn at Lawn, J. E., S. N. Cousens, and K. Wilczynska. Forthcoming. Estimating the Risk." Journal of Tropical Pediatrics 34 (1): 16­19. Cause of Death for 4 Million Neonates in the Year 2000. Darmstadt, G. L., Z. A. Bhutta, S. N. Cousens, T. Adam, L. de Bernis, and Lawn, J. E., S. N. Cousens, and J. Zupan. 2005. "Four Million Neonatal N. Walker. 2005. "Evidence-Based, Cost-Effective Interventions That Deaths: When? Where? Why?" Neonatal Series Paper 1. Lancet 365 Matter: How Many Newborns Can We Save and at What Cost?" Lancet (9462): 891­900. 365 (9463): 977­88. Datta, N. 1985. "A Study of Health Problems of Low Birth Weight Babies Lawn, J. E., and G. L. Darmstadt. Forthcoming. "A Review of Strategies to in a Rural Community and the Feasibility of Intervention Package Address Birth Asphyxia Especially for the Poor." Journal of Perinatology Likely to Improve Their Health Status." Ph.D. dissertation, (Suppl.). Postgraduate Institute of Medical Education and Research, Lawn, J. E., B. McCarthy, and S. R. Ross. 2001. The Healthy Newborn: A Chandrigarh, India. Reference Guide for Program Managers. Atlanta: Centers for Disease 548 | Disease Control Priorities in Developing Countries | Joy E. Lawn, Jelka Zupan, Geneviève Begkoyian, and others Control and Prevention and CARE. http://www.cdc.gov/reproductive- Sazawal, S., and R. E. Black. 2003. "Effect of Pneumonia Case Management health/health_newborn.htm. on Mortality in Neonates, Infants, and Preschool Children: A Lumbiganon, P., M. Panamonta, M. Laopaiboon, S. Pothinam, and Meta-analysis of Community-Based Trials." Lancet Infectious Diseases N. Patithat. 1990."Why Are Thai Official Perinatal and Infant Mortality 3 (9): 547­56. Rates So Low?" International Journal of Epidemiology 19 (4): 997­1000. Thaddeus, S., and D. Maine. 1994. "Too Far to Walk: Maternal Mortality in Manandhar, D. S., D. Osrin, B. P. Shrestha, N. Mesko, J. Morrison, K. M. Context." Social Science and Medicine 38 (8): 1091­110. Tumbahangphe, and others. 2004. "Effect of a Participatory Ticconi, C., M. Mapfumo, M. Dorrucci, N. Naha, E. Tarira, A. Pietropolli, Intervention with Women's Groups on Birth Outcomes in Nepal: and others. 2003. "Effect of Maternal HIV and Malaria Infection on Cluster-Randomised Controlled Trial." Lancet 364 (9438): 970­79. Pregnancy and Perinatal Outcome in Zimbabwe." Journal of Acquired Martines, J., V. K. Paul, Z. A. Bhutta, M. Koblinsky, A. Soucat, N. Walker, Immune Deficiency Syndrome 34 (3): 289­94. and others. 2005. "Increasing Newborn Survival: A Call to Action." Tinker A., P. ten Hoope-Bender, S. Azfar, F. Bustreo, and R. Bell. 2005. "A Lancet 365 (9465): 1189­97. Continuum of Care to Save Newborn Lives."Lancet 365 (9462): 822­52. Massawe, A., C. Kilewo, S. Irani, R. J. Verma, A. B. Chakrapam, T. Ribbe, WHO (World Health Organization). 1998a. Basic Newborn Resuscitation: and others. 1996. "Assessment of Mouth-to-Mask Ventilation in Practical Guide. WHO/RHT/MSM/98.1, 1­32. Geneva: WHO. Resuscitation of Asphyctic Newborn Babies: A Pilot Study." Tropical ------. 1998b. Evidence for the Ten Steps to Successful Breastfeeding. Medicine and International Health 1 (6): 865­73. WHO/CHD/98.9, 1­118. Geneva: WHO. Mathers C. D., C. J. L. Murray, and A. D. Lopez. 2006. "The Burden of ------. 2003a. Global Burden of Disease, 2000. Version c. Geneva: WHO. Disease and Mortality by Condition: Data, Methods and Results for the ------. 2003b. Pregnancy, Childbirth, Postpartum, and Newborn Care: A Year 2001." In Global Burden of Disease and Risk Factors, eds. Alan Guide for Essential Practice. Geneva: WHO. Lopez, Colin Mathers, Majid Ezzati, Dean Jamison, and Christopher Murray. New York: Oxford University Press. ------. 2003c. Working with Individuals, Families, and Communities. Geneva: WHO. Meegan, M. E., R. M. Conroy, S. O. Lengeny, K. Renhault, and J. Nyangole. 2001. "Effect on Neonatal Tetanus Mortality after a Culturally-Based ------. Forthcoming. Neonatal and Perinatal Mortality, Estimates for Health Promotion Programme." Lancet 358 (9284): 640­41. 2000. Geneva: WHO. O'Rourke, K., L. Howard-Grabman, and G. Seoane. 1998. "Impact of WHO (World Health Organization) Collaborative Group. 2000. "Effect of Community Organization of Women on Perinatal Outcomes in Rural Breastfeeding on Infant and Child Mortality Due to Infectious Bolivia." Revista Panamericana de Salud Pública 3: 9­14. Diseases in Less Developed Countries: A Pooled Analysis." Lancet 355: 451­55. Pattinson, R. C., ed. 2002. Saving Babies 2001: Second Perinatal Care Survey of South Africa. Pretoria: MRC Unit for Maternal and Infant Health Woods, D. L., and G. B. Theron. 1995. "The Impact of the Perinatal Care Strategies. Education Programme on Cognitive Knowledge in Midwives." South African Medical Journal 85 (3): 150­53. Peebles, D. M., and J. S. Wyatt. 2002."Synergy between Antenatal Exposure to Infection and Intrapartum Events in Causation of Perinatal Brain Yasmin, S., D. Osrin, E. Paul, and A. Costello. 2001. "Neonatal Mortality of Injury at Term." British Journal of Obstetrics and Gynaecology 109 (7): Low-Birth-Weight Infants in Bangladesh." Bulletin of the World Health 737­39. Organization 79 (7): 608­14. Peterson, S., J. Nsungwa-Sabiiti, W. Were, X. Nsabagasani, G. Magumba, Zhu, X. Y., H. Q. Fang, S. P. Zeng, Y. M. Li, H. L. Lin, and S. Z. Shi. 1997. J. Nambooze, and G. Mukasa. 2004."Coping with Paediatric Referral-- "The Impact of the Neonatal Resuscitation Program Guidelines Ugandan Parents' Experience." Lancet 363 (9425): 1955­56. (NRPG) on the Neonatal Mortality in a Hospital in Zhuhai, China." Pojda, J., and L. M. Kelley, eds. 2000. Low Birth Weight: Report of a Meeting Singapore Medical Journal 38 (11): 485­87. in Dhaka, Bangladesh, 14­17 June 1999. Nutrition Policy Paper 18. Zupan, J., and P. Garner. 2000. "Topical Umbilical Cord Care at Geneva: United Nations Subcommittee on Nutrition. Birth." Cochrane Database of Systematic Reviews (3) CD001057 Pratinidhi, A., U. Shah, A. Shrotri, and N. Bodhani. 1986. "Risk-Approach [DOI:10.1002/14651858]. Strategy in Neonatal Care." Bulletin of the World Health Organization 64: 291­97. Saugstad, O. D. 2001."Resuscitation of Newborn Infants with Room Air or Oxygen." Seminars in Neonatology 6 (3): 233­39. Newborn Survival | 549 Chapter 28 Stunting, Wasting, and Micronutrient Deficiency Disorders Laura E. Caulfield, Stephanie A. Richard, Juan A. Rivera, Philip Musgrove, and Robert E. Black Undernutrition and micronutrient deficiencies contribute sub- Growth Faltering stantially to the global burden of disease (Ezzati and others Because nutritional inputs are necessary for children's growth, 2002). Impoverished communities experience high rates of undernutrition is generally characterized by comparing the undernutrition and increased exposure to infectious diseases weights or heights (or lengths) of children at a specific age and caused by crowding and inadequate sanitation. Women of sex with the distribution of observed weights or heights in a ref- reproductive age and children experience devastating health erence population of presumed healthy children of the same consequences as a result of limited resources, cultural influ- age and sex and then calculating z-scores, that is, the difference ences, and biological vulnerabilities. Undernutrition and infec- between a child's weight or height and the median value at that tious diseases exist in a baleful synergy: undernutrition reduces age and sex in the reference population, divided by the standard immunological capacity to defend against diseases, and dis- deviation (SD) of the reference population. A child whose eases deplete and deprive the body of essential nutrients. height-for-age is less than 2 SD is considered stunted, because Undernutrition and infectious diseases further exacerbate the chances of the child's height being normal are less than poverty through lost wages, increased health care costs, and-- 3 percent. A child whose weight-for-age is less than 2 SD is most insidiously--impaired intellectual development that can considered underweight, and one whose weight-for-height significantly reduce earning potential. Health experts have is less than 2 SD is deemed wasted. Stunting results from recently recognized the long-term effects of early undernutri- chronic undernutrition, which retards linear growth, whereas tion and inadequate infant feeding for obesity and chronic dis- wasting results from inadequate nutrition over a shorter peri- eases, including diabetes and cardiovascular diseases. This od, and underweight encompasses both stunting and wasting. chapter summarizes the problems of undernutrition and vita- Typically, growth faltering begins at about six months of age, as min A, iron, zinc, and iodine deficiencies in young children and children transition to foods that are often inadequate in quan- current programmatic efforts to prevent and treat them. tity and quality, and increased exposure to the environment increases their likelihood of illness. Although knowledge about the prevalence of stunting and NATURE, CAUSES, AND BURDEN wasting is preferred, information about underweight is more OF UNDERNUTRITION available globally. The high correlation between stunting and underweight and the low prevalence of wasting mean that the The following section describes the magnitude, distribution, prevalence of underweight directly describes the magnitude of and etiology of growth faltering and specific micronutrient the problem of growth faltering and stunting in young chil- deficiencies in young children. dren. About 130 million children under the age of five are 551 underweight, with the highest prevalences in South Asia and the greatest total burden of disease (Fishman and others 2004). Sub-Saharan Africa (table 28.1). The prevalence of stunting, Children whose weight-for-age is less than 1 SD are also at underweight, and wasting is decreasing in most areas of the increased risk of death, and undernutrition is responsible for 44 world; however, in most of Africa, stunting is increasing. to 60 percent of the mortality caused by measles, malaria, pneu- Childhood malnutrition diminishes adult intellectual monia, and diarrhea. Overall, eliminating malnutrition would ability and work capacity, causing economic hardships for prevent 53 percent of deaths in young children, with most of individuals and their families. Malnourished women tend to those deaths occurring in South Asia and Sub-Saharan Africa deliver premature or small babies who are more likely to die or (table 28.2). suffer from suboptimal growth and development (Allen and Morbidity attributable to undernutrition depends on the Gillespie 2001). Poor early nutrition leads to poor school readi- nature of the illness. Susceptibility to a highly infectious disease ness and performance, resulting in fewer years of schooling, such as measles is unlikely to be affected by nutritional status: reduced productivity, and earlier childbearing. Thus, poverty, all individuals are equally likely to become infected if they are undernutrition, and ill-health are passed on from generation to unvaccinated and naive. However, 5 to 16 percent of pneumo- generation. Undernutrition impedes economic progress in all nia, diarrhea, and malaria morbidity is attributable to moder- developing countries. ate to severe underweight (Fishman and others 2004). As Undernutrition raises the likelihood that a child will become table 28.3 shows, the number of disability-adjusted life years sick and will then die from the disease. Morbidity and mortality (DALYs) attributable to undernutrition is high and, as with are highest among those most severely malnourished; however, mortality, is concentrated in South Asia and Sub-Saharan given the high prevalence of mild to moderate underweight, Africa. The tremendous costs associated with the care and the mildly or moderately underweight individuals experience treatment of childhood diseases that could be partially Table 28.1 Estimated Prevalence of Selected Nutritional Deficiencies in Children Ages Birth through Four, by Region (percent) Weight-for-age Iron Weight-for-age 2 SD through Vitamin A deficiency Zinc Region less than 2 SD less than 1 SD deficiency anemia deficiency East Asia and the Pacific 18 29 11 40 7 Eastern Europe and Central Asia 6 21 1 22 10 Latin America and the Caribbean 6 23 15 46 33 Middle East and North Africa 21 35 18 63 46 South Asia 46 44 40 76 79 Sub-Saharan Africa 32 38 32 60 50 High-income countries 2 14 0 7 5 Sources: Underweight: Fishman and others 2004; vitamin A: Rice, West, and Black 2004; iron: Stoltzfus, Mullany, and Black 2004; zinc: Caulfield and Black 2004. Table 28.2 Estimated Deaths of Children Ages Birth through Four Attributable to Selected Nutritional Deficiencies by Region (thousands) Weight-for-age Vitamin A Iron deficiency Zinc Region less than 1 SDa deficiency anemiab deficiency East Asia and the Pacific 125 11 18 15 Eastern Europe and Central Asia 14 0 3 4 Latin America and the Caribbean 22 6 10 15 Middle East and North Africa 305 70 10 94 South Asia 870 157 66 252 Sub-Saharan Africa 1,334 383 21 400 High-income countries 0 0 6 0 Sources: Underweight: Fishman and others 2004; vitamin A: Rice, West, and Black 2004; iron: Stoltzfus, Mullany, and Black 2004; zinc: Caulfield and Black 2004. a. In high-income countries, the percentage of children at each weight-for-age criterion are those expected in a healthy population. b. Considers only deaths directly attributable to iron deficiency anemia in children. Does not include perinatal deaths attributable to maternal iron deficiency anemia. 552 | Disease Control Priorities in Developing Countries | Laura E. Caulfield, Stephanie A. Richard, Juan A. Rivera, and others Table 28.3 Estimated DALYs Lost by Children Ages Birth through Four Attributable to Selected Nutritional Deficiencies by Region (thousands) Weight-for-age Vitamin A Iron deficiency Zinc Iodine Region less than 1 SD deficiency anemiaa deficiency deficiency East Asia and the Pacific 5,777 994 241 1,004 66 Eastern Europe and Central Asia 489 1 66 149 409 Latin America and the Caribbean 725 218 109 587 83 Middle East and North Africa 10,308 2,403 109 3,290 381 South Asia 27,879 4,761 704 8,510 366 Sub-Saharan Africa 45,131 13,552 596 14,094 748 High-income countries 0 0 40 2 2 Sources: Underweight: Fishman and others 2004; vitamin A: Rice, West, and Black 2004; iron: Stoltzfus, Mullany, and Black 2004; zinc: Caulfield and Black 2004. a. Only considers DALYs directly attributable to iron deficiency anemia. Not included are DALYs due to perinatal deaths attributable to maternal iron deficiency anemia. prevented through improvements in child nutrition have not inadequate intakes of fat, which facilitates the absorption of been quantified. carotenoids. Dietary sources of preformed vitamin A include Evidence is accumulating that early malnutrition increases liver, milk, and egg yolks. Dark green leafy vegetables such as the risk of numerous chronic diseases later (Caballero 2001; spinach, as well as yellow and orange noncitrus fruits (man- Gluckman and Hanson 2004). Associations of early undernu- goes, apricots, papayas) and vegetables (pumpkins, squash, trition with diabetes, hypertension, renal disease, and cardio- carrots), are common sources of carotenoids (vitamin A pre- vascular disease mean that child undernutrition also leads to cursors), which are generally less bioavailable than preformed high adult health care costs. vitamin A but tend to be more affordable. Table 28.1 shows recent estimates of the prevalence of VAD in young children (Rice, West, and Black 2004). Of those Vitamin A Deficiency affected, 250,000 to 500,000 each year will lose their sight as a Vitamin A deficiency (VAD) is a common cause of preventable result. The overall prevalence of VAD is decreasing markedly blindness and a risk factor for increased severity of infectious because of increased awareness of VAD as a public health disease and mortality (Rice, West, and Black 2004). One of the problem and increased measles immunization and vitamin A first symptoms of marginal VAD is night blindness. If VAD wors- supplementation or fortification programs. However, the ens, additional symptoms of xerophthalmia arise, eventually prevalence of VAD is increasing or is unknown in some regions resulting in blindness. A child who becomes blind from VAD has because of political instability, high rates of infectious disease, only a 50 percent chance of surviving the year. Even if children and increasing poverty. survive, blindness severely diminishes their economic potential. VAD may cause anemia in some regions, but it does not appear to impair children's growth (Ramakrishnan and others 2004). Iron Deficiency Increased mortality is associated with VAD, most likely More than 2 billion people, mostly women and young children, because of the detrimental effects on the immune system, are thought to be iron deficient (Stoltzfus and Dreyfuss 1998). which result in increased severity of illness (Sommer and West Iron is found in all plant foods but is more plentiful and 1996). According to Rice, West, and Black (2004), VAD is bioavailable in meat. Deficiency results from insufficient responsible for almost 630,000 deaths each year from infectious absorption of iron or excess loss. Absorption is tightly regulated disease (table 28.2), accounting for 20 to 24 percent of the mor- in the intestines, depending on the iron status of the individual, tality from measles, diarrhea, and malaria (Rice, West, and the type of iron, and other nutritional factors. Once iron is Black 2004). Attributable fractions are highest where VAD is absorbed, it is well conserved. Iron is depleted primarily prevalent and mortality is high. Linking morbidity with VAD is through blood loss, including from parasitic infections such as far more difficult. Vitamin A supplementation decreases the schistosomiasis and hookworm. severity of diarrhea and complications from measles, but in Mainly found in hemoglobin, iron is essential for the bind- some trials, supplementation has been associated with ing and transport of oxygen, as well as for the regulation of cell increased lower respiratory infections. growth and differentiation (Beard 2001). Iron deficiency is the VAD results from inadequate intakes of vitamin A because primary cause of anemia, although vitamin A deficiency, folate of low intakes of animal foods; inadequate intakes of nonani- deficiency, malaria, and HIV also result in anemia. Iron defi- mal sources of carotenoids that are converted to vitamin A; and ciency anemia is most prevalent in South Asia and Sub-Saharan Stunting, Wasting, and Micronutrient Deficiency Disorders | 553 Africa, but it is not limited to developing countries (table 28.1). global coverage, and control of IDD through salt iodation rep- Iron deficiency results in neurological impairment, which may resents a great achievement in international public health. not be fully reversible (Grantham-McGregor and Ani 1999). Nevertheless, significant numbers of people remain at risk. Finally, iron deficiency is known to decrease immune function, but some investigators have also hypothesized that deficiency Zinc Deficiency protects against infectious disease or that iron supplementation Zinc is ubiquitous within the body and is vital to protein syn- increases infectious disease (Caulfield, Richard, and Black thesis, cellular growth, and cellular differentiation. Studies in 2004). Iron deficiency and anemia do not appear to contribute children have demonstrated important roles for zinc in relation to growth faltering (Ramakrishnan and others 2004). to immune function, growth, and development (Brown and Stoltzfus, Mullany, and Black (2004) find that iron deficiency others 2002; Shankar and Prasad 1998). anemia was an underlying factor in 841,000 deaths per year Zinc deficiency results from inadequate intakes and, to resulting from maternal and perinatal causes, and it directly some extent, increased losses. Only animal flesh, particularly causes the deaths of 134,000 young children annually oysters and shellfish, is a good source of zinc, and fiber and (table 28.2). Worldwide, iron deficiency is a substantial contrib- phytates inhibit absorption. Thus, as with iron deficiency, utor to DALY losses (table 28.3). populations consuming a primarily plant-based diet are sus- ceptible. Deficiency can also result from losses during diar- Iodine Deficiency rheal illness. Iodine is necessary for the thyroid hormones that regulate Consensus is currently lacking on how to measure zinc growth, development, and metabolism and is essential to pre- deficiency in individuals. The International Zinc Nutrition vent goiter and cretinism. Inadequate intake can result in Consultative Group recommended using serum or plasma zinc impaired intellectual development and physical growth. A concentrations to identify the risk of deficiency at the popula- range of impairments resulting from iodine deficiency are tion level. In addition, the group used information on referred to as iodine deficiency disorders (IDD) (Hetzel 1983) absorbable zinc in the food supplies of 176 countries to esti- and can include fetal loss, stillbirth, congenital anomalies, and mate the proportion of each national population at risk of hearing impairment. The vast majority of deficient individuals inadequate intake (table 28.1). This information was used to experience mild mental retardation. This decrease in mental calculate the burden of disease (table 28.2) associated with zinc ability and work capacity may have significant economic con- deficiency in young children. Prevalence is not expected to sequences. Iodine deficiency has not, however, been associated decrease unless the implementation of zinc-related interven- with the incidence or severity of infectious disease, and studies tions increases substantially (Caulfield and Black 2004). implicating deficiency as an underlying cause of mortality are The health consequences of severe zinc deficiency have been limited. Because of this, few child deaths can be attributed to elucidated over the past 40 years, whereas the health risks of iodine deficiency, but the directly attributable DALY losses mild to moderate deficiency have been described only recently. remain considerable (table 28.3). Clinical presentations of severe deficiency include growth The prevalence of iodine deficiency is often estimated from retardation, impaired immune function, skin disorders, hypo- the prevalence of palpable goiter, but this method is not gonadism, anorexia, and cognitive dysfunction. Mild to mod- sensitive to milder expressions of deficiency. Iodine deficiency erate deficiency increases susceptibility to infection, and the is thought to be a public health problem in a community if goi- benefits of zinc supplementation on the immune system ter is detected in more than 5 percent of the school-age popu- are well documented (Shankar and Prasad 1998). Zinc can pre- lation. A prevalence greater than 30 percent means that the vent and palliate diarrhea and pneumonia (Zinc Investigators' deficiency is severe. According to World Health Organization Collaborative Group and others 1999, 2000) and also may (WHO) estimates, goiter rates among school-age children reduce malaria morbidity in young children (Caulfield, exceed 5 percent in 130 countries, putting 2,225,000 people at Richard, and Black 2004). Improvements in growth have been risk of IDD. A high prevalence of IDD occurs in Eastern Europe demonstrated (Brown and others 2002), which may operate and Central Asia, the Eastern Mediterranean and North Africa, directly or indirectly through increased immune function and South Asia, and Sub-Saharan Africa (WHO 1999). Iodized salt decreased infectious disease. programs are decreasing iodine deficiency in many regions; Zinc deficiency is estimated to be responsible for about however, this reduction is offset by apparent increases in other 800,000 deaths annually from diarrhea, pneumonia, and regions, where public health officials are now aware of the malaria in children under five (table 28.2). Sub-Saharan Africa, problem because of increased surveillance. the Eastern Mediterranean, and South Asia bear the heaviest Switzerland and the United States embarked on iodine forti- attributable burden of pneumonia and diarrhea, with Sub- fication programs in earnest in the early 1920s. Success resulted Saharan Africa accounting for nearly the entire attributable in enthusiastic political and financial support for increased malaria burden. 554 | Disease Control Priorities in Developing Countries | Laura E. Caulfield, Stephanie A. Richard, Juan A. Rivera, and others INTERVENTIONS Promotion of Optimal Feeding of Infants and Young Children. Much of the early focus on optimal feeding was on Clearly, growth faltering and micronutrient deficiency disor- breastfeeding, which should be immediate and exclusive until ders are prevalent, have deleterious consequences for children's six months of age. At that time nutritious and safe foods should health and development, and are primary contributors to the be added to a diet that is still based on breast milk until early in global burden of disease. Economic development is not the the second year of life. A consensus has been reached that six only path to solving childhood undernutrition. Improvements months is the recommended duration of exclusive breastfeed- in family income may not translate into increased food intakes ing (WHO 2002) and that the total duration is a decision left to because the income elasticity for caloric intake is relatively low. the mother. The effects on micronutrient deficiencies might be greater if Multiple approaches exist to promote the initiation of the food sources of those nutrients (meat, seafood, eggs, forti- breastfeeding and to prolong exclusive breastfeeding--health fied food products) were more sensitive to income increases education; professional support; lay support; health sector and if children had access to those foods. Price subsidies may changes (for example, infant friendly hospitals); and media reduce undernutrition in young children if targeted to foods campaigns--through health facilities and community pro- consumed by them; the potential contribution of price subsi- grams. A recent Cochrane review estimates the potential effec- dies to family nutrition is discussed elsewhere (see chapter 11). tiveness of these approaches (Sikorski and others 2002). This chapter focuses on specific public health measures that are Women who received any form of support for breastfeeding intended to address the problems directly. Progress has been were 22 percent less likely to stop exclusive breastfeeding, and made in some areas, but the current magnitude of the prob- women who received lay support, in particular, were 34 percent lems and of the associated disease burden underscore the need less likely to stop exclusive breastfeeding. Substantial evidence for more investment in nutritional interventions. indicates that interventions can be effective in prolonging breastfeeding and exclusive breastfeeding and that operational Growth Faltering and Childhood Stunting research is needed for program implementation and sustain- ability. If such programs were fully successful, they would Infants and young children falter in their growth because of reduce deaths in children under five by 13 percent (Jones and inadequate dietary intakes and recurrent infectious diseases, others 2003). which reduce appetite, increase metabolic requirements, and Complementary feeding is the process of introducing other increase nutrient loss. Even though this problem is understood, foods and liquids into the child's diet when breast milk alone progress to reduce malnutrition has been slow. Over time, is no longer sufficient to meet nutritional requirements. thinking on how to reduce growth faltering and childhood According to Brown, Dewey, and Allen (1998), complementary stunting has shifted. Whereas previous efforts focused almost feeding practices are suboptimal from several perspectives: exclusively on identifying and rehabilitating severely malnour- ished children, current efforts emphasize prevention through · Complementary foods are introduced too early or too late. combined nutritional and disease prevention and treatment · Foods are served too infrequently or in insufficient amounts, interventions. or their consistency or energy density is inappropriate. Initially, these efforts to prevent undernutrition focused on · The micronutrient content of foods is inadequate to meet diseases rather than on improved child feeding practices as the child's needs, or other factors in the diet impair the such. However, according to Becker, Black, and Brown (1991), absorption of foods. despite the devastating effects of illness on nutritional status, · Microbial contamination may occur. improving dietary intakes is more effective than disease prevention efforts in reducing undernutrition. Because of dra- In addition, because children often do not eat all the food matic reductions in appetite during illness, efforts to improve offered to them, interaction between the caregiver and the dietary intakes initially focused on maintaining energy intakes child, along with other psychosocial aspects of care during despite anorexia and on increasing intakes during recupera- feeding, requires attention. The amount of complementary tion, when appetite may be normal or high. More recent food a child needs depends on breast milk intake. Guidelines interventions aim at feeding healthy children optimal diets, are available for determining energy and nutrient intakes from which includes paying attention to dietary quality. Finally, complementary foods, given breast milk intakes (Dewey and some have argued that, for nutritional advice to be effective, it Brown 2003). needs to be provided alongside growth monitoring and pro- Several reviews of the multiple approaches to improving motion; however, it is increasingly recognized that messages infant and young child feeding practices are available (Allen for prevention are largely universal and that integrated growth and Gillespie 2001; Caulfield, Huffman, and Piwoz 1999; monitoring and promotion are not the only model for service Dewey 2002; Hill, Kirkwood, and Edmond 2004; Swindale and delivery. others 2004). Caulfield, Huffman, and Piwoz (1999) review Stunting, Wasting, and Micronutrient Deficiency Disorders | 555 16 programs in 14 countries to improve dietary intakes of principles for optimal feeding of the breastfed child (PAHO and infants 6 to 12 months of age. The programs were designed to WHO 2003). These principles, outlined in box 28.1, build on promote exclusive breastfeeding and appropriate feeding dur- lessons from previous programmatic efforts such as those ing illness up to age three, and the content and approaches reviewed here and provide a basis for designing comprehensive reflected current thinking regarding nutrition and behavior programs to reduce malnutrition. The international public change. The approaches employed included using the mass health community faces the challenge of implementing and media to reach both caregivers and the population as a whole evaluating these approaches. to change cultural norms about complementary feeding and using one-on-one or small group interactions with community Disease Control and Prevention. Interventions to prevent or health workers to provide individualized information and decrease malnutrition or infectious disease are expected to support. decrease child mortality, and interventions that accomplish Most of the projects achieved good coverage (50 to 70 per- both will have the greatest effect (Pelletier, Frongillo, and cent), with rates varying depending on the communication Habicht 1993). This subsection considers the potential for dis- strategy. They resulted in large shifts in maternal knowledge ease control and prevention efforts to reduce undernutrition in and attitudes and changes in infant feeding practices. In the few young children. programs assessing dietary intakes, intakes improved by 70 to Malaria is responsible for a large portion of childhood mor- 165 kilocalories per day. Differences in nutritional status at tality in Sub-Saharan Africa. The effect of undernutrition on 12 months indicated weight-for-age and height-for-age gains susceptibility to malaria has been discussed at length elsewhere of 0.24 to 0.87 SD. Even with a 50 percent overestimation of the (Caulfield, Richard, and Black 2004), but the nutritional defi- effects, the effect of such programs could translate into tangible ciencies resulting from malaria have been insufficiently reductions in malnutrition and attributable mortality. In addi- explored. Insecticide-treated bednets have been shown to pre- tion, these calculations do not consider the cumulative reduc- vent clinical episodes of malaria and decrease the prevalence of tion in malnutrition from programs that benefit children's anemia in children (Lengeler 2003). Improvements in growth growth into the second and third years of life. Jones and others have also been documented. (2003) use the results of the analysis, along with knowledge of Water, sanitation, and hygiene interventions decrease child- the relationship between underweight and child mortality, to hood malnutrition primarily by preventing diarrheal disease estimate that programs to promote complementary feeding (Checkley and others 2004). Hand-washing interventions can could reduce by 6 percent the deaths of children under five in reduce the risk of diarrheal diseases by about 45 percent. Hand- developing countries. washing interventions can be included in water and sanitation Many programs provide supplemental food to participants programs or can exist as a single intervention, and they are either to provide them an incentive for participating in other both effective and cost-effective (Borghi and others 2002). activities (to offset time costs and increase consumer demand for preventive services) or to rehabilitate severely malnourished children. Although the latter approach is traditionally consid- Vitamin A Deficiency ered for supplemental food programs, the former approach is Even though the consequences of VAD had been defined by more common. Indeed, India's Integrated Child Development 1920, it was 1986 when vitamin A interventions were rigorously Services Program, the world's largest supplemental food pro- studied in a large, controlled community trial (Sommer and gram, plans to shift from rehabilitation to the use of supple- West 1996). A number of other community trials soon also mental food as a "magnet" for providing other integrated child demonstrated a significant decrease in child mortality with development services (Kapil 2002). No consensus exists on vitamin A supplementation (Beaton and others 1993). when or how to include supplemental food to reduce under- Supplementation can alleviate acute VAD quickly, whereas nutrition, and inefficient targeting is frequently a key con- long-term strategies incorporate fortification and dietary straint to effectiveness. Swindale and others' (2004) review of diversification. the effectiveness of food-assisted child survival programs con- Supplementation can be either a curative or a preventive cludes that such programs are reducing malnutrition by 2.0 to measure. If an individual presents with ocular symptoms of 2.5 percent per year. VAD, supplementation is part of the usual standard of care. Despite evidence of the effectiveness of nutritional interven- Beyond the use of supplementation for symptoms that result tions in improving feeding practices and preventing undernu- directly from deficiency, its use as part of the treatment regi- trition, few programs take a comprehensive approach toward men for measles or severe malnutrition can improve health optimizing infant feeding, perhaps because of a lack of consen- outcomes. In deficient areas, high-dose oral supplementation is sus on the key components of a comprehensive strategy. In recommended every four to six months for children under five 2002, participants at a WHO consultation developed 10 guiding and is highly efficacious in reducing ocular effects as well as 556 | Disease Control Priorities in Developing Countries | Laura E. Caulfield, Stephanie A. Richard, Juan A. Rivera, and others Box 28.1 Guiding Principles for Complementary Feeding of the Breastfed Child 1. Practice exclusive breastfeeding from birth to six ries per day at 9 to 11 months, and 550 kilocalories per months of age and introduce complementary foods at day at 12 to 23 months. six months of age (180 days) while continuing to 6. Increase food consistency and variety gradually as the breastfeed. infant gets older, adapting to the infant's requirements 2. Continue frequent, on-demand breastfeeding until and abilities. Infants can eat pureed, mashed, and two years of age or beyond. semisolid foods beginning at six months. By eight 3. Practice responsive feeding, applying the principles of months most infants can also eat finger foods--that is, psychosocial care. Specifically, do the following: snacks that they can eat unaided. By 12 months, most · Feed infants directly and assist older children when children can eat the same types of foods that the rest of they feed themselves, being sensitive to their hunger the family consumes, keeping in mind the need for and satiety cues. nutrient-dense foods. Avoid foods that cause choking. · Feed slowly and patiently; encourage children to 7. Increase the frequency with which the child is fed eat, but do not force them. complementary foods as he or she gets older. The · Experiment with different food combinations, appropriate number of feedings depends on the tastes, textures, and methods of encouragement if energy density of local foods and the usual amounts children refuse many foods. consumed at each feeding. For the average healthy, · Minimize distractions during meals if the child breastfed infant, meals should be provided two or loses interest easily. three times a day at 6 to 8 months of age and three or · Remember that feeding times are periods of learn- four times a day at 9 to 23 months of age, with addi- ing and love, and talk to children during feeding, tional snacks. including making eye contact. 8. Feed a variety of foods to ensure that nutrient needs 4. Practice good hygiene and proper food handling: are met. The child should eat meat, poultry, fish, or · Wash hands before food preparation and eating eggs daily, or as often as possible. Vegetarian diets (both caregivers and children). cannot meet nutrient needs at this age unless nutrient · Store foods safely and serve foods immediately after supplements or fortified products are used. preparation. 9. Use fortified complementary foods or vitamin and · Use clean utensils to prepare and serve food. mineral supplements for the infant, as needed. In some · Use clean cups and bowls when feeding children. populations, breastfeeding mothers may also need · Avoid the use of feeding bottles, which are difficult vitamin and mineral supplements or fortified prod- to keep clean. ucts for their own health and to ensure normal con- 5. Start at six months of age with small amounts of food centrations of certain nutrients in their breast milk. and increase the quantity as the child gets older, while 10. Increase fluid intake during illness, including more maintaining frequent breastfeeding.According to aver- frequent breastfeeding, and encourage the child to eat age breast milk intakes in developing countries, infants' soft, varied, appetizing, favorite foods. After illness, needs from complementary foods are approximately give food more often than usual, and encourage the 200 kilocalories per day at 6 to 8 months, 300 kilocalo- child to eat more. Source: PAHO and WHO 2003. mortality (Sommer and West 1996). A meta-analysis of con- A variety of foodstuffs have been fortified with vitamin A, trolled trials in children demonstrated a 23 percent reduction including oil, monosodium glutamate, butter, wheat flour, in mortality (Beaton and others 1993). High-dose vitamin A sugar, and rice. Fortified white sugar has been successful in supplements are considered safe for infants younger than reducing VAD prevalence in Central America. In El Salvador six months. Several studies suggest that giving vitamin A within and Guatemala, where fortified sugar is the primary source of 48 hours of birth reduces mortality in the first three months by vitamin A, it accounts for approximately 30 percent of the rec- 21 to 74 percent (D. Ross 2002). ommended dietary intake (RDI). Fortification of monosodium Stunting, Wasting, and Micronutrient Deficiency Disorders | 557 glutamate with vitamin A has been demonstrated to be biolog- infections such as malaria and hookworm to anemia does not ically efficacious. Even though program implementation was negate the usefulness of iron supplements; rather it underscores flawed by unacceptable cost, discoloration of the monosodium the need for multiple inputs to prevent severe anemia, given the glutamate, and packaging problems, indicators of VAD risks of transfusion. Although current recommendations indi- declined significantly during periods of fortification in both cate daily supplements, less frequent delivery, such as intermit- Indonesia and the Philippines (Dary and Mora 2002). tently or weekly, is commanding interest. Beaton and McCabe's Vitamin A intakes can also be improved through dietary (1999) meta-analysis concludes that both daily and weekly sup- diversification, either by educating communities about impor- plementation are efficacious if adherence is good. tant sources of vitamin A and beta-carotene that are available In many countries, iron fortification of foods is the principal in the local diet or by increasing economic prosperity so that strategy for reducing iron deficiency and anemia. Fortified individuals have additional funds to spend on a wider variety foodstuffs include wheat and maize flours, noodles, sugar, of food. Education alone has not been demonstrated to affect condiments, and complementary foods and milk for infants and the degree of VAD in a community, but it can be a powerful children. Efficacy studies indicate the potential of fortification tool when incorporated in a broader strategy that also includes to increase iron intakes and reduce anemia, and effectiveness supplementation and fortification (Sommer and West 1996). trials in Chile (dry milk for infants), Ghana (complementary food for young children), Guatemala (sugar), India (salt), Mexico (fortified weaning food and dry milk), and República Iron Deficiency Bolivariana de Venezuela (maize and wheat) have found Despite the public health community's enduring interest in improvements in hemoglobin concentration or reductions in preventing and treating iron deficiency anemia, little evidence anemia prevalence (Allen and Gillespie 2001; Rivera and others suggests that the problem has been reduced. Indeed, in some 2004). Nevertheless, few national iron fortification programs regions the opposite may be true. From the 1970s to the 1980s, have evaluation results that are without controversy. Yip and the iron density of people's diets decreased in every region Ramakrishnan (2002) argue that the strongest examples of the except the Near East and North Africa as iron-poor cereals dis- potential for fortification are found in the Chilean program of placed legumes. During much of this period, iron deficiency fortified dry milk for infants and in the U.S. program of iron- anemia increased in South Asia and Sub-Saharan Africa, where fortified infant cereals. A randomized trial in Mexico of a the problem is most severe (Stoltzfus, Mullany, and Black poverty alleviation program that distributes a complementary 2004). Goals for reducing iron deficiency anemia were articu- food fortified with multiple micronutrients, including iron, lated for the 1990s at the 1990 World Summit for Children, and found positive effects on anemia rates (Rivera and others 2004). many countries adopted policies for providing supplementa- Evaluations of newly implemented iron fortification programs tion for young children; however, few large programs have been should gauge their contribution to anemia prevention. developed to eliminate the problem. Newer strategies, such as sprinkles (powders), spreads, or The explanations for this failure to act include doubts foodlets (a hybrid of a food and a tablet), appear promising, among both scientific program planners and policy makers particularly for regions where the infrastructure will not sup- about the causes and consequences of iron deficiency and port more traditional forms of fortification (Zlotkin and others anemia; lack of political commitment; inadequate program 2003). Processed complementary foods and beverages offer planning, including mobilization and training of health staff additional vehicles for reducing iron and other micronutrient members; insufficient community involvement; and, in partic- deficiencies and promoting well-being (Solon and others ular, inherent difficulties with prolonged adherence to daily 2003). Implementing such strategies and documenting their supplementation (Stoltzfus, Mullany, and Black 2004). Despite cost-effectiveness are important activities for the next few years. this bleak picture, guidelines for supplementation have been In many settings, promoting iron-rich organ meats and ani- formulated for children ages 6 to 24 months and for low birth- mal products and undertaking other food-based strategies may weight infants beginning at 2 months (Stoltzfus and Dreyfuss increase iron intakes and contribute to anemia reduction. Such 1998). Also, various scientific documents synthesize and com- approaches have been promoted for many years, but research is municate current knowledge about the consequences of iron still needed to document their efficacy and effectiveness (Ruel deficiency anemia and programming efforts. and Levin 2000). Ample evidence indicates that iron deficiency is the principal cause of anemia in children; that iron supplements are effica- cious in preventing and treating iron deficiency anemia, increas- Iodine Deficiency Disorders ing hemoglobin concentrations by about 1 gram per deciliter on Interventions to diminish iodine deficiency using either average in controlled trials; and that supplements reduce severe supplementation or fortification are both efficacious and inex- anemia even in malarious areas. The contribution of parasitic pensive, and WHO, the United Nations Children's Fund, and 558 | Disease Control Priorities in Developing Countries | Laura E. Caulfield, Stephanie A. Richard, Juan A. Rivera, and others the International Council for the Control of Iodine Deficiency Fortification interventions include the traditional method Disorders have pledged to eliminate iodine deficiency and the of adding zinc to a commercial food, consumer fortification spectrum of IDD. using sprinkles, and plant-breeding techniques. For example, For regions with severe endemic iodine deficiency, high- Mexico has introduced several large-scale programs, including dose iodine supplementation is indicated while longer-term the fortification of maize and wheat flours and the distribution solutions are put into place. Iodized oil and iodide tablets are of fortified complementary food and fortified milk to low- the most common means of direct administration. Injections income children (Rivera and Sepulveda 2003). Researchers are of iodized oil have been used with much success to decrease investigating the possibility of home fortification of food using the prevalence of IDD and have been shown to be effective for sprinkles containing iron and zinc (Zlotkin and others 2003), three to four years, depending on the dosage (Hetzel 1989). but further research is needed to determine whether sprinkles Although injected oil is effective, it is also expensive, requires are a viable option. Through plant breeding and genetic engi- trained personnel to administer, and carries the risk of infec- neering, staple crops may be made to contain more zinc or less tious disease transmission from contaminated needles. phytate, resulting in increased zinc bioavailability (Ruel and Because of those drawbacks, researchers began exploring oral Bouis 1998). Other dietary strategies target food preparation administration as an alternative. Oral administration of techniques, such as fermentation of unrefined flour to increase iodized oil in liquid and tablet form has been successful in the zinc bioavailability. long-term correction of clinical deficiency, and in Indonesia, oral administration was associated with a reduction in infant mortality (Cobra and others 1997). INTERVENTION COSTS AND Iodized or iodated salt is the primary strategy for correct- COST-EFFECTIVENESS ing iodine deficiency because of the nearly universal con- sumption of salt regardless of socioeconomic status; the lack Multiple strategies exist for preventing malnutrition in young of an effect on consistency, color, or taste from the addition children in the short and long term. This section considers the of iodine; and the limited number of producers in many costs and cost-effectiveness of these interventions for prevent- countries. Large-scale salt fortification has been highly suc- ing malnutrition or deaths attributable to each nutritional cessful in many countries, and of the 130 countries with iodine problem. Table 28.4 presents a compendium of cost informa- deficiency, 75 percent have laws mandating salt iodization. The tion, including, where possible, the costs of preventing a child goal of universal salt iodization for consumption by both death or saving a DALY. humans and livestock in all countries with endemic iodine Horton and others (1996) use data from Brazil, Honduras, deficiency was set at the 1990 World Summit for Children and Mexico to estimate the costs and cost-effectiveness of (WHO, UNICEF, and ICCIDD 2001). Some populations do hospital-based programs to promote breastfeeding. Using not easily embrace salt iodization because of cultural prefer- standard costing methods, they examine the costs of breast- ences or because they have an ample supply of unprocessed feeding promotion activities in each program and the addi- salt, so other means of fortification are needed. One promising tional inputs, as well as the savings. Savings accrued from the option is to add potassium iodate to irrigation water. removal of infant formula where it was currently used. Using data on infant feeding practices and morbidity and mortality from Brazil, they estimated the costs of the programs per Zinc Deficiency birth, per diarrhea case averted, and per death averted. As Although zinc deficiency is likely widespread and even mild table 28.4 shows, the costs of such programs range from deficiency probably has significant health consequences, few US$0.30 to US$0.40 per child, and from US$100 to US$200 per interventions have been developed to combat it in developing death averted, making them comparable in cost-effectiveness to countries. Possible interventions include supplementation, measles and rotavirus vaccination. Assuming that deaths would fortification, and dietary diversification or modification. The otherwise have occurred around age one, and using average strong evidence that the use of zinc supplements given during Latin American life expectancy at that age, yields a cost per and for a short time after diarrhea improves the outcome of that DALY gained of only US$3 to US$7. episode and prevents future episodes has led to the rec- In many community-based strategies, multiple organiza- ommendation that zinc, along with increased fluids and contin- tions work through a variety of communication channels to ued feeding,be used to treat all episodes of acute diarrhea (WHO promote exclusive breastfeeding. Two studies in Ghana and and UNICEF 2004). Substantial efforts are under way to initiate Madagascar provide costs estimates for such programs programs in developing countries. Prophylactic zinc supple- (Chee, Makinen, and Sakagawa 2002; Chee and others 2003). mentation also improves growth and reduces diarrhea incidence The programs cost US$4 to US$16 per child, and given the (International Zinc Nutrition Consultative Group 2004). effect on mothers' practices, the cost ranged from US$5 to Stunting, Wasting, and Micronutrient Deficiency Disorders | 559 Table 28.4 Costs and Cost-Effectiveness of Nutrition Interventions Costs (US$) Type of deficiency Per child or per Per death Per DALY and intervention Source Year Country outcome averted gained Underweight Breastfeeding support Horton and others 1996 1996 Brazil, 0.30­0.40 per birth; 100­200 3­7 Honduras, 0.65­1.10 per diarrhea case Mexico averted Breastfeeding promotion Ross, Loening, and 1987 Mali 2­3 per child 282 11 Mbele 1987 Breastfeeding promotion Chee, Makinen, and 2002 Ghana 16 per child; 5­58 per 203a 7.80 Sakagawa 2002 adopter of exclusive breastfeeding Chee and others 2003 2003 Madagascar 4.41 per child; 10­17 per -- -- adopter of exclusive breastfeeding Child survival program J. Ross 1997; WHO 2002 1997 Across 76­101 per undernourished 1,200 41­43 with nutrition component programs child averted Nutrition programs Less intensive 2­5 per child More intensive 5­10 per child Growth monitoring and Fiedler 2003 2003 Honduras 4 per child; 20 per under- 240­320b 8­11 counseling nourished child averted Vitamin A deficiency Capsule distribution Rassas, Hottor, and others 2004 2004 Ghana 0.90 per child 277 11 Rassas, Nakamba, and others 2004 2004 Zambia 1.23 per child 162 6­7 Fiedler 2000 2000 Nepal 1.25 per child 327 11­12 Fiedler and others 2000 1994 Fortification Institute of Medicine 1998; Guatemala 0.17 per child 1,000 33­35 Sugar World Bank 1994 0.05­0.15 per child Other Iron deficiency Supplements Institute of Medicine 1998; World 1994 3.17­5.30 per child -- -- Bank 1994 Fortification Salt World Bank 1994 1994 India 0.12 per child -- -- Sugar World Bank 1994 1994 Guatemala 0.20­1.00 per child -- -- Cereal World Bank 1994 0.09 per child 2,000 66­70 Iodine deficiency Institute of Medicine 1998; World Bank 1994 Oil injection 1994 Peru 2.75 per child -- -- Zaire 0.80 per child -- -- 1.25 per child -- -- Fortification Water Indonesia 0.05 per child Salt Italy 0.02­0.05 per child 1,000 34­36 Salt India 0.05 per child -- -- Zinc deficiency Supplements with oral Robberstad and others 2004 2004 n.a. 0.47 per child 2,100 73 rehydration salts -- Source did not include data from which to estimate deaths averted (and DALYs gained). Note: Deaths prevented by promoting or supporting breastfeeding are assumed to occur around age one. Deaths prevented by other programs to reduce underweight and all programs to reduce micronu- trient deficiency are assumed to occur between ages one and five. Authors' estimates of costs per DALY (in parentheses) using region-specific life expectancies at ages one and five, reflect this range. a. Assumes that all the DALY gains come from preventing deaths. b. Assumes that an undernourished child has a chance of 1 in 16 to 1 in 12 (6 to 8 percent) of dying before age five, the same as estimated for child survival programs. US$58 per adopter of exclusive breastfeeding. In Ghana, an in Central America. In 1994, estimates indicated that a pro- estimated 883 deaths were averted, yielding a program cost of gram in Guatemala cost US$0.17 per child, and US$1,000 per US$7.80 per DALY gained or US$203 per death prevented. death averted. Counting only the losses from mortality, the cost The range of costs within each program depended on the of saving a DALY was US$33 to US$35. However, for each death baseline prevalence of the behavior, the population density, prevented, there were probably several cases of eye damage and the characteristics of the implementing organizations prevented and of improved general health; thus, taking full themselves. Programs will be more cost-effective when the account of nonfatal effects would reduce the cost per DALY baseline prevalence is lower; the population density is higher; somewhat. and the organizations involved are focused, highly motivated, Iron supplementation is more costly than distribution of and well organized. vitamin A capsules, as it involves a daily supplement over an Less information is available on the costs of community- extended period. Estimates indicate that such programs cost based nutrition programs to prevent growth faltering, to con- US$3.17 to US$5.30 per child. Numerous cost estimates are trol morbidity, and to improve survival. The costs of a program available for iron fortification programs, because these pro- in Mali (Ross, Loening, and Mbele 1987), which included pro- grams have been the principal strategy to prevent and control motion of breastfeeding, counseling, and education on optimal iron deficiency anemia. Such programs have traditionally cost child feeding; prevention of diarrheal disease; and growth US$0.09 to US$1.00 per child, depending on the country and monitoring, were estimated to be US$282 per death averted the vehicle for fortification. These estimates are based on ele- and US$11 per DALY gained. This estimate is consistent with mental iron as the fortifier. Even though this is the cheapest others that nutrition programs cost US$2 to US$10 per child, form available, critics have questioned the bioavailability of ele- depending on the intensity of nutrition counseling, including mental iron, and many researchers now advocate using other Fiedler's (2003) study of the Integrated Community Child Care forms of iron. Program in Honduras, which had an estimated cost of US$4 Iodine fortification programs cost little, about US$0.02 to per child. (For a fuller analysis of such programs, including US$0.05 per child. Iodized oil injections are more costly at contextual and programmatic characteristics that affect out- US$0.80 to US$2.75 per beneficiary, but these programs may comes, see chapter 56.) be recommended for settings where people consume little In the past five years, investigators have undertaken several commercialized and easily fortified food. cost analyses of national programs to distribute vitamin A Currently, no examples of zinc intervention programs are capsules. Two reports from Ghana and Zambia are particularly available from which to estimate cost-effectiveness. However, informative (Rassas, Hottor, and others 2004; Rassas, Robberstad and others' (2004) simulation analysis examines Nakamba, and others 2004). As table 28.4 shows, such pro- the potential costs and cost-effectiveness of providing zinc as grams cost US$0.90 to US$1.23 per child, with the costs per an adjunct to oral hydration salts in treating diarrhea in young death averted ranging from US$162 to US$277. (Deaths from children. Providing zinc as part of case management carries an micronutrient deficiencies are assumed to occur between ages estimated incremental cost of US$0.47 per treatment, ranging one and five, and estimates of cost per DALY ranging from from US$0.33 to US$0.62. Given the relationship between zinc US$6 to US$11 reflect this range, as well as region-specific life provision and mortality risk, this addition to current manage- expectancies at those ages.) These costs are comparable with ment programs would cost, on average, US$2,100 per death estimates of a vitamin A program in Nepal that cost US$1.25 adverted and US$73 per DALY gained. per child and US$327 per death averted (Fiedler 2000). Ching Despite the enormity of the nutritional problems, the and others (2000) examine the costs of incorporating vita- associated loss of DALYS, and the existence of programs to min A capsule distribution into immunization campaigns in combat malnutrition, surprisingly little data on the costs or 50 countries in 1998 and 1999. Their analysis finds that the cost-effectiveness of nutritional programs are available. This total costs per death averted ranged from about US$150 to problem represents a serious gap in information for health US$600, with the incremental costs for vitamin A distribution planning, implementation, and advocacy. Nonetheless, con- amounting to only about US$30 to US$150 per death averted. siderable evidence indicates that when programs to promote The costs per death averted depended on the country setting, breastfeeding or child growth or to correct micronutrient defi- the program's coverage, the delivery of vitamin A (one or two ciencies are delivered to populations with a relatively high doses), and the underlying level of mortality. The incremental prevalence of malnutrition, the cost per participating child is cost per DALY gained could be as low as US$1 or as high as usually so low that deaths can be averted at a cost per DALY US$6. that is less than US$100, and often less than US$10, even in Fewer examples of vitamin A fortification programs are regions with low life expectancy. Few health interventions are available, with the only clear example being sugar fortification comparably cost-effective. Stunting, Wasting, and Micronutrient Deficiency Disorders | 561 ECONOMIC BENEFITS OF INTERVENTION and Remley's (1999) meta-analysis estimates gains of 3.5 IQ points, adjusting for important covariates. The previous sections outlined the costs to society in terms of Iron deficiency has long been associated with developmental deaths and disabilities resulting from growth faltering and delays, and iron supplementation studies have demonstrated micronutrient malnutrition, as well as the costs and cost- improvements in cognitive function. Whether the negative effectiveness of options for their alleviation and prevention; effects of iron deficiency and anemia on development are however, DALYs do not capture the full range of potential reversible remains controversial, which implies the need for benefits to society from effective nutrition programs. For strong preventive measures. More research is needed to learn example, even though the effect of iron deficiency on mental about the effects of iron deficiency on development and to retardation in children contributes to the attributable DALYs develop measures for evaluating programs that provide iron. (Stoltzfus, Mullany, and Black 2004), the negative effects of Multiple lines of evidence indicate that zinc influences iron deficiency on cognition that do not constitute retarda- development (Black 1998). Despite a clear biological role, epi- tion are not considered. Other effects of malnutrition on cog- demiological studies provide insufficient evidence to draw con- nitive and physical functioning that ultimately affect labor clusions on the gain in human capital if zinc deficiency were productivity are also not considered, nor are other long-term reduced through public health interventions. Research to health consequences of child undernutrition. Finally, because address this gap is under way. undernutrition increases the frequency and severity of dis- The public health community has long recognized that ease, undernutrition is associated with considerable health iodine deficiency is the most common cause of preventable care costs, which are also not captured in burden estimates. mental retardation. Even though the problem of maternal iodine deficiency and cretinism in the offspring is well recog- nized, evidence also suggests that deficiency in children is neg- Malnutrition and Human Capital Formation atively associated with cognitive abilities. Bleichrodt and Born's Researchers have studied cognitive function using global meas- (1994) meta-analysis finds losses of 13.5 IQ points in those ures of development and intelligence, such as IQ, along with with iodine deficiency. Some of these effects occur in the school performance and more narrowly defined intellectual, absence of goiter, the hallmark of IDD. More research is needed psychomotor, and behavioral skills. A large body of research to fully understand the human consequences of milder forms has examined whether undernutrition causes lasting cognitive of iodine deficiency that are probably still prevalent in devel- deficits in later life and whether potential deficits are amenable oping countries. to subsequent nutritional interventions. Acute malnutrition is associated with negative neuroanatomical, emotional, and behavioral effects on children's development. After recovery, Malnutrition and Loss of Productivity results of behavioral and developmental tests generally Abundant evidence demonstrates that both anemia and iron improve, but the long-term developmental implications deficiency decrease fitness and capacity for aerobic work by remain unclear. Many studies find IQ scores 8 to 18 points decreasing oxygen transport and respiratory efficiency in mus- lower in children who suffered from severe malnutrition cles. The consequences of iron deficiency are thus measurable (Fishman and others 2004). Studies of chronic undernutrition in terms of loss of economic productivity. Aguayo, Scott, and also report deficits in IQ and school performance with stunting Ross's (2003) case study in Sierra Leone estimates that anemia during early childhood. Evidence from nutritional interven- among women is associated with agricultural productivity tions among high-risk or undernourished children suggests losses of US$19 million per year. For children, the economic that early supplementary feeding (but no sooner than two years costs are not as clear, but those costs may be substantial depend- of age) improves developmental scores during the intervention, ing on the children's ages and the types of work they perform. with some evidence of long-term benefits. For example, follow- Growth faltering that leads to stunting in early childhood up of Guatemalan children exposed to prenatal and early translates into shortened adult stature. Adult height is related postnatal supplementation demonstrated long-term cognitive not only to total food consumption but also to protein intake benefits even after adjusting for socioeconomic factors and (Jamison, Leslie, and Musgrove 2003), which reinforces the educational experience (Pollitt and others 1995). These results importance of dietary quality. Multiple levels of evidence link argue strongly for preventing acute severe malnutrition and adult stature and worker productivity (Martorell 1996). generalized growth faltering that leads to stunting in children. Haddad and Bouis (1991) estimate that a 1 percent decrease in Nutritional interventions may preserve or improve cogni- adult stature is associated with a 1.4 percent decrease in pro- tive function through mechanisms other than preventing ductivity. Others find that a 1 percent increase in adult stature growth faltering or acute malnutrition. For example, breast- is associated with a 2.0 to 2.4 percent increase in wages or earn- feeding confers some cognitive benefits. Anderson, Johnstone, ings. Other things being equal, current programs to prevent 562 | Disease Control Priorities in Developing Countries | Laura E. Caulfield, Stephanie A. Richard, Juan A. Rivera, and others stunting in early childhood can deliver about a third to a half of PROGRAM IMPLEMENTATION: that 1 percent increase in adult stature. Thus, a lifetime of eco- LESSONS OF EXPERIENCE nomic loss results from a failure to prevent stunting in early childhood and accompanying deficits in adult stature, and For decades, countries have implemented programs to alleviate strategies to reduce this tremendous loss are available. growth faltering and micronutrient deficiencies in children; In addition, the impacts of malnutrition on cognitive devel- therefore, it is timely to consider what has been accomplished opment translate indirectly into deficits in productivity in and what can be learned from successes and failures. The task adulthood. Children who are malnourished are more likely to is difficult, because nutrition programs are diverse, ranging start school late, to perform less well, and to stay in school for from the simple fortification of salt with iodine to multifaceted a shorter time (Behrman, Alderman, and Hoddinott 2004). programs to improve dietary intakes and prevent growth fal- Studies suggest that improvements in nutrition within the cur- tering. Nonetheless, some general statements about the state of rent range of benefits of programs for young children can lead programming in this area are possible. to substantial increases in rates of school initiation and to more Success in conceptualizing and implementing programs to years of schooling (Alderman, Hoddinott, and Kinsey 2003; reduce growth faltering by combining disease control strategies Alderman and others 2001; Behrman and others 2003). Both with the promotion of breastfeeding and optimal complemen- years of schooling and school performance affect wages and eco- tary feeding has been demonstrable. The focus has shifted away nomic productivity. Alderman, Hoddinott, and Kinsey (2003) from growth monitoring and promotion (counseling) strate- calculate that the effects of malnutrition during early childhood, gies to population-based assessment with more generalized with the accompanying effects on schooling, lead to a 12 percent dissemination of key messages for behavior change. Available reduction in lifetime earnings in Zimbabwe. Current program- data suggest high cost-effectiveness for such programs. Key ming could restore a significant proportion of those lost wages. challenges involve scaling up and sustainability, as well as strengthening of monitoring and evaluation systems. A gap in this knowledge concerns optimal feeding in the presence of Resource Allocation HIV infection, and testing options and designing programs in Malnutrition increases the likelihood that a child will be sick such settings are of the highest priority. and, when sick, will become seriously ill. Thus, resources must be Iodine fortification has been a clear success over decades, allocated to health care services to deal with the increased fre- which underscores the need for continued and consistent fund- quency and severity of illness caused by undernutrition and ing and advocacy for such programs. Even when universal micronutrient deficiencies. To our knowledge, this increase in access to iodine becomes a reality, policy and programmatic likelihood and severity of illness has never been quantified, but supports will be necessary to maintain it. it is likely to be high, considering not only the costs of health care The success of iodine fortification contrasts with other exam- infrastructure, but also the time costs and costs of lost wages ples of fortification that have made slow and uneven progress. or schooling borne by the family for each episode of illness. Fortification of foodstuffs with vitamin A is limited geo- Furthermore,to the extent that undernutrition or micronutrient graphically, and even though many countries have embarked on deficiencies lead to deficits in cognitive development, resources iron fortification, these programs lag because of controversies need to be allocated to special education, rehabilitation, and about the effectiveness of existing programs, the evaluation vocational services. The costs associated with not providing such methods used, and the lack of infrastructure for fortification in services are ultimately paid in mortality and economic statistics. some settings. Concerted efforts to address the controversies and to provide evidence of the effectiveness of fortification in controlling iron deficiency are under way. In addition, recogni- Adult Disease and Disability tion that fortification should address multiple micronutrient In the past 10 years, a growing literature has identified associa- deficiencies, chiefly the B vitamins and zinc, has grown. tions between small size at birth; early patterns of postnatal Programs to distribute vitamin A capsules twice a year are a growth; and adult conditions as diverse as diabetes, cardiovas- reality in many areas characterized by VAD. Jones and others cular disease, and schizophrenia. More research is needed to (2003) estimate that current coverage of supplementation for provide evidence of causality for such associations and to children living in areas with VAD is 55 percent. In the past few create the evidence base for attributing those effects to mal- years, studies have provided solid data on the costs and cost- nutrition in burden-of-disease calculations. Given current effectiveness of such programs in diverse settings. knowledge, health care budgets in developing countries will In contrast, despite concerns about the health and develop- likely be strained to deal with the burden of chronic diseases of mental consequences of iron deficiency and anemia, few exam- adulthood caused by the failure to prevent maternal and child ples are available of even small-scale iron supplementation undernutrition. programs for young children. Supply and adherence continue to Stunting, Wasting, and Micronutrient Deficiency Disorders | 563 constrain progress, with adherence depending on program Despite decades of nutrition programs, with identifiable suc- workers' and families' perceptions of benefits and their reluc- cesses, uncertainty about their effectiveness persists. The value tance to continue the long-term use in children of what are often of publishing solid process and outcome evaluations in the considered to be medicines. Similar constraints apply to pro- scientific literature in addition to project reports has only grams to provide iron supplements to pregnant women, but recently been recognized and cannot be overstated. Whereas operational research has overcome many obstacles,and the hope outcome evaluations provide data on program effects, process is that the lessons learned can inform the design and implemen- evaluations provide key information to maintain quality assur- tation of iron supplementation programs for young children. ance and to support the plausibility of key outcomes. Food-based strategies, particularly dietary diversification Consensus is growing on the need to evaluate a package of and the promotion of specific food groups for preventing services rather than use complex strategies to tease apart the micronutrient malnutrition, are less advanced than other pro- effects of specific program elements. Well-designed programs grams. In part, this lack of advancement reflects the diverse with process evaluation efficiently provide this information. nature of the behaviors to be changed and of the available Although following standard scientific approaches to estab- options. Given this diversity and the fact that such strategies lish program effectiveness has enabled progress in many are more setting specific than, for example, capsule distribu- interventions, alternative designs can and should be used for tion, the lack of summary estimates of effectiveness is not sur- this purpose. Scientists traditionally argue that randomized prising. Consensus is growing that improving dietary intakes controlled trials are needed to establish causal evidence of through agricultural innovations and dietary diversification effectiveness and that multiple trials are needed in diverse set- represents long-term answers to micronutrient malnutrition, tings, perhaps followed by pooled or meta-analyses to provide but progress is slow because of the urgency of alleviating defi- summary estimates. Others argue that designs that provide ciencies in the short term. More research is needed to define the plausible evidence of program effects or adequate information policies that will promote these strategies. to support continued funding should be recognized as valid by Research over the past decade has articulated a strong funders and publishers in refereed journals and should be case for interventions to prevent zinc deficiency, with sup- implemented more broadly (Victora, Habicht, and Bryce 2004). plementation and fortification identified as important approaches. Experiences with zinc supplementation or forti- fication programs are needed to provide estimates of costs RESEARCH AND DEVELOPMENT AGENDA and cost-effectiveness. If the costs of providing zinc supple- ments to young children are in line with those reported Despite progress, much work remains unfinished. Other chap- earlier, then such programs would be highly cost-effective, ters focus on research and development needs in relation to considering the prevalence and burden of disease associated packaging services, scaling up, and ensuring sustainability. with zinc deficiency. Here the focus is on research for strengthening the database for Child malnutrition results from multiple factors, and even policy making. though each context has its own unique features, the etiology Gaps in knowledge remain with respect to recognized has many more commonalities. Thus, for program planners strategies for intervention programs. Often information on and policy makers intent on alleviating malnutrition to begin intervention efficacy exists, yet little scientific literature on pro- designing and implementing programs in their particular set- gram effectiveness is available. Key gaps include the following: tings from scratch is strikingly inefficient. In the past decade, this point has been recognized, and documents that articulate · evaluation of the effectiveness of national iron fortification processes for program implementation and evaluation have programs to reduce iron deficiency anemia mushroomed. These "road maps" permit policy makers and · implementation and evaluation of the effectiveness of iron program planners to capitalize rapidly on interest in address- supplementation programs for young children ing nutrition problems. The road maps also communicate a · evaluation of the effect on child mortality of multifaceted sense of feasibility by streamlining the complex processes of programs to reduce child undernutrition program design and evaluation. Thus, their use can reduce the · evaluation of the effectiveness of programs based on the likelihood that programs will be diffuse (too many inputs), new guiding principles for reducing undernutrition and will be culturally inappropriate, will have unrealistic expecta- micronutrient malnutrition in young children tions, and will have no possibility of sustainability and no · implementation and evaluation of the effectiveness of food- plans for process or impact evaluation. based strategies to reduce micronutrient malnutrition Demonstrating that nutrition programs are effective is key · implementation and evaluation of the effectiveness of early to translating scientific findings into policies and programs as postnatal vitamin A supplementation to reduce infant well as to ensuring the continuity and expansion of funding. mortality. 564 | Disease Control Priorities in Developing Countries | Laura E. Caulfield, Stephanie A. Richard, Juan A. Rivera, and others Costing studies should accompany the evaluations to allow Allen, L., and S. Gillespie. 2001. What Works? A Review of the Efficacy and estimates of cost-effectiveness for decision making. Effectiveness of Nutrition Interventions. Geneva: United Nations, Administrative Committee on Coordination and Subcommittee on Because of the logistical difficulties in developing fortifica- Nutrition in collaboration with the Asian Development Bank. tion approaches in settings with little industry infrastructure, Anderson, J. W., B. M. Johnstone, and D. T. Remley. 1999. "Breast-Feeding alternative fortification approaches are needed, such as and Cognitive Development: A Meta-Analysis." American Journal of micronutrient sprinkles or foodlets. In addition, operational Clinical Nutrition 70 (4): 525­35. research is needed to develop, implement, and evaluate pro- Beard, J. L. 2001. "Iron Biology in Immune Function, Muscle Metabolism, and Neuronal Functioning." Journal of Nutrition 131 (2 Suppl. 2): grams to improve zinc status. Never has so much evidence been S568­79. amassed on the consequences of a deficiency disorder without Beaton, G. H., R. Martorell, K. J. Aronson, B. Edmonston, G. McCabe, A. C. programmatic application. The challenge now is to develop and Ross, and others. 1993."Effectiveness of Vitamin A Supplementation in implement programs for preventing and treating zinc deficiency the Control of Young Child Morbidity and Mortality in Developing Countries." Geneva: United Nations, Administrative Committee on and to evaluate their effectiveness for child growth, health, and Coordination and Subcommittee on Nutrition. survival. The International Zinc Nutrition Consultative Group Beaton, G. H., and G. McCabe. 1999. "Efficacy of Intermittent Iron (2004) has laid out a research agenda with these aims in mind. Supplementation in the Control of Iron Deficiency Anaemia in Developing Countries: An Analysis of Experience--Final Report to the Micronutrient Initiative." Montreal: Micronutrient Initiative. Becker S., R. E. Black, and K. H. Brown. 1991."Relative Effects of Diarrhea, CONCLUSIONS Fever, and Dietary Energy Intake on Weight Gain in Rural Bangladeshi Children." American Journal of Clinical Nutrition 53 (6): 1499­503. Undernutrition is a major cause of death and disability in Behrman, J., H. Alderman, and J. Hoddinott. 2004. Hunger and young children. When ranked among other causes, growth fal- Malnutrition. Copenhagen: Copenhagen Consensus. tering and micronutrient deficiencies figure prominently, both Behrman, J., J. Hoddinott, J. A. Maluccio, A. Quisumbing, R. Martorell, because they are prevalent and because their consequences are and A. D. Stein. 2003. The Impact of Experimental Nutritional Interventions on Education into Adulthood in Rural Guatemala: devastating. Not included in the numbers, however, are the Preliminary Longitudinal Analysis. Philadelphia: University of losses of lifetime productivity associated with early malnutri- Pennsylvania; Atlanta: Emory University; Washington, DC: Inter- tion and the resources that must be allocated to confront the national Food Policy Research Institute. developmental and morbidity consequences of child malnutri- Black, M. M. 1998. "Zinc Deficiency and Child Development." American Journal of Clinical Nutrition 68 (Suppl. 2): S464­69. tion, which last a lifetime. Bleichrodt, N., and M. Born. 1994. "A Meta-Analysis of Research into Success has been achieved in preventing and controlling Iodine and Its Relationship to Cognitive Development." In The iodine deficiency, and palpable progress has been made in the Damaged Brain of Iodine Deficiency, ed. J. B. Stanbury, 195­200. New past 20 years in correcting vitamin A deficiency and promoting York: Communication Corporation. breastfeeding; however, for iron, articulated goals have not Borghi, J., L. Guinness, J. Ouedraogo, and V. Curtis. 2002. "Is Hygiene Promotion Cost-Effective? A Case Study in Burkina Faso." Tropical been translated into programs, and the problem has remained Medicine and International Health 7 (11): 960­69. the same or worsened. Zinc deficiency is now recognized as an Brown, K. H., K. G. Dewey, and L. H. Allen. 1998. Complementary Feeding important new challenge. of Young Children in Developing Countries: A Review of Current As shown here, solid evidence shows that nutrition pro- Scientific Knowledge. WHO/NUT/98.1. Geneva: World Health Organization. grams can be effective at addressing nutritional problems in Brown, K. H., J. M. Peerson, J. Rivera, and L. H. Allen. 2002. "Effect of young children. Increasingly available cost data, when com- Supplemental Zinc on the Growth and Serum Zinc Concentrations bined with outcome evaluations, demonstrate that nutritional of Prepubertal Children: A Meta-Analysis of Randomized Controlled interventions rank favorably in terms of cost-effectiveness Trials." American Journal of Clinical Nutrition 75 (6): 1062­71. when compared with competing interventions. The case that Caballero, B. 2001. "Early Nutrition and Risk of Disease in the Adult." Public Health Nutrition 4 (6A): 1335­36. further investment in nutrition interventions is warranted is Caulfield, L. E., and R. E. Black. 2004. "Zinc Deficiency." In Comparative thus compelling. Quantification of Health Risks: Global and Regional Burden of Disease Attributable to Selected Major Risk Factors, ed. M. Ezzati, A. D. Lopez, A. Rodgers, and C. J. L. Murray, vol. 1, 257­9. Geneva: World Health REFERENCES Organization. Caulfield, L. E., S. L. Huffman, and E. G. Piwoz. 1999. "Interventions to Aguayo, V. M., S. Scott, and J. Ross. 2003. "Sierra Leone: Investing in Improve the Complementary Food Intakes of 6­12 Month Old Infants Nutrition to Reduce Poverty--A Call for Action." Public Health in Developing Countries: Impact on Growth, Prevalence of Nutrition 6 (7): 653­57. Malnutrition, and Potential Contribution to Child Survival." Food and Alderman, H., J. Behrman, D. Ross, and R. Sabot. 2001. "Child Health and Nutrition Bulletin 20 (2): 183­200. School Enrollment: A Longitudinal Analysis." Journal of Human Caulfield, L. E., S. A. Richard, and R. E. Black. 2004. "Undernutrition as Resources 36 (1): 185­205. an Underlying Cause of Malaria Morbidity and Mortality in Alderman, H., J. Hoddinott, and B. Kinsey. 2003. Long Term Consequences Children." American Journal of Tropical Medicine Hygiene 71 of Early Childhood Malnutrition. Washington, DC: World Bank. (Suppl. 2): S55­63. Stunting, Wasting, and Micronutrient Deficiency Disorders | 565 Checkley, W., P. R. Gilman, P. R. Black, L. D. Epstein, L. Cabrera, P. C. Horton, S., T. Sanghvi, M. Phillips, J. Fiedler, R. Perez-Escamilla, C. Lutter, Sterling, and L. H. Moulton. 2004. "Effect of Water and Sanitation on and others. 1996. "Breastfeeding Promotion and Priority Setting in Childhood Health in a Poor Peruvian Peri-Urban Community." Lancet Health." Health Policy and Planning 11 (2): 56­68. 363 (9403): 112­18. Institute of Medicine. 1998. Prevention of Micronutrient Deficiencies: Tools Chee, G., M. Makinen, and B. Sakagawa. 2002. Cost and Effectiveness for Policy Makers and Public Health Workers. Washington, DC: National Analysis of LINKAGES' Breastfeeding Interventions in Ghana. Bethesda, Academy Press. MD: Abt Associates. International Zinc Nutrition Consultative Group. 2004. "Developing Zinc Chee, G., K. Smith, M. Makinen, and Z. Rambeloson. 2003. Cost and Intervention Programs." Food and Nutrition Bulletin 24 (1): 163­86. Effectiveness Analysis of LINKAGES' Infant and Young Child Feeding Jamison, D. T., J. Leslie, and P. Musgrove. 2003. "Malnutrition and Dietary Program in Madagascar. Bethesda, MD: Abt Associates. Protein: Evidence from China and from International Comparisons." Ching, P., M. Birmingham, T. Goodman, R. Sutter, and B. Loevinsohn. Food and Nutrition Bulletin 24 (2): 145­54. 2000. "Childhood Mortality Impact and Costs of Integrating Vitamin Jones, G., R. W. Steketee, R. E. Black, Z. A. Bhutta, and S. S. Morris. 2003. A Supplementation into Immunization Campaigns." American Journal "How Many Child Deaths Can We Prevent This Year?" Lancet 362 of Public Health 90 (10): 1526­29. (9377): 65­71. Cobra, C., Muhilal, K. Rusmil, D. Rustama, Djatnika, S. S. Suwardi, and Kapil, U. 2002. "Integrated Child Development Services (ICDS) Scheme: others. 1997. "Infant Survival Is Improved by Oral Iodine A Program for Holistic Development of Children in India." Indian Supplementation." Journal of Nutrition 127 (4): 574­78. Journal of Pediatrics 69 (7): 597­601. Dary, O., and J. O. Mora. 2002. "Food Fortification to Reduce Vitamin A Lengeler, C. 2004. "Insecticide-Treated Bednets and Curtains for Deficiency: International Vitamin A Consultative Group Recom- Preventing Malaria." Cochrane Library (2) CD000363. mendations." Journal of Nutrition 132 (Suppl. 9): S2927­33. Martorell, R. 1996. "The Role of Nutrition in Economic Development." Dewey, K. G. 2002. "Successful Intervention Programs to Promote Nutrition Reviews 54 (4 Part 2): S66­71. Complementary Feeding." In Public Health Issues in Infant and Child PAHO and WHO (Pan American Health Organization and World Health Nutrition, Nestle Nutrition Workshop Series Pediatric Program, Organization). 2003. Guiding Principles for Complementary Feeding of ed. R. E. Black and K. F. Michaelsen, vol. 48, 199­216. Philadelphia: the Breastfed Child. Washington, DC: PAHO and WHO. Lippincott, Williams, and Wilkins. Pelletier, D. L., E. A. Frongillo Jr., and J. P. Habicht. 1993. "Epidemiologic Dewey, K. G., and K. H. Brown. 2003. "Update on Technical Issues Evidence for a Potentiating Effect of Malnutrition on Child Mortality." Concerning Complementary Feeding of Young Children in American Journal of Public Health. 83 (8): 1130­33. Developing Countries and Implications for Intervention Programs." Food and Nutrition Bulletin 24 (1): 5­28. Pollitt, E., K. S. Gorman, P. L. Engle, J. A. Rivera, and R. Martorell. 1995. "Nutrition in Early Life and the Fulfillment of Intellectual Potential." Ezzati, M., A. D. Lopez, A. Rodgers, S. Vander Hoorn, C. J. Murray, and Journal of Nutrition 125 (Suppl. 4): S1111­18. Comparative Risk Assessment Collaborating Group. 2002. "Selected Major Risk Factors and Global and Regional Burden of Disease." Ramakrishnan, U., N. Aburto, G. McCabe, and R. Martorell. 2004. Lancet 360 (9343): 1342­43. "Multimicronutrient Interventions but Not Vitamin A or Iron Interventions Alone Improve Child Growth: Results from Three Meta- Fiedler, J. L. 2000. "The Nepal National Vitamin A Program: Prototype to Analyses." Journal of Nutrition 134 (10): 2592­602. Emulate or Donor Enclave?" Health Policy Planning 15 (2): 145­56. Rassas, R., J. K. Hottor, O. A. Anerkai, M. M. Kwame, M. M. Agble, ------. 2003. A Cost Analysis of the Honduras Community-Based A. Nyaku, and T. Taylor. 2004. Cost Analysis of the National Vitamin A Integrated Child Care Program. Washington, DC: World Bank. Program in Ghana. Arlington, VA: International Science and Fiedler, J. L., D. R. Dado, H. Maglalang, N. Juban, M. Capistrano, and Technology Institute. M. V. Magpantay. 2000. "Cost Analysis as a Vitamin A Program Design Rassas, R., P. M. Nakamba, C. M. Mwela, R. Mutemwa, B. Mulenga, W. and Evaluation Tool: A Case Study of the Philippines." Social Science Siamusantu, and T. Taylor. 2004. Cost Analysis of the National Vitamin and Medicine 51 (2): 223­42. A Program in Zambia. Arlington, VA: International Science and Fishman, S., L. Caulfield, M. de Onis, M. Blossner, A. Hyder, L. Mullany, Technology Institute. and R. Black. 2004. "Childhood and Maternal Underweight." In Rice, A. L., K. P. West Jr., and R. E. Black. 2004. "Vitamin A Deficiency." In Comparative Quantification of Health Risks: Global and Regional Comparative Quantification of Health Risks: Global and Regional Burden of Disease Attributable to Selected Major Risk Factors, ed. Burden of Disease Attributable to Selected Major Risk Factors, ed. M. Ezzati, A. D. Lopez, A. Rodgers, and C. J. L. Murray, . vol. 1, 39­162. M. Ezzati, A. D. Lopez, A. Rodgers, and C. J. L. Murray, vol. 1, 211­56. Geneva: World Health Organization. Geneva: World Health Organization. Gluckman, P. D., and M. A. Hanson. 2004."Living with the Past: Evolution, Rivera, J. A., and A. J. Sepulveda. 2003. "Conclusions from the Mexican Development, and Patterns of Disease." Science 305 (5691): 1733­36. National Nutrition Survey 1999: Translating Results into Nutrition Grantham-McGregor, S. M., and C. C. Ani. 1999. "The Role of Policy." Salud Publica Mexico. 45 (Suppl. 4): S565­75. Micronutrients in Psychomotor and Cognitive Development." British Rivera, J. A., D. Sotres-Alvarez, J. P. Habicht, T. Shamah, and S. Villalpando. Medical Bulletin 55 (3): 511­27. 2004. "Impact of the Mexican Program for Education, Health, and Haddad, L. J., and H. E. Bouis. 1991. "The Impact of Nutritional Status on Nutrition (Progresa) on Rates of Growth and Anemia in Infants and Agricultural Productivity: Wage Evidence from the Philippines." Young Children: A Randomized Effectiveness Study." Journal of the Oxford Bulletin of Economics and Statistics. 53: 45­68. American Medical Association 291 (21): 2563­70. Hetzel, B. S. 1983. "Iodine Deficiency Disorders (IDD) and Their Robberstad, B., T. Strand, R. E. Black, and H. Sommerfelt. 2004. "Cost- Eradication." Lancet 2 (8359): 1126­29. Effectiveness of Zinc as Adjunct Therapy for Acute Childhood ------. 1989. "The Prevention and Control of Iodine Deficiency Diarrhea in Developing Countries." Bulletin of the World Health Disorders." Nutrition Policy Discussion Paper 3, United Nations, Organization 82 (7): 523­31. New York. Ross, D. A. 2002. "Recommendations for Vitamin A Supplementation." Hill, Z., B. Kirkwood, and K. Edmond. 2004. Family and Community Journal of Nutrition 132 (Suppl. 9): S2902­6. Practices That Promote Child Survival, Growth, and Development: A Ross, J. S. 1997. Cost-Effectiveness of the Nutrition Communication Project Review of the Evidence. Geneva: World Health Organization. in Mali. Washington, DC: Academy for Educational Development. 566 | Disease Control Priorities in Developing Countries | Laura E. Caulfield, Stephanie A. Richard, Juan A. Rivera, and others Ross, S. M.,W. E. Loening, and B. E. Mbele. 1987."Breast-Feeding Support." Victora, C. G., J. P. Habicht, and J. Bryce. 2004. "Evidence-Based Public South African Medical Journal 72 (5): 357­58. Health: Moving Beyond Randomized Trials." American Journal of Ruel, M. T., and H. E. Bouis. 1998. "Plant Breeding: A Long-Term Strategy Public Health 94 (3): 400­5. for the Control of Zinc Deficiency in Vulnerable Populations." WHO (World Health Organization). 1999. Progress Towards Elimination of American Journal of Clinical Nutrition 68 (Suppl. 2): S488­94. Iodine Deficiency Disorders. WHO/NHD/99.4. Geneva: WHO. Ruel, M. T., and C. Levin. 2000. Assessing the Potential for Food-Based ------. 2002. World Health Report 2002: Reducing Risks, Promoting Strategies to Alleviate Vitamin A and Iron Deficiencies: A Review of Healthy Life. Geneva: WHO. Recent Evidence. Discussion paper 92, International Food Policy WHO and UNICEF (United Nations Children's Fund). 2004. Research Institute, Washington, DC. WHO/UNICEF Joint Statement on the Clinical Management of Acute Shankar, A. H., and A. S. Prasad. 1998. "Zinc and Immune Function: The Diarrhea. WHO/FCH/CAH/04.7. Geneva: WHO and UNICEF. Biological Basis of Altered Resistance to Infection." American Journal of WHO, UNICEF, and ICCIDD (International Council for the Control of Clinical Nutrition 68 (Suppl. 2): S447­63. Iodine Deficiency Disorders). 2001. Assessment of the Iodine Deficiency Sikorski, J., M. J. Renfrew, S. Pindoria, and A. Wade. 2002. "Support for Disorders and Monitoring Their Elimination. WHO/NHD/01.1. Breastfeeding Mothers." Cochrane Database System Reviews (1): Geneva: WHO. CD001141. World Bank. 1994. Enriching Lives: Overcoming Vitamin and Mineral Solon, F. S., J. N. Sarol, A. B. I. Bernardo, J. A. A. Solon, H. Mehansho, L. E. Malnutrition in Developing Countries. Washington, DC: World Bank. Sanchez-Fermin, and others. 2003. "Effect of a Multiple-Micronutrient Yip, R., and U. Ramakrishnan. 2002. "Experiences and Challenges in Fortified Fruit Powder Beverage on the Nutrition Status, Physical Developing Countries." Journal of Nutrition 132 (Suppl. 4): S827­30. Fitness, and Cognitive Performance of Schoolchildren in the Zinc Investigators' Collaborative Group, Z. A. Bhutta, S. M. Bird, R. E. Philippines." Food and Nutrition Bulletin 24 (Suppl. 4): S129­40. Black, K. H. Brown, J. M. Gardner, and others. 2000. "Therapeutic Sommer, A., and K. P. West. 1996. Vitamin A Deficiency Health, Survival, Effects of Oral Zinc in Acute and Persistent Diarrhea in Children in and Vision. New York: Oxford University Press. Developing Countries: Pooled Analysis of Randomized Controlled Stoltzfus, R. J., and M. L. Dreyfuss. 1998. Guidelines for the Use of Iron Trials." American Journal of Clinical Nutrition 72 (6): 1516­22. Supplements to Prevent and Treat Iron Deficiency Anemia. Washington, Zinc Investigators' Collaborative Group, Z. A. Bhutta, R. E. Black, K. H. DC: ILSI Press. Brown, J. M. Gardner, S. Gore, and others. 1999. "Prevention of Stoltzfus, R. J., L. Mullany, and R. E. Black. 2004."Iron Deficiency Anemia." Diarrhea and Pneumonia by Zinc Supplementation in Children in In Comparative Quantification of Health Risks: Global and Regional Developing Countries: Pooled Analysis of Randomized Controlled Burden of Disease Attributable to Selected Major Risk Factors, ed. Trials." Journal of Pediatrics 135 (6): 689­97. M. Ezzati, A. D. Lopez, A. Rodgers, and C. J. L. Murray, vol. 1, 163­209. Zlotkin, S., P. Arthur, C. Schauer, K. Y. Antwi, G. Yeung, and A. Piekarz. Geneva: World Health Organization. 2003. "Home-Fortification with Iron and Zinc Sprinkles or Iron Swindale, A., M. Deitchler, B. Cogill, and T. Marchione. 2004. "The Impact Sprinkles Alone Successfully Treats Anemia in Infants and Young of Title II Maternal and Child Health and Nutrition Programs on the Children." Journal of Nutrition 133 (4): 1075­80. Nutritional Status of Children." Occasional Paper 4, Academy for Educational Development, Washington, DC. Stunting, Wasting, and Micronutrient Deficiency Disorders | 567 Chapter 29 Health Service Interventions for Cancer Control in Developing Countries Martin L. Brown, Sue J. Goldie, Gerrit Draisma, Joe Harford, and Joseph Lipscomb INTRODUCTION earlier onset of the tobacco epidemic, the earlier exposure to occupational carcinogens, and the Western diet and lifestyle in Cancer imposes a major disease burden worldwide, with such countries. In contrast, up to one-fourth of cancers in considerable geographic variations in incidence; mortality; developing countries are associated with chronic infections. survival; overall disease burden; causative environmental fac- Liver cancer is often causally associated with infection by the tors; and mix of prevention, detection, treatment, and palliative hepatitis B virus (HBV), cervical cancer is associated with infec- programs that make up a country's cancer control strategy. tion by certain types of human papillomavirus (HPV), and Unless cancer prevention and screening interventions effec- stomach cancer is associated with Helicobacter pylori infection. tively reduce the incidence of cancer, the number of new can- This chapter focuses on interventions for controlling seven cer cases will increase from an estimated 10 million cases in cancers that impose a particularly heavy burden of disease on 2000 to 15 million in 2020, 9 million of which would be in devel- developing countries: cervical cancer, liver cancer, stomach oping countries. By 2050, the cancer burden could reach 24 mil- cancer, esophageal cancer, lung cancer, colorectal cancer, and lion cases per year worldwide, with 17 million cases occurring in breast cancer. In 2000, these seven types of cancer accounted developing countries (Parkin, Bray, and Devesa 2001). for approximately 60 percent of all newly diagnosed cancer Researchers have made numerous efforts to quantify the cases and cancer deaths in developing countries (Ferlay and global burden of cancer and to estimate site-specific cancer others 2001). Four of the seven cancers--cervical, liver, stom- mortality and morbidity (see, for example, Ferlay and others ach, and esophageal--have elevated incidence and mortality 2004; Parkin, Bray, and Devesa 2001). A recent report from the rates in developing countries. The other three--lung, colorec- International Agency for Research on Cancer provides esti- tal, and breast--have lower incidence and mortality rates than mates of cancer incidence for Africa by site and country (Parkin the other four cancers, but they nonetheless impose a heavy and others 2003). In general, however, data on cancer incidence, disease burden and are increasing because of demographic and prevalence, and mortality are less complete and less accurate in industrial transitions. Pediatric cancers and HIV-related can- developing countries than in developed countries, because the cers, two topics that are of great importance and concern, are latter have more resources to invest in population-based cancer beyond the scope of this chapter. registries and the infrastructure to maintain such registries. Despite the limitations of current data for developing coun- tries, the epidemiology of cancer in developing countries clearly BURDEN OF CANCER IN DEVELOPING COUNTRIES differs from that in developed countries in some important respects. Developed countries often have relatively high rates Data from Ferlay and others (2004) clearly illustrate the differ- of lung, colorectal, breast, and prostate cancer because of the ing patterns of cancer incidence in developing and developed 569 Male Female Male Female Lung Lung Stomach Colon and rectum Breast Breast Liver Prostate Cervix uteri Stomach Esophagus Bladder Colon and rectum Non-Hodgkin's lymphoma Oral cavity Kidney Leukemia Corpus uteri Prostate Pancreas Non-Hodgkin's lymphoma Melanoma of skin Bladder Leukemia Brain, nervous system Liver Ovary Ovary Pancreas Oral cavity Larynx Cervix uteri Other pharynx Brain, nervous system Thyroid Esophagus Nasopharynx Larynx Kidney Thyroid Corpus uteri Multiple myeloma Hodgkin's lymphoma Other pharynx Multiple myeloma Testis Melanoma of skin Hodgkin's lymphoma Cases Deaths Cases Deaths Testis Nasopharynx 6,000 4,000 2,000 0 2,000 4,000 6,000 8,000 6,000 4,000 2,000 0 2,000 4,000 6,000 8,000 Globocan 2002, IARC Globocan 2002, IARC Source: Ferlay and others 2004. Source: Ferlay and others 2004. Figure 29.1 Estimated Number of Cancer Cases of All Ages, Figure 29.2 Number of Cancer Cases of All Ages, Developed Developing Regions, 2002 Regions, 2002 (hundreds) (hundreds) countries (figures 29.1 and 29.2). In developing countries, the For some cancers, including esophageal, liver, lung, and top five female cancers in rank order of incidence are breast, pancreatic cancer, survival rates vary little between developing cervical, stomach, lung, and colorectal cancer; however, cervical and developed countries (Sankaranarayanan, Black, and Parkin cancer still accounts for more deaths than breast cancer in 1998). Currently available methods of early detection and developing countries. The top five male cancers are lung, stom- treatment have not been demonstrated to be effective for these ach, liver, esophageal, and colorectal cancer (figure 29.1). The cancers, so primary prevention remains the most practical incidence of cancers of the lung and breast is relatively high in intervention for control. For a second group of cancers, includ- both developed and developing countries. Colorectal cancer ing large bowel, breast, ovarian, and cervical cancer, proven accounts for a smaller share of the burden in developing coun- methods of early detection, diagnosis, and treatment are avail- tries than in developed countries, but cancer of the stomach able that can, in principle, be delivered through district health accounts for a higher share. Some cancers that are more care facilities. For these cancers, survival rates vary both common in developing than in developed countries, including between developing and developed countries as a whole and stomach, liver, and cervical cancer, are related to the absence of between specific countries within each of these groups. For a a well-developed public health infrastructure for the control of third group of cancers, including testicular cancer, leukemia, cancer-causing infectious agents and contaminants, the lack and lymphoma, the variability in survival between developing of basic preventive health care and screening services for much and developed countries is tremendous. Even though relatively of the population, and the poor-quality diets available to the effective treatments are available for these cancers, they are most economically disadvantaged members of society in many multimodal treatments that require a relatively high level of developing countries. Cancer of the esophagus, also relatively medical resources, a good health care infrastructure, and a level common in developing countries, may reflect, in part, the con- of sophisticated knowledge, which low- and middle-income sumption of traditional beverages at extremely high tempera- developing countries may not have. tures. Some cancers that are increasingly common in develop- Table 29.1 shows estimated cancer deaths and the estimated ing countries, including lung, breast, and colorectal cancer, may disease burden in terms of disability-adjusted life years reflect the increasing Westernization of lifestyles, longer life (DALYs) lost as a result of various types of cancers in develop- expectancy, and globalization of markets for tobacco products. ing and developed countries and by region in 2001. As the table 570 | Disease Control Priorities in Developing Countries | Martin L. Brown, Sue J. Goldie, Gerrit Draisma, and others Table 29.1 Number of Cancer Deaths and DALYs Lost to Cancer, by World Bank Region and Country Income Level, 2001 East Asia and Europe and Latin America and Middle East and Sub-Saharan Low- and middle- High-income the Pacific Central Asia the Caribbean North Africa South Asia Africa income countries countries Cancer site Deaths DALYs lost Deaths DALYs lost Deaths DALYs lost Deaths DALYs lost Deaths DALYs lost Deaths DALYs lost Deaths DALYs lost Deaths DALYs lost Trachea, bronchus, 387,000 5,333,000 165,000 2,323,000 55,000 728,000 20,000 283,000 129,000 1,807,000 15,000 225,000 771,000 10,701,000 456,000 5,397,000 and lung cancers Stomach cancer 442,000 6,134,000 101,000 1,376,000 57,000 735,000 18,000 252,000 45,000 629,000 33,000 487,000 696,000 9,616,000 146,000 1,628,000 Liver cancer 373,000 5,923,000 28,000 379,000 21,000 277,000 9,000 138,000 27,000 464,000 46,000 762,000 505,000 7,945,000 102,000 1,223,000 Esophageal cancer 232,000 3,217,000 21,000 288,000 16,000 215,000 5,000 72,000 80,000 1,116,000 24,000 343,000 380,000 5,252,000 58,000 702,000 Colorectal cancer 159,000 2,334,000 96,000 1,290,000 37,000 485,000 10,000 164,000 35,000 499,000 20,000 291,000 357,000 5,060,000 257,000 3,175,000 Breast cancer 93,000 1,730,000 63,000 1,058,000 37,000 642,000 14,000 273,000 76,000 1,246,000 34,000 574,000 317,000 5,527,000 155,000 2,509,000 Mouth and 66,000 1,064,000 27,000 426,000 14,000 204,000 5,000 78,000 140,000 2,020,000 19,000 284,000 271,000 4,078,000 41,000 576,000 oropharyngeal cancers Cervix uteri cancer 47,000 805,000 19,000 356,000 26,000 494,000 5,000 93,000 83,000 1,423,000 38,000 627,000 218,000 3,799,000 17,000 319,000 Lymphomas and 42,000 753,000 23,000 375,000 24,000 383,000 12,000 232,000 82,000 1,401,000 34,000 622,000 216,000 3,770,000 115,000 1,362,000 multiple myeloma Health Leukemia 76,000 1,652,000 27,000 462,000 22,000 444,000 14,000 307,000 38,000 851,000 14,000 245,000 190,000 3,965,000 73,000 919,000 Service Prostate cancer 16,000 164,000 25,000 283,000 37,000 340,000 6,000 64,000 21,000 210,000 40,000 416,000 145,000 1,479,000 119,000 1,212,000 Pancreatic cancer 37,000 544,000 35,000 481,000 20,000 248,000 4,000 55,000 13,000 176,000 8,000 117,000 117,000 1,621,000 110,000 1,232,000 Interventions Bladder cancer 30,000 348,000 24,000 300,000 9,000 100,000 15,000 214,000 30,000 408,000 10,000 133,000 117,000 1,504,000 59,000 670,000 Ovarian cancer 25,000 464,000 21,000 350,000 9,000 152,000 2,000 42,000 21,000 327,000 9,000 152,000 86,000 1,488,000 46,000 651,000 Corpus uteri cancer 8,000 175,000 17,000 349,000 12,000 254,000 1,000 22,000 4,000 66,000 3,000 41,000 44,000 908,000 27,000 586,000 for Cancer Melanoma and 5,000 66,000 11,000 160,000 7,000 97,000 1,000 19,000 3,000 41,000 8,000 118,000 35,000 501,000 30,000 409,000 other skin cancers Control Other malignant 104,000 1,640,000 123,000 1,901,000 82,000 1,263,000 26,000 440,000 26,000 1,444,000 55,000 844,000 490,000 7,538,000 257,000 3,316,000 neoplasms in Total (all malignant 2,142,000 32,346,000 826,000 12,157,000 485,000 7,061,000 167,000 2,748,000 853,000 14,128,000 410,000 6,281,000 4,955,000 74,752,000 2,068,000 25,886,000 Developing neoplasms) Source: Mathers and others 2006. Countries Note: For an explanation of how DALYs are computed, see chapter 15 of this volume. | 571 shows, the seven types of cancer that are the focus of this There is increasing emphasis worldwide on the develop- chapter account for seven of the first eight cancer sites ranked ment of specialized cancer centers that apply evidence- by number of deaths in developing countries. Considerable based multimodal therapies, including rehabilitation and heterogeneity in the pattern of cancer burden across the six palliative care. regions is apparent, and additional heterogeneity is apparent · Palliative care. The scope of palliative care has been within these regions. Deaths from liver cancer are relatively expanded in recent years to encompass the alleviation of high in East Asia and the Pacific and in Sub-Saharan Africa, symptoms and treatment during all phases of disease-- probably because of the high prevalence of chronic HBV from diagnosis to death--and to address matters related to infection and the lack of adequate resources for food storage the psychological and quality-of-life aspects of disease, as and preservation in those regions (Parkin and others 2003). well as the physiological aspects. Furthermore, palliative The number of deaths from colorectal and breast cancer, as a care has been expanded to include consideration for the proportion of all cancer deaths, is relatively high in Europe well-being of the patient's family members as well as for and Central Asia and in Latin America and the Caribbean, the patient (Singer and Bowman 2002). probably because those regions have increasingly adopted more Western lifestyle patterns of reproductive behavior, diet, The discussion in this chapter focuses primarily on health and physical activity. The number of deaths from oral cancer service interventions for controlling the seven cancers that are is particularly high in South Asia, where the use of betel quid is the subject of this chapter. Other chapters deal with broad pub- common. lic health interventions involving the control of occupational and environmental exposures; health education; policy inter- ventions such as regulation, labeling, and taxation related to TYPES OF INTERVENTIONS FOR tobacco consumption; diet; and physical activity. CANCER CONTROL The World Health Organization (WHO) emphasizes that, COST-EFFECTIVENESS OF CANCER when developing national strategies for controlling cancer, CONTROL INTERVENTIONS countries should consider the following four broad approaches There is a growing literature on the cost-effectiveness of inter- (WHO 2002): ventions within each of the four categories above. In this sec- tion, we review published studies of the cost-effectiveness of · Primary prevention. The goal of primary prevention is to health services­based cancer control interventions, and we reduce or eliminate exposure to cancer-causing factors, present new analyses of the cost-effectiveness of screening which include environmental carcinogens and lifestyle fac- interventions for cervical and breast cancer. tors related to nutrition and physical activity. For the seven cancers considered here, approaches to primary prevention include immunization against, or treatment of, infectious Primary Prevention agents that cause certain cancers; use of tobacco control This subsection reviews studies of the effectiveness and cost- programs; reduction of excessive alcohol consumption; effectiveness of several interventions for the primary preven- dietary intervention; and pharmacological intervention. tion of cancer. · Early detection and secondary prevention. The main objective of early detection or secondary prevention through Immunization against--or Treatment of--Infectious Agents population-based screening programs is detection at a stage That Cause Certain Cancers. Infectious agents are causally at which curative treatment is possible. Interventions for the associated with three of the seven cancers that are the focus of early detection of cancer can help reduce mortality from this chapter--liver cancer (HBV), cervical cancer (HPV infec- cancer only if they are part of a wider cancer control strategy tion), and stomach cancer (H. pylori infection)--so eliminat- that includes effective diagnostic follow-up procedures and ing these agents through immunization or other means offers treatment (Anderson and others 2003). For cervical, col- hope for preventing such cancers. orectal, and breast cancer, effective methods of early detec- The HBV vaccine was designed to prevent liver cancer and tion and treatment are available, but their implementation is currently the only such vaccine in widespread use. Long-term has been uneven (Sankaranarayanan, Black, and Parkin protection against acute and chronic infection has been 1998). demonstrated with the HBV vaccine in a wide range of settings · Diagnosis and treatment. The primary modalities of cancer (Coursaget and others 1994; Viviani and others 1999), and treatment are surgery, chemotherapy, and radiotherapy, and recent data support a reduction in hepatocellular carcinoma these modalities may be used alone or in combination. (Lee, Hsieh, and Ko 2003). 572 | Disease Control Priorities in Developing Countries | Martin L. Brown, Sue J. Goldie, Gerrit Draisma, and others Infection with specific high-risk types of HPV plays a key role Several studies, most of them in developed countries, have in causing cervical cancer. A double-blind placebo-controlled assessed the potential cost-effectiveness of screening individu- trial of an HPV 16 vaccine reported encouraging efficacy results als for infection with H. pylori and then eradicating H. pylori in young female volunteers who had been fully vaccinated (three with antibiotic therapy as a means of preventing the later doses of vaccine or placebo) over a 1.7-year follow-up period occurrence of stomach cancer. Roderick and others (2003) (Koutsky and others 2002). In a more recent study, a bivalent examine the cost-effectiveness of an H. pylori screening pro- HPV 16/18 vaccine prevented approximately 95 percent of per- gram conducted in the United Kingdom. Discounting costs sistent infections with HPV 16 and 18 (Harper and others 2004). and benefits at 6 percent, they find that the cost-effectiveness Several modeling studies have explored the potential bene- ratio for screening for H. pylori, initiated at age 40, is approxi- fits of HPV vaccination at the population level (Goldie and mately US$28,000 per year of life saved (YLS). Optimal cost- others 2003; Hughes, Garnett, and Koutsky 2002; Kulasingam effectiveness was not achieved until the H. pylori screening and Myers 2003) and have elucidated several priorities for program had run for at least 40 years. Harris and others (1999) future research, including a better understanding of the hetero- estimate the cost-effectiveness ratio associated with one-time geneity of vaccine response and the effects of type-specific vac- screening for H. pylori at age 50 to be approximately cination on other HPV types. US$50,000 per YLS (in 1995 dollars, 3 percent discount rate) Hughes, Garnett, and Koutsky (2002) evaluate the potential when treatment for H. pylori infection results in a 15 percent effectiveness of HPV vaccination using a dynamic transmission reduction in stomach cancer risk. Assuming a 30 percent model and find that, when both men and women were reduction, the figure was US$25,000 per YLS for the United vaccinated--assuming 90 percent coverage, 75 percent effec- States, but only a few hundred dollars per YLS in Colombia, tiveness, and 10-year immunity--type-specific HPV prevalence which has a much higher rate of stomach cancer and lower was reduced by 44 percent. When only women were vaccinated, health care costs. the reduction was 30 percent. The authors show that, if the vac- cine targeted only certain types of high-risk HPV, cervical can- Tobacco and Alcohol Control Programs. Tobacco consump- cer incidence was not reduced proportionally because other tion is the most important cause of lung and other cancers of high-risk types of HPV progressed to invasive cancer. the respiratory system, as well as of esophageal cancer, and may Goldie and others (2003) assess the impact of a type-specific be a contributing factor for several other cancers. The most HPV 16/18 vaccine calibrated to population-based data for effective national tobacco control programs combine health Costa Rica. They find that a vaccine that prevented 98 percent promotion, education, and health service interventions of persistent HPV 16/18 was associated with an approximate with policies. Policy instruments include regulating tobacco equivalent reduction in HPV 16/18­associated cancer and a advertising and promotion; enacting smoking bans in work- 51 percent reduction in total cervical cancer. The effect on total places, restaurants, and public buildings and on public cancer was attenuated because of the competing risks associ- transportation; and increasing excise taxes on tobacco products ated with oncogenic types of HPV other than HPV 16/18. (Fiore, Hatsukami, and Baker 2002; WHO 2002). Decreased Three studies have evaluated the potential cost-effectiveness rates of smoking uptake by children and adolescents would of HPV vaccination in countries with cervical cancer screening result in the greatest potential gain in life years. The WHO programs (Goldie, Kohli, and others 2004; Kulasingam and Framework Convention on Tobacco Control (WHO 2003b) Myers 2003; Sanders and Taira 2003). In general, these studies summarizes tobacco control policies and programs related to indicate that a program of HPV vaccination that permits a later regulation, taxation, and education. Da Costa e Silva (2003) age of screening initiation and a less frequent screening inter- shows prioritized treatment approaches for tobacco cessation, val is likely to be a cost-effective use of health care resources in based on countries' levels of resources. developed countries. Excessive alcohol use accounts for 20 to 30 percent of liver In Fujian province, China, a region of high mortality attrib- and esophageal cancer (WHO 2001b). Interventions to reduce utable to stomach cancer, a recently completed randomized excessive consumption of alcohol have many principles in controlled trial of H. pylori eradication with antibiotics pro- common with tobacco control, including the effectiveness of vides some evidence that this approach may be effective in pre- regulatory and taxation measures along with health promotion venting stomach cancer in the subgroup of H. pylori carriers and addiction treatment programs. without precancerous lesions at the time of treatment (Wong and others 2004). A recent randomized trial of H. pylori eradication in Chiapas, Mexico, which used preneoplastic con- Dietary and Related Interventions. The dietary ingestion of ditions as surrogate markers for the development of gastric substances produced by the mold Aspergillus flavus, specifically cancer, found some evidence for the effectiveness of this treat- aflatoxin B1, is causally associated with hepatocellular carci- ment (Ley and others 2004). noma. Exposure to aflatoxins may be synergistic with HBV Health Service Interventions for Cancer Control in Developing Countries | 573 infection in the development of this cancer. Effective means are Early Detection and Secondary Prevention available for preventing the contamination of grains and other This subsection looks at studies of the effectiveness and cost- types of food with aflatoxin during the growth, harvest, stor- effectiveness of several interventions for the early detection and age, and processing of such products (Kensler and others 2003; secondary prevention of cancer. Turner and others 2002). Furthermore, chlorophyllin supple- ments have been found to reduce the carcinogenic properties of aflatoxin. That finding provides additional evidence for current Screening for Liver Cancer. Screening methods for early dietary guidelines that meals should contain foods rich in detection of liver cancer include serum assays for alpha- chlorophylls--for example, spinach and other green, leafy veg- fetoprotein and, potentially, ultrasound. A recently completed etables (Kensler and others 2003). randomized controlled trial of liver cancer screening in China Among those infected with H. pylori, diet is thought to play evaluated the use of two or six alpha-fetoprotein assays over a a critical role in the progression of superficial gastritis to period of four years among men age 30 to 69 with chronic HBV chronic atrophic gastritis. Prolonged consumption of foods (Chen and others 2003). Screening resulted in earlier diagnosis rich in salted, pickled, and smoked products increases the risk of liver cancer, but because treatment for established liver of stomach cancer, and increased consumption of fresh fruit cancer is largely ineffective, screening did not reduce overall and vegetables likely decreases the risk. Obesity is also a well- mortality. established risk factor for several cancers (Vainio and Bianchini Randomized trials that include ultrasound screening for 2002b). For that reason, WHO recommends that governments liver cancer and that incorporate recent advances in antiviral seeking to ensure compliance with nutritional objectives con- preventive treatment have yet to be conducted. Sarasin, duct appropriate school and public education campaigns on Giostra, and Hadengue's (1996) model-based cost-effectiveness diet and work with the food and agriculture sectors (WHO analysis explores whether biannual screening of patients with 2002). Child-Pugh class A cirrhosis, under a set of assumptions sys- tematically favorable to screening, would be cost-effective. The Pharmacological Interventions. Chemoprevention is defined authors conclude that, even under best-case conditions, screen- as the reduction of the risk of cancer development through the ing for liver cancer is not likely to be cost-effective. use of micronutrients or pharmaceuticals. Clinical trials among high-risk individuals to establish the efficacy of chemopreven- Screening for Stomach Cancer. Mass screening programs for tion via micronutrients (for instance, carotenoids and the early detection of invasive stomach cancer using radiologi- retinoids) and dietary fiber have been mainly negative (Alberts cal or endoscopic techniques have been widely implemented in and others 2000; ATBC 1994; Omenn and others 1996; Japan, where incidence rates of stomach cancer are high. Schatzkin and others 2000). However, several ongoing clinical Babazono and Hillman (1995) compare the cost- studies are examining the potential cancer preventive effects effectiveness of three methods for the early detection of stom- of calcium, vitamin D, folic acid, selenium, and vitamin E ach cancer in the context of mass screening programs in Japan: (Christensen 2004). indirect radiology (barium meal plus photofluoroscopy), direct Both case-control and cohort studies show a reduced risk radiology, and endoscopy. When screening for stomach cancer for colorectal cancer after prolonged use of aspirin (Vainio and was started late in life, indirect radiology was the most cost- Morgan 1999). Additional evidence indicates that aspirin has effective screening method. This analysis supports an increase a preventive effect on several other types of cancer, including in the recommended age for initiating screening for stomach hormone receptor­positive breast cancer (Terry and others cancer from age 40 to 50. 2004), but questions remain about the balance between the clinical benefits and adverse side effects of long-term aspirin therapy, including gastrointestinal bleeding and hemorrhagic Screening for Lung Cancer. Investigators have carried out sev- stroke (Imperiale 2003). eral cost-effectiveness analyses of the screening of high-risk Some evidence suggests that the antiestrogen drug tamox- individuals, such as current and former smokers, for lung can- ifen may reduce the risk of breast cancer (Gail and others cer using helical computed tomography (Chirikos and others 1999), but there is also conflicting evidence (Powles and others 2002; Mahadevia and others 2003; Marshall and others 2001). 1998; Veronesi and others 1998). The potential for primary The results of these studies vary widely from quite favorable prevention using other selective estrogen receptor modulators (US$19,000 per YLS) to extremely unfavorable (more than is a topic of current clinical research (Lippman, Lee, and US$100,000 per YLS). The main reason for the wide variation Sabichi 1998). Preliminary analyses indicate that the use of in these studies is different assumptions about the clinical tamoxifen to prevent breast cancer could be cost-effective in nature of early lung lesions detected by helical computed the United States (T. Smith and Hillner 2000). tomography--specifically, whether a large proportion of these 574 | Disease Control Priorities in Developing Countries | Martin L. Brown, Sue J. Goldie, Gerrit Draisma, and others Table 29.2 Estimates of the Cost-Effectiveness of Colorectal Cancer Screening Interventions, United States (cost-effectiveness ratios expressed as 2000 US$/YLS) Wagner and Frazier and Khandker and Sonnenberg, Delco, Vijan and Colorectal screening test others 1996 others 2000 others 2000 and Inadomi 2000 others 2001 Annual fecal occult blood test 11,725 17,805 13,656 10,463 5,691 Flexible sigmoidoscopy every 5 years 12,477 15,630 12,804 39,359 19,058 Flexible sigmoidoscopy every 5 years 13,792 22,518 18,693 n.a. 17,942 and annual fecal occult blood test Double-contrast barium enema every 5 years 11,168 21,712 25,624 n.a. n.a. Colonoscopy every 10 years 10,933 21,889 22,012 11,840 9,038 Source: Pignone and others 2002. n.a. not applicable. Note: All costs and life years are discounted at 3 percent, except in the study by Wagner and others (1996), who use a discount rate of 5 percent. small lung nodules represents "pseudo-disease" that will never with simple visual screening methods, such as visual inspection progress to clinical lung cancer (Marcus and others 2000). The after application of an acetic acid solution (VIA), or with HPV National Lung Cancer Screening Trial, currently under way DNA testing (Denny and others 2000; Sankaranarayanan and (van Meerbeeck and Tournoy 2004), hopes to answer this others 1999; Schiffman and others 2000; Wright 2003; Wright question. Until results from the trial are available, no definitive and others 2000; Zimbabwe Project 1999). These newer statement can be made about the effectiveness or cost- options also eliminate colposcopy, potentially allowing screen- effectiveness of lung cancer screening. ing and treatment to be performed during the same visit. In middle-income countries where cytology screening is available Screening for Colorectal Cancer. Screening methods for early but cervical cancer mortality has not been reduced, key ques- detection of colorectal cancer include fecal occult blood testing, tions center around improving the quality of cytology-based sigmoidoscopy,barium enema,and colonoscopy.Several studies programs; such improvement includes having adequate col- of the cost-effectiveness of colorectal cancer screening in devel- poscopy and biopsy facilities and accessible treatment oped countries have been published (Pignone and others 2002). (Lazcano-Ponce and others 1999); making use of HPV DNA Table 29.2 presents estimates of the cost-effectiveness of col- testing technology in a cost-effective manner; and targeting the orectal cancer screening in the United States. Cost-effectiveness appropriate age group for cervical cancer screening more accu- ratios for various modalities of colorectal cancer screening range rately. The vast majority of published cost-effectiveness analy- from almost US$6,000 to about US$40,000 perYLS. Using mod- ses of population-based cervical cancer screening performed els closely linked to European trials of biennial fecal occult blood during 1980­2003 focused on high-income countries. (A list of testing to screen for colorectal cancer, Whynes and Nottingham the 39 studies reviewed is available from the authors.) The Faecal Occult Blood Screening Trial (2004) report favorable detailed results of each study are somewhat difficult to com- cost-effectiveness ratios ranging from US$2,500 to US$4,000 pare. The types of costs included in each study varied substan- per YLS. Studies of the cost-effectiveness of colorectal cancer tially (patient time costs and programmatic costs often were screening in developed countries consistently conclude that omitted), studies frequently did not discount costs and benefits such screening is cost-effective, but they do not totally agree on or did not note the discount rate used, and sensitivity analyses the relative rankings of different colorectal screening strategies were not conducted consistently on all relevant variables. (Pignone and others 2002). Despite those limitations, several themes emerge. The incremental cost-effectiveness of screening in the general pop- Screening for Cervical Cancer. Cytology-based screening ulation becomes increasingly less favorable as programs are using the Papanicolaou smear has been the main screening intensified by shortening the screening interval. For example, method used for the secondary prevention of cervical cancer Mandelblatt and others (2002) reported that for conventional worldwide. In many low-income countries, however, cytology cytology and HPV testing, compared with cytology alone, the screening has proved difficult to sustain because of its reliance incremental cost was more than US$300,000 when conducted on highly trained cytotechnologists; good-quality laboratories; annually compared to US$15,400 per YLS when conducted and infrastructure to support up to three visits for screening, every 10 years. Maxwell and others (2002) reported that liquid- evaluation of cytologic abnormalities with colposcopy, and based cytology and HPV testing for equivocal results cost treatment (Sankaranarayanan, Budukh, and Rajkumar 2001). US$231,300 per YLS if conducted annually incremental to Two alternative screening approaches replace the Pap smear 14,300 per YLS if conducted every three years. Kim, Wright, Health Service Interventions for Cancer Control in Developing Countries | 575 and Goldie (2002) reported similar results for this same strate- tive but more costly than VIA. They found that, in comparison gy (US$20,300 per YLS conducted every five years, US$59,600 with no screening, a single lifetime VIA screen at age 35, cou- per YLS every three years, and US$174,200 every two years). pled with immediate treatment of women with positive results, The analyses, which included strategies that employed both fre- resulted in a cost saving of US$39 per YLS as compared with a quent screening and screening tests with higher sensitivity, two-visit HPV, although programmatic costs were not consid- often found the cost-effectiveness of frequent screening to be ered. Using sensitivity analysis, the authors find the choice even less attractive. For example, Goldie, Kim, and Wright between using HPV DNA testing or VIA depended on the rel- (2004) reported annual screening with combined cytology and ative costs and sensitivity of the two tests and on the percentage HPV DNA testing in women over age 30 exceeded US$1 mil- of women lost to follow-up between the first and second visit. lion per YLS compared to every two years. Although many Mandelblatt, Lawrence, Gaffikin, and others (2002) used a analyses find that extending the age range to the very young, simulation model to compare seven cervical cancer screening the very old, or both can be less cost-effective, for certain techniques in Thailand. Comparing each strategy to the next women in high-risk groups, including older, uninsured women less expensive alternative, the authors found that VIA per- who have never been screened, screening for cervical cancer at formed at five-year intervals in women age 35 to 55, followed older ages can be cost-effective. by immediate treatment of abnormalities, was the least expen- The analyses conducted in low-income countries focused sive option and saved the greatest number of lives. on assessing the cost-effectiveness of an expanded set of strate- The Alliance for Cervical Cancer Prevention used primary gies that included alternatives to conventional cytology. In data from studies conducted in India, Kenya, Peru, South addition, these analyses--unlike those in developed regions-- Africa, and Thailand to develop a series of standardized, often raised issues of feasibility, affordability, cultural context, country-specific cost-effectiveness analyses. The costs and ben- accessibility, and equity. efits associated with alternative strategies to reduce cervical In one of the earliest stochastic modeling evaluations of cer- cancer mortality were estimated for these five countries with vical cancer screening programs in developing regions, different epidemiological profiles by integrating country- Sherlaw-Johnson, Gallivan, and Jenkins (1997) explored the specific data from each site and using a standardized set of effectiveness of cytology and HPV testing in the context of assumptions agreed on by an expert panel with experience in infrequent screening. They reported that the most efficient use each country (Goldie, Gaffikin, and others 2004). In all five of resources would be to concentrate cervical cancer screening countries, lifetime cancer risk was reduced by approximately 25 efforts on women age 30 to 59 at least once per lifetime, to 35 percent with a single lifetime screen using either one-visit because such blanket screening would reduce the incidence of VIA or two-visit HPV DNA testing targeted at women age 35 invasive cervical cancer by up to 30 percent. to 40. Risk was reduced by more than 50 percent if screening In an analysis focused on cervical cancer control in was performed two or three times per lifetime. Although the Vietnam, Suba and others (2001) reported that, because of the cost of screening differed considerably between the countries, low direct medical costs associated with Vietnam's cervical strategies were identified that, when performed two or three cytology program, such a program appeared to be attractive for times per lifetime, would be considered extremely cost-effective that country. They found that total costs to establish a nation- depending on the individual country's per capita gross domes- wide Pap screening program based on five-year intervals aver- tic product. aged less than US$148,000 annually during the 10 years the We conducted an exploratory analysis to evaluate the poten- authors assumed would be necessary to develop the program. tial cost-effectiveness of cervical cancer screening strategies in Assuming 70 percent participation in the program, the authors Brazil, Madagascar, and Zimbabwe using computer-based sim- found the cost-effectiveness ratio for cervical cytology screen- ulation models calibrated to age-specific cervical cancer inci- ing, compared with no screening, to be US$725 per discounted dence and mortality in each country, along with published YLS. data. We evaluated once-in-a-lifetime screening between age 35 Goldie and others (2001) assessed the cost-effectiveness of and 40 with (a) one-visit VIA, with screening and treatment several cervical cancer screening strategies in previously conducted during the same visit; (b) two-visit HPV DNA unscreened 30-year-old South African women. Screening tests screening, with HPV DNA testing during the first visit followed included VIA, cytology, and HPV DNA testing. Strategies dif- by treatment of screen-positive women during the second visit; fered by the number of clinic visits required, frequency of and (c) three-visit cervical cytology screening, with a cytology screening and individual's age at the time of screening, and sample obtained during the first visit, colposcopy for screen- response to a positive test result. The authors found that when positive women conducted during the second visit, and treat- all strategies were considered to be equally available and were ment provided during the third visit. We assumed that for the compared incrementally, HPV DNA testing was always more one- and two-visit strategies, women who screened positive effective and less costly than cytology and generally more effec- and were eligible for cryotherapy were treated immediately, but 576 | Disease Control Priorities in Developing Countries | Martin L. Brown, Sue J. Goldie, Gerrit Draisma, and others Table 29.3 Economic Outcomes of Once-in-a-Lifetime Cervical Cancer Screening Programs, Brazil, Madagascar, and Zimbabwe Category No screening One-visit VIA Two-visit HPV DNA testing Three-visit cytology Brazil Lifetime cost (international $) 68.41 75.08 77.43 121.12 Cost-effectiveness ratio (international $/YLS)* n.a. 113 155 1430 Cost-effectiveness ratio (US$/YLS) n.a. 54 118 572 Life expectancy gain per 1 million screened n.a. 59,100 58,200 36,900 Number of deaths averted per 1 million screened n.a. 10,399 10,235 6,411 Number of DALYs averted per 1 million screened n.a. 56,646 55,751 35,174 Madagascar Lifetime cost (international $) 25.22 32.98 40.41 51.91 Cost-effectiveness ratio (international $/YLS)* n.a. 167 332 921 Cost-effectiveness ratio (US$/YLS) n.a. 52 162 368 Life expectancy gain per 1 million women screened n.a. 46,500 45,800 29,000 Number of deaths averted per 1 million women screened n.a. 8,815 8,676 5,438 Number of DALYs averted per 1 million women screened n.a. 42,424 41,754 26,352 Zimbabwe Lifetime cost (international $) 31.10 39.69 44.81 61.93 Cost-effectiveness ratio (international $/YLS)* n.a. 140 227 803 Cost-effectiveness ratio (US$/YLS) n.a. 42 114 321 Life expectancy gain per 1 million screened n.a. 61,300 60,400 38,400 Number of deaths averted per 1 million screened n.a. 10,412 10,248 6,419 Number of DALYs averted per 1 million screened n.a. 53,770 52,921 33,472 Source: Authors' calculations. n.a. not applicable. *Converted from national currency, using purchasing power parity (PPP) exchange rates. those ineligible for cryotherapy were referred for colposcopy VIA, followed by immediate treatment, or HPV DNA testing or and diagnostic workup. cervical cytology followed by treatment at a second visit. Note We estimated direct medical costs using data from the liter- that all screening tests may not be equally available in low- ature and unit costs provided by the volume editors and WHO. resource settings and that certain screening tests may be selected All costs for the analysis are presented in 2000 dollars. We esti- because of cultural preferences or for programmatic reasons. mated patients' time costs and direct nonmedical costs using Implementing cervical cancer screening programs on the basis of our own previous work and wage estimates based on World VIA, HPV DNA testing, or cytology requires different types of Bank data on per capita gross national income (WHO n.d.) and resources,and the relative availability of these resources in differ- wage estimate regressions developed by the U.S. Department of ent settings will affect the choice of strategy. Commerce. Table 29.3 presents the results of our analysis. Lifetime costs per individual screened are given in interna- Screening for Breast Cancer. Methods for early detection of tional dollars. Cost-effectiveness ratios are provided in U.S. dol- breast cancer include screening by mammography, clinical lars as well as international dollars to facilitate comparison to breast examination (CBE), and breast self-examination. other studies.The available data show that cervical cancer screen- Screening by mammography, CBE, or both may decrease breast ing conducted once, twice, or three times in a lifetime can have a cancer mortality, but uncertainty about the magnitude of the significant effect on the lifetime risk of cervical cancer compared benefit remains because the quality of the evidence varies and with no screening. For countries with limited resources, screen- results are inconsistent (Humphrey and others 2002). Recent ing efforts should target women age 35 or older; strategies should controlled studies of organized breast self-examination pro- focus on screening all women at least once in their lifetime before grams indicate that this approach is not effective (Semiglazov increasing the frequency of screening; and countries should con- and others 1999; Thomas and others 2002). sider alternative approaches to the conventional three-visit cer- A randomized controlled trial of CBE screening for breast vical cytology screening techniques--for example, single-visit cancer began in Manila in 1995, but the intervention was Health Service Interventions for Cancer Control in Developing Countries | 577 discontinued after the first round because compliance with in 2000 (United Nations Population Division 2003). We referral among women who were found to have a breast lump assumed that the screening program would last for 25 years was extremely low (21 percent) and attempts to improve com- and would have an attendance rate of 100 percent. We pliance failed. Analysis of the incidence of cancer cases in 1999 expressed the effects of screening as the reduction in the num- shows that the screening intervention succeeded in detecting ber of deaths caused by breast cancer and the number of life more localized breast tumors, but the low compliance with years gained because of the screening program. Costs and referral and low yield of early cancers meant that the early effects were discounted at a rate of 3 percent. detection program could not succeed in preventing deaths We estimated the model's parameters using data from Dutch from breast cancer (International Agency for Research on screening projects (Collette and others 1992; Vervoort and oth- Cancer n.d.). ers 2004). We used trial results to estimate the effectiveness of Numerous cost-effectiveness studies of breast cancer mammography in reducing breast cancer mortality (de Koning screening programs have been conducted in developed coun- and others 1995). We based sensitivity estimates of CBE on data tries (Vainio and Bianchini 2002a). Most cost-effectiveness from Rijnsburger and others (2004) and based alternative studies of mammography screening in Europe yield cost- (lower) estimates on data from Bobo, Lee, and Thames (2000) effectiveness ratios in the range of US$3,000 to US$10,000 per and Rijnsburger and others (forthcoming). We calibrated the YLS, whereas those in the United States yield far less favorable model so that it would correctly predict the age-specific inci- cost-effectiveness ratios, ranging from US$20,000 to dence and mortality of breast cancer in India (Ferlay and US$100,000 per YLS (table 29.4). others 2001) and its stage distribution at clinical diagnosis To investigate the potential cost-effectiveness of CBE and (Sankaranarayanan, Black, and Parkin 1998). Details of these mammography for India, we used a microsimulation model of methods are available elsewhere (Lamberts and others 2004). breast cancer screening (van Oortmarssen and others 1990). We calculated total costs by comparing the differential costs The model simulates individual life histories of disease states, of breast cancer screening, diagnosis, initial therapy, adjuvant and consequences of screening are calculated by comparing the therapy, follow-up, and advanced disease in the case of screen- histories with and without screening intervention for each ing versus no screening. We calculated component costs by individual. For our purposes, we assumed a population of 1 mil- multiplying the estimated resource use by the estimated costs lion Indian women with the age distribution of the country per unit for each health care input. Reliable cost data for India were limited, so we extrapolated estimates from Dutch unit costs (Mulligan and others 2003). For the analysis discussed Table 29.4 Estimates of the Cost-Effectiveness of Breast above, we calculated costs based on a market-basket approach. Cancer Screening Every Two Years for Women in Selected Developed Countries The overall incidence of breast cancer is lower in India than in Western countries. The relationship between the incidence Age of Cost- of breast cancer and age also differs: in Western countries, the women being effectiveness incidence of breast cancer increases with age, whereas in India, Country screened (years) ratio (US$/YLS)a it decreases with age, beginning at age 50. Investigators have Australia (de Koning 2000) 50­69 7,680 generally attributed this finding to a cohort effect: breast can- France (de Koning 2000) 50­69 4,580 cer is more common among younger cohorts than older Germany (de Koning 2000) 50­69 8,880 cohorts. The stage at which breast cancer is diagnosed is much Netherlands (de Koning 2000) 50­69 3,140 less favorable in India than in Western countries. Norway (Norum 1999) 50­69 14,790 Table 29.5 presents the results of our exploratory cost- Spain (de Koning 2000) 50­69 6,590 effectiveness analysis of various breast cancer screening pro- grams involving CBE or mammography for a population of Spain, Catalonia (de Koning 2000) 50­69 4,400 1 million women in India. As the table shows, biennial CBE Spain, Navarra (de Koning 2000) 45­65 2,450 from age 40 to 60 costs US$2.6 million, averts 358 breast can- United Kingdom (de Koning 2000) 50­69 2,680 cer deaths, prevents the loss of 4,896 life years, and has a cost- United Kingdom (northwest) 50­64 3,650 effectiveness ratio of US$522 per YLS in comparison with no (de Koning 2000) screening. Biennial CBE from age 50 to 70 is less favorable in United States (M. Brown and 50­69 34,600 terms of cost-effectiveness: US$582 per YLS. Fintor 1993) The cost-effectiveness ratios for biennial mammography United States (Simpson and 50­64 20,611 Snyder 1991) screening are not as favorable as those for biennial CBE screen- ing. Annual CBE screening results in almost the same number Source: M. Brown and Fintor 1993; de Koning 2000; Norum 1999; Simpson and Snyder 1991. of life years saved as biennial mammography screening at Note: The discount rate used was 5 percent. a. Converted from euros to U.S. dollars, using the exchange rate 1 US$0.925. 36 percent of the cost. 578 | Disease Control Priorities in Developing Countries | Martin L. Brown, Sue J. Goldie, Gerrit Draisma, and others Table 29.5 Cost-Effectiveness Analysis of Various Breast Cancer Screening Programs Involving Either CBE or Mammography for a Population of 1 Million Women, Compared with No Screening, India Base model: CBE once Biennial Biennial One lifetime biennial CBE, Annual CBE, Biennial CBE, every 5 years, mammography, mammography, mammogram, Category ages 40­60 ages 40­60 ages 50­70 ages 40­60 ages 40­60 ages 50­70 age 50 Effectiveness Number of screening 2,319,839 4,426,854 1,620,568 1,056,544 2,318,641 1,619,051 212,008 tests performed Number of cancers 1,689 2,330 1,683 938 2,561 2,649 465 detected by screening Number of deaths averted 358 528 313 184 599 557 105 Number of life years saved 4,896 7,242 3,464 2,462 7,955 6,180 1,422 Percentage reduction in 7.8 11.4 6.8 4.0 13.0 12.1 2.3 mortality Number of screening tests 6,473 8,385 5,170 5,730 3,868 2,909 2,028 per death averted Number of screening tests 474 611 468 429 291 262 149 per life year saved Number of screening tests 1,373 1,900 963 1,127 906 611 456 per cancer detected Cost-effectiveness Differential costs (2001 US$) 2,553,425 5,230,303 2,017,186 1,108,883 14,681,387 10,559,356 1,282,024 Cost per death prevented 7,125 9,907 6,435 6,014 24,493 18,970 12,262 (2001 US$) Cost per life year saved 522 722 582 450 1,846 1,709 902 (2001 US$) Source: Authors' calculations. Note: The discount rate used was 3 percent. Table 29.6 shows the results of our sensitivity analysis for the Cancer Treatment and Palliative Care exploratory cost-effectiveness analysis of breast cancer screening Barnum and Greenberg (1993) used an indirect approach to in India. Cost-effectiveness ratios are lower when the incidence of estimate the cost-effectiveness of initial cancer treatment in devel- cancer is higher, as in Bombay. Cost-effectiveness ratios are 32 oping countries. They assumed that they could estimate the effec- and 16 percent higher, respectively, with a lower sensitivity of tiveness of initial cancer treatment by comparing cancer survival CBE and when the averted costs of palliative treatment are not in the United States for the period 1975­80 with the period included. Using alternative approaches to estimate screening pro- 1940­45.The logic of such a comparison is that major advances in gram costs has a major effect, resulting in cost-effectiveness esti- cancer diagnosis, surgery, radiation, and chemotherapy occurred mates 6 to 11 times higher than the base case analysis. This result during the intervening period, and thus survival in the 1940­45 underlines the need for economic studies that can obtain reliable period could be equated to outcomes expected to result from no data from primary sources on the true resource costs of cancer treatment or ineffective treatment. Barnum and Greenberg's control interventions in developing countries. With data from results indicated a cost-effectiveness ratio of the following: such studies, researchers would not have to continue to rely on extrapolating cost estimates from data in developed countries. · US$1,300 to US$6,200 per YLS for initial treatment of the These results depend critically on assumptions about the more treatable cancers, that is, cervical, breast, oral cavity, efficacy of CBE, for which the evidence is limited, highlighting and colorectal cancer the need for controlled studies of CBE in developing countries. · US$53,000 to US$163,000 per YLS for initial treatment of Our estimates indicate that the cost-effectiveness of screening the less treatable cancers, that is, liver, lung, stomach, and mammography in India compares favorably, in absolute terms, esophageal cancer. with breast cancer screening in developed countries. Nevertheless, screening mammography for breast cancer is The following subsections review cost-effectiveness studies likely to be less cost-effective in a country such as India than performed on selected adjuvant or palliative cancer treatments is screening for cervical cancer. that have been studied extensively in controlled clinical trials. Health Service Interventions for Cancer Control in Developing Countries | 579 Table 29.6 Sensitivity Analysis for Changes in Breast Cancer Incidence and Attendance Rate, CBE Sensitivity, No Palliative Treatment, and Alternative Cost Estimates for a Population of 1 Million Women, Compared with No Screening, India Base model: Incidence Alternative cost estimation biennial CBE, of breast cancer, Attendance CBE No palliative Category ages 40­60 Bombay rate, 70% sensitivitya treatment Method 1b Method 2c Effectiveness Number of screening 2,319,839 2,319,991 1,624,401 2,320,051 2,319,839 2,319,839 2,319,839 tests performed Number of cancers 1,689 1,921 1,229 1,370 1,689 1,689 1,689 detected by screening Number of deaths averted 358 405 255 286 358 358 358 Number of life years saved 4,896 5,400 3,483 3,893 4,896 4,896 4,896 Percentage reduction in mortality 7.8 6.9 5.5 6.2 7.8 7.8 7.8 Number of screening tests 6,473 5,727 6,358 8,119 6,473 6,473 6,473 per death averted Number of screening tests per 474 430 466 596 474 474 474 life year saved Number of screening tests per 1,373 1,208 1,322 1,693 1,373 1,373 1,373 cancer detected Cost-effectiveness Differential costs (2001 US$) 2,553,425 2,505,274 1,798,662 2,684,628 2,983,754 28,814,056 16,532,879 Cost per death prevented (2001 US$) 7,125 6,184 7,040 9,395 8,325 80,396 46,130 Cost per life year saved (2001 US$) 522 464 516 690 609 5,885 3,377 Source: Authors' calculations. Note: The discount rate used was 3 percent. a. From Rijnsburger and others forthcoming. b. Costs using 2001 prices in the Netherlands. c. Costs using 2001 prices in the Netherlands multiplied by the ratio of gross domestic product shares spent on health care in India and the Netherlands, respectively. Table 29.7 Cost-Effectiveness of Selected Breast Cancer Treatments for a Hypothetical Cohort of 45-Year-Old Premenopausal Women with Early-Stage Breast Cancer, United States (cost in 2000 US$/quality-adjusted life year) Node-negative, Node-negative, Node-positive, Node-positive, estrogen receptor­ estrogen receptor­ estrogen receptor­ estrogen receptor­ Treatment positive negative positive negative Tamoxifen 17,400 326,800 6,600 88,300 Chemotherapy 17,400 7,600 14,000 7,500 Tamoxifen and chemotherapy 50,400 131,600 22,600 123,200 Source: T. Smith and Hillner 1993. Breast Cancer Treatment Interventions. The following para- cost-effectiveness of various breast cancer treatments. Smith graphs review studies of the cost-effectiveness of adjuvant sys- and Hillner's cost-effectiveness estimates for single-modality temic therapy for early-stage breast cancer and of radiation systemic adjuvant therapy for breast cancer are about the same therapy following mastectomy and chemotherapy to treat order of magnitude as Barnum and Greenberg's (1993) esti- node-positive breast cancer in premenopausal women. mates of cost-effectiveness for initial therapy of breast cancer T. Smith and Hillner (2000), relying on results from the (about US$7,300 per YLS in 2000 dollars). Other studies Early Breast Cancer Trialists' Collaborative Group (EBCTCG (Malin and others 2002; Norum 2000) have yielded cost- 1998), modeled the natural history of breast cancer in pre- effectiveness estimates for chemotherapy and hormonal therapy menopausal 45-year-old women in the United States who were two to three times more favorable than Smith and Hillner's diagnosed with early-stage breast cancer and treated with estimates. The more favorable estimates are probably the result tamoxifen, chemotherapy, or both. Table 29.7 summarizes the of the investigators' use of a discount rate of 3 percent instead 580 | Disease Control Priorities in Developing Countries | Martin L. Brown, Sue J. Goldie, Gerrit Draisma, and others of 5 percent and their assumption that the benefits of treat- chemoradiation--is more effective than standard therapy ment continue over a longer period of time. using radiation alone in the treatment of advanced cervical Two U.S. studies (Lee and others 2002; Marks and others cancer (Rose and Lappas 2000). Using an economic model, 1999) have estimated the cost-effectiveness ratio for radiation Rose and Lappas apply unit costs to resource allocation data therapy following mastectomy and chemotherapy for node- derived from the cisplatin-based chemoradiation arms of the positive breast cancer in premenopausal women to be in the five randomized trials and examine the benefits in terms of range of US$22,600 to US$43,000 per quality-adjusted life year increased median survival time. Costs per YLS for cisplatin- (adjusted to 2000 U.S. dollars, with a discount rate of 3 per- based chemoradiation regimens varied from US$2,384 to cent). Results were sensitive to treatment costs, survival benefit, US$28,770 on the basis of published survival and from US$308 and patient time costs. to US$3,712 on the basis of estimated survival. Although The clinical trials of postmastectomy radiation on which the chemoradiation for advanced cervical cancer would probably two U.S. studies are based compared radiation following sur- be considered cost-effective in most developed countries, gery plus chemotherapy with surgery plus chemotherapy analyses that take local treatment settings into account are alone. Love and others (2003), however, offer observational needed to determine if this result also holds for developing evidence that radiation treatment may also extend survival countries. for Chinese and Vietnamese women when administered to patients with one to three positive nodes following mastectomy Palliative Care Interventions. The most basic approach to pal- alone or mastectomy combined with oophorectomy and liative care for terminally ill cancer patients, especially in low- tamoxifen. If these benefits were confirmed, postmastectomy resource settings, involves using inexpensive oral analgesics, radiation might be cost-effective in developing countries, ranging from aspirin to opiates, depending on individual where the cost of radiation treatment is lower than in most patients' needs. Unfortunately, sufficient supplies of opioid developed countries. drugs for use in palliative care are often not available in develop- ing countries because of regulatory or pricing obstacles, igno- Colorectal Cancer Treatment Interventions. As concerns col- rance, or false beliefs (for more information see http://www. orectal cancer, investigators have carried out cost-effectiveness medsch.wisc.edu/painpolicy/index.htm and chapter 52). studies on surgical techniques, adjuvant treatment, follow-up Appropriate palliative care for cancer patients may involve monitoring for recurrence, and treatment of advanced disease a variety of other treatment modalities, including antiemetic (van den Hout and others 2002). Brown, Nayfield, and Shibley drugs to relieve the side effects of chemotherapy, radiation to (1994) estimate that the cost-effectiveness of adjuvant effect temporary tumor regression, and physical therapy to chemotherapy for stage three colon cancer ranges from alleviate disability related to lymphedema following breast can- US$3,000 to US$7,000 per YLS (adjusted to 2000 U.S. dollars, cer surgery. Berthelot and others' (2000) study combines infor- with a discount rate of 6 percent). R. Smith and others' (1993) mation from several clinical trials and Canadian treatment cost study conducted in the Australian health care setting obtains information to perform cost-effectiveness analyses of different similar results in terms of cost per YLS but yields substantially ambulatory chemotherapy regimens used for patients with higher costs per quality-adjusted life year. metastatic non-small-cell lung cancer to palliate symptoms and Dahlberg and others' (2002) cost-effectiveness study, which modestly improve survival. They report that vinblastine plus relies on cost and clinical outcome data from the Swedish cisplatin resulted in both better survival and lower health care Rectal Cancer Trial (1997), demonstrates that rectal cancer expenditures than best supportive care because it resulted in patients receiving preoperative radiation therapy had improved fewer episodes of rehospitalization. cancer-specific and overall survival rates, as well as reduced Van den Hout and others' (2003) study examines the cost- local rectal cancer recurrence rates. They estimate the overall effectiveness of single-fraction versus multiple-fraction radio- cost-effectiveness of preoperative radiation therapy for rectal therapy for palliative treatment of cancer patients with painful cancer patients to be US$3,654 per YLS (in 2001 U.S. dollars, bone metastases. They find that overall medical and social costs using a discount rate of 3 percent). In a sensitivity analysis, for single-fraction radiotherapy for palliative therapy-- which varied the rates of local rectal cancer recurrence and the US$1,144 per patient in medical costs and US$1,753 per survival advantage with and without radiation treatment, cost- patient in total social costs--were lower than comparable costs effectiveness ratios for preoperative radiation therapy for for multiple-fraction radiotherapy, despite the higher rate of patients with rectal cancer ranged from US$908 to US$15,228 retreatment associated with single-fraction radiotherapy. per YLS. Whether those results are directly applicable to radiation treatment in developing countries, where single-fraction radia- Cervical Cancer Treatment Interventions. Five recent phase 3 tion treatment may be relatively less effective, is unknown. trials indicate that a new alternative therapy--cisplatin-based Nonetheless, the results strongly suggest that single-fraction Health Service Interventions for Cancer Control in Developing Countries | 581 radiotherapy may be an acceptable, if not preferred, choice of countries. As the figure shows, the incidence of cervical cancer palliative treatment in settings where resources for radiation in developing countries is relatively high in comparison with treatment are relatively scarce and the need for palliative treat- the incidence of these cancers in developed countries, whereas ment is relatively high. the incidence of breast cancer is relatively low in developing countries compared with that in developed countries. Given the relatively high incidence of cervical cancer in developing countries, interventions for cervical cancer prevention and APPLICABILITY OF COST-EFFECTIVENESS screening are likely to be more cost-effective in developing STUDIES FROM DEVELOPED TO countries rather than developed countries, compared with DEVELOPING COUNTRIES interventions for breast cancer, all else being equal. Many of the cost-effectiveness studies of cancer control inter- ventions (prevention, screening, and treatment) have been per- Factors Affecting the Applicability of Cost-Effectiveness formed in the context of high-income, developed countries; Studies of Treatment thus, the question arises whether such studies are applicable to Many of the treatments for breast, colorectal, and lung cancer health care delivery settings in developing countries. No simple that have been shown to be efficacious in controlled clinical tri- rule is available to indicate how the results of cost-effectiveness als have been estimated to have cost-effectiveness ratios in the studies in developed countries might translate to health care range of a few thousand U.S. dollars to a few tens of thousands delivery settings in developing countries, but disease incidence and time horizon are major pertinent considerations in rela- tion to cancer prevention and screening interventions. In rela- tion to cancer treatment, other considerations have to be taken a. Incidence of cervix uteri into account. Rate (per 100,000) 80 Factors Affecting the Applicability of Cost-Effectiveness Studies of Prevention and Screening 60 Cost-effectiveness analyses of cancer prevention and screening 40 interventions are complex.Several parameters have a large influ- ence on the results of these studies, including the following: 20 · age-specific cancer incidence · all-cause life expectancy and temporal trends of major 0 0­ 15­ 45­ 55­ 65­ epidemics age · population age structure · availability, effectiveness, and costs of cancer treatment b. Incidence of breast cancer · health system costs of the prevention or screening Rate (per 100,000) intervention. 250 As illustrated by the several examples described in this 200 chapter, those parameters are likely to vary widely between 150 developed and developing countries. For example, age-specific cancer incidence in the absence of 100 a preventive or screening intervention can have a major influ- ence on the potential cost-effectiveness of a cancer prevention 50 or screening intervention. Generally, the higher the back- 0 ground incidence of the cancer, the more cost-effective the can- 0­ 15­ 45­ 55­ 65­ age cer prevention or screening intervention will be. For that rea- son, the relative cancer incidence patterns in developed and More developed countries Less developed countries developing countries for the cancer screening interventions Source: Ferlay and others 2001. described earlier need to be considered. Figure 29.3 shows age-specific cancer incidence patterns for Figure 29.3 Age-Specific Incidence of Cervical and Breast Cancer, cervical and breast cancer for developed and developing Developed and Developing Countries, 2000 582 | Disease Control Priorities in Developing Countries | Martin L. Brown, Sue J. Goldie, Gerrit Draisma, and others of U.S. dollars per YLS. This range is considered quite favorable single patient could approach US$300,000. Those estimates do in developed countries but might be viewed as less favorable in not consider the additional costs of chemotherapy preparation, low- and middle-income countries that face stringent con- administration, and supervision and supportive care (Schrag straints on health care resources. Disease incidence and time 2004). The situation is similar for other common cancers. horizon do not loom as major considerations in the case of the Clearly, low- and middle-income countries cannot afford to cost-effectiveness of cancer treatment, because the cost of make the newest cancer drugs widely available to cancer treatment applies only to those individuals already diagnosed patients; however, this example illustrates the need for periodic with cancer and considered eligible for a specific treatment, not updating of available chemotherapy options along with evalu- to a broader population considered to be at risk for developing ations of the incremental costs and benefits associated with cancer. them. Thus, in low-income, low-cost countries with high mortality rates, because of the lack of primary treatment, the provision of basic cancer treatment may be a cost-effective first RESEARCH AGENDA step toward cancer control, especially for highly treatable can- cers with relatively low incidence in developing countries. For Knowledge about the feasibility, effectiveness, and cost- example, using a generalized cost-effectiveness approach, effectiveness of cancer control interventions by health services Ginsberg and others (2004) conclude that the provision of in developing countries is extremely limited, partly because of basic treatment for colorectal cancer in low-income African the relative paucity of active research in this area. Work in the countries is likely to be a cost-effective first step toward cancer area of descriptive epidemiology, especially work based on control. cancer registry data, dominates the research literature on can- Nevertheless, issues of economies of scale and scope may be cer in developing countries. A second body of literature con- associated with fixed investments in specialized medical equip- sists of comparative epidemiology and case-control studies ment and skilled human capital. The centralization and region- designed to assess the importance of various risk factors for alization of cancer treatment may be associated with a higher cancer. technical quality of care and might also be associated with the Although information from such studies is an essential first need to use these resources at economically efficient levels. step for characterizing the nature and extent of the cancer bur- Some cost elements, such as local labor and the availability of den and for monitoring the ultimate effect of cancer control generic drugs since initial clinical trials were conducted, will interventions, it does not provide a sufficient knowledge base clearly be lower in the contemporary setting of developing for designing and implementing cancer control programs. For countries than in many of the cost-effectiveness studies progress to be made for developing countries, much more work reviewed earlier. is needed in the following areas: Finally, developments in cancer treatment, especially in rela- tion to chemotherapy, are extremely dynamic. For example, the · Clinical evaluation studies of cancer control interventions in 1999 WHO list of essential drugs for cancer therapy (Sikora developing countries. Clinical evaluation studies of preven- and others 1999; WHO 2003a), includes 5-fluorouracil as a pri- tive, screening, and treatment interventions that are specifi- ority one (essential) drug and irinotecan as a priority three cally tailored to the needs and conditions of developing (palliative benefit only) drug for the treatment of colorectal countries would be useful, including controlled clinical tri- cancer. Just five years later, in many developed countries the als where possible. following drugs, in addition to irinotecan, have been added · Health services research in developing countries. Health serv- to the basic regimen of 5-fluorouracil plus leucovorin for ices research designed to characterize the amount, distribu- the treatment of colorectal cancer: oxaliplatin, bevacizumab, tion, and organizational structure of health sector resources and cetuximab. Whereas 5-fluorouracil-based treatment of in developing countries would be helpful, along with metastatic colorectal cancer increased median survival from 8 research to fill the gaps between current resource endow- to 12 months, the newer drugs increase median survival to 21 ments and the amount of funding that would be needed to months or more, at a significantly increased economic cost. implement the minimally acceptable level of effective cancer In the United States, the drug cost of 5-fluorouracil-based control. In developing countries, shortages of the equipment therapy ranges from US$63 to US$263 for the initial eight and personnel needed to administer radiotherapy for can- weeks of therapy. Adding irinotecan or oxaliplatin increases the cer, for example, have been well documented (Levin, drug cost to about US$10,000, and adding bevacizumab or Meghzifene, and Tatsuzaki 2001). However, no systematic cetuximab adds another US$20,000 to US$30,000 to the cost of analyses are available outside developed countries (Owen, initial treatment. If the latter drugs are used over the longer Coia, and Hanks 1997) that project radiotherapy resource term as envisioned, the average cost of supplying the drugs to a needs in terms of clinically effective applications of Health Service Interventions for Cancer Control in Developing Countries | 583 radiotherapy, both by cancer site and by the known effec- A useful way to draw inferences about the relative cross- tiveness of radiotherapy for primary treatment, adjuvant country affordability of interventions is to translate cost- therapy, and palliative care. Similarly, even though effectiveness ratios into percentage of per capita gross national researchers have carried out patterns of care studies that product (GNP) per YLS (WHO 2001a). Our preliminary characterize the dissemination of radiation, chemotherapy, analysis of breast cancer screening in India, for example, sug- and hormonal therapy in many developed countries, com- gesting an absolute cost-effectiveness level for screening mam- parable information for developing countries is generally mography of about US$2,000 per YLS, compared with about unavailable. Health services research studies could also con- US$3,000 per YLS in the Netherlands. At about 10 percent per tribute important information about the current structure capita GNP per YLS, screening mammography might be con- and organization of primary, secondary, and tertiary care in sidered to be extremely cost-effective for the Netherlands. In specific developing countries, with the ultimate aim of mod- India, however, we found a CE estimate equal to 400 percent eling and implementing cancer control delivery systems that per capita GNP per YLS suggesting that national policy makers either are integrated with or supplement existing care deliv- would be much less likely to consider screening mammography ery systems. Studies of this type are needed to ensure that as a viable intervention given India's health care budget con- there is a balance, for example, between resources devoted to straints. However, they might well consider a CBE breast can- screening and those devoted to diagnostic follow-up and cer screening program, at about 200 percent per capita GNP treatment. The disappointing performance of cervical can- per YLS in India, to be moderately affordable if the program cer screening programs in many developing countries has were definitively established to be effective. been due in part to the lack of effective diagnostic follow-up For middle-income developing countries that have cancer and treatment following screening. incidence rates similar to those in high-income developed · Country-specific economic evaluation studies. Country- countries, the results of cost-effectiveness analyses from the specific studies need to be done that assess resource require- developed countries may be more relevant, although further ments, economic costs, effectiveness, and ultimately cost- analysis clearly is needed. The case study of cervical cancer con- effectiveness of cancer control interventions adapted or trol that was cited earlier suggests that for low-income coun- tailored to the needs and requirements of low- and middle- tries tailored cancer control interventions may need to be income settings. Heuristic extrapolation is a first analytical developed that would be both cost-effective and affordable. step in this direction, but such studies can indicate only However, that suggestion does not imply that low-tech whether more direct and realistic studies are needed. approaches should be uncritically embraced and assumed to be · Studies of innovative health care information and communica- cost-effective. Until recently, education campaigns to promote tions technology. More research is needed to determine if tech- breast self-examination were widely advocated as the low-tech nological advances, such as computerized image reading or alternative to screening mammography for breast cancer con- long-distance consultation by oncology specialists, facilitated trol in low-income countries; however, the best current evi- by telemedicine communications technology, might alter dence now indicates that such campaigns have no effect on the cost-effectiveness equation by raising quality, by lowering breast cancer mortality (Semiglazov and others 1999; Thomas costs, or both. For remote localities or small, low-income and others 2002). developing countries, training and employing local expertise In cancer treatment interventions, the cost-effectiveness of or advanced equipment for every aspect of cancer control initial surgical treatment for treatable cancers, such as breast, may not be necessary if advanced communication and cervical, and colorectal cancer, may be in the relatively favor- information technology could be used to facilitate virtual able range of a few to several thousand dollars per YLS, which collaboration. indicates that such interventions are likely to be cost-effective for middle-income countries and are possibly cost-effective for low-income countries. Although cost-effectiveness ratios for CONCLUSIONS some of the approaches to adjuvant therapy that use conven- tional radiation and drugs also fall within this relatively favor- Our ability to draw any conclusions about the cost-effectiveness able range, others are in the range of tens of thousands of of cancer control interventions for low- and middle-income dollars for each YLS. Thus, these forms of treatment would developing countries is limited, because most cost-effectiveness likely be considered potentially cost-effective and affordable in studies in this area have been conducted in high-income, devel- middle-income countries but not in low-income countries; oped countries. Cancer control interventions that appear to be however, more detailed examinations of specific cost condi- cost-effective in high-income countries may not be cost- tions and available resource endowments for the delivery of effective in low-income countries, even when the lower cost of cancer treatment services are needed to confirm these prelimi- providing health services is taken into account. nary impressions. As with the case of cervical cancer control, 584 | Disease Control Priorities in Developing Countries | Martin L. Brown, Sue J. Goldie, Gerrit Draisma, and others treatment interventions that are tailored to the conditions of Developing countries should consider scaling up their regional low-income countries might be shown to be efficacious and or national programs only after the pilot programs have been more economically attractive than treatment approaches that shown to perform well. are transported directly from developed countries; however, Starting small also might entail applying an initial pilot research in this area is lacking. program to a limited age range that is estimated to yield the most benefits per resource use or to a limited group of high-risk individuals defined by various risk characteristics, such as first- Time Horizon and a Balanced Approach to Cancer degree relatives of people with cancer. Indeed, various versions Control Programs of this approach have characterized the dissemination of many The time horizon for cancer prevention and screening inter- cancer control interventions in developed countries. Organized ventions is highly relevant to policy makers and health system breast cancer screening programs in some European countries, planners, yet reports on the cost-effectiveness of such interven- for example, were first implemented as pilot programs in spe- tions often omit information about time horizons. For exam- cific regions and evaluated against control communities ple, interventions that involve cancer control agents that pre- (Fracheboud and others 2001; Olsson and others 2000; van der vent cancer cases that would have otherwise occurred many Maas 2001), and regional and national programs were initially years after the preventive action, such as HPV vaccination, have limited to the age groups, screening procedures, and screening a long time horizon. Similarly, the favorable cost-effectiveness frequencies estimated to be the most cost-effective. The pro- of preventive screening for stomach cancer is not apparent grams were later extended, in terms of more intensive proce- until four decades following the initiation of the intervention. dures, more frequent screening intervals, and wider age groups, In the case of the 25-year program of CBE in India analyzed after monitoring and analysis of initial program performance earlier, only about 10 percent of the benefits in terms of breast indicated that the incremental cost-effectiveness of these exten- cancer deaths prevented would have been realized after 10 years sions would be favorable (Boer and others 1995; Shapiro and of program operation. Decision makers must understand and others 1998). The United Kingdom has taken a similar approach take these time horizons into account when interpreting and to colorectal cancer screening (Steele and others 2001). acting on cost-effectiveness ratios; however, the long time hori- zon for cancer prevention and screening interventions is, in itself, not an argument against the application of such inter- ACKNOWLEDGMENTS ventions. In some cases, countries that are more recent entrants into the field of cancer control may be able to benefit from the The authors would like to thank Rachel Ballard-Barbash, M.D.; experience of developed countries and from the dynamic tech- Ted Trimble, M.D.; and Stephen Taplin, M.D., of the National nical progress in this area to go directly to new innovations. For Cancer Institute, and Deborah Schrag, M.D., of Memorial example, they might be able to implement HPV testing right Sloan Kettering Cancer Institute for reading and commenting away as the basis for cervical cancer screening, bypassing on early versions of this chapter. We thank Kerry Kemp and cervical cytology. Achieving the optimal temporal balance in Penny Randall-Levy for editorial assistance. The exploratory comprehensive cancer control represents a daunting challenge analysis of breast cancer screening is the joint work of Quirine to planning, evaluation, and implementation. J. Lamberts, M.D., M.Sc.; Arno J. Der Kinderen, M.Sc.; Gerrit Draisma, Ph.D.; and Harry J. de Koning, M.D., of the Department of Public Health, Erasmus University Medical Start Small, Scale Up Smart Center, Rotterdam. Steven Sweet, Jane Kim, and Jeremy Because the current understanding of the effectiveness, optimal Goldhaber-Fiebert of the Harvard Initiative for Global Health resource mix, and cost of many cancer control interventions is made invaluable contributions to the section on cervical cancer. incomplete and uncertain, especially in relation to low- and middle-income countries, developing countries should start small. By starting small, they can gain knowledge from pilot REFERENCES programs that are well documented with regard to organiza- Alberts, D. S., M. E. Martinez, D. J. Roe, J. M. Guillen-Rodriguez, J. R. tional and process factors; that are conducted in controlled set- Marshall, J. B. van Leeuwen, and others. 2000. "Lack of Effect of a tings, if possible; and that are monitored for efficiency, per- High-Fiber Cereal Supplement on the Recurrence of Colorectal formance, and effectiveness. Thus, for example, new screening Adenomas: Phoenix Colon Cancer Prevention Physicians' Network." or treatment programs can be initiated in focused geographical New England Journal of Medicine 342 (16): 1156­62. areas or specific facilities with known and well-characterized Anderson, B. O., S. Braun, S. Lim, R. A. Smith, S. Taplin, and D. B. Thomas (Global Summit Early Detection Panel). 2003. "Early Detection of target populations, and their performance and outcomes can Breast Cancer in Countries with Limited Resources." Breast Journal 9 be compared with matched control areas or facilities. (Suppl. 2): S51­59. Health Service Interventions for Cancer Control in Developing Countries | 585 ATBC (Alpha-Tocopherol, Beta Carotene Cancer Prevention Study EBCTCG (Early Breast Cancer Trialists' Collaborative Group). 1998. Group). 1994. "The Effect of Vitamin E and Beta Carotene on the "Polychemotherapy for Early Breast Cancer: An Overview of the Incidence of Lung Cancer and Other Cancers in Male Smokers." New Randomised Trials." Lancet 352 (9132): 930­42. England Journal of Medicine 330 (15): 1029­35. Ferlay, J., F. Bray, P. Pisani, and D. M. Parkin. 2001. GLOBOCAN 2000: Babazono, A., and A. L. Hillman. 1995. "Declining Cost-Effectiveness of Cancer Incidence, Mortality, and Prevalence Worldwide, Version 1.0, Screening for Disease: The Case of Gastric Cancer in Japan." IARC CancerBase No. 5. Lyon, France: International Agency for International Journal of Technology Assessment in Health Care 11 (2): Research on Cancer and World Health Organization, IARC Press. 354­64. ------. 2004. GLOBOCAN 2002: Cancer Incidence, Mortality, and Barnum, H., and E. R. Greenberg. 1993. "Cancers." In Disease Control Prevalence Worldwide, Version 2.0, IARC CancerBase No. 5. Lyon, Priorities in Developing Countries, ed. D. T. Jamison, W. H. Mosley, France: International Agency for Research on Cancer and World A. R. Measham, and J. L. Bobadilla, 529­59. New York: Oxford Health Organization, IARC Press. University Press. Fiore, M. C., D. K. Hatsukami, and T. B. Baker. 2002. "Effective Tobacco Berthelot, J. M., B. P. Will, W. K. Evans, D. Coyle, C. C. Earle, and L. Dependence Treatment." Journal of the American Medical Association Bordeleau. 2000. "Decision Framework for Chemotherapeutic 288 (14): 1768­71. Interventions for Metastatic Non-Small-Cell Lung Cancer." Journal of Fracheboud, J., H. J. de Koning, R. Boer, J. H. Groenewoud, A. L. Verbeek, the National Cancer Institute 92 (16): 1321­29. M. J. Broeders, and others (National Evaluation Team for Breast Bobo, J. K., N. C. Lee, and S. F. Thames. 2000. "Findings from 752,081 Cancer Screening in the Netherlands). 2001. "Nationwide Breast Clinical Breast Examinations Reported to a National Screening Cancer Screening Programme Fully Implemented in the Netherlands." Program from 1995 through 1998." Journal of the National Cancer Breast 10 (1): 6­11. Institute 92 (12): 971­76. Frazier, A. L., G. A. Colditz, C. S. Fuchs, and K. M. Kuntz. 2000. "Cost- Boer, R., H. J. de Koning, G. J. van Oortmarssen, and P. J. van der Maas. Effectiveness of Screening for Colorectal Cancer in the General 1995. "In Search of the Best Upper Age Limit for Breast Cancer Population." Journal of the American Medical Association 284 (15): Screening." European Journal of Cancer 31A (12): 2040­43. 1954­61. Brown, M. L., and L. Fintor. 1993. "Cost Effectiveness of Breast Cancer Gail, M. H., J. P. Costantino, J. Bryant, R. Croyle, L. Freedman, Screening: Preliminary Results of a Systematic Review of the K. Helzlsouer, and V. Vogel. 1999. "Weighing the Risks and Benefits Literature." Breast Cancer Research and Treatment 25 (2): 113­18. of Tamoxifen Treatment for Preventing Breast Cancer." Journal of the Brown, M. L., S. G. Nayfield, and L. M. Shibley. 1994. "Adjuvant Therapy National Cancer Institute 91 (21): 1829­46. for Stage III Colon Cancer: Economic Returns to Research and Cost- Ginsberg, G. M., S. Lim, J. Lauer, C. Sepulveda, and T. Tantorres-Edeger. Effectiveness of Treatment." Journal of the National Cancer Institute 86 2004. Prevention, Screening, and Treatment of Colorectal Cancer: A (6): 424­30. Global and Regional Generalized Cost Effectiveness Analysis. Geneva: Chen, J. G., D. M. Parkin, Q. G. Chen, J. H. Lu, Q. J. Shen, B. C. Zhang, and World Health Organization. Y. R. Zhu. 2003. "Screening for Liver Cancer: Results of a Randomised Goldie, S., L. Gaffikin, A. Gordillo-Tobar, C. Levin, C. Mahé, J. Goldhaber- Controlled Trial in Qidong, China." Journal of Medical Screening Fiebert, and T. Wright. 2004. "A Comprehensive Policy Analysis of 10 (4): 204­9. Cervical Cancer Screening in Peru, India, Kenya, Thailand, and South Chirikos, T. N., T. Hazelton, M. Tockman, and R. Clark. 2002. "Screening Africa." Paper presented for the Alliance for Cervical Cancer for Lung Cancer with CT: A Preliminary Cost-Effectiveness Analysis." Prevention at the 21st International Papillomavirus Conference, Chest 121 (5): 1507­14. Mexico City, February 20­26. Christensen, D. 2004. "Dietary Prevention of Cancer: A Smorgasbord of Goldie, S. J., D. Grima, M. Kohli, T. C. Wright, M. Weinstein, and E. Options for Moving Ahead." Journal of the National Cancer Institute Franco. 2003. "A Comprehensive Natural History Model of HPV 96 (11): 822­24. Infection and Cervical Cancer to Estimate the Clinical Impact of a Collette, C., H. J. Collette, J. Fracheboud, B. J. Slotboom, and F. de Waard. Prophylactic HPV-16/18 Vaccine." International Journal of Cancer 106 1992. "Evaluation of a Breast Cancer Screening Programme--The (6): 896­904. DOM Project." European Journal of Cancer 28A (12): 1985­88. Goldie, S. J., J. J. Kim, and T. C. Wright. 2004. "Cost-Effectiveness of Coursaget, P., D. Leboulleux, M. Soumare, P. le Cann, B. Yvonnet, J. P. Human Papillomavirus DNA Testing for Cervical Cancer Screening in Chiron, and others. 1994. "Twelve-Year Follow-up Study of Hepatitis B Women Aged 30 Years or More." Obstetrics and Gynecology 103 (4): Immunization of Senegalese Infants." Journal of Hepatology 21 (2): 619­31. 250­54. Goldie, S. J., M. Kohli, D. Grima, M. C. Weinstein, T. C. Wright, X. C. da Costa e Silva, V. 2003. Policy Recommendations for Smoking Cessation Bosch, and E. Franco. 2004. "Projected Clinical Benefits and and Treatment of Tobacco Dependence: Tools for Public Health. Geneva: Cost-Effectiveness of a Human Papillomavirus 16/18 Vaccine." Journal World Health Organization. of the National Cancer Institute 96 (8): 604­15. Dahlberg, M., A. Stenborg, L. Pahlman, and B. Glimelius. 2002. "Cost- Goldie, S. J., L. Kuhn, L. Denny, A. Pollack, and T. C. Wright. 2001. "Policy Effectiveness of Preoperative Radiotherapy in Rectal Cancer: Results Analysis of Cervical Cancer Screening Strategies in Low-Resource from the Swedish Rectal Cancer Trial." International Journal of Settings: Clinical Benefits and Cost-Effectiveness." Journal of the Radiation Oncology, Biology, Physics 54 (3): 654­60. American Medical Association 285 (24): 3107­15. de Koning, H. J. 2000. "Breast Cancer Screening: Cost-Effective in Harper, D. M., E. L. Franco, C. Wheeler, D. G. Ferris, D. Jenkins, Practice?" European Journal of Radiology 33 (1): 32­37. A. Schuind, and others. 2004. "Efficacy of a Bivalent L1 Virus-Like de Koning, H. J., R. Boer, P. G. Warmerdam, P. M. Beemsterboer, and Particle Vaccine in Prevention of Infection with Human P. J. van der Maas. 1995. "Quantitative Interpretation of Age-Specific Papillomavirus Types 16 and 18 in Young Women: A Randomised Mortality Reductions from the Swedish Breast Cancer Screening Controlled Trial." Lancet 364 (9447): 1757­65. Trials." Journal of the National Cancer Institute 87 (16): 1217­23. Harris, R. A., D. K. Owens, H. Witherell, and J. Parsonnet. 1999. Denny, L., L. Kuhn, A. Pollack, H. Wainwright, and T. C. Wright Jr. 2000. "Helicobactor pylori and Gastric Cancer: What Are the Benefits of "Evaluation of Alternative Methods of Cervical Cancer Screening for Screening Only for the CagA Phenotype of H. pylori?" Helicobactor 4 Resource-Poor Settings." Cancer 89 (4): 826­33. (2): 69­76. 586 | Disease Control Priorities in Developing Countries | Martin L. Brown, Sue J. Goldie, Gerrit Draisma, and others Hughes, J. P., G. P. Garnett, and L. Koutsky. 2002. "The Theoretical Premenopausal Vietnamese and Chinese Women with Breast Cancer Population-Level Impact of a Prophylactic Human Papilloma Virus Treated in an Adjuvant Hormonal Therapy Study." International Vaccine." Epidemiology 13 (6): 631­39. Journal of Radiation Oncology, Biology, Physics 56 (3): 697­703. Humphrey, L. L., M. Helfand, B. K. Chan, and S. H. Woolf. 2002. "Breast Mahadevia, P. J., L. A. Fleisher, K. D. Frick, J. Eng, S. N. Goodman, and Cancer Screening: A Summary of the Evidence for the U.S. Preventive N. R. Powe. 2003. "Lung Cancer Screening with Helical Computed Services Task Force."Annals of Internal Medicine 137 (5, Part 1): 347­60. Tomography in Older Adult Smokers: A Decision and Cost- Imperiale, T. F. 2003. "Aspirin and the Prevention of Colorectal Cancer." Effectiveness Analysis." Journal of the American Medical Association 289 New England Journal of Medicine 348 (10): 879­80. (3): 313­22. International Agency for Research on Cancer. No date. "Cancer Mondial: Malin, J. L., E. Keeler, C. Wang, and R. Brook. 2002. "Using Cost- DEP Scientific Programmes." http://www-dep.iarc.fr/thisunit/ Effectiveness Analysis to Define a Breast Cancer Benefits Package for depproge.htm. the Uninsured." Breast Cancer Research and Treatment 74 (2): 143­53. Kensler, T. W., G. S. Qian, J. G. Chen, and J. D. Groopman. 2003. Mandelblatt, J. S., W. F. Lawrence, L. Gaffikin, K. K. Limpahayom, "Translational Strategies for Cancer Prevention in Liver." Nature P. Lumbiganon, S. Warakamin, and others. 2002. "Costs and Benefits Reviews: Cancer 3 (5): 321­29. of Different Strategies to Screen for Cervical Cancer in Less- Khandker, R. K., J. D. Dulski, J. B. Kilpatrick, R. P. Ellis, J. B. Mitchell, and Developed Countries." Journal of the National Cancer Institute 94 (19): W. B. Baine. 2000. "A Decision Model and Cost-Effectiveness Analysis 1469­83. of Colorectal Cancer Screening and Surveillance Guidelines for Marcus, P. M., E. J. Bergstralh, R. M. Fagerstrom, D. E. Williams, Average-Risk Adults." International Journal of Technology Assessment in R. Fontana, W. F. Taylor, and P. C. Prorok. 2000. "Lung Cancer Health Care 16 (3): 799­810. Mortality in the Mayo Lung Project: Impact of Extended Follow-Up." Kim, J. J., T. Wright, and S. Goldie. 2002. "Cost Effectiveness of Alternative Journal of the National Cancer Institute 92 (16): 1308­16. Triage Strategies for Atypical Squamous Cells of Undetermined Marks, L. B., P. H. Hardenbergh, E. T. Winer, and L. R. Prosnitz. 1999. Significance." Journal of the American Medical Association 287 (18): "Assessing the Cost-Effectiveness of Postmastectomy Radiation 2382­90. Therapy." International Journal of Radiation Oncology, Biology, Physics Koutsky, L. A., K. A. Ault, C. M. Wheeler, D. R. Brown, E. Barr, F. B. Alvarez, 44 (1): 91­98. and others (Proof of Principle Study Investigators). 2002. "A Controlled Trial of a Human Papillomavirus Type 16 Vaccine." New Marshall, D., K. N. Simpson, C. C. Earle, and C. W. Chu. 2001. "Economic England Journal of Medicine 347 (21): 1645­51. Decision Analysis Model of Screening for Lung Cancer." European Journal of Cancer 37 (14): 1759­67. Kulasingam, S. L., and E. R. Myers. 2003. "Potential Health and Economic Impact of Adding a Human Papillomavirus Vaccine to Screening Pro- Mathers, C. D., A. D. Lopez, and C. J. L. Murray. "The Burden of Disease grams." Journal of the American Medical Association 290 (6): 781­89. and Mortality by Condition: Data, Methods, and Results for 2001." In Lamberts, Q. J., A. J. der Kinderen, G. Draisma, and H. J. de Koning. 2004. Global Burden of Disease and Risk Factors, eds. A. D. Lopez, C. D. "Breast Cancer Screening in Developing Countries: A Cost-Effective- Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. New York: ness Analysis for India." Working Paper, Erasmus University Medical Oxford University Press. Center, Department of Public Health, Rotterdam, the Netherlands. Maxwell, G. L., J. W. Carlson, M. Ochoa, T. Krivak, G. S. Rose, and E. R. Lazcano-Ponce, E. C., S. Moss, P. Alonso de Ruiz, J. Salmeron Castro, and Myers. 2002. "Costs and Effectiveness of Alternative Strategies for M. Hernandez Avila. 1999. "Cervical Cancer Screening in Developing Cervical Cancer Screening in Military Beneficiaries." Obstetrics and Countries: Why Is It Ineffective? The Case of Mexico." Archives of Gynecology 100 (4): 740­48. Medical Research 30 (3): 240­50. Mulligan, J., J. A. Fox-Rushby, T. Adam, B. Johns, and A. Mills. 2003. "Unit Lee, C. L., K. S. Hsieh, and Y. C. Ko. 2003. "Trends in the Incidence of Costs of Delivering Health Interventions in Low- and Middle-Income Hepatocellular Carcinoma in Boys and Girls in Taiwan after Large- Countries: Tertiary Unit Costs of Delivering Health Interventions in Scale Hepatitis B Vaccination." Cancer Epidemiology, Biomarkers, and Low- and Middle-Income Countries." Working Paper 9, Disease Prevention 12 (1): 57­9. Control Priorities Project, Bethesda, MD. Lee, J. H., H. A. Glick, J. A. Hayman, and L. J. Solin. 2002. "Decision- Norum, J. 1999. "Breast Cancer Screening by Mammography in Norway. Analytic Model and Cost-Effectiveness Evaluation of Postmastectomy Is It Cost-Effective?" Annals of Oncology 10 (2): 197­203. Radiation Therapy in High-Risk Premenopausal Breast Cancer Patients." Journal of Clinical Oncology 20 (11): 2713­25. --------. 2000. "Adjuvant Cyclophosphamide, Methotrexate, Fluorouracil (CMF) in Breast Cancer: Is It Cost-Effective?" Acta Levin, V., A. Meghzifene, and H. Tatsuzaki. 2001. "Improving Cancer Care: Oncologica 39 (1): 33­39. Increased Need for Radiotherapy in Developing Countries." IAEA (International Atomic Energy Agency) Bulletin 43: 25­32. Olsson, S., I. Andersson, I. Karlberg, N. Bjurstam, E. Frodis, and S. Hakansson. 2000. "Implementation of Service Screening with Ley, C., A. Mohar, J. Guarner, R. Herrera-Goepfert, L. S. Figueroa, D. Mammography in Sweden: From Pilot Study to Nationwide Halperin, and others. 2004. "Helicobacter pylori Eradication and Programme." Journal of Medical Screening 7 (1): 14­18. Gastric Preneoplastic Conditions: A Randomized, Double-Blind, Placebo-Controlled Trial." Cancer Epidemiology, Biomarkers, and Omenn, G. S., G. E. Goodman, M. D. Thornquist, J. Balmes, M. R. Cullen, Prevention 13 (1): 4­10. A. Glass, and others. 1996. "Effects of a Combination of Beta Carotene Lippman, S. M., J. J. Lee, and A. L. Sabichi. 1998. "Cancer Chemo- and Vitamin A on Lung Cancer and Cardiovascular Disease." New prevention: Progress and Promise." Journal of the National Cancer England Journal of Medicine 334 (18): 1150­55. Institute 90 (20): 1514­28. Owen, J. B., L. R. Coia, and G. E. Hanks. 1997. "The Structure of Radiation Lopez, Alan D., Colin D. Mathers, Majid Ezzati, Dean T. Jamison, and Oncology in the United States in 1994." International Journal of Christopher J. L. Murray, eds. 2006. Global Burden of Disease and Risk Radiation Oncology, Biology, Physics 39 (1): 179­85. Factors. New York: Oxford University Press. Parkin, D. M., F. I. Bray, and S. S. Devesa. 2001. "Cancer Burden in the Year Love, R. R., N. Ba Duc, N. Cong Binh, P. A. Mahler, B. R. Thomadsen, 2000: The Global Picture." European Journal of Cancer 37 (Suppl. 8): N. Hong Long, and others. 2003. "Postmastectomy Radiotherapy in S4­66. Health Service Interventions for Cancer Control in Developing Countries | 587 Parkin, D. M., J. Ferlay, M. Hamdi-Cherif, F. Sitas, J. O. Thomas, Sherlaw-Johnson, C., S. Gallivan, and D. Jenkins. 1997. "Evaluating H. Wabinga, and S. L. Whelan. 2003. Cancer in Africa: Epidemiology Cervical Cancer Screening Programmes for Developing Countries." and Prevention. Lyon, France: International Agency for Research on International Journal of Cancer 72 (2): 210­16. Cancer and World Health Organization. Sikora, K., S. Advani, V. Koroltchouk, I. Magrath, L. Levy, H. Pinedo, and Pignone, M., S. Saha, T. Hoerger, and J. Mandelblatt. 2002. "Cost- others. 1999. "Essential Drugs for Cancer Therapy: A World Health Effectiveness Analyses of Colorectal Cancer Screening: A Systematic Organization Consultation." Annals of Oncology 10 (4): 385­90. Review for the U.S. Preventive Services Task Force." Annals of Internal Simpson, K. N., and L. B. Snyder. 1991. "Informing the Mammography Medicine 137 (2): 96­104. Coverage Debate. Results of Meta-Analysis, Computer Modeling, and Powles, T., R. Eeles, S. Ashley, D. Easton, J. Chang, M. Dowsett, and others. Issue Analysis." International Journal of Technology Assessment in 1998. "Interim Analysis of the Incidence of Breast Cancer in the Royal Health Care 7 (4): 616­31. Marsden Hospital Tamoxifen Randomised Chemoprevention Trial." Singer, P. A., and K. W. Bowman. 2002. "Quality End-of-Life Care: A Lancet 352 (9122): 98­101. Global Perspective." BMC Palliative Care 1 (1): 4­13. Rijnsburger, A. J., G. J. van Oortmarssen, R. Boer, C. Baines, A. B. Miller, Smith, R. D., J. Hall, H. Gurney, and P. R. Harnett. 1993. "A Cost-Utility and H. J. de Koning. Forthcoming. "Clinical Breast Exams as a Approach to the Use of 5-Fluorouracil and Levamisole as Adjuvant Screening Tool: Cost-Effectiveness." Chemotherapy for Dukes' C Colonic Carcinoma." Medical Journal of Rijnsburger, A. J., G. J. van Oortmarssen, R. Boer, G. Draisma, T. To, A. B. Australia 158 (5): 319­22. Miller, and H. J. de Koning. 2004. "Mammography Benefit in the Smith, T. J., and B. E. Hillner. 1993."The Efficacy and Cost-Effectiveness of Canadian National Breast Screening Study-2: A Model Evaluation." Adjuvant Therapy of Early Breast Cancer in Premenopausal Women." International Journal of Cancer 110 (5): 756­62. Journal of Clinical Oncology 11 (4): 771­76. Roderick, P., R. Davies, J. Raftery, D. Crabbe, R. Pearce, P. Patel, and ------. 2000. "Tamoxifen Should Be Cost-Effective in Reducing Breast P. Bhandari. 2003. "Cost-Effectiveness of Population Screening for Cancer Risk in High-Risk Women." Journal of Clinical Oncology 18 (2): Helicobactor pylori in Preventing Gastric Cancer and Peptic Ulcer 284­86. Disease, Using Simulation." Journal of Medical Screening 10 (3): 148­56. Sonnenberg, A., F. Delco, and J. M. Inadomi. 2000. "Cost-Effectiveness of Rose, P. G., and P. T. Lappas. 2000. "Analysis of the Cost-Effectiveness of Colonoscopy in Screening for Colorectal Cancer." Annals of Internal Concurrent Cisplatin-Based Chemoradiation in Cervical Cancer: Medicine 133 (8): 573­84. Implications from Five Randomized Trials." Gynecologic Oncology 78 Steele, R. J., R. Parker, J. Patnick, J. Warner, C. Fraser, N. A. Mowat, and oth- (1): 3­6. ers (United Kingdom Colorectal Screening Pilot Group). 2001. "A Sanders, G. D., and A. V. Taira. 2003. "Cost-Effectiveness of a Potential Demonstration Pilot Trial for Colorectal Cancer Screening in the Vaccine for Human Papillomavirus." Emerging Infectious Diseases 9 (1): United Kingdom: A New Concept in the Introduction of Healthcare 37­48. Strategies." Journal of Medical Screening 8 (4): 197­202. Sankaranarayanan, R., R. J. Black, and D. M. Parkin, eds. 1998. Cancer Suba, E. J., C. H. Nguyen, B. D. Nguyen, and S. S. Raab (Viet/American Survival in Developing Countries, IARC Scientific Publication 145. Cervical Cancer Prevention Project). 2001. "De Novo Establishment Lyon, France: International Agency for Research on Cancer Press and and Cost-Effectiveness of Papanicolaou Cytology Screening Services in World Health Organization. the Socialist Republic of Vietnam." Cancer 91 (5): 928­39. Sankaranarayanan, R., A. M. Budukh, and R. Rajkumar. 2001. "Effective Swedish Rectal Cancer Trial. 1997. "Improved Survival with Preoperative Screening Programmes for Cervical Cancer in Low- and Middle- Radiotherapy in Resectable Rectal Cancer." New England Journal of Income Developing Countries." Bulletin of the World Health Medicine 336 (14): 980­87. Organization 79 (10): 954­62. Terry, M. B., M. D. Gammon, F. F. Zhang, H. Tawfik, S. L. Teitelbaum, J. A. Sankaranarayanan, R., B. Shyamalakumary, R. Wesley, N. Sreedevi Amma, Britton, and others. 2004. "Association of Frequency and Duration of D. M. Parkin, and M. K. Nair. 1999. "Visual Inspection with Acetic Acid Aspirin Use and Hormone Receptor Status with Breast Cancer Risk." in the Early Detection of Cervical Cancer and Precursors." Journal of the American Medical Association 291 (20): 2433­40. International Journal of Cancer 80 (1): 161­63. Thomas, D. B., D. L. Gao, R. M. Ray, W. W. Wang, C. J. Allison, F. L. Chen, Sarasin, F. P., E. Giostra, and A. Hadengue. 1996. "Cost-Effectiveness of and others. 2002. "Randomized Trial of Breast Self-Examination in Screening for Detection of Small Hepatocellular Carcinoma in Western Shanghai: Final Results." Journal of the National Cancer Institute 94 Patients with Child-Pugh Class A Cirrhosis." American Journal of (19): 1445­57. Medicine 101 (4): 422­34. Turner, P. C., A. Sylla, M. S. Diallo, J. J. Castegnaro, A. J. Hall, and C. P. Wild. Schatzkin, A., E. Lanza, D. Corle, P. Lance, F. Iber, B. Caan, and others. 2002. "The Role of Aflatoxins and Hepatitis Viruses in the 2000. "Lack of Effect of a Low-Fat, High-Fiber Diet on the Recurrence Etiopathogenesis of Hepatocellular Carcinoma: A Basis for Primary of Colorectal Adenomas." New England Journal of Medicine 342 (16): Prevention in Guinea-Conakry, West Africa." Journal of 1149­55. Gastroenterology and Hepatology 17 (Suppl.): S441­48. Schiffman, M., R. Herrero, A. Hildesheim, M. E. Sherman, M. Bratti, S. United Nations Population Division. 2003."India, Population by Five-Year Wacholder, and others. 2000. "HPV DNA Testing in Cervical Cancer Age Group and Sex (Thousands), Medium Variant, 2000­2004." In Screening: Results from Women in a High-Risk Province of Costa World Population Prospects: The 2002 Revision Population Database, Rica." Journal of the American Medical Association 283 (1): 87­93. Tertiary World Population Prospects. United Nations. http:// Schrag, D. 2004. "The Price Tag on Progress: Chemotherapy for Colorectal esa.un.org/unpp. Cancer." New England Journal of Medicine 351 (4): 317­19. Vainio, H., and F. Bianchini. 2002a. Breast Cancer Screening. IARC Semiglazov, V. F., V. M. Moiseenko, A. G. Manikhas, S. A. Protsenko, R. S. Handbooks of Cancer Prevention, Vol. 7. Lyon, France: International Kharikova, R. T. Popova, and others. 1999. "Interim Results of a Agency for Research on Cancer Press and World Health Organization. Prospective Randomized Study of Self-Examination for Early ------. 2002b. Weight Control and Physical Activity. IARC Handbooks of Detection of Breast Cancer." Voprosy Onkologii 45 (3): 265­71. Cancer Prevention, Vol. 6. Lyon, France: International Agency for Shapiro, S., E. A. Coleman, M. Broeders, M. Codd, H. de Koning, Research on Cancer Press and World Health Organization. J. Fracheboud, and others. 1998. "Breast Cancer Screening Vainio, H., and G. Morgan. 1999. "Mechanisms of Aspirin Programmes in 22 Countries: Current Policies, Administration, and Chemoprevention of Colorectal Cancer." European Journal of Drug Guidelines." International Journal of Epidemiology 27 (5): 735­42. Metabolism and Pharmacokinetics 24 (4): 289­92. 588 | Disease Control Priorities in Developing Countries | Martin L. Brown, Sue J. Goldie, Gerrit Draisma, and others van den Hout, W. B., M. van den Brink, A. M. Stiggelbout, C. J. van de WHO (World Health Organization). 2001a. Macroeconomics and Health: Velde, and J. Kievet. 2002. "Cost-Effectiveness Analysis of Colorectal Investing in Health for Economic Development: Report of the Cancer Treatments." European Journal of Cancer 38 (7): 953­63. Commission on Macroeconomics and Health. Geneva: WHO. van den Hout, W. B., Y. M. van der Linden, E. Steenland, R. G. Wiggenraad, ------. 2001b. World Health Report 2002: Reducing Risks, Promoting J. Kievit, H. de Haes, and J. W. Leer. 2003. "Single- Versus Multiple- Healthy Life. Geneva: WHO. http://www.who.int/whr/en/. Fraction Radiotherapy in Patients with Painful Bone Metastases: ------. 2002. National Cancer Control Programmes, Policies, and Cost-Utility Analysis Based on a Randomized Trial." Journal of the Managerial Guidelines, 2nd ed. Geneva: WHO. National Cancer Institute 95 (3): 222­29. ------. 2003a. "Essential Drugs and Medicines Policy: 13th Expert van der Maas, P. J. 2001. "Breast Cancer Screening Programme in the Committee on the Selection and Use of Essential Medicines, 31 March Netherlands: An Interim Review." Breast 10 (1): 12­14. to 3 April 2003." Geneva: WHO. http://www.who.int/medicines/ organization/par/edl/expertcomm.shtml. van Meerbeeck, J. P., and K. G. Tournoy. 2004. "Screening and Diagnosis of NSCLC." Annals of Oncology 15 (Suppl. 4): iv 65­70. ------. 2003b. "WHO Framework Convention on Tobacco Control." Geneva: WHO. http://www.who.int/tobacco/fctc/en/fctc_booklet_ van Oortmarssen, G. J., J. D. Habbema, P. J. van der Maas, H. J. de Koning, english.pdf. H. J. Collette, A. L. Verbeek, and others. 1990. "A Model for Breast Cancer Screening." Cancer 66 (7): 1601­12. ------. No date. "WHO Statistical Information System." Geneva: WHO. http://www3.who.int/whosis. Veronesi, U., P. Maisonneuve, A. Costa, V. Sacchini, C. Maltoni, C. Whynes, D. K., and Nottingham Faecal Occult Blood Screening Trial. 2004. Robertson, and others. 1998. "Prevention of Breast Cancer with "Cost-Effectiveness of Screening for Colorectal Cancer: Evidence from Tamoxifen: Preliminary Findings from the Italian Randomised Trial the Nottingham Faecal Occult Blood Trial." Journal of Medical among Hysterectomised Women--Italian Tamoxifen Prevention Screening 11 (1): 11­15. Study." Lancet 352 (9122): 93­97. Wong, B. C., S. K. Lam, W. M. Wong, J. S. Chen, T. T. Zheng, R. E. Feng, and Vervoort, M. M., G. Draisma, J. Fracheboud, L. V. van de Poll-Franse, and others (China Gastric Cancer Study Group). 2004. "Helicobacter pylori H. J. de Koning. 2004. "Trends in the Usage of Adjuvant Systemic Eradication to Prevent Gastric Cancer in High-Risk Region of China: Therapy for Breast Cancer in the Netherlands and Its Effect on A Randomized Controlled Trial." Journal of the American Medical Mortality." British Journal of Cancer 91 (2): 242­47. Association 291 (2): 187­94. Vijan, S., E. W. Hwang, T. P. Hofer, and R. A. Hayward. 2001."Which Colon Wright, T. C. Jr. 2003. "Chapter 10: Cervical Cancer Screening Using Cancer Screening Test? A Comparison of Costs, Effectiveness, and Visualization Techniques." Journal of the National Cancer Institute Compliance." American Journal of Medicine 111 (8): 593­601. Monographs (31): 66­71. Viviani, S., A. Jack, A. J. Hall, N. Maine, M. Mendy, R. Montesano, and H. Wright, T. C. Jr., L. Denny, L. Kuhn, A. Pollack, and A. Lorincz. 2000. "HPV C. Whittle. 1999. "Hepatitis B Vaccination in Infancy in The Gambia: DNA Testing of Self-Collected Vaginal Samples Compared with Protection against Carriage at 9 Years of Age." Vaccine 17 (23­24): Cytologic Screening to Detect Cervical Cancer." Journal of the American 2946­50. Medical Association 283 (1): 81­86. Wagner, J., S. Tunis, M. Brown, A. Ching, and R. Almeida. 1996. "Cost- Zimbabwe Project. 1999. "Visual Inspection with Acetic Acid for Cervical- Effectiveness of Colorectal Cancer Screening in Average-Risk Adults." Cancer Screening: Test Qualities in a Primary-Care Setting: University In Prevention and Early Detection of Colorectal Cancer, ed. G. Young, of Zimbabwe/JHPIEGO Cervical Cancer Project." Lancet 353 (9156): P. Rozen, and B. Levin, 321­56. London: Saunders. 869­73. Health Service Interventions for Cancer Control in Developing Countries | 589 Chapter 30 Diabetes: The Pandemic and Potential Solutions K. M. Venkat Narayan, Ping Zhang, Alka M. Kanaya, Desmond E. Williams, Michael M. Engelgau, Giuseppina Imperatore, and Ambady Ramachandran NATURE AND DISTRIBUTION OF DIABETES people age 20 to 79 (table 30.1). The prevalence of diabetes was higher in developed countries than in developing coun- Diabetes is a metabolic disease characterized by hyperglycemia tries. In the developing world, the prevalence was highest in resulting from defects in insulin secretion, insulin action, or Europe and Central Asia and lowest in Sub-Saharan Africa. both (American Diabetes Association 2004). Some of these variations may reflect differences in the age structures and level of urbanization of the various popula- Classification of Diabetes tions. By 2025, the worldwide prevalence is projected to be 6.3 percent, a 24 percent increase compared with 2003. The Diabetes takes three major forms. Type 1 diabetes results from largest increase in prevalence by 2025 is expected to be in East destruction of the beta cells in the pancreas, leading to absolute Asia and the Pacific, and the smallest in Sub-Saharan Africa. insulin deficiency. It usually occurs in children and young In terms of those affected, the biggest increase in the devel- adults and requires insulin treatment. Type 2 diabetes, which oping countries is projected to take place among adults of accounts for approximately 85 to 95 percent of all diagnosed working age. cases, is usually characterized by insulin resistance in which In 2003, 194 million people worldwide ages 20 to 79 had target tissues do not use insulin properly. A third type of dia- diabetes, and by 2025, this number is projected to increase to betes, gestational diabetes, is first recognized during pregnancy. 333 million, a 72 percent increase (table 30.1). The developing Other rare types of diabetes include those caused by genetic world accounted for 141 million people with diabetes (72.5 per- conditions (for example, maturity-onset diabetes of youths), cent of the world total) in 2003. During the same period, the surgery, drug use, malnutrition, infections, and other illnesses. number of people with diabetes is projected to double in three of the six developing regions: the Middle East and North Africa, The Burden of Diabetes South Asia, and Sub-Saharan Africa. Diabetes affects persons of all ages and races. The disease Diabetes-Related Mortality and Disability. The death rate reduces both a person's quality of life and life expectancy and of men with diabetes is 1.9 times the rate for men without imposes a large economic burden on the health care system and diabetes, and the rate for women with diabetes is 2.6 times on families. that for women without diabetes (W. L. Lee and others 2000). Premature mortality caused by diabetes results in an Secular Trend and Projections. In 2003, the worldwide estimated 12 to 14 years of life lost (Manuel and Schultz prevalence of diabetes was estimated at 5.1 percent among 2004; Narayan and others 2003). Cardiovascular disease 591 Table 30.1 Estimated Numbers of People Age 20 to 79 with Diabetes, Mortality, DALYs, and Direct Medical Costs Attributable to Diabetes, by Regions Direct medical costs, Disability- Number of people Prevalence 2003 (US$ million) Deaths, adjusted life (thousands) (percent) Low High 2001 years, 2001 Region 2003 2025 2003 2025 estimate estimate (thousands) (thousands) Developing countries 140,849 264,405 4.5 5.9 12,304 23,127 757 15,804 East Asia and the Pacific 31,363 60,762 2.6 3.9 1,368 2,656 234 4,930 Europe and Central Asia 25,764 33,141 7.6 9.0 2,884 5,336 51 1,375 Latin America and the Caribbean 19,026 36,064 6.0 7.8 4,592 8,676 163 2,775 Middle East and North Africa 10,792 23,391 6.4 7.9 2,347 4,340 31 843 South Asia 46,309 94,848 5.9 7.7 840 1,589 196 4,433 Sub-Saharan Africa 7,595 16,199 2.4 2.8 273 530 82 1,448 Developed countries 53,337 68,345 7.8 9.2 116,365 217,760 202 4,192 World 194,186 332,750 5.1 6.3 128,669 240,887 959 19,996 Source: Number of persons with diabetes, prevalence of diabetes, and direct medical costs of diabetes, International Diabetes Federation 2003b; all other information, WHO 2004. (CVD) causes up to 65 percent of all deaths in developed Economic Burden of Diabetes countries of people with diabetes (Geiss, Herman, and Smith Diabetes imposes large economic burdens on national 1995). health care systems and affects both national economies and The World Health Organization (WHO) estimates that, in individuals and their families. Direct medical costs include 2001, 959,000 deaths worldwide were caused by diabetes, resources used to treat the disease. Indirect costs include lost accounting for 1.6 percent of all deaths, and approximately productivity caused by morbidity, disability, and premature 3 percent of all deaths caused by noncommunicable diseases. mortality. Intangible costs refer to the reduced quality of life for More recent estimates by WHO suggest that the actual num- people with diabetes brought about by stress, pain, and anxiety. ber may be triple this estimate and that about two-thirds of these deaths occur in developing countries (WHO 2004). Direct Medical Costs. Good data on the direct medical costs Within the developing regions, most deaths caused by diabetes of diabetes are not available for most developing countries. occurred in East Asia and the Pacific and the fewest in Sub- Extrapolation from developed countries suggests that, in 2003, Saharan Africa (table 30.1). the direct costs of diabetes worldwide for people age 20 to 79 Diabetes-related complications include microvascular dis- totaled at least US$129 billion and may have been as high as eases (for example, retinopathy, blindness, nephropathy, and US$241 billion (table 30.1). In the developing world, the costs kidney failure) and macrovascular diseases (coronary heart dis- were highest in Latin America and the Caribbean and lowest in ease, stroke, peripheral vascular disease, and lower-extremity Sub-Saharan Africa. The direct health care costs of diabetes amputation). Those complications result in disability. In the range from 2.5 to 15.0 percent of annual health care budgets, United States, a much higher proportion of people with dia- depending on local prevalence and sophistication of the treat- betes than of people without diabetes have physical limitations: ments available (International Diabetes Federation 2003b). 66 percent compared with 29 percent (Ryerson and others 2003). Disabilities are even more pronounced among older Indirect and Intangible Costs. In developing countries, the people (Gregg and others 2000). indirect costs of diabetes are at least as high, or even higher, The World Health Organization estimated that, in 2001, dia- than the direct medical costs (Barcelo and others 2003). betes resulted in 19,996,000 disability-adjusted life years Because the largest predicted rise in the number of people with (DALYs) worldwide. More than 80 percent of the DALYs result- diabetes in the next three decades will be among those in the ing from diabetes were in developing countries (table 30.1). economically productive ages of 20 to 64 (King, Aubert, and East Asia and the Pacific had the largest burden, and the Middle Herman 1998), the future indirect costs of diabetes will be even East and North Africa had the smallest burden. DALYs result- larger than they are now. ing from diabetes increased by 250 percent worldwide from Diabetes lowers people's quality of life in many ways, 1990 to 2001 and by 266 percent for low- and middle-income including their physical and social functioning and their countries (Mathers and others 2000). perceived physical and mental well-being. With a value of 592 | Disease Control Priorities in Developing Countries | K. M. Venkat Narayan, Ping Zhang, Alka M. Kanaya, and others 1 representing the health-related quality of life without illness indigenous populations and in developing economies (Rowley and 0 representing death, people with type 2 diabetes had a and others 1997; Williams and others 2001). Conversely, in value of 0.77 in the population of the United Kingdom developed countries, those in lower socioeconomic groups prospective diabetes study (Clarke, Gray, and Holman 2002). have a higher risk of obesity and consequently of type 2 diabetes (Everson and others 2002). Surrogates for socioeconomic sta- tus, such as level of education attained and income (Paeratakul Risk Factors for Diabetes and others 2002; Robbins and others 2001) are inversely asso- Risk factors for diabetes vary by disease type. ciated with diabetes in high-income countries. Type 1 Diabetes. Type 1 diabetes is most likely a polygenic dis- ease, and a number of potential environmental risk factors have INTERVENTIONS AND DELIVERY MODES been implicated--including dietary factors; breastfeeding; initi- Interventions against diabetes include those for preventing the ation of bovine milk; infectious agents (for example, enterovirus, disease, those for detecting the disease in its asymptomatic stage, rotavirus, and rubella); chemicals; and toxins--but the results and those for managing the disease to reduce its complications. have been inconclusive (Akerblom and Knip 1998). Type 2 Diabetes. The risk for type 2 diabetes is higher in Preventing Type 1 Diabetes monozygotic twins and people with a family history of diabetes Not enough scientific evidence is available to indicate that (Rich 1990). This finding strongly suggests that genetic deter- type 1 diabetes can be prevented, although various interven- minants play a role, but so far few genes have been associated tions have been explored. Examples of tested interventions with type 2 diabetes. include eliminating or delaying exposure to bovine protein and Environmental factors include prenatal factors, obesity, using insulin or nicotinamide for people at high risk of devel- physical inactivity, and dietary and socioeconomic factors oping the disease. (Qiao and others 2004). Exposure to diabetes in utero increases the risk of developing type 2 diabetes in early adulthood (Dabelea and others 2000). Disproportionate growth and low Preventing Type 2 Diabetes birthweight increase the risk of developing diabetes and insulin Four major trials--in China, Finland, Sweden, and the United resistance. In the postnatal environment, breastfeeding protects States--have demonstrated that intensive lifestyle interventions against the development of obesity, insulin resistance, and dia- involving a combination of diet and physical activity can delay betes (Pettitt and others 1997; Young and others 2002). or prevent diabetes among people at high risk (Eriksson and The strongest and most consistent risk factors for diabetes Lindgarde 1991; Knowler and others 2002; Pan and others 1997; and insulin resistance among different populations are obesity Tuomilehto and others 2001). In the largest randomized, con- and weight gain (Haffner 1998): for each unit increase in body trolled trial to date, the Diabetes Prevention Program (Knowler mass index, the risk of diabetes increases by 12 percent (Ford, and others 2002), the goals of the intensive lifestyle intervention Williamson, and Liu 1997). The distribution of fat around the were weight loss of 7 percent of baseline bodyweight through trunk region, or central obesity, is also a strong risk factor for a low-calorie diet and moderate physical activity for at least diabetes (Yajnik 2001). Diabetes risk may be reduced by increas- 150 minutes per week. After 2.8 years of follow-up, the average ing physical activity. Conversely, a sedentary lifestyle and physi- weight loss was 4.5 kilograms for those in the lifestyle interven- cal inactivity are associated with increased risks of developing tion group and less than 0.3 kilograms for those in the placebo diabetes (Hu and others 2003). Some studies report a positive group. The lifestyle intervention reduced the incidence of dia- relationship between dietary fat and diabetes, but specific types betes by 58 percent. of fats and carbohydrates may be more important than total fat Pharmacological studies of diabetes prevention have been or carbohydrate intake. Polyunsaturated fats and long-chain reviewed in detail elsewhere (Kanaya and Narayan 2003). In omega-3 fatty acids found in fish oils (Adler and others 1994) summary, a variety of specific medications have been tested may reduce the risk of diabetes, and saturated fats and trans (for example, metformin, acarbose, orlistat, troglitazone, fatty acids may increase the risk of diabetes (Hu, van Dam, and angiotensin-converting enzyme [ACE] inhibitors, statins, Liu 2001). Sugar-sweetened beverages are associated with an estrogens, and progestins) and have been found to lower dia- increased risk of diabetes (Schulze and others 2004). High betes incidence, but the expense, side effects, and cumulative intakes of dietary fiber and of vegetables may reduce the risk of years of drug intervention are practical concerns. Except for the diabetes (Fung and others 2002; Stevens and others 2002). Diabetes Prevention Program (Knowler and others 2002), no Increased affluence and Westernization have been associated trial of medication intervention has directly compared the with an increase in the prevalence of diabetes in many effectiveness of a drug to that of lifestyle modification. Diabetes: The Pandemic and Potential Solutions | 593 Screening for People with Diabetes or Prediabetes Managing Diabetes The benefits of early detection of type 2 diabetes through High-quality evidence exists for the efficacy of several current screening are not clearly documented, nor is the choice of the treatments in reducing morbidity and mortality in people with appropriate screening test established. Questionnaires used diabetes. These interventions are summarized in table 30.2. alone tend to work poorly; biochemical tests alone or in com- In addition, a review of previous studies (Norris, bination with assessment of risk factors are a better alternative Engelgau, and Narayan 2001) found positive effects for short (Engelgau, Narayan, and Herman 2000). follow-up (less than six months) of self-management training Table 30.2 Effectiveness and Cost-Effectiveness of Interventions for Preventing and Treating Diabetes in Developed Countries Quality of Cost-effectiveness ratio Strategy Benefit evidencea (US$/QALY)b Preventing diabetes · Lifestyle interventions for preventing Reduction of 35­58 percent in incidence among I 1,100 (Diabetes Prevention Program Research type 2 diabetes people at high risk Group forthcoming) · Metformin for preventing type 2 diabetes Reduction of 25­31 percent in incidence among I 31,200 (Diabetes Prevention Program Research people at high risk Group forthcoming) Screening for diabetes · Screening for type 2 diabetes in general Reduction of 25 percent in microvascular III 73,500 (CDC Diabetes Cost-Effectiveness Study population disease Group 1998) Treating diabetes and its complications · Glycemic control in people with HbA1c Reduction of 30 percent in microvascular disease I Cost saving (CDC Diabetes Cost-Effectiveness greater than 9 percent per 1 percent drop in HbA1c Study Group 1998) · Glycemic control in people with HbA1c Reduction of 30 percent in microvascular disease I 34,400 (CDC Diabetes Cost-Effectiveness Study greater than 8 percent per 1 percent drop in HbA1c Group 1998; Klonoff and Schwartz 2000) · Blood pressure control in people whose Reduction of 35 percent in macrovascular and I Cost saving (CDC Diabetes Cost-Effectiveness pressure is higher than 160/95 mmHg microvascular disease per 10 mmHg drop in Study Group 1998) blood pressure · Cholesterol control in people with total Reduction of 25­55 percent in coronary heart II-1 63,200 (CDC Diabetes Cost-Effectiveness Study cholesterol greater than 200 milligrams/ diseases events; 43 percent fall in death rate Group 1998) deciliter · Smoking cessation with recommended 16 percent quitting rate I 12,500 (CDC Diabetes Cost-Effectiveness Study guidelines Group 1998) · Annual screening for microalbuminuria Reduction of 50 percent in nephropathy using III 47,400 (Klonoff and Schwartz 2000) ACE inhibitors for identified cases · Annual eye examinations Reduction of 60 to 70 percent in serious I 6,000 (Klonoff and Schwartz 2000; Vijan, Hofer, vision loss and Hayward 2000) · Foot care in people with high risk of ulcers Reduction of 50 to 60 percent in serious foot I Cost saving (Ragnarson and Apelqvist 2001) disease · Aspirin use Reduction of 28 percent in myocardial infarctions, I Not available reduction of 18 percent in cardiovascular disease · ACE inhibitor use in all people with Reduction of 42 percent in nephropathy; I 8,800 (Golan, Birkmeyer, and Welch 1999) diabetes 22 percent drop in cardiovascular disease · Influenza vaccinations among the elderly Reduction of 32 percent in hospitalizations; II-2 3,100 (Sorensen and others 2004) for type 2 diabetes 64 percent drop in respiratory conditions and death · Preconception care for women of Reduction of 30 percent in hospital charges and II-2 Cost saving (Klonoff and Schwartz 2000) reproductive age 25 percent in hospital days Source: Authors. Note: mmHg millimeters of mercury; QALY quality-adjusted life year. a. I indicates evidence from at least one randomized, controlled trial; II-1 indicates evidence from a well-designed, controlled trial without randomization; II-2 indicates evidence from cohort or case con- trol studies; and III indicates opinions of respected authorities (U.S. Preventive Services Task Force 1996). b. We adjusted cost-effectiveness ratios to 2002 U.S. dollars using the consumer price index for medical care. In cases in which multiple studies evaluated the cost-effectiveness of an intervention, we report the median cost-effectiveness ratio. 594 | Disease Control Priorities in Developing Countries | K. M. Venkat Narayan, Ping Zhang, Alka M. Kanaya, and others on knowledge, frequency, and accuracy of self-monitoring of ratios--each weighted by its share (Barcelo and others 2003)-- blood glucose; self-reported dietary habits; and glycemic con- for outpatient care, inpatient care, drugs and laboratory tests, trol. Effects on lipids, physical activity, weight, and blood and treatment for diabetic complications. The cost ratio for each pressure varied. cost component was calculated as the cost of medical services or drugs in the United States divided by the cost of the same serv- ices or drugs in Latin America and the Caribbean. U.S. data for COST-EFFECTIVENESS OF INTERVENTIONS medical services and drugs for routine diabetes care, plus treat- AND PRIORITIES ment cost for diabetes complications, were obtained from a 1998 cost-effectiveness Markov model of the U.S. Centers for Most of the interventions to prevent and treat diabetes and its Disease Control and Prevention (CDC). Data for laboratory complications significantly affect the use of health services. The service were obtained from the 2001 Clinical Diagnostic limitations of clinical trials include their failure in most cases Laboratory Fee Schedule from the U.S. Centers for Medicare to capture the entire intervention effect over a lifetime and to Services (available from http://www.cms.gov). Data for Latin include all segments of a population to whom the intervention America and the Caribbean were obtained from three coun- may apply. Evaluating the cost-effectiveness of interventions tries--Argentina (Gagliardino and others 1993), Brazil (Health often requires the use of computer simulation models, but data Policy Division of the Brazilian Ministry of Health), and Mexico availability, technical complexity, and resource needs present a (Villarreal-Rios and others 2000). significant barrier to constructing such models for developing We applied Mulligan and others' framework (2003) to esti- countries. Furthermore, data on interventions are often avail- mate the costs of intervention and diabetes care in each devel- able only from developed countries, and these data are often oping region. Assuming that cost estimates are available for one extrapolated to developing countries. of the regions, this framework allows the development of a rel- ative cost index for health care services that can then be used to obtain cost estimates for the other five regions. Using costs esti- Estimating the Cost-Effectiveness of Interventions mated by Mulligan and others (2003), we first estimated three in Developing Countries health service indexes, including hospital bed days, outpatient To assess the cost-effectiveness of interventions in developing and inpatient services, and laboratory tests and procedures. We countries, we updated the results from Klonoff and Schwartz's then combined the three indexes into one overall index for dia- (2000) comprehensive review by including studies that were betes care in accordance with the share of each component in published up to 2003. Table 30.2 summarizes the cost- developing countries (Barcelo and others 2003). Finally, we effectiveness of interventions for the developed countries, estimated the costs of intervention and diabetes care in the mainly in the United States. The results show that the cost- other five developing regions by multiplying the cost of care in effectiveness of interventions varies greatly--from cost saving the Latin America region by the overall regional relative cost (an intervention is both more effective and less expensive than index. the comparator) to US$73,500 per quality-adjusted life year (QALY) gained. We estimated the cost-effectiveness ratio of diabetes interven- Ranking Implementation Priorities tions for the six developing regions shown in table 30.3. We We assessed the implementation priority and feasibility of assumed that the effectiveness of these interventions, as interventions, as explained in table 30.3. measured in QALYs, was the same as in developed countries but that the cost of interventions and other diabetes care differed Level 1 Interventions. All three interventions in this category between developed and developing countries and also among the are cost saving and are also feasible in terms of all four aspects six developing regions. Using this assumption, we estimated considered. The barrier to implementing these interventions the cost-effectiveness ratio for a developing region as the cost- may be a short-term hike in intervention costs. effectiveness ratio in the developed country, mainly represented Glycemic control in a population with poor control (hemo- by the United States,multiplied by the ratio of costs in the devel- globin A1c greater than 9 percent or another measure of oping region to the cost in the developed countries, which we glucose control in situations where HbA1c tests may be unaf- calculated as follows. These cost-effectiveness ratios are based on fordable) is cost saving because the reduction in medical care costs and benefits over a lifetime, except for preconception care costs associated with both short-term and long-term complica- for women of reproductive age. tions is greater than is the cost of intervention. Glycemic con- We estimated that the cost of intervention and other diabetes trol for people with type 1 diabetes involves insulin use and, for care in the United States was 8.6 times the cost in Latin America people with type 2 diabetes, depending on the stage and sever- and the Caribbean. This cost ratio was an average of four cost ity of the disease, consists of diet and physical activity, oral Diabetes: The Pandemic and Potential Solutions | 595 Table 30.3 Cost-Effectiveness of Interventions for Preventing and Treating Diabetes and Its Complications in Developing Regions Cost/QALY (2001 US$) East Asia Latin America Middle East and Europe and and the and Sub-Saharan Implementing Intervention the Pacific Central Asia Caribbean North Africa South Asia Africa Feasibilitya priorityb Level 1 Glycemic control in Cost saving Cost saving Cost saving Cost saving Cost saving Cost saving ++++ 1 people with HbA1c higher than 9 percent Blood pressure control in Cost saving Cost saving Cost saving Cost saving Cost saving Cost saving ++++ 1 people with pressure higher than 160/95 mmHg Foot care in people with Cost saving Cost saving Cost saving Cost saving Cost saving Cost saving ++++ 1 a high risk of ulcers Level 2 Preconception care for women Cost saving Cost saving Cost saving Cost saving Cost saving Cost saving ++ 2 of reproductive age Lifestyle interventions for 80 100 130 110 60 60 ++ 2 preventing type 2 diabetes Influenza vaccinations among 220 290 360 310 180 160 ++++ 2 the elderly for type 2 diabetes Annual eye examination 420 560 700 590 350 320 ++ 2 Smoking cessation 870 1,170 1,450 1,230 730 660 ++ 2 ACE inhibitor use for people 620 830 1,020 870 510 460 +++ 2 with diabetes Level 3 Metformin intervention for 2,180 2,930 3,630 3,080 1,820 1,640 ++ 3 preventing type 2 diabetes Cholesterol control for people 4,420 5,940 7,350 6,240 3,680 3,330 +++ 3 with total cholesterol higher than 200 milligrams/deciliter Intensive glycemic control for 2,410 3,230 4,000 3,400 2,000 1,810 ++ 3 people with HbA1c higher than 8 percent Screening for undiagnosed 5,140 6,910 8,550 7,260 4,280 3,870 ++ 3 diabetes Annual screening for 3,310 4,450 5,510 4,680 2,760 2,500 ++ 3 microalbuminuria Source: Authors. a. Feasibility was assessed based on difficulty of reaching the intervention population (the capacity of the health care system to deliver an intervention to the targeted population), technical complexity (the level of medical technologies or expertise needed for implementing an intervention), capital intensity (the amount of capital required for an intervention), and cultural acceptability (appropriateness of an intervention in terms of social norms and/or religious beliefs). ++++ indicates feasible for all four aspects, +++ indicates feasible for three of the four, ++ indicates feasible for two of the four, and + indicates feasible for one of the four. b. Implementing priority was assessed by combining the cost-effectiveness of an intervention and its implementation feasibility; 1 represents the highest priority and 3 represents the lowest priority. glucose-lowering agents, and insulin. Patient education is an For example, the mean HbA1c level for people with diabetes essential component of these interventions to encourage in India was 8.9 percent in 1998 (Raheja and others 2001). patients to comply with medication regimes and to change to A survey conducted by the International Diabetes Federation and maintain healthy lifestyles. in 1997 (2003b) showed that no country in Africa had 100 per- Glucose is generally poorly controlled in people with both cent accessibility to insulin. Ensuring adequate access to type 1 and type 2 diabetes, mostly because of lack of access to insulin should be an important priority for developing insulin and other diabetes supplies in developing countries. countries. 596 | Disease Control Priorities in Developing Countries | K. M. Venkat Narayan, Ping Zhang, Alka M. Kanaya, and others Blood pressure control for people with diabetes and hyper- the prevalence of diabetes in all regions. The expertise tension reduces the incidence of both microvascular and required for the intervention, such as dietitians and exercise macrovascular diseases. Major medication interventions physiologists, and the capacity of health care systems to han- include an ACE inhibitor, thiazide diuretics, or a beta blocker. dle the large populations eligible for the intervention may Blood pressure control is cost saving mainly because of its large present a barrier to implementing the intervention in many health benefits and relatively low intervention costs. Even in the developing countries. United States, moderate blood pressure control costs less than People with diabetes are at higher risk of complications US$250 per patient per year. Because many blood pressure from influenza and pneumococcal infections than those medications are generic drugs, the costs are much lower in without diabetes. Influenza vaccinations are a relatively cost- developing countries. In addition, the prevalence of people effective intervention, mainly because of the low intervention with poor control of blood pressure may be high in developing cost. However, the level of adoption for the intervention would countries. For example, in Latin America and the Caribbean, depend on a country's ability to deliver the intervention to the 60 percent of people with type 2 diabetes in 2000 had blood targeted population. pressure higher than 140/90 mmHg (Gagliardino, de la Hera The detection of proliferative diabetic retinopathy and mac- and Siri 2001). ular edema by dilated eye examination followed by appropriate Complications related to foot problems are common among laser photocoagulation therapy prevents blindness. Annual diabetics in developing countries. For example, in India, screening and treatment programs for diabetic retinopathy cost 43 percent of diabetes patients had foot-related complications US$700 or less per QALY gained in developing countries. The (Raheja and others 2001). Interventions for foot care are low intervention is more cost-effective among older people, those tech and require little capital. Interventions for foot care in who require insulin (Klonoff and Schwartz 2000), or those with developing countries should include educational programs for poor glucose control (Vijan, Hofer, and Hayward 2000). In patients and professionals (for example, on foot hygiene, treat- addition, screening less frequently, such as every two years, may ment of calluses, awareness of functional infections, and care be more cost-effective than screening every year (Vijan, Hofer, for skin injuries); access to appropriate footwear; and multidis- and Hayward 2000). Eye complications among people with ciplinary clinics. All three interventions could be cost saving, diabetes are common in developing countries; for example, mainly because the cost of the interventions is low and the 39 percent of people with diabetes in India had eye-related com- interventions can reduce the risk of foot ulceration and ampu- plications (Rajala and others 1998). Although laser treatment is tation, which are costly. Applying these interventions for high- an effective intervention, such treatment may not be available in risk patients, such as those with at least one previous foot ulcer many developing countries or may be extremely costly. or amputation, would yield even larger savings (Klonoff and ACE inhibitors can lower the blood pressure of those with Schwartz 2000). hypertension and delay the onset or prevent further progres- sion of renal disease for those with diabetes. Compared with Level 2 Interventions. The six interventions in this category screening for microalbuminuria and treating only those who are either cost saving and not feasible in one or more aspects or have the condition, offering ACE inhibitors to all people with cost less than US$1,500 per QALY and are at least moderately diabetes was more cost-effective at less than US$1,020 per feasible. Thus, interventions in this category represent good QALY gained. This intervention was more cost-effective among value for money but may present some difficulties in terms of younger people and was sensitive to the cost of drug. Thus, feasibility. lowering the cost of the medication is a key factor for the suc- Preconception care among women of reproductive age cess of this intervention in developing countries. includes patient education and intensive glucose management. Smoking cessation includes both counseling and using This intervention reduces short-term hospital costs for both medication such as a nicotine patch. Smoking cessation mothers and infants and improves birth outcomes. However, appears to be the least cost-effective among the level 2 inter- the intervention may not be feasible in some developing ventions. However, the benefits of smoking cessation may be countries because of the resources needed for the intervention underestimated because our calculations only took the reduced and the difficulty of reaching the target population. risk of CVD into account (Earnshaw and others 2002). Adding The lifestyle intervention for preventing type 2 diabetes the health benefits derived from preventing cancer and pul- costs US$60 to US$130 per QALY over a lifetime, depending monary diseases would improve the cost-effectiveness of on the region. The potential population eligible for a lifestyle smoking cessation. Considering the high prevalence of smok- intervention (those with impaired glucose tolerance or ing in developing countries, smoking cessation should be a impaired fasting glucose) is large in developing countries. The high-priority intervention, but the availability of the nicotine International Diabetes Federation (2003b) estimates that the patch may be a barrier to implementing this intervention in prevalence of impaired glucose tolerance was at least as high as developing countries. Diabetes: The Pandemic and Potential Solutions | 597 Level 3 Interventions. The five interventions included in this antihypertensive medications--aspirin, statin, and folic acid category cost at least US$1,640 per QALY but could cost as (see also chapter 33). Currently, neither is it available for use, much as US$8,550 per QALY. Compared with the level 1 and 2 nor have estimates of its benefits and adverse effects been con- interventions, those in this category are also less feasible. In firmed in a formal, randomized, controlled trial. The idea is general, depending on cost-effectiveness and feasibility, these thus still theoretical. The cost-effectiveness of this hypothetical interventions may not always be justifiable for all people in pill was, however, simulated using a computer model of people developing countries, given the limited health care resources. with newly diagnosed diabetes in the United States (Sorensen However, these interventions may be reasonable for selected and others 2004), and the assessment found that a polypill subpopulation groups, such as those who can afford them. intervention would cost US$11,000 per QALY gained. The Metformin therapy for preventing type 2 diabetes among intervention would be cost saving if such a pill cost US$1.28 people at high risk, such as those with prediabetes, is feasible or less per day. We estimated that the cost-effectiveness ratio of because the drug is affordable in many developing countries; the polypill ranged from US$560 to US$1,280 per QALY however, the intervention may not be good value for money. gained for the six developing regions. This result was sensitive Cholesterol control intervention for people with diabetes falls to changes in the cost of the intervention, but the intervention into the same category. The cost-effectiveness of both these remained cost-effective within the most likely ranges of its cost interventions would improve if the costs of the drug could be (Sorensen and others 2004). A barrier to this intervention, in lowered. addition to the feasibility of producing such pill, is that its The aim of intensive glucose control is to lower the glucose benefits and side effects would still have to be established in a level of a person with diabetes to a level close to that of a per- randomized clinical trial. son without diabetes. Implementing this intervention is a lower priority, mainly because of its relatively low cost-effectiveness in the context of the limited health care resources in develop- Cost-Effectiveness of Diabetes Education ing countries. Although the U.K. Prospective Diabetes Study People with diabetes play a central role in managing their dis- clearly demonstrates that lowering glucose levels can prevent or ease. Thus, diabetes education is an integral part of diabetes delay long-term diabetes complications (UKPDS Group 1998), care. The goal of diabetes education is to support the efforts the marginal return on very intensive glucose control in devel- of people with diabetes to understand the nature of their oping countries was relatively small. illness and its treatment; to identify emergency health prob- Screening for undiagnosed diabetes is a low-priority inter- lems at early, reversible stages; to adhere to self-care practices; vention mainly because of its relatively high cost per QALY. and to make necessary changes to their health habits However, screening for undiagnosed diabetes can be a worth- (International Diabetes Federation 2003b). Health providers while intervention for subpopulation groups, such as those that can deliver diabetes education programs in various settings. have a high prevalence of undiagnosed diabetes. In the United Evaluating the effectiveness of health education is challenging States, for example, screening for undiagnosed diabetes among because of the difficulty of separating out its effect from that African Americans was estimated to be 10 times more cost- of other interventions. Nevertheless, a review of literature effective than screening among other population groups (CDC published in the United States suggests that self-management Diabetes Cost-Effectiveness Study Group 1998). In addition, diabetes education may be cost-effective (Klonoff and screening for undiagnosed diabetes may be a worthwhile inter- Schwartz 2000). vention for patients with risk factors for other chronic diseases, Training in diabetes self-management reduces medical such as hypertension, high lipid profiles, and prediabetes. costs for diabetes care in developing countries in the short Annual screening for microalbuminuria was a low-priority term. A multicenter intervention study in 10 Latin American intervention because screening added costs with no significant countries demonstrated that an education program could benefits. Treating all persons with diabetes with ACE inhibitors reduce the cost of drugs by 62 percent (International Diabetes was a better treatment option than screening for microalbu- Federation 2003b), and another program in Argentina found a minuria and treating only those who have the condition. reduction in diabetes-related costs of 38 percent (Gagliardino and Etchegoyen 2001). Because the costs of education pro- grams are generally low, the intervention may be cost-effective. Cost-Effectiveness of a Polypill to Prevent CVD Training patients to better manage their diabetes is also A meta-analysis estimated that a hypothetical polypill could feasible because of its low technical complexity, low capital reduce the risk of CVD by 80 percent among all people over requirements, and cultural acceptability. Thus, diabetes educa- 55 or people with diabetes of any age (Wald and Law 2003). tion should be a high-priority intervention for all developing This hypothetical pill is a combination of three half-dose regions. 598 | Disease Control Priorities in Developing Countries | K. M. Venkat Narayan, Ping Zhang, Alka M. Kanaya, and others LESSONS AND EXPERIENCE had hypertension, 53 percent had high cholesterol, and 45 per- cent had abnormal triglycerides (Gagliardino, de la Hera, and A number of lessons can be learned from the experiences Siri 2001). in countries where the interventions described have been implemented. Quality of Diabetes Care Small, single-site studies indicate that several interventions to Prevention improve quality of care at the patient, provider, or system lev- Data are sparse on community- or population-based strategies els are promising (Narayan and others 2004). A systematic for preventing diabetes along with other chronic diseases such review (Renders and others 2001) found that multifaceted pro- as CVD. Available studies on preventing type 2 diabetes have fessional interventions may enhance providers' performance in used clinic-based approaches targeted at high-risk groups, and managing diabetes care; that organizational interventions researchers generally agree that type 2 diabetes can be prevent- involving regularly contacting and tracking patients by means ed or its onset delayed. Putting these results into practice, how- of computerized tracking systems or through nurses can also ever, is fraught with difficulties and unanswered questions, improve diabetes management; that patient-oriented interven- such as the following: tions can improve patients' outcomes; and that nurses can play an important role in patient-oriented interventions by educat- · Who would benefit from diabetes prevention? ing patients and facilitating patients' adherence to treatment · How can those who may benefit be identified? regimes. (See also chapter 70.) · What are the costs and cost-effectiveness of diabetes pre- Interventions that could modify providers' behavior include vention at a population level? education as part of more complex interventions that also · How should results be extrapolated from developed countries focus on systems and on the organization of practices--for to developing countries, whose priorities and approaches example, feedback on performance, reminder systems, consen- may be different? sus development, and clinical practice guidelines. Potential systemic interventions include the use of continuous quality improvement techniques; feedback on performance; physician Treatment incentives for quality; nurses to provide diabetes care (which is The quality of diabetes care generally remains suboptimal typically provided by physicians); computerized reminder sys- worldwide, regardless of a particular country's level of devel- tems for providers, alone or in combination with a perform- opment, health care system, or population (Engelgau and oth- ance feedback program; patient-tracking or other reminder ers 2003; Garfield and others 2003). The Costs of Diabetes in systems to improve regular follow-up; dedicated blocks of time Europe--Type 2 study, conducted in eight European countries, set aside for diabetes patients in primary care practices; team found suboptimal diabetes care in each country (Liebl, Mata, care; electronic medical record systems; and other methods, and Eschwege 2002). In the United States, population-based such as telephone and mailing reminders, chart stickers, and surveys in the 1990s among adults age 18 to 75 with diabetes flow sheets to prompt both providers and patients. found that only 63 percent of them had had a dilated eye exam- Interventions that empower patients can be successful com- ination and only 55 percent had had a foot examination with- ponents of diabetes programs. A systems-oriented approach in the past year, 18 percent had poor glycemic control, 42 per- using manual or computerized systems that remind patients to cent had good cholesterol control, and 66 percent had a blood make follow-up appointments and that prompt staff members pressure within the normal range (Saaddine and others 2002). to generate reminder cards for patients can improve compli- The Diabcare-Asia project was conducted in the late 1990s. ance with follow-up and enhance efficiency of office practices. Results from India, Singapore, and Taiwan (China) found that In addition, comprehensive implementation of multiple risk- in 1998, 32 to 50 percent of the diabetic population had poor factor interventions in real-life settings has been shown to glycemic control (equivalent to HbA1c 8 percent), 43 to reduce vascular events by more than 50 percent among people 67 percent had high cholesterol (greater than 5.2 millimoles per with diabetes (Gaede and others 2003). deciliter), and 47 to 54 percent had an abnormal level of triglyc- The Institute of Medicine Committee on Quality of Health eride (greater than 1.7 millimoles per deciliter) (W. R. Lee and Care in America (2001) argues strongly that newer systems of others 2001; Raheja and others 2001). Data from Latin America care and newer ways of thinking are needed to tackle complex and the Caribbean showed that 41 percent of people with type diseases such as diabetes. Furthermore, the model of the process 1 diabetes and 57 percent of those with type 2 diabetes had of change in a simple mechanical system is woefully inadequate poor glucose control. Of those with type 2 diabetes, 56 percent for dealing with the complex, interactive, and interconnected Diabetes: The Pandemic and Potential Solutions | 599 adaptive systems in which diabetes is prevented and treated. Medicine Committee on Quality of Health Care in America Applied research, designed to encompass the system as a whole 2001).Computer models suitable for assessing cost-effectiveness and not simply its component parts, can enhance our under- and for forecasting the burden in developing countries are need- standing of complex health care dynamics for chronic diseases ed. Operational research aimed at understanding the tradeoffs (Fraser and Greenhalgh 2001; Plsek and Greenhalgh 2001). and the best mix of resource allocation for diabetes and chronic disease care in developing countries is also needed. RESEARCH AND DEVELOPMENT AGENDA Basic Research The following subsections discuss the major issues for research Further strategic unraveling of the genetic basis of type 2 dia- and development. betes and gene-environment interactions may help explain the diabetes epidemic and provide better understanding of the pathophysiology of the disease. It may also may lead to better Prevention prevention and treatment strategies. Understanding the role of Well-designed community-based studies of primary preven- prenatal influences, especially in developing countries, may tion for type 2 diabetes are needed, especially as part of multi- offer productive opportunities for interventions. Because of the factorial interventions, in developing countries. Research is also increasing occurrence of type 2 diabetes in children, as well as needed into safer and cheaper drugs to prevent diabetes when the role of obesity in accelerating the onset of type 1 diabetes, lifestyle intervention either is not feasible or has failed. In addi- further research into the typology and classification of diabetes tion, we need to know the long-term effects of diabetes preven- is vital. The rapid industrialization and economic development tion on CVD and other outcomes. More effective and cheaper being experienced by several developing countries may make ways to prevent the major complications of diabetes are also research into the role of socioeconomic factors, urban stress, needed. Other areas also deserving of research include nonin- and lifestyle factors on the causation of diabetes productive. vasive methods for monitoring blood glucose and more effec- tive and efficient ways of screening for prediabetes, diabetes, and early diabetes complications. Evidence of the benefits of CONCLUSIONS diabetes education on outcomes is lacking, and organized research to assess effective components of diabetes education A growing diabetes pandemic is unfolding with rapid increases and their impact on control of risk factors and long-term out- in the prevalence of type 2 diabetes. The direct health care costs comes should be a priority. of diabetes worldwide amount to 2003 US$129 billion per year. Estimates indicate that developing countries spend between Epidemiological and Economics Research 2.5 and 15.0 percent of their annual direct health budgets on diabetes care, and families with diabetic members spend 15 to Scant data are available on the future burden of diabetes and its 25 percent of their incomes on diabetes care. complications in developing countries. Data on trends in and A whole array of effective interventions to prevent diabetes the effects of risk factors for diabetes in developing countries-- and its complications is available, and we have attempted to obesity; birthweight; physical inactivity; television viewing; assess their potential cost-effectiveness in developing regions. dietary factors; fast foods; socioeconomic factors; and effects of Using these estimations and a qualitative assessment of the fea- urbanization, industrialization, globalization, and stress--are sibility of implementation, we have prioritized available inter- also sparse. Low-cost ways to obtain such data in a standardized ventions into the following three categories: manner may be worth considering. More data are also needed on the costs of diabetes, the impact of the disease on quality of · level 1--cost saving and highly feasible life, and the cost-effectiveness of various interventions in the · level 2--cost saving or cost less than US$1,500 per QALY context of developing countries (International Diabetes but pose some feasibility challenges Federation 2003a). · level 3--cost between US$1,640 and US$8,550 per QALY and pose significant feasibility challenges. Health Systems and Operational Research Greater emphasis on translation research is needed. Well- Table 30.4 presents a summary of all major diabetes inter- designed and standardized studies of quality of care and out- ventions, major health effects of the interventions, and level of comes will help (TRIAD Study Group 2002). Research aimed at implementation priority. understanding system-level complexity and finding ways to In addition, we propose diabetes education as an essential deliver chronic disease care that takes such complexity into intervention. However, more organized research into the account is also likely to yield profitable results (Institute of precise components of diabetes education and its effect on 600 | Disease Control Priorities in Developing Countries | K. M. Venkat Narayan, Ping Zhang, Alka M. Kanaya, and others Table 30.4 Key Cost-Effective Interventions for Preventing and Treating Diabetes and Its Complications Intervention Description Applicable population Major effect Level 1a · Glycemic control in people with Insulin, oral glucose-lowering agents, diet People with diabetes, all ages, HbA1c Reduction in microvascular disease poor control and exercise greater than 9 percent · Blood pressure control Blood pressure control medications People with diabetes, hypertensive, Reduction in macrovascular disease, all ages microvascular disease, and mortality · Foot care Patient and provider education, foot People with diabetes, middle-aged Reduction in serious foot diseases and examination, foot hygiene, and appropriate or older amputations footwear Level 2b · Preconception care for women Patient self-management Women with diabetes who plan to Reduction in HbA1c level and hospital of reproductive age become pregnant expenses of the mother and baby · Lifestyle intervention to prevent Behavioral change, including diet and People who are at high risk (for example, Reduction in type 2 diabetes incidence diabetes physical activity, to reduce bodyweight prediabetes for type 2 diabetes) by 58 percent · Influenza vaccination Vaccination Elderly people with diabetes Reduction in hospitalizations, respiratory conditions, and mortality · Detection and treatment of eye Eye examination to screen for and treat eye People with diabetes, middle-aged Reduction in serious vision loss diseases diseases or older · ACE inhibitors Angiotensin-converting enzyme medication People with diabetes Reduction in nephropathy, cardiovascular disease, and death · Smoking cessation Physician counseling and nicotine People with diabetes, all ages, Increase in quitting rate and reduction replacement therapy smokers in cardiovascular disease Level 3c · Metformin therapy for Metformin medication People who are at high risk Reduction in type 2 diabetes incidence preventing diabetes (for example, prediabetes for by 33 percent type 2 diabetes) · Intensive glucose control Insulin, oral glucose-lowering agents, Diabetes, all ages, with HbA1c less Reduction in microvascular disease or both than 9 percent · Lipid control Cholesterol-lowering medication Diabetes, all ages, with high Reduction in cardiovascular disease cholesterol events and mortality · Screening for microalbuminuria Screening for microalbuminuria and treating Diabetes, all ages Reduction in kidney diseases those who test positive · Screening for undiagnosed Screening for undiagnosed diabetes and People who are at high risk for type 2 Reduction in microvascular disease diabetes treating those who test positive diabetes Essential background interventiond Diabetes education Patient self-management Diabetes, all ages Reduction in HbA1c level and better compliance with lifestyle changes Other promising interventione Polypill Hypothetical pill combining low doses of Diabetes, all ages Reduction in cardiovascular disease antihypertensive medication, aspirin, statin, and folate Source: Authors. a. Level 1 interventions are cost saving and highly feasible. b. Level 2 interventions are cost saving or cost less than US$1,500 per quality-adjusted life year but pose feasibility challenges. c. Level 3 interventions cost between US$1,640 and US$8,550 per quality-adjusted life year and pose significant feasibility challenges. d. Diabetes education is the backbone on which many diabetes interventions depend, but empirical data on the effectiveness of diabetes education on outcomes and on the precise components of diabetes education are still lacking. e. An intervention that appears promising but needs further research to document its effectiveness and/or safety. The polypill is only a theoretical concept at this time and is not available for implementation. Diabetes: The Pandemic and Potential Solutions | 601 long-term outcomes is needed. We also propose that further Fung, T. T., F. B. Hu, M. A. Pereira, S. Liu, M. J. Stampfer, G. A. Colditz, and research be launched in relation to the novel and potentially others. 2002. "Whole-Grain Intake and the Risk of Type 2 Diabetes: A Prospective Study in Men." American Journal of Clinical Nutrition 76 promising polypill. (3): 535­40. Finally, this chapter suggests a number of interventions at Gaede, P., P. Vedel, N. Larsen, G. V. Jensen, H. H. Parving, and O. Pedersen. the level of the patient, provider, and system that could help 2003. "Multifactorial Intervention and Cardiovascular Disease in address the overall suboptimal quality of diabetes care; notes Patients with Type 2 Diabetes." New England Journal of Medicine 348 (5): 383­93. the possible benefits of making important drugs available at Gagliardino, J. J., H. M. de la Hera, and F. Siri. 2001. "Evaluation of the cheaper costs in developing countries; and suggests some Quality of Care for Diabetic Patients in Latin America" (in Spanish). research priorities for developing regions. Revista Panamericana de Salud Pública 10 (5): 309­17. Gagliardino, J. J., and G. Etchegoyen. 2001."A Model Educational Program for People with Type 2 Diabetes: A Cooperative Latin American REFERENCES Implementation Study (PEDNID-LA)." Diabetes Care 24 (6): 1001­7. Gagliardino, J. J., E. M. Olivera, H. Barragan, and R. A. Puppo. 1993. "A Adler, A. I., E. J. Boyko, C. D. Schraer, and N. J. Murphy. 1994. "Lower Simple Economic Evaluation Model for Selecting Diabetes Health Care Prevalence of Impaired Glucose Tolerance and Diabetes Associated Strategies." Diabetic Medicine 10 (4): 351­54. with Daily Seal Oil or Salmon Consumption among Alaska Natives." Garfield, S. A., S. Malozowski, M. H. Chin, K. M. Venkat Narayan, R. E. Diabetes Care 17 (12): 1498­1501. Glasgow, L. W. Green, and others. 2003. "Considerations for Diabetes Akerblom, H. K., and M. Knip. 1998. "Putative Environmental Factors in Translational Research in Real-World Settings." Diabetes Care 26 (9): Type 1 Diabetes." Diabetes/Metabolism Review 14 (1): 31­67. 2670­74. American Diabetes Association. 2004. "Diagnosis and Classification of Geiss, L. S., W. H. Herman, and P. J. Smith. 1995. "Mortality among Diabetes Mellitus." Diabetes Care 27 (Suppl. 1): S5­10. Persons with Non-Insulin Dependent Diabetes." In Diabetes in Barcelo, A., C. Aedo, S. Rajpathak, and S. Robles. 2003. "The Cost of America, 2nd ed., ed. National Diabetes Data Group, 233­58. Bethesda, Diabetes in Latin America and the Caribbean." Bulletin of the World MD: National Institutes of Health. Health Organization 81 (1): 19­27. Golan, L., J. D. Birkmeyer, and H. G.Welch. 1999."The Cost-Effectiveness of CDC (U.S. Centers for Disease Control and Prevention) Diabetes Cost- Treating All Patients with Type 2 Diabetes with Angiotensin-Converting Effectiveness Study Group. 1998. "The Cost-Effectiveness of Screening Enzyme Inhibitors." Annals of Internal Medicine 131 (9): 660­67. for Type 2 Diabetes." Journal of the American Medical Association 280 Gregg, E. W., G. L. Beckles, D. F. Williamson, S. G. Leveille, J. A. Langlois, (20): 1757­63. M. M. Engelgau, and others. 2000. "Diabetes and Physical Disability Clarke, P., A. Gray, and R. Holman. 2002. "Estimating Utility Values for among Older U.S. Adults." Diabetes Care 23 (9): 1272­77. Health States of Type 2 Diabetic Patients Using the EQ-5D (UKPDS Haffner, S. M. 1998. "Epidemiology of Type 2 Diabetes: Risk Factors." 62)." Medical Decision Making 22 (4): 340­49. Diabetes Care 21 (Suppl. 3): C3­6. Dabelea, D., R. L. Hanson, R. S. Lindsay, D. J. Pettitt, G. Imperatore, M. M. Hu, F. B., T. Y. Li, G. A. Colditz, W. C. Willett, and J. E. Manson. 2003. Gabir, and others. 2000. "Intrauterine Exposure to Diabetes Conveys "Television Watching and Other Sedentary Behaviors in Relation to Risks for Type 2 Diabetes and Obesity: A Study of Discordant Risk of Obesity and Type 2 Diabetes Mellitus in Women." Journal of the Sibships." Diabetes 49 (12): 2208­11. American Medical Association 289 (14): 1785­91. Diabetes Prevention Program Research Group. Forthcoming. "The Cost- Hu, F. B., R. M. van Dam, and S. Liu. 2001. "Diet and Risk of Type II Effectiveness of Diet and Physical Activity or Metformin in the Diabetes: The Role of Types of Fat and Carbohydrate." Diabetologia 44 Prevention of Type 2 Diabetes among Adults with Impaired Glucose (7): 805­17. Tolerance." Annals of Internal Medicine. Institute of Medicine Committee on Quality of Health Care in America. Earnshaw, S. R., A. Richter, S. W. Sorensen, T. J. Hoerger, K. A. Hicks, M. 2001. Crossing the Quality Chasm: A New Health System for the 21st Engelgau, and others. 2002. "Optimal Allocation of Resources across Century. Washington, DC: National Academy Press. Four Interventions for Type 2 Diabetes." Medical Decision Making 22 (Suppl. 5): S80­91. International Diabetes Federation. 2003a. Cost-Effective Approaches to Diabetes Care and Prevention. Brussels: International Diabetes Engelgau, M. M., K. M. Narayan, and W. H. Herman. 2000. "Screening for Federation. Type 2 Diabetes." Diabetes Care 23 (10): 1563­80. . 2003b. Diabetes Atlas. 2nd ed. Brussels: International Diabetes Engelgau, M. M., K. M. Narayan, J. B. Saaddine, and F. Vinicor. 2003. Federation. "Addressing the Burden of Diabetes in the 21st Century: Better Care and Primary Prevention." Journal of the American Society of Nephrology Kanaya, A. M., and K. M. Narayan. 2003. "Prevention of Type 2 Diabetes: 14 (7 Suppl. 2): S88­91. Data from Recent Trials." Primary Care 30 (3): 511­26. Eriksson, K. F., and F. Lindgarde. 1991. "Prevention of Type 2 (Non- King, H., R. E. Aubert, and W. H. Herman. 1998. "Global Burden of Insulin-Dependent) Diabetes Mellitus by Diet and Physical Exercise. Diabetes, 1995­2025: Prevalence, Numerical Estimates, and The 6-Year Malmo Feasibility Study." Diabetologia 34 (12): 891­98. Projections." Diabetes Care 21 (9): 1414­31. Everson, S. A., S. C. Maty, J. W. Lynch, and G. A. Kaplan. 2002. Klonoff, D. C., and D. M. Schwartz. 2000. "An Economic Analysis of "Epidemiologic Evidence for the Relation between Socioeconomic Interventions for Diabetes." Diabetes Care 23 (3): 390­404. Status and Depression, Obesity, and Diabetes." Journal of Knowler, W. C., E. Barrett-Connor, S. E. Fowler, R. F. Hamman, J. M. Psychosomatic Research 53 (4): 891­95. Lachin, E. A. Walker, and others. 2002. "Reduction in the Incidence Ford, E. S., D. F. Williamson, and S. Liu. 1997. "Weight Change and of Type 2 Diabetes with Lifestyle Intervention or Metformin." New Diabetes Incidence: Findings from a National Cohort of US Adults." England Journal of Medicine 346 (6): 393­403. American Journal of Epidemiology 146 (3): 214­22. Lee, W. L., A. M. Cheung, D. Cape, and B. Zinman. 2000. "Impact of Fraser, S. W., and T. Greenhalgh. 2001. "Coping with Complexity: Diabetes on Coronary Artery Disease in Women and Men: A Meta- Educating for Capability." British Medical Journal 323 (7316): 799­803. analysis of Prospective Studies." Diabetes Care 23 (7): 962­68. 602 | Disease Control Priorities in Developing Countries | K. M. Venkat Narayan, Ping Zhang, Alka M. Kanaya, and others Lee, W. R., H. S. Lim, A. C. Thai, W. L. Chew, S. Emmanuel, L. G. Goh, and National Health and Nutrition Examination Survey." American Journal others. 2001. "A Window on the Current Status of Diabetes Mellitus in of Public Health 91 (1): 76­83. Singapore--The Diabcare-Singapore 1998 Study." Singapore Medical Rowley, K. G., J. D. Best, R. McDermott, E. A. Green, L. S. Piers, and Journal 42 (11): 501­507. K. O'Dea. 1997. "Insulin Resistance Syndrome in Australian Aboriginal Liebl, A., M. Mata, and E. Eschwege. 2002. "Evaluation of Risk Factors for People." Clinical and Experimental Pharmacology and Physiology 24 Development of Complications in Type II Diabetes in Europe." (9­10): 776­81. Diabetologia 45 (7): S23­28. Ryerson, B., E. F. Tierney, T. J. Thompson, M. M. Engelgau, J. Wang, E. W. Manuel, D. G., and S. E. Schultz. 2004. "Health-Related Quality of Life and Gregg, and others. 2003. "Excess Physical Limitations among Adults Health-Adjusted Life Expectancy of People with Diabetes in Ontario, with Diabetes in the U.S. population, 1997­1999." Diabetes Care 26 (1): Canada, 1996­1997." Diabetes Care 27 (2): 407­14. 206­10. Mathers, C. D., C. Stein, D. Ma Fat, C. Rao, M. Inoue, N. Tomijima, and Saaddine, J. B., M. M. Engelgau, G. L. Beckles, E. W. Gregg, T. J. Thompson, others. 2000. Global Burden of Disease 2000: Version 2 Methods and and K. M. Narayan. 2002."A Diabetes Report Card for the United States: Results. Global Programme on Evidence for Health Policy Discussion Quality of Care in the 1990s."Annals of Internal Medicine 136 (8): 565­74. Paper Series. Geneva: World Health Organization. Schulze, M. B., J. E. Manson, D. S. Ludwig, G. A. Colditz, M. J. Stampfer, Mulligan, J., J. A. Fox-Rushby, T. Adam, B. Johns, and A. Mills. 2003. "Unit W. C. Willett, and others. 2004. "Sugar-Sweetened Beverages, Weight Costs of Health Care Inputs in Low and Middle Income Regions." Gain, and Incidence of Type 2 Diabetes in Young and Middle-Aged Disease Control Priorities Project Working Paper 9, Fogarty Women." Journal of the American Medical Association 292 (8): 927­34. International Center, National Institutes of Health, Bethesda, MD. Sorensen, S., M. Engelgau, T. Hoerger, K. Hicks, K. Narayan, D. Williamson, Narayan, K. M., E. Benjamin, E. W. Gregg, S. L. Norris, and M. M. Engelgau. and others. 2004. "Assessment of the Benefits from a Polypill to Reduce 2004."Diabetes Translation Research: Where Are We and Where Do We Cardiovascular Disease among Persons with Type 2 Diabetes Mellitus." Want to Be?" Annals of Internal Medicine 140 (11): 958­63. Poster presented at the 64th Annual Scientific Sessions of the American Diabetes Association, Orlando, Florida, June 4­8, 2004. Narayan, K. M., J. P. Boyle, T. J. Thompson, S. W. Sorensen, and D. F. Williamson. 2003. "Lifetime Risk for Diabetes Mellitus in the United Stevens, J., K. Ahn, Juhaeri, D. Houston, L. Steffan, and D. Couper. 2002. States." Journal of the American Medical Association 290 (14): 1884­90. "Dietary Fiber Intake and Glycemic Index and Incidence of Diabetes in African-American and White Adults: The ARIC Study." Diabetes Care Norris, S. L., M. M. Engelgau, and K. M. Narayan. 2001. "Effectiveness of 25 (10): 1715­21. Self-Management Training in Type 2 Diabetes: A Systematic Review of Randomized Controlled Trials." Diabetes Care 24 (3): 561­87. TRIAD Study Group. 2002. "The Translating Research into Action for Diabetes (TRIAD) Study: A Multicenter Study of Diabetes in Managed Paeratakul, S., J. C. Lovejoy, D. H. Ryan, and G. A. Bray. 2002."The Relation Care." Diabetes Care 25 (2): 386­89. of Gender, Race, and Socioeconomic Status to Obesity and Obesity Comorbidities in a Sample of U.S. Adults." International Journal of Tuomilehto, J., J. Lindstrom, J. G. Eriksson, T. T. Valle, H. Hamalainen, Obesity and Related Metabolic Disorders 26 (9): 1205­10. P. Ilanne-Parikka, and others. 2001. "Prevention of Type 2 Diabetes Mellitus by Changes in Lifestyle among Subjects with Impaired Glucose Pan, X. R., G. W. Li, Y. H. Hu, J. X. Wang, W. Y. Yang, Z. X. An, and others. Tolerance." New England Journal of Medicine 344 (18): 1343­50. 1997. "Effects of Diet and Exercise in Preventing NIDDM in People UKPDS (U.K. Prospective Diabetes Study) Group. 1998. "Intensive Blood- with Impaired Glucose Tolerance: The Da Qing IGT and Diabetes Glucose Control with Sulphonylureas or Insulin Compared with Study." Diabetes Care 20 (4): 537­44. Conventional Treatment and Risk of Complications in Patients with Pettitt, D. J., M. R. Forman, R. L. Hanson, W. C. Knowler, and P. H. Bennett. Type 2 Diabetes (UKPDS 33)." Lancet 352 (9131): 837­53. 1997. "Breastfeeding and Incidence of Non-Insulin-Dependent U.S. Preventive Services Task Force. 1996. Guide to Clinical Preventive Diabetes Mellitus in Pima Indians." Lancet 350 (9072): 166­68. Services: Report of the U.S. Preventive Services Task Force, 2nd ed. Plsek, P. E., and T. Greenhalgh. 2001."Complexity Science: The Challenge of Washington, DC: Office of Disease Prevention and Health Promotion, Complexity in Health Care."British Medical Journal 323 (7313): 625­28. U.S. Government Printing Office. Qiao, Q., D. E. Williams, G. Imperatore, K. M. Venkat Narayan, and Vijan, S., T. P. Hofer, and R. A. Hayward. 2000. "Cost-Utility Analysis of J. Tuomilehto. 2004. "Epidemiology and Geography of Type 2 Diabetes Screening Intervals for Diabetic Retinopathy in Patients with Type 2 Mellitus." In International Textbook of Diabetes Mellitus, 3rd ed., ed. Diabetes Mellitus." Journal of the American Medical Association 283 (7): R. A. DeFronzo and others, 33­56. Chichester, U.K.: John Wiley & Sons. 889­96. Ragnarson, T. G., and J. Apelqvist. 2001. "Prevention of Diabetes-Related Villarreal-Rios, E., A. M. Salinas-Martinez, A. Medina-Jauregui, M. E. Foot Ulcers and Amputations: A Cost-Utility Analysis Based on Garza-Elizondo, G. Nunez-Rocha, and E. R. Chuy-Diaz. 2000. "The Markov Model Simulations." Diabetologia 44 (11): 2077­87. Cost of Diabetes Mellitus and Its Impact on Health Spending in Raheja, B. S., A. Kapur, A. Bhoraskar, S. R. Sathe, L. N. Jorgensen, S. R. Mexico." Archives of Medical Research 31 (5): 511­14. Moorthi, and others. 2001. "DiabCare Asia--India Study: Diabetes Wald, N. J., and M. R. Law. 2003. "A Strategy to Reduce Cardiovascular Care in India--Current Status." Journal of the Association of Physicians Disease by More Than 80%." British Medical Journal 326 (7404): 1419. of India 49: 717­22. Williams, D. E., W. C. Knowler, C. J. Smith, R. L. Hanson, J. Roumain, A. Rajala, U., M. Laakso, Q. Qiao, and S. Keinanen-Kiukaanniemi. 1998. Saremi, and others. 2001. "The Effect of Indian or Anglo Dietary "Prevalence of Retinopathy in People with Diabetes, Impaired Glucose Preference on the Incidence of Diabetes in Pima Indians." Diabetes Tolerance, and Normal Glucose Tolerance." Diabetes Care 21 (10): Care 24 (5): 811­16. 1664­69. WHO (World Health Organization). 2004. "Global Burden of Disease for Renders, C. M., G. D. Valk, S. J. Griffin, E. H. Wagner, V. J. Eijk, and W. J. the Year 2001 by World Bank Region, for Use in Disease Control Assendelft. 2001. "Interventions to Improve the Management of Priorities in Developing Countries." 2nd ed. http://www.fic.nih.gov/ Diabetes in Primary Care, Outpatient, and Community Settings: A dcpp/gbd.html. Systematic Review." Diabetes Care 24 (10): 1821­33. Yajnik, C. S. 2001."The Insulin Resistance Epidemic in India: Fetal Origins, Rich, S. S. 1990. "Mapping Genes in Diabetes. Genetic Epidemiological Later Lifestyle, or Both?" Nutrition Reviews 59 (1, part 1): 1­9. Perspective." Diabetes 39 (11): 1315­19. Young, T. K., P. J. Martens, S. P. Taback, E. A. Sellers, H. J. Dean, M. Cheang, Robbins, J. M., V. Vaccarino, H. Zhang, and S. V. Kasl. 2001. and others. 2002. "Type 2 Diabetes Mellitus in Children: Prenatal and "Socioeconomic Status and Type 2 Diabetes in African American and Early Infancy Risk Factors among Native Canadians." Archives of Non-Hispanic White Women and Men: Evidence from the Third Pediatrics and Adolescent Medicine 156 (7): 651­55. Diabetes: The Pandemic and Potential Solutions | 603 Chapter 31 Mental Disorders Steven Hyman, Dan Chisholm, Ronald Kessler, Vikram Patel, and Harvey Whiteford Mental disorders are diseases that affect cognition, emotion, Twin studies make it clear that environmental risk factors and behavioral control and substantially interfere both with the also play an important role in mental disorders; concordance for ability of children to learn and with the ability of adults to disease among identical twins, although substantially higher function in their families, at work, and in the broader society. than among nonidentical twins, is still well below 100 percent Mental disorders tend to begin early in life and often run a (Kendler and others 2003). However, as is the case for genetic chronic recurrent course. They are common in all countries factors, investigation of environmental risk factors has proved where their prevalence has been examined. Because of the difficult. For schizophrenia, where nongenetic components of combination of high prevalence, early onset, persistence, and risk may include obstetrical complications and season of birth impairment, mental disorders make a major contribution to (Mortensen and others 1999), perhaps as a proxy for infections total disease burden. Although most of the burden attributable early in life, research has been hampered by the modest proven to mental disorders is disability related, premature mortality, effect of the nongenetic risk factors identified to date. For especially from suicide, is not insignificant. Table 31.1 summa- depression, anxiety, and substance use disorders, where envi- rizes discounted disability-adjusted life years (DALYs) for ronmental risk factors are more robust, adverse circumstances selected psychiatric conditions in 2001. associated with risk, such as early childhood abuse, violence, Mental disorders have complex etiologies that involve inter- poverty, and stress (Patel and Kleinman 2003) correlate with actions among multiple genetic and nongenetic risk factors. multiple disorders and could be affected by selection bias as well Gender is related to risk in many cases: males have higher rates as by bias associated with self-reporting. Generalizable, prospec- of attention deficit hyperactivity disorder, autism, and sub- tive cross-cultural studies are needed to delineate nongenetic stance use disorders; females have higher rates of major depres- risk factors more clearly. Posttraumatic stress disorder (PTSD) sive disorder, most anxiety disorders, and eating disorders. is the mental disorder for which clear environmental triggers are Biochemical and morphological abnormalities of the brain best documented. Even here, though, enormous interindividual associated with schizophrenia, autism, mood, and anxiety dis- variability occurs in the threshold of stress severity associated orders are being identified using approaches such as post- with PTSD as well as in the evidence from twin studies of genetic mortem analysis and noninvasive neuroimaging. Major world- influences on stress reactivity in triggering PTSD. wide efforts under way to identify risk-conferring genes for The last half of the 20th century saw enormous progress in mental disorders are proving challenging, but initial results are the development of treatments for mental disorders. Beginning promising. Identifying the gene or genes causing or creating in the early 1950s, effective psychotropic drugs were discovered vulnerability for a disorder should help us understand what that treated the symptoms of schizophrenia, bipolar disorder, goes wrong in the brain to produce mental illness and should major depression, anxiety disorders, obsessive-compulsive have a clinical effect by contributing to improved diagnostics disorder, attention deficit hyperactivity disorder, and others. and therapeutics (Hyman 2000). The safety and efficacy of antipsychotic, mood-stabilizing, 605 Table 31.1 Disease Burden of Selected Major Psychiatric Disorders, by World Bank Region World Bank region Sub-Saharan Latin America and Middle East and Europe and East Asia and High-income Africa the Caribbean North Africa Central Asia South Asia the Pacific countries World Total population (millions) 668 526 310 477 1,388 1,851 929 6,159 Total disease burden 344,754 104,287 65,570 116,502 408,655 346,941 149,161 1,535,870 (thousands of DALYs) Total neuropsychiatric 15,151 18,781 8,310 14,106 37,734 42,992 31,230 168,304 disease burden (thousands of DALYs) Total burden (thousands of discounted DALYs per year) Schizophrenia 1,146 1,078 696 778 2,896 3,934 1,115 11,643 Bipolar disorder 1,204 883 567 668 2,237 3,118 1,056 9,733 Depression 3,275 5,219 2,027 4,268 14,582 14,054 8,408 51,833 Panic disorder 519 409 264 340 1,081 1,401 536 4,550 Total burden (DALYs per year per 1 million population) Schizophrenia 1,716 2,049 2,247 1,630 2,087 2,126 1,201 1,894 Bipolar disorder 1,803 1,678 1,830 1,400 1,612 1,685 1,137 1,583 Depression 4,905 9,919 6,544 8,944 10,507 7,594 9,054 8,431 Panic disorder 777 777 852 713 779 757 577 740 Percentage of total disease burden Schizophrenia 0.33 1.03 1.06 0.67 0.71 1.13 0.75 0.76 Bipolar disorder 0.35 0.85 0.86 0.57 0.55 0.90 0.71 0.63 Depression 0.95 5.00 3.09 3.66 3.57 4.05 5.64 3.37 Panic disorder 0.15 0.39 0.40 0.29 0.26 0.40 0.36 0.30 Percentage of neuropsychiatric disease burden Schizophrenia 7.56 5.74 8.38 5.52 7.67 9.15 3.57 6.92 Bipolar disorder 7.95 4.70 6.82 4.74 5.93 7.25 3.38 5.78 Depression 21.62 27.79 24.39 30.26 38.64 32.69 26.92 30.80 Panic disorder 3.43 2.18 3.18 2.41 2.86 3.26 1.72 2.70 Source: WHO Global Burden of Disease 2001 estimates recalculated by World Bank region (http://www.fic.nih.gov/dcpp/gbd.html). antidepressant, anxiolytic, and stimulant drugs have been edge exists to guide treatment. It is particularly unfortunate, established through a large number of randomized clinical therefore, that timely diagnoses and the application of trials. Psychosocial treatments have been developed and tested research-based treatments significantly lag behind the state of using modern methodologies. Brief, symptom-focused psy- knowledge in industrial and developing countries alike. As a chotherapies such as cognitive-behavioral therapies have been result, substantial opportunities exist to decrease the enormous shown to be efficacious for panic disorder, phobias, obsessive- burden attributable to mental disorders worldwide by closing compulsive disorder, and major depression. the gap between what we know and what we do. There is, however, an important caveat about the current Mental disorders are stigmatized in many countries and knowledge base for treatment. As is the case for almost all of cultures (Weiss and others 2001). Stigma has been facilitated medicine, randomized clinical trials have been performed by the slow emergence of convincing scientific explanations for largely with highly selected populations in specialized research the etiologies of mental disorders and by the mistaken belief settings in industrial countries. A need exists to subject existing that symptoms are caused by a lack of will power or reflect some treatments to effectiveness trials in more representative popu- moral taint. Recent scientific findings combined with educa- lations and diverse settings, especially in developing countries. tional efforts in some countries have begun to reduce the stigma That limitation notwithstanding, a substantial body of knowl- (Rahman and others 1998), but shame and fear associated with 606 | Disease Control Priorities in Developing Countries | Steven Hyman, Dan Chisholm, Ronald Kessler, and others mental illness remain substantial obstacles to help seeking, to fixed false beliefs that are not explained by the person's culture diagnosis, and to treatment worldwide. The stigmatization of and that the patient holds despite all reasonable evidence to the mental illness has resulted in disparities, compared with other contrary. illnesses, in the availability of care, in research, and in abuses of Patients also exhibit negative symptoms--that is, deficits in the human rights of people with these disorders. normal capacities, such as marked social deficits, impoverish- This chapter focuses on the attributable and avoidable ment of thought and speech, blunting of emotional responses, burden of four leading contributors to mental ill health globally: and lack of motivation. Additionally, patients typically have schizophrenia and related nonaffective psychoses, bipolar cognitive symptoms, such as disorganized or illogical thinking affective disorder (manic-depressive illness), major depressive and an inability to hold goal information in mind to make disorder, and panic disorder. The choice of these disorders is decisions or plan actions. determined not only by their contribution to disease burden, but also by the availability of data for the cost-effectiveness Natural History and Course analyses. Even where such data are available, they are often from Schizophrenia, as defined in current diagnostic manuals, is industrial countries and extrapolation has been necessary. The almost certainly heterogeneous, but still does not comprise all exclusion of other mental disorders, such as childhood disor- nonaffective psychoses (NAPs). In addition to schizophrenia, ders, from analysis is not because the authors consider these dis- NAPs include schizophreniform disorder,characterized by schiz- orders unimportant but because of the paucity of data.Also, this ophrenia-like symptoms of inadequate duration to qualify as chapter does not specifically deal with the important issue of schizophrenia. Because they cannot be readily disentangled in suicide. A background paper on suicide in developing countries community epidemiological surveys, schizophrenia and other has been developed as part of the Disease Control Priorities NAPs are considered together.Because of the data available,how- Project (DCPP) and is available (Vijayakumar, Nagaraj, and ever, the cost-effectiveness analyses reported below are restricted John 2004). The economic analysis presented in this chapter to schizophrenia. Despite likely etiological heterogeneity, schizo- uses the cost-effectiveness analysis methodology specifically phrenia exhibits consistency in its symptom pattern across those developed for the DCPP. The authors recognize that mental countries and cultures studied (Jablensky and others 1992). disorders impose costs and burdens on families as well as Incidence studies show that onset of schizophrenia and individuals that are not captured by the DALY. Treatment will other NAPs is typically in middle to late adolescence for males alleviate some of this burden in addition to alleviating symp- and late adolescence to early adulthood for females, although toms and disability. later onsets are observed. Childhood-onset cases are quite rare A description of the major clinical features, natural course, but particularly severe (Nicolson and Rapoport 1999). Often, epidemiology, burden, and treatment effectiveness for each schizophrenia is first diagnosed with the occurrence of an acute group of disorders is given in the next section. For diagnostic episode of florid psychotic symptoms. The first psychotic criteria, readers are referred to The ICD-10 Classification of episode is often preceded by prodromal symptoms such as social Mental and Behavioral Disorders (ICD-10) (WHO 1992) or withdrawal, irritability or dysphoria, increasing academic or Diagnostic and Statistical Manual of Mental Disorders (DSM- work-related difficulties, and increasing eccentricity. However, IVTR) (American Psychiatric Association 2000). A discussion such symptoms are not specific; studies of whether early diag- follows of population-level costs and cost-effectiveness of inter- nosis and intervention can improve outcomes are under way ventions capable of reducing the current burden associated (McGorry and others 2002). with four disorders in different developing regions of the world The course of schizophrenia is typically one of acute exacer- (tables 31.2­31.6), before moving to a discussion of key issues bations of severe psychotic symptoms, followed by full or par- and implications for mental health policy and improvement of tial remission. Psychotic episodes may be followed by a full services in developing regions of the world. remission after the first and occasionally other early episodes, but over time, residual symptoms and disability typically con- tinue between relapses (Robinson and others 1999). The time SCHIZOPHRENIA AND NONAFFECTIVE between relapses is markedly extended by maintenance treat- PSYCHOSES ment with antipsychotic drugs, generally at lower doses than are needed to treat acute episodes. Cognitive and occupational Schizophrenia is a chronic disorder punctuated by episodes of functioning tends to decline over the first years of the illness florid psychotic symptoms, such as hallucinations and delu- and then to plateau at a level that is generally well below what sions. Hallucinations are sensory perceptions that occur in the would have been expected for the individual. Residual impair- absence of appropriate stimuli. Hallucinations may occur in ment, though, has substantial cross-cultural variation for any sensory modality but in schizophrenia are most commonly reasons that are not well understood. Schizophrenia has consis- auditory--for example, hearing voices or noises. Delusions are tently been found in epidemiological surveys to be highly Mental Disorders | 607 comorbid, usually with anxiety disorders, mood disorders, and behavioral approaches to managing specific symptoms and substance use disorders (Kendler and others 1996). improving medication adherence, group therapy, and family interventions all have demonstrated efficacy in improving clinical outcomes. Community-based models of mental health Epidemiology and Burden care delivery with case management and assertive outreach A great many studies of NAP incidence have been carried out in programs have been shown in health systems of industrial clinical samples. In a review of these studies, Jablensky (2000) countries to be effective ways of managing schizophrenia in the found incidence estimates to be in the range of 0.002 to 0.011 community, for example, by reducing the need for hospital percent per year for schizophrenia and 0.016 to 0.042 percent admissions. However, the applicability of these models to per year for overall NAP. Those annual estimates can be multi- developing countries, as is discussed later, is hard to estimate plied by the number of birth cohorts at risk to yield an estimate because of differences in health system characteristics. Long- of lifetime risk in any one cohort. Assuming conservatively that term remission rates for schizophrenia in developing countries the main age range of risk is between ages 15 and 55, researchers appear to be significantly higher than those reported in indus- estimate lifetime risk is in the range of 0.08 to 0.44 percent for trial countries (Harrison and others 2001), likely resulting from schizophrenia and in the range of 0.64 to 1.68 percent for NAPs. such factors as strong family social support. Lifetime prevalence estimates from community epidemio- Despite their clear usefulness, current treatments do not logical surveys of NAPs are quite consistent with those from prevent schizophrenia, and no clear evidence demonstrates that clinical studies, in the range of 0.3 to 1.6 percent (see, for exam- they induce full recovery or prevent premature mortality. ple, Hwu, Yeh, and Cheng 1989; Kendler and others 1996). Instead, treatment reduces time in episode of florid psychosis Although schizophrenia is a relatively uncommon disorder, and increases time between episodes; thus treatment effects can aggregate estimates of disease burden are high--around 2,000 be understood in terms of improvements in disability. Reported DALYs lost per 1 million total population (table 31.1)-- treatment effect sizes from meta-analyses in the literature, con- because the condition is associated with early onset, long dura- verted into improvements in the average level of disability tion, and severe disability. (Andrews and others 2003; Sanderson and others 2004), show improvements (compared with no treatment) of 18 to 19 per- cent (antipsychotic drugs alone) and 30 to 31 percent (antipsy- Interventions chotic drugs with adjunctive psychosocial treatment). Placed A substantial body of evidence exists on the efficacy of various on a disability scale of 0 to 1, where 0 equals no disability, an treatments for schizophrenia and NAP and on the effectiveness "average" case of schizophrenia moves from a disability level of of various models of health care delivery for persons with these 0.63 (untreated weight from the Global Burden of Disease disorders. This evidence comes primarily from industrial coun- study, Murray and Lopez 1996) to 0.43 to 0.54 (treated). tries. The efficacy data show conclusively that antipsychotic drugs reduce severity of the episodes, hasten resolution of florid symptoms, and reduce duration of hospitalization. MOOD DISORDERS Maintenance treatment with antipsychotic drugs prolongs the The cardinal features of mood disorders are pervasive abnor- period between relapses (Joy, Adams, and Lawrie 2001). malities in the predominant emotional state of the person, such A second generation of antipsychotic medications (also as depressed, elated, or irritable. In mood disorders, these core called atypical) is replacing older neuroleptic antipsychotic emotional symptoms are accompanied by abnormalities in drugs throughout the industrial world. In some clinical trials, physiology, such as changes in patterns of sleep, appetite, and second-generation drugs show small advantages in efficacy energy, and by changes in cognition and behavior. In develop- over first-generation drugs, but their widespread adoption ing countries, concurrent somatic symptoms are also com- results from marked improvement in tolerability. Their relative monly reported and may be the chief complaint. A generally lack of side effects compared with first-generation drugs has accepted subclassification of mood disorders distinguishes led to improved quality of life and improved treatment adher- unipolar depressive disorders from bipolar disorder (defined ence. Second-generation drugs are not without side effects, by the occurrence of mania). This distinction is based on however; for example, some are associated with substantial symptoms, course of illness, patterns of familial transmission, weight gain and increased risk of diabetes. One drug, clozapine, and treatment response. has greater efficacy than other antipsychotic drugs, but because of a 1 percent risk of agranulocytosis, its use requires weekly blood counts and is cumbersome and expensive. Bipolar Disorder Psychosocial interventions also play an important role in Bipolar disorder is characterized by episodes of mania and managing schizophrenia (Bustillo and others 2001). Cognitive- depression, often followed by relative periods of healthy mood 608 | Disease Control Priorities in Developing Countries | Steven Hyman, Dan Chisholm, Ronald Kessler, and others (euthymia). Mixed states with symptoms of both mania and disorder also exhibit chronic psychotic symptoms superim- depression also occur. Mania is typically characterized by posed on their mood syndrome. These individuals are said to euphoria or irritability, a marked increase in energy, and a have schizoaffective disorder. Their prognosis tends to be less decreased need for sleep. Individuals with mania often exhibit favorable than for the usual bipolar patient, although somewhat intrusive, impulsive, and disinhibited behaviors. They may be better than for individuals with schizophrenia. Schizoaffective excessively involved in goal-directed behaviors characterized disorder may also be diagnosed when chronic psychotic symp- by poor judgment; for example, a person might spend all toms are superimposed on unipolar depression. Individuals funds to which he or she has access and more. Self-esteem is with this combination of symptoms have outcomes similar to typically inflated, frequently reaching delusional proportions. patients with schizophrenia (Tsuang and Coryell 1993). Speech is often rapid and difficult to interrupt. Individuals with mania also may exhibit cognitive symptoms; patients can- Epidemiology and Burden. Lifetime and 12-month preva- not stick to a topic and may jump rapidly from idea to idea, lence estimates of bipolar disorder have been reported from a making comprehension of their train of thought difficult. number of community psychiatric epidemiological surveys. Psychotic symptoms are common during manic episodes. The Lifetime prevalence estimates are in the range 0.1 to 2.0 percent depressive episodes of people with bipolar disorder are symp- (Vega and others 1998; Vicente and others 2002), with a tomatically indistinguishable from those who have unipolar weighted mean across surveys of 0.7 percent. Prevalence esti- depressions alone. Unlike anxiety and unipolar mood disor- mates for past-year episodes have a similarly wide range (0.1 to ders, which are more common in women, bipolar disorder has 1.3 percent) (Vega and others 1998) and a weighted mean of an equal gender ratio of lifetime prevalence, although the ratio 0.5 percent. It is important to note that good evidence exists of depressive-to-manic episodes is higher among bipolar suggesting that bipolar disorder has a wide subthreshold spec- women than men. trum that includes people who are often seriously impaired even though they do not meet full DSM or ICD criteria for the Natural History and Course. Retrospective reports from com- disorder (Perugi and Akiskal 2002). This spectrum might munity epidemiological surveys consistently show that bipolar include as much as 5 percent of the general population. The disorder has an early age of onset (in the late teens through mid- ratio of recent-to-lifetime prevalence of bipolar disorder in 20s). Onset in childhood is increasingly recognized, although community surveys is quite high (0.71), indicating that bipolar it remains controversial. Late onset is less common. The vast disorder is persistent. majority of patients with bipolar disorder have recurrent Epidemiological data show that bipolar disorder is associ- episodes of illness, both mania and depression. Classic descrip- ated with substantial impairments in both productive and tions of bipolar disorder suggest recovery to baseline function- social roles (Das Gupta and Guest 2002). Epidemiological evi- ing between episodes, but many patients have residual symp- dence documents consistent delays in patients initially seeking toms that may cause significant impairment (Angst and Sellaro professional treatment (Olfson and others 1998), especially 2000). These states of mania, depression, and lesser (or absent) among early-onset cases, as well as substantial undertreatment symptoms are used in the intervention analysis below. of current cases. Each of these characteristics--chronic, recur- The rate of cycling between mania and depression varies rent course; significant impairments to functioning; modest widely among individuals. One common pattern of illness is treatment rates--contributes to estimates of aggregate disease for episodes initially to be separated by a relatively long period, burden that approach those for schizophrenia (1,200 to 1,800 perhaps a year, and then to become more frequent with age. A DALYs lost per 1 million population, making up more than minority of patients with four or more cycles per year, termed 5 percent of the burden attributable to neuropsychiatric disor- rapid cyclers, tend to be more disabled and less responsive to ders as a whole--see table 31.1). existing treatments. Once cycles are established, most acute episodes start without an identifiable precipitant; the best doc- umented exception is that manic episodes may be initiated by Interventions. Analyses of the primary treatment approaches sleep deprivation, making a regular daily sleep schedule and for bipolar disorder are based on the three health states that avoidance of shift work important in management (Frank, characterize the disorder--mania, depression, and euthymia. Swartz, and Kupfer 2000). Robust evidence from controlled trials shows that antipsychot- Bipolar disorder has consistently been found in epidemio- ic drugs and some benzodiazepines produce a relatively rapid logical surveys to be highly comorbid with other psychiatric reduction in symptoms of a manic phase. Mood-stabilizing disorders, especially anxiety and substance use disorders drugs act more slowly, but they reduce the severity and dura- (ten Have and others 2002). The extent of comorbidity is much tion of acute manic episodes. Maintenance treatment with two greater than for unipolar depressive disorders or anxiety mood-stabilizing drugs--lithium and valproic acid (adminis- disorders. Some individuals with classic symptoms of bipolar tered as sodium valproate)--has been shown to have Mental Disorders | 609 significant, albeit partial, efficacy in reducing rates of both trating, slow thinking, and poor memory. Psychotic symptoms manic and depressive relapses. The drawback of lithium is that occur in a minority of cases. toxic levels are not much greater than therapeutic levels; thus, serum-level monitoring is required. Natural History and Course. Major depression is an episodic For the cost-effectiveness analyses, lithium and valproic disorder that generally begins early in life (median age of onset acid, which have empirical data supporting their efficacy in in the mid to late 20s in community epidemiological surveys), treating and preventing manic and depressive episodes, were although new onsets can be observed across the lifespan. considered. Because evidence suggests that psychosocial Childhood onset is being increasingly recognized, although not approaches enhance compliance with medication (Huxley, all childhood precursors of adult depression take the form of a Parikh, and Baldessarini 2000), adjuvant strategies also were clear depressive disorder. Most individuals suffering from assessed. The primary treatment effect was a change in the a depressive episode will have a recurrence (Mueller and others population-level disability associated with bipolar disorder (a 1999), with recurrence risk greater among those with early- weighted average of time spent in a manic, depressed, or onset disease. Many individuals do not recover completely euthymic phase of illness). Both an acute treatment effect-- from their acute episodes and have chronic milder depression calculated as the product of initial response and reduced punctuated by acute exacerbations (Judd and others 1998). The episode duration--and a prophylactic treatment effect were current term for chronic, milder depression lasting more than ascribed to lithium and valproic acid, resulting in an estimated two years is dysthymia. Although the symptoms of minor improvement of close to 50 percent over the untreated com- depression are, by definition, less severe than those of a major posite disability weight of 0.445 (Chisholm and others forth- depressive episode, chronicity ultimately makes even this lesser coming). This estimate then was adjusted for expected nonad- form of the illness very disabling in many cases (Judd, Schettler, herence to treatment in real-world clinical settings--slightly and Akiskal 2002). Depression has consistently been found in lower for lithium than for valproic acid (Bowden and others epidemiological surveys to be highly comorbid with other 2000). A secondary effect of treatment--reduction of the case mental disorders, with roughly half the people who have a fatality rate by two-thirds--was also ascribed to lithium, history of depression also having a lifetime anxiety disorder. though, because of an absence of current evidence, not to val- Comorbidities of depression and anxiety disorders are genera- proic acid (Goodwin and others 2003). This reduction was lly strongest with generalized anxiety disorder and panic derived through a change in the standardized mortality ratio disorder (Kessler and others 1996). from 2.5 to 1.5, estimated on the basis of natural history stud- ies reported for the prelithium era (for example, Astrup, Epidemiology and Burden. Prevalence of nonbipolar depres- Fossum, and Holmboe 1959; Helgason 1964) to the postlithium sion has been estimated in a number of large-scale community era (for example, Goodwin and others 2003). epidemiological surveys. Lifetime prevalence estimates of hav- ing either major depressive disorder or dysthymia in these sur- veys are in the range 4.2 to 17.0 percent (Andrade and others Major Depressive Disorder 2003; Bijl and others 1998), with a weighted mean of 12.1 per- The core symptom of major depression is a disturbance of cent. Six- to 12-month prevalence estimates have a similarly mood; sadness is most typical, but anger, irritability, and loss of wide range (1.9 to 10.9 percent) (Andrade and others 2003; interest in usual pursuits may predominate. Often the affected Robins and Regier 1991), with a weighted mean of 5.8 percent. person is unable to experience pleasure (anhedonia) and may These wide differences in prevalence likely represent the difficul- feel hopeless. In many countries of the developing world, ties inherent in self-reporting of conditions that are invariably patients will not complain of such emotional symptoms, but stigmatized across cultures.Prevalence estimates are consistently rather of physical symptoms, such as fatigue or multiple aches highest in North America and lowest in Asia (with prevalence and pains. estimates of major depressive disorders generally a good deal Typical physiological symptoms that occur across cultures higher than those of dysthymia). include sleep disturbance (most often insomnia with early Epidemiological data document consistent delays in morning awakening, but occasionally excessive sleeping); patients initially seeking professional treatment for depression, appetite disturbance (usually loss of appetite and weight loss, especially among early-onset cases (Olfson and others 1998), as but occasionally excessive eating); and decreased energy. well as substantial undertreatment. For example, World Mental Behaviorally, some individuals with depression exhibit slowed Health surveys in six Western European countries found that motor movements (psychomotor retardation), whereas others only 36.6 percent of people with active nonbipolar depression may be agitated. Cognitive symptoms may include thoughts of in the 12 months before the survey received any professional worthlessness and guilt, suicidal thoughts, difficulty concen- treatment for this disorder during the subsequent year 610 | Disease Control Priorities in Developing Countries | Steven Hyman, Dan Chisholm, Ronald Kessler, and others (ESEMeD/MHEDEA 2000 Investigators 2004). The situation is depressive episode was a reduction in the duration of time even worse in developing countries, where the vast majority of depressed, equivalent to an increase in the remission rate (25 to people with depression who seek help do so in general health 40 percent improvement over no treatment; Malt and others care settings and complain of nonspecific physical symptoms. 1999; Solomon and others 1997). In addition, all interventions Such individuals receive a correct diagnosis in less than were attributed a modest improvement in the level of disability one-quarter of cases and typically are treated with medicines of for an unremitted depressive episode (10 to 15 percent), doubtful efficacy (Linden and others 1999). resulting from increased proportions of cases moving from Depression is consistently found in community surveys to more to less severe health states. For the estimated 56 percent of be associated with substantial impairments in both productive prevalent cases eligible for maintenance treatment (at least two and social roles (Wang, Simon, and Kessler 2003). As with lifetime episodes), an additional effect of efficacious mainte- bipolar depression, but exacerbated by its high incidence, the nance treatment was incorporated into the analysis by reducing recurrent nature and disabling consequences of (unipolar) the incidence of recurrent episodes by 50 percent (Geddes and depression mean that overall disease burden estimates are high others 2003). Estimates of intervention effectiveness include the in all regions of the world (5,000 to 10,000 DALYs per 1 million positive change that would occur naturally and also incorporate population, as much as 5 percent of the total burden of disease any placebo effect, which, in the treatment of depression, is not from all causes; table 31.1). Depression is, in fact, ranked as the inconsiderable (Andrews 2001). fourth leading cause of disease burden globally and represents the single largest contributor to nonfatal burden (Ustun and others 2004). ANXIETY DISORDERS Interventions. Efficacy has been demonstrated for several Anxiety disorders are a group of disorders that have as their classes of antidepressant drugs and for two psychosocial treat- central feature the inability to regulate fear or worry. Although ments for depression (Paykel and Priest 1992). The older tri- anxiety in itself is likely to feature in the clinical presentation of cyclic antidepressants (TCAs) and newer drugs, including the most patients, somatic complaints such as chest pain, palpita- selective serotonin reuptake inhibitors (SSRIs), have similar tions, respiratory difficulty, headaches, and the like are also efficacy. The newer drugs have milder side-effect profiles and common, and these symptoms may be more common in are consequently more likely to be tolerated at therapeutic developing countries. A number of different types of anxiety doses (Pereira and Patel 1999). SSRIs have not been widely used disorder exist, some of which are now briefly described. in developing countries because of their higher cost, although The central feature of panic disorder is an unexpected panic as the patent protection expires, this situation is likely to attack, which is a discrete period of intense fear accompanied change (Patel 1996). Of the psychosocial treatments with by physiologic symptoms such as a racing heart, shortness of demonstrated efficacy, the most widely accepted are cognitive- breath, sweating, or dizziness. The person may have an intense behavioral approaches. Alone or in combination, drug and psy- fear of losing control or of dying. Panic disorder is diagnosed chosocial treatments speed recovery from acute episodes. when panic attacks are recurrent and give rise to anticipatory Maintenance treatment with drugs decreases relapse risk anxiety about additional attacks. People with panic disorder (Geddes and others 2003). Some evidence suggests that a may progressively restrict their lives to avoid situations in course of psychotherapy may also delay relapses. Although which panic attacks occur or situations from which it might be most of the clinical trials have been carried out in industrial difficult to escape should a panic attack occur. They common- countries, at least three high-quality trials have demonstrated ly avoid crowds, traveling, bridges, and elevators, and ultimate- the efficacy of antidepressants, group therapy, or both in devel- ly some individuals may stop leaving home altogether. oping countries (Araya and others 2003; Bolton and others Pervasive phobic avoidance is described as agoraphobia. 2003; Patel and others 2003). Generalized anxiety disorder is characterized by chronic For the cost-effectiveness analyses, depression was modeled unrealistic and excessive worry. These symptoms are accompa- as an episodic disorder with a high rate of remission and nied by specific anxiety-related symptoms such as sympathetic subsequent recurrence, and with excess mortality from suicide nervous system arousal, excessive vigilance, and motor tension. (Chisholm and others 2004). None of the selected depression Posttraumatic stress disorder follows serious trauma. It is interventions was accorded a reduction in case fatality, however, characterized by emotional numbness, punctuated by intrusive owing to the lack of robust clinical evidence that antidepressants reliving of the traumatic episode, generally initiated by envi- or psychotherapy in themselves alter the relative risk of death by ronmental cues that act as reminders of the trauma; by dis- suicide (Storosum and others 2001). The main modeled impact turbed sleep; and by hyperarousal, such as exaggerated startle of intervention targeted toward episodic treatment of a new responses. Mental Disorders | 611 Social anxiety disorder (social phobia) is characterized by a disorder typically also have a second anxiety disorder, while persistent fear of social situations or performance situations that more than half the people with a history of either anxiety or expose a person to potential scrutiny by others. The affected mood disorder typically have both types of disorder. person has intense fear that he or she will act in a way that will Retrospective reports from community surveys consistently be humiliating. Separating social anxiety disorder from show that anxiety disorders have early average ages of onset. extremes of normal temperament, such as shyness, is difficult. An impressive cross-national consistency can be seen in these Nonetheless, social anxiety disorder can be quite disabling. patterns, with an estimated median age of onset of anxiety at Simple phobias are extreme fear in the presence of discrete stim- approximately 15. uli or cues, such as fear of heights. Epidemiological surveys have also looked at the treatment The core features of obsessive-compulsive disorder are obses- of anxiety disorders. As with depression, consistent evidence in sions (intrusive, unwanted thoughts) and compulsions (per- these surveys suggests that delays in initially seeking profes- formance of highly ritualized behaviors intended to neutralize sional treatment for an anxiety disorder are widespread after the negative thoughts and emotions resulting from the obses- first onset (Olfson and others 1998). This finding is especially sions). One symptom pattern might be repetitive hand washing true among early-onset cases. Epidemiological data also show beyond the point of skin damage to neutralize fears of that only a minority of current cases receive any formal treat- contamination. ment in Western countries, whereas treatment of anxiety disor- ders is virtually nonexistent in many developing countries. The most recently published surveys, the World Mental Health Natural History and Course surveys in six Western European countries, found that only The anxiety disorders differ in their age of onset, course of ill- 26.3 percent of people with an active anxiety disorder in the ness, and symptom triggers. One of these disorders, PTSD, is 12 months before the survey received any professional treat- dependent for its etiology on one or more powerfully negative ment (ESEMeD/MHEDEA 2000 Investigators 2004). life events. Although the anxiety disorders are discussed as a Anxiety disorders have consistently been found to be associ- group, panic disorder is chosen because of the available data for ated with substantial impairments in both productive roles (for the purposes of the cost-effectiveness analysis. example, work absenteeism, work performance, unemploy- Prevalence estimates of anxiety disorders based on commu- ment, and underemployment) and social roles (social isolation, nity epidemiological surveys vary widely, from a low of 2.2 per- interpersonal tensions, and marital disruption, among others) cent (Andrade and others 2003) to a high of 28.5 percent (see, for example, Kessler and Frank 1997). As noted earlier, for (Kessler and others 1994), with a weighted mean across surveys the purposes of this chapter, one of the anxiety disorders-- of 15.6 percent. Prevalence estimates for anxiety disorders in panic disorder--has been chosen to describe interventions and the past 6 to 12 months have a similarly wide range (1.2 to undertake cost-effectiveness analysis. Panic disorder is as dis- 19.3 percent) (Andrade and others 2003; Kessler and others abling as obsessive-compulsive disorder and PTSD, accounts 1994), with a weighted mean of 9.4 percent. Despite wide for about one-third of all seriously impairing anxiety disorders, variation in overall prevalence, several clear relative prevalence is one of the most common anxiety disorders presenting for patterns can be seen across surveys. Specific phobia is generally treatment, and imposes an estimated burden of 600 to 800 the most prevalent lifetime anxiety disorder, with social phobia DALYs per 1 million population. generally the second most prevalent lifetime anxiety disorder. Good evidence exists that both drug and psychosocial Panic disorder and obsessive-compulsive disorder are generally treatments are effective for managing anxiety disorders. the least prevalent. Antidepressant drugs (both older TCAs and SSRIs) have been These surveys also provide evidence about the persistence of shown to be effective for the treatment of several anxiety disor- anxiety disorders, indirectly defined as the ratio of 6-month or ders, including panic disorder, reducing the duration and 12-month to lifetime prevalence. This ratio averages approxi- intensity of the disorder. Although high-potency benzodi- mately 60 percent for overall anxiety disorders, indicating a azepines are efficacious for panic disorder, these drugs carry a high rate of persistence across the life course. The highest risk of dependence and are not considered the first line of persistence is generally found for social phobia, and the lowest treatment. Psychosocial treatments, especially cognitive- for agoraphobia. These estimates of high persistence are con- behavioral therapy, are also effective in diminishing both panic sistent with results obtained from longitudinal studies of attacks and phobic avoidance. patients (Yonkers and others 2003). Anxiety disorders have consistently been found in epidemio- logical surveys to be highly comorbid both among themselves Interventions for Panic Disorder and with mood disorders (for example, de Graaf and others Although evidence-based interventions for panic disorder have 2003). The vast majority of people with a history of one anxiety yet to be evaluated or made widely available in developing 612 | Disease Control Priorities in Developing Countries | Steven Hyman, Dan Chisholm, Ronald Kessler, and others countries, the potential population-level impact of a number of (reflecting a societal preference for health benefits to be realized interventions--including older and newer antidepressants, sooner), but no age-weighting was used. anxiolytic drugs (benzodiazepines), and psychosocial Estimation of the baseline epidemiological situation that treatments--was examined. Interventions reduce the severity would prevail without treatment used incidence and preva- of panic attacks and improve the probability of making a full lence estimates from the Global Burden of Disease 2000 study recovery. Effect sizes for symptom improvement were drawn of the World Health Organization (WHO) (see online Global from a meta-analysis of the long-term effects of intervention of Burden of Disease documentation for the four disorders at panic disorder (Bakker and others 1998) and converted into an http://www.who.int/evidence/bod). Current pharmacolog- equivalent change in disability weight (Sanderson and others ical or psychosocial treatments do not exert a primary preven- 2004). Concerning remission, a number of controlled and tive effect on the onset of the four conditions (although naturalistic studies (for example, Faravelli, Paterniti, and some evidence exists that treating depression in parents may Scarpato 1995; Yonkers and others 2003) reveal a consistent reduce risk for offspring), indicating that currently observed remission rate of 12 to 13 percent for pharmacological and incidence rates coincide with those that would pertain under combination strategies--except for benzodiazepine use, for no treatment. Prevention of recurrences of acute episodes (sec- which the evidence is that longer-term recovery is actually ondary prevention) has been demonstrated for maintenance worse than placebo (Katschnig and others 1995)--which repre- treatments for major depression and bipolar disorder. sents a 62 percent improvement in efficacy over the untreated Maintenance treatment with antipsychotic drugs decreases the remission rate (7.4 percent). risk of recurrent acute episodes of schizophrenia. For each con- dition, a range of treatment strategies was considered and assessed, including older (and widely available) psychothera- COST-EFFECTIVENESS METHODS AND RESULTS peutic drugs, newer pharmacotherapies, psychosocial treat- ments, and combination treatments (see table 31.2 for a list of This section estimates the burden attributed to schizophrenia, interventions included). bipolar disorder, depression, and panic disorder that could be averted (through scaling up) by proven, efficacious treatments. It is followed by calculations of the expected cost and cost- Estimation of Population-Level Treatment Costs effectiveness of such treatments. Analysis is conducted at the Cost estimation followed the principles and procedures level of six low- and middle-income geographical World Bank described in chapter 7 for carrying out economic analyses of dis- regions. ease control priorities in developing countries. For depression and panic disorder, treatment was assumed to occur in a pri- Estimation of Population-Level Effectiveness of Treatments mary care setting, whereas for schizophrenia and bipolar disor- In modeling the impact of mental health interventions, we der, which often produce highly disruptive behaviors, both hos- used a state-transition model (Lauer and others 2003) that pital- and community-based outpatient service models were traces the development of a population, taking into account derived and compared. Both program- and patient-level costs births, deaths, and the disease in question. In addition to pop- were identified and estimated. Program-level costs included the ulation size and structure, the model makes use of a number of infrastructure and administrative support for implementing epidemiological parameters (incidence and prevalence, remis- mental health treatments, as well as training inputs (for exam- sion, and cause-specific and residual rates of mortality) and ple, two to three days per trainee were estimated for training assigns age- and gender-specific disability weights to both the primary care doctors and case managers in psychotropic med- disease in question and the general population. The output of ication management). Patient-level resource inputs included the model is an estimate of the total healthy life years experi- medication regimens (for example, fluoxetine, 20 milligrams enced by the population over a lifetime period (100 years). The daily), laboratory tests (for example, lithium blood levels), model was run for a number of possible scenarios, including no primary care visits (including any contacts with a case manager), treatment at all (natural history), current treatment coverage, and hospital outpatient and inpatient care. Estimated patient- and scaled-up coverage of current as well as potential new level resource inputs for each of the four disorders were informed interventions. For the treatment scenarios, an implementation by empirical economic evaluative studies (for example, Patel and period of 10 years was used (thereafter, epidemiological rates others 2003; Srinivasa Murthy and others 2005) as well as a multi- and health state valuations return to natural history levels). The national Delphi consensus study of resource use for psychiatric model derived the number of additional healthy years gained disorders in seven developing countries (Ferri and others 2004). (equivalent to DALYs averted) each year in the population Region-specific unit costs or prices were applied to all resource compared with the outcome for no treatment at all. DALYs inputs (see Mulligan and others 2003) to give an annual cost for averted in future years were discounted at a rate of 3 percent each case as well as for all cases at the specified level of treatment Mental Disorders | 613 Table 31.2 Interventions for Reducing the Burden of Major Psychiatric Disorders in Developing Countries Disorder Intervention Example Schizophrenia Older (neuroleptic) antipsychotic drug Haloperidol Treatment setting: hospital outpatient Newer (atypical) antipsychotic drug Risperidone Treatment coverage (target): 80 percent Older antipsychotic drug and psychosocial treatment Haloperidol plus family psychoeducation Newer antipsychotic drug and psychosocial treatment Risperidone plus family psychoeducation Bipolar affective disorder Older mood-stabilizing drug Lithium carbonate Treatment setting: hospital outpatient Newer mood-stabilizing drug Sodium valproate Treatment coverage (target): 50 percent Older mood-stabilizing drug and psychosocial treatment Lithium plus family psychoeducation Newer mood-stabilizing drug and psychosocial treatment Valproate plus family psychoeducation Depression Episodic treatment Treatment setting: primary health care Older TCA Imipramine or amitriptyline Treatment coverage (target): 50 percent Newer antidepressant drug (SSRI; generic) Fluoxetine Psychosocial treatment Group psychotherapy Older antidepressant drug and psychosocial treatment Amitriptyline plus group psychotherapy Newer antidepressant drug and psychosocial treatment Fluoxetine plus group psychotherapy Maintenance treatment Older antidepressant drug and psychosocial treatment Imipramine plus group psychotherapy Newer antidepressant drug and psychosocial treatment Fluoxetine plus group psychotherapy Panic disorder Benzodiazepines Alprazolam Treatment setting: primary health care Older TCA Amitriptyline Treatment coverage (target): 50 percent Newer antidepressant drug (SSRI; generic) Fluoxetine Psychosocial treatment Cognitive therapy Older antidepressant drug and psychosocial treatment Amitriptyline plus cognitive therapy Newer antidepressant drug and psychosocial treatment Fluoxetine plus cognitive therapy Source: Authors' own estimates and recommendations. Note: Interventions in bold are the most cost-effective treatments of choice. coverage. Costs incurred over the 10-year implementation peri- evidence-based pharmacological and psychosocial treatments od were discounted at 3 percent and expressed in U.S. dollars (Ferri and others 2004; Kohn and others 2004), plus those in (rather than international dollars, which attempt to adjust for contact with traditional healers (the effectiveness of which was differences in purchasing power between countries). conservatively approximated by ascribing a placebo effect size for each disorder). Coverage In each World Bank region, treatment costs and effects were Results ascribed to the population in need, both at current levels Tables 31.3 through 31.6 provide estimates of the population- of intervention coverage and at a scaled-up, target level of level effects (measured in DALYs averted), costs, and cost- coverage (80 percent for schizophrenia, 50 percent for the other effectiveness of each intervention by world region for the four conditions). Target coverage levels were predicated on the basis types of psychiatric disorder considered in this chapter. A num- of what could feasibly be achieved given existing rates of treat- ber of key findings emerge from this analysis. ment (Ferri and others 2004; Kohn and others 2004), as well as on prerequisites for increased coverage, such as recognition of Treatment Effectiveness. Results for schizophrenia and bipo- common mental disorders in primary care. Estimation of cur- lar disorder are similar (albeit at differing coverage levels), rang- rent regional levels of effective coverage is hampered by lack of ing from less than 100 DALYs averted per 1 million population data; nevertheless, an attempt was made to approximate the under the current situation in Sub-Saharan Africa and South expected proportion of the diseased population receiving Asia to 350 to 400 DALYs averted per 1 million population for 614 | Disease Control Priorities in Developing Countries | Steven Hyman, Dan Chisholm, Ronald Kessler, and others Table 31.3 Cost-Effectiveness Results: Schizophrenia Model definition: World Bank region Treatment setting: (a) hospital- based; (b) community-based Sub-Saharan Latin America Middle East and Europe and East Asia and Treatment coverage: 80 percent Africa and the Caribbean North Africa Central Asia South Asia the Pacific Total effect (DALYs averted per year per 1 million population) Current situation 74 136 115 258 87 148 Older (neuroleptic) antipsychotic drug 149 219 214 254 177 231 Newer (atypical) antipsychotic drug 160 235 230 273 190 248 Older antipsychotic drug plus 254 373 364 353 300 392 psychosocial treatment Newer antipsychotic drug plus 261 383 373 364 308 403 psychosocial treatment Total cost (US$ million per year per 1 million population) Current situation 0.42 2.07 1.31 3.13 0.51 1.11 Hospital-based service model Older (neuroleptic) antipsychotic drug 0.60 3.09 2.40 2.24 0.74 1.18 Newer (atypical) antipsychotic drug 2.80 6.33 5.41 6.16 3.36 4.63 Older antipsychotic drug plus 0.67 3.27 2.56 2.36 0.81 1.26 psychosocial treatment Newer antipsychotic drug plus 2.87 6.56 5.61 6.31 3.44 4.73 psychosocial treatment Community-based service model Older (neuroleptic) antipsychotic drug 0.40 1.58 1.42 1.17 0.44 0.66 Newer (atypical) antipsychotic drug 2.59 4.85 4.45 5.11 3.07 4.12 Older antipsychotic drug plus 0.47 1.81 1.61 1.32 0.52 0.75 psychosocial treatment Newer antipsychotic drug plus 2.67 5.09 4.66 5.28 3.16 4.22 psychosocial treatment Cost-effectiveness (US$ per DALY averted) Current situation 5,695 15,192 11,400 12,134 5,900 7,533 Hospital-based service model Older (neuroleptic) antipsychotic drug 4,047 14,123 11,205 8,793 4,164 5,120 Newer (atypical) antipsychotic drug 17,433 26,893 23,543 22,530 17,702 18,700 Older antipsychotic drug plus 2,623 8,781 7,040 6,685 2,693 3,212 psychosocial treatment Newer antipsychotic drug plus 10,996 17,146 15,027 17,329 11,164 11,746 psychosocial treatment Community-based service model Older (neuroleptic) antipsychotic drug 2,668 7,230 6,618 4,595 2,499 2,855 Newer (atypical) antipsychotic drug 16,174 20,583 19,352 18,685 16,178 16,622 Older antipsychotic drug plus 1,839 4,847 4,431 3,745 1,743 1,917 psychosocial treatment Newer antipsychotic drug plus 10,232 13,313 12,485 14,481 10,239 10,484 psychosocial treatment Source: Authors' own estimates. Note: Intervention data in bold are the most cost-effective treatments of choice. Mental Disorders | 615 Table 31.4 Cost-Effectiveness Results: Bipolar Disorder Model definition: World Bank region Treatment setting: (a) hospital- based; (b) community-based Sub-Saharan Latin America Middle East and Europe and East Asia and Treatment coverage: 50 percent Africa and the Caribbean North Africa Central Asia South Asia the Pacific Total effect (DALYs averted per year per 1 million population) Current situation 79 128 97 199 93 153 Older mood-stabilizing drug (lithium) 292 336 296 381 319 389 Newer mood-stabilizing drug 211 300 273 331 278 351 (valproate) Older mood-stabilizing drug plus 312 365 322 413 346 422 psychosocial treatment Newer mood-stabilizing drug plus 232 330 300 365 306 386 psychosocial treatment Total cost (US$ million per year per 1 million population) Current situation 0.31 1.22 0.74 1.27 0.42 0.67 Hospital-based service model Older mood-stabilizing drug (lithium) 0.61 2.77 1.92 2.03 0.82 1.30 Newer mood-stabilizing drug 0.79 2.87 2.04 2.20 1.03 1.53 (valproate) Older mood-stabilizing drug plus 0.63 2.79 1.95 2.05 0.84 1.32 psychosocial treatment Newer mood-stabilizing drug plus 0.81 2.90 2.08 2.22 1.06 1.55 psychosocial treatment Community-based service model Older mood-stabilizing drug (lithium) 0.46 1.78 1.20 1.37 0.59 0.93 Newer mood-stabilizing drug 0.64 1.91 1.36 1.57 0.82 1.17 (valproate) Older mood-stabilizing drug plus 0.48 1.80 1.23 1.39 0.62 0.95 psychosocial treatment Newer mood-stabilizing drug plus 0.67 1.95 1.39 1.59 0.85 1.19 psychosocial treatment Cost-effectiveness (US$ per DALY averted) Current situation 3,967 9,518 7,668 6,398 4,463 4,373 Hospital-based service model Older mood-stabilizing drug (lithium) 2,091 8,246 6,478 5,341 2,553 3,348 Newer mood-stabilizing drug 3,727 9,579 7,501 6,648 3,709 4,358 (valproate) Older mood-stabilizing drug plus 2,016 7,644 6,036 4,957 2,424 3,119 psychosocial treatment Newer mood-stabilizing drug plus 3,480 8,800 6,937 6,100 3,459 4,016 psychosocial treatment Community-based service model Older mood-stabilizing drug (lithium) 1,587 5,295 4,068 3,608 1,862 2,394 Newer mood-stabilizing drug 3,057 6,386 4,971 4,727 2,943 3,338 (valproate) Older mood-stabilizing drug plus 1,545 4,928 3,823 3,359 1,787 2,241 psychosocial treatment Newer mood-stabilizing drug plus 2,874 5,908 4,645 4,359 2,765 3,092 psychosocial treatment Source: Authors' own estimates. Note: Intervention data in bold are the most cost-effective treatments of choice. 616 | Disease Control Priorities in Developing Countries | Steven Hyman, Dan Chisholm, Ronald Kessler, and others Table 31.5 Cost-Effectiveness Results: Depression Model definition: World Bank region Treatment setting: primary health care Sub-Saharan Latin America Middle East and Europe and East Asia and Treatment coverage: 50 percent Africa and the Caribbean North Africa Central Asia South Asia the Pacific Total effect (DALYs averted per year per 1 million population) Current situation 133 264 218 308 218 243 Episodic treatment: older 599 995 920 874 987 891 antidepressant drug (TCA) Episodic treatment: newer 632 1,049 971 925 1,042 941 antidepressant drug (SSRI) Episodic psychosocial treatment 624 1,036 958 936 1,028 927 Episodic psychosocial treatment 745 1,237 1,144 1,100 1,228 1,107 plus older antidepressant Episodic psychosocial treatment 745 1,237 1,144 1,100 1,228 1,107 plus newer antidepressant Maintenance psychosocial treatment 1,174 1,953 1,806 1,789 1,937 1,747 plus older antidepressant Maintenance psychosocial treatment 1,174 1,953 1,806 1,789 1,937 1,747 plus newer antidepressant Total cost (US$ million per year per 1 million population) Current situation 0.36 0.90 0.63 0.74 0.56 0.67 Episodic treatment: older 0.30 1.28 0.96 0.81 0.47 0.47 antidepressant drug (TCA) Episodic treatment: newer 0.66 1.86 1.47 1.39 1.04 0.99 antidepressant drug (SSRI) Episodic psychosocial treatment 0.37 1.67 1.27 0.97 0.55 0.53 Episodic psychosocial treatment 0.50 1.96 1.53 1.21 0.77 0.72 plus older antidepressant Episodic psychosocial treatment 0.90 2.60 2.10 1.85 1.40 1.29 plus newer antidepressant Maintenance psychosocial treatment 0.96 3.44 2.77 2.19 1.45 1.38 plus older antidepressant Maintenance psychosocial treatment 1.80 4.80 3.99 3.56 2.81 2.59 plus newer antidepressant Cost-effectiveness (US$ per DALY averted) Current situation 2,692 3,414 2,905 2,391 2,546 2,777 Episodic treatment: older 505 1,288 1,039 929 478 533 antidepressant drug (TCA) Episodic treatment: newer 1,042 1,771 1,516 1,501 1,003 1,048 antidepressant drug (SSRI) Episodic psychosocial treatment 592 1,611 1,330 1,035 537 570 Episodic psychosocial treatment 674 1,586 1,335 1,104 627 653 plus older antidepressant Episodic psychosocial treatment 1,203 2,101 1,834 1,682 1,140 1,161 plus newer antidepressant Maintenance psychosocial treatment 817 1,760 1,533 1,226 749 788 plus older antidepressant Maintenance psychosocial treatment 1,535 2,459 2,211 1,990 1,449 1,481 plus newer antidepressant Source: Authors' own estimates. Note: Intervention data in bold are the most cost-effective treatments of choice. Mental Disorders | 617 Table 31.6 Cost-Effectiveness Results: Panic Disorder Model definition: World Bank region Treatment setting: primary health care Sub-Saharan Latin America Middle East and Europe and East Asia and Treatment coverage: 50 percent Africa and the Caribbean North Africa Central Asia South Asia the Pacific Total effect (DALYs averted per year per 1 million population) Current situation 49 94 64 88 57 90 Anxiolytic drug (benzodiazepine) 144 182 170 183 168 195 Older antidepressant drug (TCA) 232 290 272 290 269 312 Newer antidepressant drug (SSRI; 245 307 287 307 284 330 generic) Psychosocial treatment 233 292 273 292 270 313 (cognitive-behavioral therapy) Older antidepressant plus 262 329 308 329 304 353 psychosocial treatment Newer antidepressant plus 276 346 324 346 320 372 psychosocial treatment Total cost (US$ million per year per 1 million population) Current situation 0.06 0.13 0.08 0.07 0.05 0.10 Anxiolytic drug (benzodiazepine) 0.10 0.20 0.15 0.15 0.10 0.12 Older antidepressant drug (TCA) 0.09 0.18 0.14 0.14 0.08 0.11 Newer antidepressant drug 0.15 0.27 0.21 0.23 0.16 0.20 (SSRI; generic) Psychosocial treatment (cognitive- 0.11 0.27 0.21 0.17 0.09 0.11 behavioral therapy) Older antidepressant plus 0.15 0.32 0.26 0.23 0.13 0.17 psychosocial treatment Newer antidepressant plus 0.22 0.41 0.34 0.32 0.22 0.26 psychosocial treatment Cost-effectiveness (US$ per DALY averted) Current situation 1,192 1,378 1,208 824 948 1,109 Anxiolytic drug (benzodiazepine) 681 1,075 892 842 572 629 Older antidepressant drug (TCA) 369 619 508 474 305 339 Newer antidepressant drug (SSRI; 630 865 747 741 567 606 generic) Psychosocial treatment (cognitive- 468 927 786 594 338 365 behavioral therapy) Older antidepressant plus 556 977 844 685 443 474 psychosocial treatment Newer antidepressant plus 788 1,188 1,050 918 671 709 psychosocial treatment Source: Authors' own estimates. CBT cognitive behavioral therapy Note: Intervention data in bold are the most cost-effective treatments of choice. combination drug and psychosocial interventions in Europe tolerability and adherence); lithium was considered modestly and Central Asia and East Asia and the Pacific. Second- more effective as a mood-stabilizing drug than valproate (on the generation (atypical) antipsychotic drugs were considered basis of its additional positive effect on suicide rates). Adjuvant slightly more effective than first-generation drugs (on the basis psychosocial treatment in combination with pharmacotherapy of a modest intrinsic efficacy difference and differences in significantly added to expected population-level health gain. 618 | Disease Control Priorities in Developing Countries | Steven Hyman, Dan Chisholm, Ronald Kessler, and others With the exception of Europe and Central Asia, less than 10 per- program using newer antidepressants, three times more costly cent of the disease burden currently is being averted, whereas the than episodic treatment with newer antidepressant drugs only. implementation of combined interventions at a scaled-up level Patient-level resource inputs for panic disorder interventions of coverage is expected to avert 14 to 22 percent of the burden of cost US$50 to US$200 per case per year, and overall costs schizophrenia (coverage level, 80 percent) and 17 to 29 percent including program costs of training and administration of the burden of bipolar disorder (coverage level, 50 percent). amounted to US$0.10 to US$0.30 per capita. For primary care treatment of common mental disorders, including depression and panic disorder, current levels of effec- Cost-Effectiveness. Compared with both the current situation tive coverage avert only 3 to 8 percent of the existing disease and the epidemiological situation of no treatment (natural his- burden, whereas scaling up of the most effective interventions tory), the most cost-effective strategy for averting the burden of to a coverage level of 50 percent could be expected to avert psychosis and severe affective disorders in developing countries more than 20 percent of the burden of depression and up to is expected to be a combined intervention of first-generation one-third of the burden of panic disorder. Considered at a pop- antipsychotic or mood-stabilizing drugs with adjuvant psy- ulation level, episodic treatments for depressive episodes did chosocial treatment delivered through a community-based not differ substantially within regions (averting 10 to15 percent outpatient service model, with a cost-effectiveness ratio of of current burden); more substantial health gain is expected below US$2,000 in Sub-Saharan Africa and South Asia, rising by providing maintenance treatment to individuals with recur- to US$5,000 in Latin America and the Caribbean (equivalent to rent depression (approximately 1,200 to 1,900 DALYs averted more than 500 DALYs averted per US$1 million expenditure in per 1 million population; 18 to 23 percent of burden). Such an Sub-Saharan Africa and South Asia and 200 DALYs averted in approach has been found to reduce the risk of relapse by half. Latin America and the Caribbean). Currently, the high acquisi- Although the evidence to date from developing regions is mea- tion price of second-generation antipsychotic drugs makes ger, our results suggest that SSRIs such as fluoxetine, alone or in their use in developing regions questionable on efficiency combination with psychosocial treatment, are the most effec- grounds, although this situation can be expected to change as tive treatments for panic disorder, with health gains consider- these drugs come off patent. By contrast, evidence indicates ably better than those estimated for benzodiazepine anxiolytic that the relatively modest additional cost of adjuvant psy- drugs such as alprazolam. chosocial treatment reaps significant health gains, thereby making such a combined strategy for schizophrenia and Treatment Costs. Community-based service models for bipolar disorder treatment more cost-effective than pharma- schizophrenia and bipolar disorder were found to be apprecia- cotherapy alone. bly less costly than hospital-based service models (for example, For more common mental disorders treated in primary care interventions for bipolar disorder were 25 to 40 percent less settings (depressive and anxiety disorders), the single most costly). The total cost per capita of community-based outpa- cost-effective strategy is the scaled-up use of older antidepres- tient treatment with first-generation antipsychotic or mood- sants (because of their lower cost but similar efficacy compared stabilizing drugs, including all patient-level resource needs as with newer antidepressants). However, as the price margin well as infrastructural support, ranged from US$0.40 to between older and generic newer antidepressants continues to US$0.50 in Sub-Saharan Africa and South Asia to US$1.20 to diminish, generic SSRIs--which have milder side effects and US$1.90 in Latin America and the Caribbean and in Europe are more likely to be taken at a therapeutic dose (Pereira and and Central Asia (equivalent patient costs per year, US$170 to Patel 1999)--can be expected to be at least as cost-effective US$300 and US$300 to US$800, respectively). The cost per and, therefore, the pharmacological treatment of choice in the capita for interventions using second-generation (atypical) future. Because depression is often a recurring condition, antipsychotic drugs still under patent is much higher (US$2.50 proactive care management, including long-term maintenance to US$5.00). By contrast, some of the newer antidepressant treatment with antidepressant drugs, represents a cost-effective drugs (SSRIs) are now off patent, and their use in treating way of significantly reducing the enormous burden of depres- depression and panic disorder was accordingly costed at their sion that exists in developing regions now (400 to 1,300 DALYs generic, nonbranded price. The patient-level cost of treating a averted per US$1 million expenditure). 6-month episode of depression ranged from as little as US$30 (older antidepressants in Sub-Saharan Africa or South Asia) to US$150 (newer antidepressants in combination with brief psy- POLICY AND SERVICE IMPLICATIONS chotherapy in Latin America and the Caribbean). Total annual costs for all incidents of depressive episodes receiving treat- Many attempts have been made during the past 50 years to ment, including training and other program-level costs, were as have mental health care placed higher on national and interna- much as US$2 to US$5 per capita for a maintenance treatment tional agendas. In 1974, a WHO Expert Committee on the Mental Disorders | 619 Organization of Mental Health Services in Developing work closely (and apparently effectively) with conventional Countries (WHO 1975) made the following recommendations: mental health services (Thara, Padmavati, and Srinivasan 2004). Alternatively, animosity and competition can exist, and recent · Develop a national mental health policy and create a unit examples of human rights violations by traditional healers within the Health Ministry to implement it. demonstrate the heterogeneity of this group of providers. · Budget for workforce development, essential drug procure- The formal diagnosis and treatment of mental disorders ment, infrastructure development, data collection, and occur in both primary and specialist health services. Examples research. in nearly a dozen countries now show it is feasible and practi- · Decentralize service provision and integrate mental health cable to treat common mental disorders in primary health care into primary health care. settings (for example, Chisholm and others 2000; De Jong · Train and supervise primary health care providers in mental 1996; Mohit and others 1999). The challenge is to enhance sys- health using specialist mental health staff. tems of care by taking effective local models and disseminating them throughout a country. Thirty years later, international agencies, nongovernmental Concern has been expressed that the more sophisticated organizations, and professional bodies continue to make those psychotherapies used in mental health care are beyond the exact recommendations. One reason for the lack of action in human resources of developing countries. However, basic psy- mental health has been the paucity of information on the cost- chological therapies can be effective, though there is some evi- effectiveness of mental health interventions. Advocacy without dence, at least for depression, that the newer drug therapies are the necessary science can readily be ignored in countries with more cost-effective than psychological therapies (Patel and massive health problems and meager resources. This chapter others 2003). Psychoeducational family intervention has been aims to address this deficiency. shown to be suitable for rehabilitation in schizophrenia in rural Symptoms of mental disorders are often attributed to other China (Ran and others 2003) and to be cost-effective compared illnesses, and mental disorders are often not considered health with other standard treatment (Xiong and others 1994). problems (Jacob 2001). Many nonscientific explanations for Evidence also shows that nurses can replace physicians as pri- mental illness exist, and stigma exists to varying degrees every- mary health care providers in certain circumstances without where (Weiss and others 2001) with widespread delays or fail- loss of effectiveness (Climent and others 1978). Primary care ure to seek appropriate care (James and others 2002). practitioners need support to develop skills and experience in When care is sought, a hierarchy of interventions comes into diagnosing and treating mental disorders: they need a sustain- play, ranging from self-help, informal community support, tra- able supply of medicines, access to supervision, and incentives ditional healers, primary health care, specialist community to see patients with mental illness (Abas and others 2003). mental health care, and psychiatric units in general hospitals to Community approaches using low-cost, locally available specialist long-stay mental hospitals. The mix of interventions resources may improve treatment adherence and clinical out- depends on the availability of resources within a country or comes even in rural and underresourced settings (Chatterjee region (Saxena and Maulik 2003). The more resource- and others 2003; Srinivasa Murthy and others 2005). constrained the country or region is, the greater is the reliance In most countries, acute inpatient beds are being moved on self-help, informal community support (especially family- from mental hospitals into general or district hospitals. based), and primary health care. Although this policy potentially improves accessibility and Traditional healers are often the first source individuals with increases the links with, and support provided to, primary mental illness and their families turn to for professional assis- mental health care, concerns can be raised as to whether gen- tance (see, for example, Abiodun 1995). A recent review of com- eral hospitals can adapt to provide adequate services to people mon mental disorders among primary health clinics and tradi- with severe mental disorders. However, such services have tional healers in urban Tanzania showed that the prevalence of been effectively established in a number of countries (see, for common mental disorders among those attending traditional example, Alem and others 1999; Kilonzo and Simmons 1998), healers was double that of patients at primary health care centers showing this form of service delivery to be feasible when it is (Ngoma, Prince, and Mann 2003). Traditional healers are a het- clinically indicated. erogeneous group and include faith healers, spiritual healers, Nongovernmental organizations are important providers of religious healers, and practitioners of indigenous or alternative mental health care. An estimated 93 percent of African and systems of medicine. In some countries, they are part of the 80 percent of Southeast Asian countries have nongovernmental informal health sector, but in other countries, traditional healers organizations in the mental health sector. They provide diverse charge for their services and should be considered part of the services--including advocacy, informal support, housing, private health care sector. Often, traditional healers have high suicide prevention, substance misuse counseling, dementia acceptability and are accessible; at times, traditional healers support, rehabilitation, research, and other programs--that 620 | Disease Control Priorities in Developing Countries | Steven Hyman, Dan Chisholm, Ronald Kessler, and others complement, or in some cases substitute for, public and private developing countries (Tansella and Thornicroft 1998; clinical services (Levkoff, Macarthur, and Bucknall 1995; Patel Townsend and others 2004; WHO 2003). and Thara 2003). Services for children and adolescents, the majority of the population in many developing countries, are even more defi- CONCLUSION: PUBLIC SUPPORT FOR A cient than those for adults. Priority needs to be given to these COST-EFFECTIVE INTERVENTION PACKAGE services (Rahman and others 2000). At the other end of the life spectrum, many developing countries are facing aging popula- In developing countries, much of the mental health care spend- tions with grossly underdeveloped aged care services (Levkoff, ing is reported to be out of pocket. Individuals purchase mod- Macarthur, and Bucknall 1995). The high level of civil conflict ern and traditional treatments if they can afford to do so. and natural disasters requires attention to postconflict and Although a large private health sector exists in low-income posttrauma mental health conditions. The prevalence of these countries (Mills and others 2002), the quality and cost vary. disorders is demonstrated by a recent study (Livanou, Basoglu, Although unregulated markets fail in health, they fail even and Kalendar 2002) showing that, of 1,000 survivors of the more in mental health. It is unlikely that a country will be able August 1999 earthquake in Turkey, the incidence of PTSD was to rely on an unregulated private sector to deliver services that 63 percent and of depression was 42 percent. will reduce the burden of mental disorders. Specialist mental health providers, especially mental hospi- In addition to being a large and growing component of dis- tals, tend to focus the services they provide on the lower- ease burden, mental disorders meet virtually all the criteria by prevalence, higher-disability disorders, such as schizophrenia which we determine the need for government involvement in and bipolar disorder. Modern treatments, if available and used, health care (Beeharry and others 2002). They affect the poor, allow most patients to be treated effectively out of hospital. cause externalities, and inflict catastrophic costs; moreover, pri- Specifically, the use of antipsychotic and mood-stabilizing vate demand is inadequate. Indeed, the authors recognize that drugs and the development of strategies for community-based the main measure of outcome used in this and other chapters-- treatment have led to the closing of large numbers of psychi- the disability-adjusted life year--is limited to capturing change atric inpatient beds in many countries and their replacement in service user­level symptoms, disability, recovery, and case- with community services and general hospital psychiatric units fatality. The DALY does not capture the positive change that (for example, Larrobla and Botega 2001). treatment may have on a number of other significant conse- However, in some countries, the majority of psychotic quences of mental disorders, including family burden (in par- patients remain in long-term inpatient facilities that engage in ticular, productive time and household resources given up in custodial care, which is often of poor quality; moreover, basic the care of the sick family member) and lost productivity, at the rights are often violated at such facilities (van Voren and level of both the individual and the household (treatment accel- Whiteford 2000). Even if the quality of care is reasonable, acces- erates return to paid work or usual household activities) and, by sibility is a problem: these hospitals are often situated in urban implication, at the level of society in general. The evidence base areas, but populations are largely rural and have limited trans- for these productivity increases, although modest in volume, portation (Saraceno and others 1995). Furthermore, the con- constitutes an important additional argument alongside "cost centration of resources in these facilities can leave little for per DALY" considerations for investing in mental health. other service components (Gallegos and Montero 1999). For The total budgetary requirements and health consequences example, in Indonesia, 97 percent of the mental health budget of a cost-effective package of mental health care can begin to be is spent on public mental hospitals (Trisnantoro 2002). For mapped out by selecting one intervention for each of the four many developing countries, the debate about the role of, or disorders considered in this chapter. Although the data avail- problems with, mental hospitals is subsumed within a gross able for this exercise have limitations and will need to be refined deficiency of psychiatric beds of any kind. with further research, table 31.7 summarizes the estimated The priority for virtually all countries is generating suffi- costs and effects of a package consisting of (a) outpatient-based cient resources for primary mental health care and deciding treatment of schizophrenia and bipolar disorder with first-gen- how to expand and best use scarce specialist resources. The eration antipsychotic or mood-stabilizing drugs and adjuvant quality of care is often very poor, and huge variations exist in psychosocial treatment, (b) proactive care of depression in pri- resource availability between countries (Saxena and Maulik mary care with generic SSRIs (including maintenance treat- 2003; WHO 2001). Very few countries have what could be con- ment of recurrent episodes), and (c) treatment of panic disor- sidered an optimal mix of these services, and there are no uni- der in primary care with generic SSRIs. The estimated benefit of versally accepted planning parameters. However, conceptual such a package would be an annual reduction of 2,000 to 3,000 models for developing national mental health policy and DALYs per 1 million population, at a cost of US$3 million to guidelines for service planning exist that can be useful in US$9 million (that is, US$3 to US$4 per capita in Sub-Saharan Mental Disorders | 621 Table 31.7 Costs and Effects of a Specified Mental Health Care Package World Bank region Sub-Saharan Latin America Middle East and Europe and East Asia and Africa and the Caribbean North Africa Central Asia South Asia the Pacific Total effect (DALYs averted per year per 1 million population) Schizophrenia: older antipsychotic drug plus 254 373 364 353 300 392 psychosocial treatment Bipolar disorder: older mood-stabilizing drug 312 365 322 413 346 422 plus psychosocial treatment Depression: proactive care with newer 1,174 1,953 1,806 1,789 1,937 1,747 antidepressant drug (SSRI; generic) Panic disorder: newer antidepressant drug 245 307 287 307 284 330 (SSRI; generic) Total effect of interventions 1,985 2,998 2,779 2,862 2,867 2,891 Total cost (US$ million per year per 1 million population) Schizophrenia: older antipsychotic drug 0.47 1.81 1.61 1.32 0.52 0.75 plus psychosocial treatment Bipolar disorder: older mood-stabilizing drug 0.48 1.80 1.23 1.39 0.62 0.95 plus psychosocial treatment Depression: proactive care with newer 1.80 4.80 3.99 3.56 2.81 2.59 antidepressant drug (SSRI; generic) Panic disorder: newer antidepressant drug 0.15 0.27 0.21 0.23 0.16 0.20 (SSRI; generic) Total cost of interventions 2.9 8.7 7.0 6.5 4.1 4.5 Cost-effectiveness (DALYs averted per US$1 million expenditure) Schizophrenia: older antipsychotic drug 544 206 226 267 574 522 plus psychosocial treatment Bipolar disorder: older mood-stabilizing drug 647 203 262 298 560 446 plus psychosocial treatment Depression: proactive care with newer 652 407 452 502 690 675 antidepressant drug (SSRI; generic) Panic disorder: newer antidepressant drug 1,588 1,155 1,339 1,350 1,765 1,649 (SSRI; generic) Source: Authors' own estimates. Africa and South Asia, and US$7 to US$9 per capita in Latin cultural, financial, and structural barriers that prevent people America and the Caribbean). Accordingly, for every US$1 mil- from seeking and receiving treatment. We need to close the gap lion invested in such a mental health care package, 350 to 700 between what we know and what we do in treating mental dis- healthy years of life would be gained over what would occur orders. We can alleviate the substantial burden of these disor- without intervention. ders and reverse or limit many of the devastating social and At a country level, data such as those presented in this chap- economic impacts. ter can be used to estimate the proportion of burden currently averted, the proportion that can be averted with current knowl- REFERENCES edge and optimal coverage, and the burden not able to be averted with current knowledge. Such modeling has been done Abas, M., L. Mbengeranwa, I. Chagwedera, P. Maramba, and J. Broadhead. for some countries (for example, Andrews and others 2004). 2003. "Primary Care Services for Depression in Harare, Zimbabwe." Harvard Review of Psychiatry 11 (3): 157­65. Although much remains to be learned about the etiology Abiodun, O. 1995."Pathways to Mental Health Care in Nigeria." Psychiatric and treatment of mental disorders, the potential clearly exists Services 46 (8): 823­26. for a considerable reduction in the burden caused by them. Alem, A., L. Jacobsson, M. Araya, D. Kebede, and G. Kullgren. 1999. "How For these gains to be made, the challenge is to overcome the Are Mental Disorders Seen, and Where Is Help Sought in a Rural 622 | Disease Control Priorities in Developing Countries | Steven Hyman, Dan Chisholm, Ronald Kessler, and others Ethiopian Community?" Acta Psychiatrica Scandinavica 100 (Suppl. Primary Care: Demonstration Cost-Outcome Study in India and 397): 40­47. Pakistan." British Journal of Psychiatry 176 (6): 581­88. American Psychiatric Association. 2000. Diagnostic and Statistical Manual Chisholm, D., M. Van Ommeren, J. L. Ayuso-Mateos, and S. Saxena. of Mental Disorders. 4th ed., text revision. Washington, DC: American Forthcoming. "Cost-Effectiveness of Clinical Interventions for Psychiatric Association. Reducing the Global Burden of Bipolar Disorder: A Global Analysis Andrade, L., J. J. Caraveo-Anduaga, P. Berglund, R. V. Bijl, E. (WHO-CHOICE)." British Journal of Psychiatry. Dragomirecka, R. Kohn, and others. 2003. "The Epidemiology of Climent, C. E., M. V. De Arango, R. Plutchick, and C. A. Leon. 1978. Major Depressive Episodes: Results from the International "Development of an Alternative, Efficient Low Cost Mental Health Consortium of Psychiatric Epidemiology (ICPE) Surveys." Delivery System in Cali, Colombia, 1: The Auxiliary Nurse." Social International Journal of Methods in Psychiatric Research 12 (1): 3­21. Psychiatry 13 (1): 29­35. Andrews, G. 2001. "Placebo Response in Depression: Bane of Research, Das Gupta, R., and J. F. Guest. 2002. "Annual Cost of Bipolar Disorder to Boon to Therapy." British Journal of Psychiatry 178 (3): 192­94. U.K. Society." British Journal of Psychiatry 180 (3): 227­33. Andrews, G., C. Issakidis, K. Sanderson, J. Corry, and H. Lapsley. 2004. de Graaf, R., R. V. Bijl, J. Spijker, A. T. Beekman, and W. A. Vollebergh. 2003. "Utilizing Survey Data to Inform Public Policy: Comparison of the "Temporal Sequencing of Lifetime Mood Disorders in Relation to Cost-Effectiveness of Treatment of Ten Mental Disorders" British Comorbid Anxiety and Substance Use Disorders--Findings from the Journal of Psychiatry 184 (6): 526­33. Netherlands Mental Health Survey and Incidence Study." Social Andrews, G., K. Sanderson, J. Corry, C. Issakidis, and H. Lapsley. 2003. Psychiatry and Psychiatric Epidemiology 38 (1): 1­11. "Cost-Effectiveness of Current and Optimal Treatment for De Jong, J. 1996. "A Comprehensive Public Mental Health Programme in Schizophrenia." British Journal of Psychiatry 183 (5): 427­35. Guinea-Bissau: A Useful Model for African, Asian, and Latin-American Countries." Psychological Medicine 26 (1): 97­108. Angst, J., and R. Sellaro. 2000. "Historical Perspectives and Natural History of Bipolar Disorder." Biological Psychiatry 48 (6): 445­57. ESEMeD/MHEDEA 2000 Investigators. 2004. "Use of Mental Health Services in Europe: Results from the European Study of Epidemiology Araya, R., G. Rojas, R. Fritsch, J. Gaete, M. Rojas, and T. J. Peters. 2003. of Mental Disorders, ESEMeD Project." Acta Psychiatrica Scandinavica "Treating Depression in Primary Care in Low-Income Women in 109 (Suppl. 420): 47­54. Santiago, Chile: A Randomised Controlled Trial." Lancet 361 (9362): 995­1000. Faravelli, C., S. Paterniti, and A. Scarpato. 1995. "5-Year Prospective, Naturalistic Follow-up Study of Panic Disorder." Comprehensive Astrup, C., A. Fossum, and R. Holmboe. 1959. "A Follow-up Study of 270 Psychiatry 36 (4): 271­77. Patients with Acute Affective Psychoses." Acta Psychiatrica Scandinavica 34 (Suppl. 135): 1­65. Ferri, C., D. Chisholm, M. Van Ommeren, and M. Prince. 2004. "Resource Utilisation for Neuropsychiatric Disorders in Developing Countries: Bakker, A., A. J. L. M. van Balkom, P. Spinhoven, B. M. Blaauw, and R. van A Multinational Delphi Consensus Study." Social Psychiatry and Dyck. 1998. "Follow-up on the Treatment of Panic Disorder with or Psychiatric Epidemiology 39 (3): 218­27. without Agoraphobia: A Quantitative Review." Journal of Nervous and Mental Disease 186 (7): 414­19. Frank, E., H. A. Swartz, and D. J. Kupfer. 2000. "Interpersonal and Social Rhythm Therapy: Managing the Chaos of Bipolar Disorder." Biological Beeharry, G., H. Whiteford, D. Chambers, and F. Baingana. 2002. Psychiatry 48 (6): 593­604. "Outlining the Scope for Public Sector Involvement in Mental Health." Health Nutrition and Population Discussion Paper, World Bank, Gallegos, A., and F. Montero. 1999. "Issues in Community Based Washington, DC. Rehabilitation for Persons with Mental Illness in Costa Rica." International Journal of Mental Health 28: 25­30. Bijl, R. V., G. van Zessen, A. Ravelli, C. de Rijk, and Y. Langendoen. 1998. "The Netherlands Mental Health Survey and Incidence Study Geddes, J., S. M. Carney, T. A. Furukawa, D. J. Kupfer, and G. M. Goodwin. (NEMESIS): Objectives and Design." Social Psychiatry and Psychiatric 2003. "Relapse Prevention with Antidepressant Drug Treatment in Epidemiology 33 (12): 581­86. Depressive Disorders: A Systematic Review."Lancet 361 (9358): 653­61. Bolton, P., J. Bass, R. Neugebauer, H. Verdeli, K. Clougherty, Goodwin, F. K., B. Fireman, G. E. Simon, E. Hunkeler, J. Lee, and P. Wickramaratne, and others. 2003. "Group Interpersonal D. Revicki. 2003. "Suicide Risk in Bipolar Disorder during Treatment Psychotherapy for Depression in Rural Uganda." Journal of the with Lithium and Divalproex." Journal of the American Medical American Medical Association 289 (23): 3117­24. Association 290 (11): 1467­73. Bowden, C. L., J. R. Calabrese, S. L. McElroy, L. Gyulai, A. Wassef, F. Petty, Harrison, G., K. Hopper, T. Craig, E. Laska, C. Diegel, J. Wanderling, and and others. 2000. "A Randomized, Placebo-Controlled 12-Month Trial others. 2001. "Recovery from Psychotic Illness: A 15- and 25-Year of Divalproex and Lithium in Treatment of Outpatients with Bipolar I International Follow-up Study." British Journal of Psychiatry 178 (6): Disorder: Divalproex Maintenance Study Group." Archives of General 506­17. Psychiatry 57 (5): 481­89. Helgason, T. 1964. "Epidemiology of Mental Disorders in Iceland: A Bustillo, J. R., J. Lauriello, W. P. Horan, and S. J. Keith. 2001. "The Psychiatric and Demographic Investigation of 5,395 Icelanders." Acta Psychosocial Treatment of Schizophrenia: An Update." American Psychiatrica Scandinavica 40 (Suppl. 173): 1­180. Journal of Psychiatry 158 (2): 163­75. Huxley, N. A., S. V. Parikh, and R. J. Baldessarini. 2000. "Effectiveness of Chatterjee, S., V. Patel, A. Chatterjee, and H. Weiss. 2003. "Evaluation of a Psychosocial Treatments in Bipolar Disorder: State of the Evidence." Community-Based Rehabilitation Model for Chronic Schizophrenia in Harvard Review of Psychiatry 8 (3): 126­40. Rural India." British Journal of Psychiatry 182 (1): 57­62. Hwu, H. G., E. K. Yeh, and L. Y. Cheng. 1989. "Prevalence of Psychiatric Chisholm, D., K. Sanderson, J. L. Ayuso-Mateos, and S. Saxena. 2004. Disorders in Taiwan Defined by the Chinese Diagnostic Interview "Reducing the Burden of Depression: A Population-Level Analysis of Schedule." Acta Psychiatrica Scandinavica 79 (2): 136­47. Intervention Cost-Effectiveness in 14 Epidemiologically Defined Sub- Hyman, S. E. 2000. "The Genetics of Mental Illness: Implications for Regions (WHO-CHOICE)." British Journal of Psychiatry 184 (5): Practice." Bulletin of the World Health Organization 78 (4): 455­63. 393­403. Jablensky, A. N. 2000. "Epidemiology of Schizophrenia: The Global Chisholm, D., K. Sekar, K. K. Kumar, K. Saeed, S. James, M. Mubbashar, Burden of Disease and Disability." European Archives of Psychiatry and and R. S. Murthy. 2000. "Integration of Mental Health Care into Clinical Neuroscience 250 (6): 274­85. Mental Disorders | 623 Jablensky, A. N., G. Sartorius, M. Ernberg, A. Anker, J. E. Korten, R. Cooper, Livanou, M., M. Basoglu, and D. Kalendar. 2002. "Traumatic Stress and others. 1992. "Schizophrenia: Manifestations, Incidence, and Responses in Treatment-Seeking Earthquake Survivors in Turkey." Course in Different Cultures: A World Health Organization Ten- Journal of Nervous and Mental Disorders 190 (12): 816­23. Country Study." Psychological Medicine (Suppl 20): 1­97. Malt, U. F., O. H. Robak, H-P. Madsbu, and M. Loeb. 1999. "The Jacob, K. 2001. "Community Care for People with Mental Disorders in Norwegian Naturalistic Treatment Study of Depression in Primary Developing Countries." British Journal of Psychiatry 178 (4): 296­98. Practice (NORDEP)--I: Randomised Double Blind Study." British James, S., D. Chisholm, R. S. Murthy, K. Sekar, K. Saeed, and Medical Journal 318 (7192): 1180­84. M. Mubbashar. 2002. "Demand for, Access to, and Use of Community McGorry, P. D., A. R. Yung, L. J. Phillips, H. P. Yuen, S. Francey, E. M. Mental Health Care: Lessons from a Demonstration Project in India Cosgrave, and others. 2002. "Randomized Controlled Trial of and Pakistan." International Journal of Social Psychiatry 48 (3): 163­76. Interventions Designed to Reduce the Risk of Progression to First- Joy, C. B., C. E. Adams, and S. M. Lawrie. 2001. "Haloperidol versus Episode Psychosis in a Clinical Sample with Subthreshold Symptoms." Placebo for Schizophrenia." Cochrane Database of Systematic Reviews Archives of General Psychiatry 59 (10): 921­28. (2) D003082. http://www.mediscope.ch/cochrane-abstracts/ab003082. Mills, A., R. Brugha, K. Hanson, and B. McPake. 2002. "What Can Be Done htm. about the Private Health Sector in Low-Income Countries." Bulletin of Judd, L. L., H. S. Akiskal, J. D. Maser, P. J. Zeller, J. Endicott, W. Coryell, and the World Health Organization 80 (4): 325­30. others. 1998. "A Prospective 12-Year Study of Subsyndromal and Mohit, A., K. Saeed, D. Shahmohamadi, and J. Bolhari. 1999. "Mental Syndromal Depressive Symptoms in Unipolar Major Depressive Health Manpower Development in Afghanistan: Report of a Training Disorders." Archives of General Psychiatry 55 (8): 694­700. Course for Primary Health Care Physicians." Eastern Mediterranean Judd, L. L., P. J. Schettler, and H. S. Akiskal. 2002. "The Prevalence, Clinical Health Journal 5 (2): 215­19. Relevance, and Public Health Significance of Subthreshold Mortensen, P. B., C. B. Pedersen, T. Westergaard, J. Wohlfahrt, H. Ewald, Depressions." Psychiatric Clinics of North America 25 (4): 685­98. O. Mors, and others. 1999. "Effects of Family History and Place and Katschnig, H., M. Amering, J. M. Stolk, G. L. Klerman, J. C. Ballenger, Season of Birth on the Risk of Schizophrenia." New England Journal of A. Briggs, and others. 1995. "Long-Term Follow-up after a Drug Trial Medicine 340 (8): 603­8. for Panic Disorder." British Journal of Psychiatry 167 (4): 487­94. Mueller, T. I., A. C. Leon, M. B. Keller, D. A. Solomon, J. Endicott, Kendler, K. S., T. J. Gallagher, J. M. Abelson, and R. C. Kessler. 1996. W. Coryell, and others. 1999. "Recurrence after Recovery from Major "Lifetime Prevalence, Demographic Risk Factors, and Diagnostic Depressive Disorder during 15 Years of Observational Follow-up." Validity of Nonaffective Psychosis as Assessed in a U.S. Community American Journal of Psychiatry 156 (7): 1000­6. Sample: The National Comorbidity Survey." Archives of General Mulligan, J-A., J. A. Fox-Rushby, T. Adam, B. Johns, and A. Mills. 2003. Psychiatry 53 (11): 1022­31. "Unit Costs of Health Care Inputs in Low and Middle Income Kendler, K. S., C. A. Prescot, J. Myers, and M. C. Neale. 2003. "The Regions." Working Paper 9, Disease Control Priorities Project, Fogerty Structure of Genetic and Environmental Risk Factors for Common International Center, National Institutes of Health, Bethesda, MD. Psychiatric and Substance Use Disorders in Men and Women." Archives http://www.fic.nih.gov/dcpp/wps.html. of General Psychiatry 60 (9): 929­37. Murray, C. J. L., and A. D. Lopez. 1996. The Global Burden of Diseases: A Kessler, R. C., and R. G. Frank. 1997. "The Impact of Psychiatric Disorders Comprehensive Assessment of Mortality and Disability from Diseases, on Work Loss Days." Psychological Medicine 27 (4): 861­73. Injuries, and Risk Factors in 1990 and Projected to 2020. Boston: Harvard School of Public Health; Geneva: World Health Organization; Kessler, R. C., K. A. McGonagle, S. Zhao, C. B. Nelson, M. Hughes, S. Washington, DC: World Bank. Eshleman, and others. 1994. "Lifetime and 12-Month Prevalence of DSM-III-R Psychiatric Disorders in the United States: Results from the Ngoma, M., M. Prince, and A. Mann. 2003. "Common Mental Disorders National Comorbidity Survey." Archives of General Psychiatry 51 (1): among Those Attending Primary Health Clinics and Traditional Healers 8­19. in Urban Tanzania." British Journal of Psychiatry 183 (4): 349­55. Kessler, R. C., C. B. Nelson, K. A. McGonagle, J. Liu, M. Swartz, and D. G. Nicolson, R., and J. L. Rapoport. 1999. "Childhood-Onset Schizophrenia: Blazer. 1996. "Comorbidity of DSM-III-R Major Depressive Disorder Rare but Worth Studying." Biological Psychiatry 46 (10): 1418­28. in the General Population: Results from the U.S. National Comorbidity Olfson, M., R. C. Kessler, P. A. Berglund, and E. Lin. 1998. "Psychiatric Survey." British Journal of Psychiatry 168 (Suppl. 30): 17­30. Disorder Onset and First Treatment Contact in the United States and Kilonzo, G., and N. Simmons. 1998. "Development of Mental Health Ontario." American Journal of Psychiatry 155 (10): 1415­22. Services in Tanzania: A Reappraisal for the Future." Social Science and Patel, V. 1996. "Influences on Cost-Effectiveness." British Journal of Medicine 47 (4): 419­28. Psychiatry 169 (3): 381. Kohn, R., S. Saxena, I. Levav, and B. Saraceno. 2004. "The Treatment Gap Patel, V., D. Chisholm, S. Rabe-Hesketh, F. Dias-Saxena, G. Andrew, and in Mental Health Care." Bulletin of the World Health Organization 82 A. Mann. 2003. "Efficacy and Cost-Effectiveness of Drug and (11): 858­66. Psychological Treatments for Common Mental Disorders in General Larrobla, C., and N. Botega. 2001. "Restructuring Mental Health: A South Health Care in Goa, India: A Randomised Controlled Trial." Lancet 361 American Survey." Social Psychiatry and Psychiatric Epidemiology 36 (9351): 33­39. (5): 256­59. Patel, V., and A. Kleinman. 2003. "Poverty and Common Mental Disorders Lauer, J. A., C. J. L. Murray, K. Roehrich, and H. Wirth. 2003. "PopMod: A in Developing Countries." International Journal of Public Health 81 (8): Longitudinal Population Model with Two Interacting Disease States." 609­15. Cost Effectiveness and Resource Allocation 1: 6. Patel, V., and R. Thara, eds. 2003. Meeting Mental Health Needs of Levkoff, S., I. Macarthur, and J. Bucknall. 1995. "Elderly Mental Health in Developing Countries: NGO Innovations in India. New Delhi: Sage. the Developing World." Journal of Social Science and Medicine 41 (7): Paykel, E. S., and R. Priest. 1992. "Recognition and Management of 983­1003. Depression in General Practice: Consensus Statement." British Medical Linden, M., Y. Lecrubier, C. Bellantuono, O. Benkert, S. Kisely, and Journal 305 (6863): 1198­202. G. Simon. 1999. "The Prescribing of Psychotropic Drugs by Primary Pereira, J., and V. Patel. 1999."Which Antidepressants Are Best Tolerated in Care Physicians: An International Collaborative Study." Journal of Primary Care? A Pilot Randomized Trial in Goa." Indian Journal of Clinical Psychopharmacology 19 (2): 132­40. Psychiatry 41 (4): 358­63. 624 | Disease Control Priorities in Developing Countries | Steven Hyman, Dan Chisholm, Ronald Kessler, and others Perugi, G., and H. S. Akiskal. 2002. "The Soft Bipolar Spectrum Redefined: Townsend, C., H. Whiteford, F. Baingana, W. Gulbinat, R. Jenkins, A. Baba, Focus on the Cyclothymic, Anxious-Sensitive, Impulse-Dyscontrol, and others. 2004. "A Mental Health Policy Template: Domains and and Binge-Eating Connection in Bipolar II and Related Conditions." Elements for Mental Health Policy Formulation." International Review Psychiatric Clinics of North America 25 (4): 713­37. of Psychiatry 16 (1­2): 18­23. http://www.qcmhr.uq.edu.au/template/. Rahman, A., M. Mubbashar, R. Gater, and D. Goldberg. 1998. Trisnantoro, L. 2002. "Decentralization Policy on Public Mental Hospitals "Randomised Trial of Impact of School Mental Health Programme in in Indonesia: A Financial Perspective." Paper presented at the Seminar Rural Rawalpindi, Pakistan." Lancet 352 (9133): 1022­25. on Mental Health and Health Policy in Developing Countries, May 15, Rahman, A., M. Mubbashar, R. Harrington, and R. Gater. 2000. Harvard University. "Annotation: Developing Child Mental Health Services in Tsuang, D., and W. Coryell. 1993. "An 8-Year Follow-up of Patients with Developing Countries." Journal of Child Psychology and Psychiatry 41 DSM-III-R Psychotic Depression, Schizoaffective Disorder, and (5): 539­46. Schizophrenia." American Journal of Psychiatry 150 (8): 1182­88. Ran, M. S., M. Z. Xiang, C. L. W. Chan, J. Leff, P. Simpson, M. S. Huang, Ustun, T. B., J. L. Ayuso-Mateos, S. Chatterji, C. Mathers, and C. J. L. and others. 2003. "Effectiveness of Psychoeducational Intervention for Murray. 2004. "Global Burden of Depressive Disorders: Methods and Rural Chinese Families Experiencing Schizophrenia." Social Psychiatry Data Sources." British Journal of Psychiatry 184 (5): 386­92. and Psychiatric Epidemiology 38 (2): 69­75. van Voren, R., and H. Whiteford. 2000. "Reform of Mental Health in Robins, L. N., and D. A. Regier. 1991. Psychiatric Disorders in America: The Eastern Europe." Eurohealth Special Issue 6 (2): 63­65. Epidemiologic Catchment Area Study. New York: Free Press. Vega, W. A., B. Kolody, S. Aguilar-Gaxiola, E. Alderete, R. Catalana, and Robinson, D., M. G. Woerner, J. M. Alvir, R. Bilder, R. Goldman, S. Geisler, J. J. Caraveo-Anduaga. 1998. "Lifetime Prevalence of DSM-III-R and others. 1999. "Predictors of Relapse Following Response from a Psychiatric Disorders among Urban and Rural Mexican Americans in First Episode of Schizophrenia or Schizoaffective Disorder." Archives of California." Archives of General Psychiatry 55 (9): 771­78. General Psychiatry 56 (3): 241­47. Vicente, B., P. Rioseco, S. Saldivia, R. Kohn, and S. Torres. 2002. "Chilean Sanderson, K., G. Andrews, J. Corry, and H. Lapsley. 2004. "Modeling Study on the Prevalence of Psychiatric Disorders (DSM-III-R/CIDI) Change in Preference Values from Descriptive Health Status Using the (ECPP)." Revista Medica de Chile 130 (5): 527­36. Effect Size." Quality of Life Research 13 (7): 1255­64. Vijayakumar, L., K. Nagaraj, and S. John. 2004. "Suicide and Suicide Saraceno, B., E. Terzian, F. Barquero, and G. Tognoni. 1995."Mental Health Prevention in Developing Countries." Working Paper 27, Disease Care in the Primary Health Care Setting: A Collaborative Study in Six Control Priorities Project, Fogerty International Center, National Countries of Central America." Health Policy and Planning 10 (2): Institutes of Health, Bethesda, MD. http://www.fic.nih.gov/dcpp/ 133­43. wps.html. Saxena, S., and P. K. Maulik. 2003. "Mental Health Services in Low- and Wang, P. S., G. E. Simon, and R. C. Kessler. 2003. "The Economic Burden Middle-Income Countries: An Overview." Current Opinion in of Depression and the Cost-Effectiveness of Treatment." International Psychiatry 16 (4): 437­42. Journal of Methods in Psychiatric Research 12 (1): 22­33. Solomon, D. A., M. B. Keller, A. C. Leon, T. I. Mueller, M. T. Shea, Weiss, M. G., S. Jadhav, R. Raguram, P. Vounatsou, and R. Littlewood. M. Warshaw, and others. 1997. "Recovery from Depression: A 10-Year 2001. "Psychiatric Stigma across Cultures: Local Validation in Prospective Follow-up across Multiple Episodes." Archives of General Bangalore and London." Anthropology and Medicine 8 (1): 71­87. Psychiatry 54 (11): 1001­6. WHO (World Health Organization). 1975. Organization of Mental Health Srinivasa Murthy, R., K. Kishore Kumar, D. Chisholm, S. Kumar, Services in Developing Countries: Sixteenth Report of the WHO Expert T. Thomas, K. Sekar, and C. Chandrashekar. 2005. "Community Committee on Mental Health. Technical Report Series 564, WHO, Outreach for Untreated Schizophrenia in Rural India: A Follow-up Geneva. Study of Symptoms, Disability, Family Burden, and Costs." ------. 1992. The ICD-10 Classification of Mental and Behavioral Psychological Medicine 35: 341­51. Disorders. Geneva: WHO. Storosum, J. G., B. J. van Zweiten, W. van den Brink, B. Gersons, and M. D. ------. 2001. "Mental Health Resources: Project Atlas." WHO, Geneva. Broekmans. 2001. "Suicide Risk in Placebo-Controlled Studies of http://www.who.int/mip/2003/other_documents/en/EAARMentalHea Major Depression." American Journal of Psychiatry 158 (8): 1271­75. lthATLAS.pdf. Tansella, M., and G. Thornicroft. 1998. "A Conceptual Framework for ------. 2003. "Mental Health Policy and Services Development Project." Mental Health Services: The Matrix Model." Psychological Medicine WHO, Geneva. http://www.who.int/mental_health/policy/en/. 28 (3): 503­8. Xiong, W., M. R. Phillips, X. Hu, R. Wang, Q. Dai, J. Kleinman, and ten Have, M., W. Vollebergh, R. Bijl, and W. A. Nolen. 2002. "Bipolar A. Kleinman. 1994. "Family-Based Intervention for Schizophrenic Disorder in the General Population in the Netherlands (Prevalence, Patients in China: A Randomised Controlled Trial." British Journal of Consequences, and Care Utilisation): Results from the Netherlands Psychiatry 165 (2): 239­47. Mental Health Survey and Incidence Study (NEMESIS)." Journal of Yonkers, K. A., S. E. Bruce, I. R. Dyck, and M. B. Keller. 2003. "Chronicity, Affective Disorders 68 (2­3): 203­13. Relapse, and Illness-Course of Panic Disorder, Social Phobia, and Thara, R., R. Padmavati, and T. Srinivasan. 2004. "Focus on Psychiatry in Generalized Anxiety Disorder: Findings in Men and Women from India." British Journal of Psychiatry 184 (4): 366­73. 8 Years of Follow-up." Depression and Anxiety 17 (3): 173­79. Mental Disorders | 625 Chapter 32 Neurological Disorders Vijay Chandra, Rajesh Pandav, Ramanan Laxminarayan, Caroline Tanner, Bala Manyam, Sadanand Rajkumar, Donald Silberberg, Carol Brayne, Jeffrey Chow, Susan Herman, Fleur Hourihan, Scott Kasner, Luis Morillo, Adesola Ogunniyi, William Theodore, and Zhen-Xin Zhang Historically, policy makers and researchers have used mortality loss of family income; the requirement for caregiving, with statistics as the principal measure of the seriousness of diseases, further potential loss of wages; the cost of medications; and the based on which countries and organizations have launched need for other medical services--can be expected to be partic- disease control programs. Mortality statistics alone, however, ularly devastating among those with limited resources. In addi- underestimate the suffering caused by diseases that may be non- tion to health costs, those suffering from these conditions are fatal but cause substantial disability. Many neurological and also frequently victims of human rights violations, stigmatiza- psychiatric conditions belong in this category. The absence of tion, and discrimination. Stigmatization and discrimination some neurological disorders from lists of leading causes of death further limit patients' access to treatment. These disorders, has contributed to their long-term neglect. When the relative therefore, require special attention in developing countries. seriousness of diseases is assessed by time lived with disability This chapter addresses Alzheimer's disease (AD) and other rather than by mortality, several neurological disorders appear dementias, epilepsy, Parkinson's disease (PD), and acute as leading causes of suffering worldwide. ischemic stroke. These conditions are current or emerging pub- World Health Organization data suggest that neurological lic health problems in developing countries, as assessed by high and psychiatric disorders are an important and growing cause prevalence, large numbers of people who are untreated, and of morbidity. The magnitude and burden of mental, neuro- availability of inexpensive but effective interventions that could logical, and behavioral disorders is huge, affecting more than be applied on a large scale through primary care. Unfortu- 450 million people globally. According to the Global Burden nately, reliable population-based data from developing coun- of Disease Report, 33 percent of years lived with disability and tries on the epidemiology of these and other neurological 13 percent of disability-adjusted life years (DALYs) are due to disorders are extremely limited. Some other important neuro- neurological and psychiatric disorders, which account for four logical conditions that cause high morbidity, such as headache, out of the six leading causes of years lived with disability are not covered because of difficulties in recommending (Mathers and others 2003). evidence-based interventions in developing countries. Unfortunately, the burden of these disorders in developing countries remains largely unrecognized. Moreover, the burden imposed by such chronic neurological conditions in general ALZHEIMER'S DISEASE AND OTHER DEMENTIAS can be expected to be particularly devastating in poor popula- tions. Primary manifestations of the impact on the poor-- Dementia is a deterioration of intellectual function and other including the loss of gainful employment, with the attendant cognitive skills that is of sufficient severity to interfere with 627 social or occupational functioning. Of the many diseases that Protective factors reported in the literature include a higher lead to dementia, AD is the most common cause worldwide level of education, a specific gene (APO E2), the intake of anti- among people age 65 and older, followed by vascular dementia, oxidants, and the use of some anti-inflammatory medications mixed dementia consisting of AD plus vascular dementia, and (Henderson and Jorm 2000). The use of estrogen supplements dementia caused by general medical conditions. Although dis- for women was believed to be a protective factor for AD tinguishing AD from other causes of dementia is important, (Henderson 1997), but a recent study of women taking a com- particularly for treatment with acetylcholinesterase inhibitors, bination of estrogen and progesterone showed that these the burden from all causes of dementia is similar. Although the women had twice the risk of developing dementia than women discussion in this chapter deals mostly with AD, the role of taking a placebo (Shumaker and others 2003). treatable dementias in developing countries is important as it Studies from developed countries have reported median can reduce the burden of caring in families. survival after the onset of dementia symptoms ranging from 5.0 years to 9.3 years (Walsh, Welch, and Larson 1990). In devel- oping countries, the reported median survival was 3.3 years for Prevalence and Incidence Rate all demented subjects and 2.7 years for those with AD (Chandra More than 100 prevalence studies of AD and other dementias and others 1998). have been reported throughout the world. The prevalence of dementia has generally been found to double with every five- Burden of Disease year increase in age, from 3 percent at age 70 to 20 to 30 percent Burden of disease estimates of AD and other dementias include at age 85 (Henderson and Jorm 2000). Studies in developing vascular dementia, unspecified dementias, and other unclassi- countries have shown a prevalence of dementia ranging from fied degenerative diseases of the nervous system. Mathers and 0.84 to 3.50 percent (Chandra and others 1998; Hendrie and others (2003) estimate DALYs for all dementias as 17,108,000, others 1995; Rajkumar, Kumar, and Thara 1997). Several stud- with the burden being almost twice as much for females ies have reported the incidence rate of AD and other dementias (11,016,000) as for males (6,092,000). Because dementia is a in Europe and the United States (Jorm and Jolley 1998). disease of older ages, the burden from dementia is generally Compared with incidence rates in developed countries, very greater in high-income countries, where life expectancy is low age-specific incidence rates of AD and other dementias higher, diagnosis is better, and better treatment leads to have been reported from developing countries (Chandra and increased longevity. Note, however, the relatively high burden others 2001; Hendrie and others 2001). in East Asia and the Pacific and South Asia relative to their level A comparison of data from developed and developing coun- of economic development (table 32.1). tries raises several important questions. The reported differ- The bulk of care for those with dementia in developing ences in the prevalence of AD and other dementias across countries is provided by the family at home, where the main countries could be due partly to methodological differences or caregivers are spouses (36 percent) and children (42 percent) could be due to genuine differences caused by variations in diet, (Prince 2000). Women in both developed and developing education, life expectancy, sociocultural factors, and other risk countries are usually the main caregivers (Prince 2000). Studies factors. The low incidence reported from Ballabgarh, India, and in developed countries indicate that caregivers' psychological Ibadan, Nigeria, raises the possibility of environmental factors well-being is a key factor in patients' admission to nursing or or gene-environment interactions in the causation of AD. At residential care (Levin, Moriarty, and Gorbach 1994). the same time, multi-infarct dementia is more common than In estimating the overall costs of care for dementia, one primary degenerative dementia in China (Li and others 1991), must emphasize the value of reducing the burden on care- which also suggests variation in risk factors across countries. givers. Caregiving can result in social isolation, psychological stress, and high rates of depression (Buck and others 1997). Risk and Protective Factors and Survivorship However, the methodology for estimating the costs of informal care needs to be standardized. Three separate genes (APP, PS1, and PS2) are linked to early- onset, familial AD. Another gene (APO E4) is a risk factor for late-onset, nonfamilial cases (Henderson and Jorm 2000). Other Interventions genes have been implicated but not confirmed in large studies. As of now, there is no cure for AD, but some measures can pro- Other risk factors reported in the literature include increasing vide symptomatic relief to patients and caregivers. age, positive family history of dementia, female gender (but this factor is controversial), lower level of education, several medical Population-Based Interventions. No firm evidence indicates conditions, and exposure to such environmental factors as that any form of population-based intervention can prevent organic solvents and aluminum (Henderson and Jorm 2000). AD or that the progression of cognitive decline in old age can 628 | Disease Control Priorities in Developing Countries | Vijay Chandra, Rajesh Pandav, Ramanan Laxminarayan, and others Table 32.1 Disability-Adjusted Life Years by Cause and Region, 2001 (thousands) Latin America Middle East Sub- High- Global total East Asia and Europe and and the and Saharan income Condition Both sexes Males Females the Pacific Central Asia Caribbean North Africa South Asia Africa countries AD and other 17,108 6,092 11,016 4,110 1,612 1,215 292 1,955 450 7,468 dementias Epilepsy 6,223 3,301 2,922 1,303 354 737 248 1,741 1,373 464 PD 2,325 1,124 1,202 435 228 90 81 303 100 1,086 Cerebrovascular 72,024 35,482 36,542 25,832 12,616 3,936 1,948 13,184 5,125 9,354 disease Source: Mathers and others 2006. be halted or reduced. However, growing inferential evidence culturally appropriate interventions that can be delivered suggests that reducing the risk of brain trauma in earlier life, within existing resources, such as supporting families in their for example, by mandating seat belt and crash helmet use, may role as caregivers. help prevent dementia in later life (Gentleman, Graham and Treating underlying disease and risk factors for cardiovascu- Roberts 1993). lar disease can help prevent future cerebrovascular disease that could lead to multi-infarct dementia. Other conditions, such as Personal Interventions. There is a reduction in brain levels of hypothyroidism or vitamin B12 deficiency, which could lead to the neurotransmitter acetylcholine in patients suffering from or aggravate dementia, are easily treatable, and the costs of AD. Drugs that inhibit acetylcholinesterase, the enzyme treatment are much lower than the costs of dementia care. responsible for metabolizing acetylcholine, cause an increase in In Western countries, the model of care for patients with brain acetylcholine. Evidence from randomized trials has con- moderate to severe dementia is based on skilled, long-term care firmed that, for patients with mild to moderate AD, cognitive in institutions. However, such long-term care institutions do performance benefits, at least in the short term, from the use of not exist in developing countries, and if they were set up, they acetylcholinesterase inhibitors (Foster and others 1996). would be extremely expensive and beyond the reach of most Despite this benefit to patients, the practical benefits of treat- patients and their families. Thus, the model of care in develop- ment with acetylcholinesterase inhibitors are mainly attributa- ing countries should be based on home care, along with pro- ble to the lowered caregiver burden. The benefits of using viding training and support for family caregivers. acetylcholinesterase inhibitors for other dementias have yet to Interventions that should not be pursued include the use of be proven. multiple medications, which can be detrimental in older age The behavioral and psychological symptoms of dementia groups, particularly unproven medications such as cerebral are a major source of stress to family members providing care activators and neurotropic agents. In addition, in many devel- to patients. Training family caregivers in behavioral manage- oping countries, dementia is still equated with "madness," and ment techniques, including problem solving, memory training, patients are sometimes taken to traditional healers. Community and reality orientation, has been shown to reduce the level of education has a role to play in eliminating such practices. agitation and anxiety in people with dementia (Brodaty and Gresham 1989; Haupt, Karger, and Janner 2000). Use of low doses of antipsychotic medications, which calm the patient and EPILEPSY reduce symptoms such as aggression and wandering, have been shown to reduce caregiver stress, but these improvements have Epilepsy is a common brain disorder characterized by two or not been quantified (Melzer and others 2004). more unprovoked seizures. Seizures are discrete events caused Interventions that have specifically targeted stress and by transient, hypersynchronous, abnormal neuronal activity. depression among caregivers and have shown positive results Seizures may occur in close temporal association with a variety include caregiver training, counseling and support for care- of acute medical and neurological diseases, such as acute givers, and cognitive and behavioral family interventions stroke, sepsis, or alcohol withdrawal. However, the vast major- (Marriott and others 2000). Limitations to the implementation ity of seizures have no immediate identifiable cause. of such strategies include the need for training by specialists, Epilepsy can be broadly divided into three categories: which makes these strategies less suitable for developing idiopathic epilepsy (for example primary generalized countries. The challenge for developing countries is to develop childhood-onset absence epilepsy), which is thought to have a Neurological Disorders | 629 genetic basis; secondary or symptomatic epilepsy, which is disabilities are not a risk factor for epilepsy in themselves, but caused by a known central nervous system injury or disorder, they may be associated with seizure disorder (Casetta and such as infection, stroke, traumatic brain injury, or cerebral others 2002; Leone and others 2002). dysgenesis; and cryptogenic epilepsy, for which there is no clear evidence of an etiological factor. Idiopathic and cryptogenic Treatment Gap cases represent approximately 70 percent of epilepsy cases; the Epilepsy affects about 50 million people worldwide, of whom remaining 30 percent are symptomatic (secondary). approximately 80 percent live in developing countries (WHO 2000). The difference between the number of people with Prevalence, Incidence Rate, Remission, and Mortality active epilepsy and the number who are being appropriately treated in a given population at a given point in time is known The generally accepted estimate of the prevalence of active as the treatment gap. Meinardi and others (2001) estimate that epilepsy globally is in the range of 5 to 8 per 1,000 population, 90 percent of people with epilepsy in developing countries are but investigators from African and Latin American countries inadequately treated. Possible reasons for the high treatment report at least double the prevalence reported elsewhere gap include fear of stigmatization, cultural beliefs, lack of (Leonardi and Ustun 2002). knowledge about the medical nature of epilepsy, illiteracy, eco- The incidence rate of epilepsy in developed countries nomic issues, distance to health facilities, inadequate supply of is approximately 43 per 100,000 (Kotsopoulos and others antiepileptic drugs (AEDs), and lack of prioritization by health 2002). In developing countries, the incidence rate of epilepsy authorities (Wang and others 2003). Even in the developed is higher, with a median of 69 per 100,000 (Kotsopoulos and world, patients who live in isolated rural regions or inner-city others 2002). slums and those who are isolated from the majority because of Based on follow-up of patients under treatment by general cultural factors may suffer a treatment gap. practitioners in the United Kingdom, Cockerell and others (1997) report that after nine years 86 percent of epilepsy patients had achieved a remission of three years, and 68 percent Faith Healers had achieved a remission of five years. Thus, data from devel- Many people with epilepsy seek treatment from faith healers, to oped countries suggest a good outcome of seizure control in whom they pay large sums in cash or in kind for treatment with most patients with treatment. In developing countries, no beneficial medical effects. Karaagac and others (1999) find although many people with new onset seizures do not receive that in Silivri, Turkey, 65 percent of 49 people with epilepsy had treatment, some proportion of patients go into spontaneous visited religious figures at the onset or during the course of the remission even without treatment (Mani and others 1993). disease. A study from rural India revealed that 44 percent of However, the actual remission rate in developing countries is children with epilepsy had sought help from traditional practi- yet to be documented in population-based studies. tioners, whereas approximately 33 percent had received help The risk of premature death in people with epilepsy is two from both qualified and traditional practitioners (Pal and oth- to three times higher than for the general population. In addi- ers 2002). Native Americans still seek traditional healing cere- tion to sudden unexplained death, which occurs in up to 1 in monies for epilepsy instead of--or in addition to--Western 100 patients with severe refractory epilepsy, additional mortal- medicine. ity results from accidents and suicide. However, the exact cause of the increased risk is not known in most cases. Patient Compliance In a study in rural Thailand, only 57 percent of people with Risk Factors epilepsy were 100 percent compliant with treatment, possibly A reported risk factor for idiopathic (presumed genetic) because of misunderstanding of the instructions (48 percent), epilepsy is family history of epilepsy. Reported risk factors for forgetfulness (16 percent), and economic limitations (13 per- symptomatic epilepsy include prenatal or perinatal causes cent) (Asawavichienjinda, Sitthi-Amorn, and Tanyanont 2003). (obstetric complications, prematurity, low birthweight, neona- To improve compliance in a rural African community, medical tal asphyxia). Recent data suggest that the effect of obstetric personnel visited the community every 6 months and provided complications or neonatal asphyxia may have been overem- a long-term supply of medications; this effort led to a sub- phasized. Prematurity, low birthweight, and neonatal seizures stantial increase in compliance at 20 months (Kaiser and may be independent risk factors as well as markers of underly- others 1998). In India, Desai and others (1998) demonstrate ing disease. Other causes include traumatic brain injuries, cen- the dependency of compliance on access to free treatment. tral nervous system infections, cerebrovascular disease, brain Inadequate communication between doctors and patients tumors, and neurodegenerative diseases. Developmental influences compliance negatively (Gopinath and others 2000). 630 | Disease Control Priorities in Developing Countries | Vijay Chandra, Rajesh Pandav, Ramanan Laxminarayan, and others Burden of Disease AEDs (such as phenobarbital, phenytoin, carbamazepine, and The burden of disease (BOD) estimates for epilepsy include valproic acid) and newer AEDs (such as lamotrigine, oxcar- epilepsy and status epilepticus. Mathers and others (2003) esti- bazepine, and topiramate) in controlling seizure frequency and mate the DALYs for epilepsy as 6,223,000, with slightly higher (b) the safety of these AEDs when prescribed alone or in com- rates for males (3,301,000) than for females (2,922,000). Many bination. Some, but not all, of the new AEDs may be better risk factors for epilepsy are linked with a lower level of eco- tolerated in monotherapy and have fewer long-term adverse nomic development; thus, the burden is highest in South Asia effects than older AEDs. However, no study has shown any dif- followed by Sub-Saharan Africa (table 32.1). A notable obser- ference in efficacy between the older and newer medications vation is the reportedly low burden in the Middle East and (Aldenkamp, De Krom, and Reijs 2003). Newer medications North Africa, despite parts of that region being relatively are more expensive and, for people in most developing coun- underdeveloped. Epilepsy imposes a large economic burden on tries, are practically impossible to access. In some low-income patients and their families. It also imposes a hidden burden countries, however, even older AEDs are not available, and associated with stigmatization and discrimination against when they are, their supply is irregular. patients and even their families in the community, workplace, Newer AEDs are generally recommended as add-on or school, and home. Social isolation, emotional distress, depend- adjunctive drugs for better seizure control in patients with ence on family, poor employment opportunities, and personal refractory epilepsy already on AEDs. The first AED will render injury add to the suffering of people with epilepsy. approximately 50 percent of patients seizure free. Approxi- mately 20 to 40 percent of patients who do not respond to the first AED will respond to the introduction of a second AED, Interventions with a greater than 50 percent decrease in seizure frequency Currently, there are no preventive measures for idiopathic or (Schapel and others 1993). cryptogenic epilepsy; however, much can be done to prevent The Global Campaign against Epilepsy, which is jointly secondary seizures. sponsored by the World Health Organization, International League against Epilepsy, and International Bureau for Epilepsy, Population-Based Interventions. Public health policies, such advocates using phenobarbital to close the high treatment as better perinatal care by well-trained birth attendants (partic- gap in low-income countries. As a first step, all patients with ularly in rural areas) and strategies to control severe head epilepsy should be given phenobarbital, so that the majority injuries (for example, by means of laws requiring motorcyclists of patients responsive to phenobarbital will be appro- to wear helmets and prohibiting drunk driving), can modify priately treated. In resource-poor countries, phenobarbital risk factors for epilepsy and thereby reduce the incidence and can be provided for as little as US$5 to US$10 per year. Pheno- prevalence of epilepsy. Policies to control neurocysticercosis barbital has extremely low abuse potential. Its side effects-- (for instance, building latrines in rural areas) can serve to pre- predominantly sedation, possible mild cognitive impairment, vent such infections. Mass deworming for neurocysticercosis and depression--have limited its use in industrial countries. In has not been shown to be effective in the long term (Pal, developing countries, however, side effects are less important Carpio, and Sander 2000) but was effective in a campaign in than uncontrolled seizures, and they can be diminished by Ecuador (M. Cruz, personal communication, 2004). using the lowest possible effective doses. Thus, phenobarbital Estimates indicate that 70 to 80 percent of people in devel- is the drug of choice for large-scale, community-based pro- oping countries live in rural and remote areas and have no easy grams, particularly in rural and remote areas of developing access to skilled medical care. Strategies that involve training countries. community-based health care providers who practice in these In recent years, some centers in both developed and devel- communities to identify and manage patients with epilepsy oping countries have been performing surgery on cases of should be considered. refractory epilepsy, that is, on patients who do not respond to Policies are needed to ensure the continuous availability of any AEDs. Before centers can undertake such surgery, however, cheap and efficacious medications, such as phenobarbital, to they must have the requisite expertise, facilities, and equip- all epilepsy patients. Campaigns to educate communities ment, including a skilled neurosurgeon. Proper selection of about the medical nature of epilepsy and to dispel myths and patients--for example, those with mesial temporal pathology misconceptions about epilepsy could reduce stigma against on MRI--is extremely important. A meta-analysis of studies of epilepsy and thereby encourage patients to seek medical people who underwent epilepsy surgery in developed countries treatment. shows that 58 percent are seizure free and 10 to 15 percent have reduced seizure frequency (Engel and others 2003). After sur- Personal Interventions. Researchers, primarily in high- gery, even if patients are seizure free, medication should be income countries, have tested (a) the efficacy of both older continued for one to two years (Engel and others 2003). Neurological Disorders | 631 PARKINSON'S DISEASE higher in females (1,202,000) than males (1,124,000). Though male gender is a risk factor for PD, the higher burden in females PD is characterized by bradykinesia, resting tremor, cogwheel may reflect their longer life span. As PD is a disease of older rigidity, postural reflex impairment, progressive course, and ages, the burden from PD is generally higher in high-income good response to dopaminergic therapy. Other distinct forms countries, where life expectancy is higher, diagnosis is better, of parkinsonism include relatively rare genetic forms and the and better treatment leads to increased longevity. However, the less common neurodegenerations with multiple system burden is high in East Asia and the Pacific and South Asia rela- involvement or significant striatal lesions (for example, pro- tive to that in other regions (table 32.1). gressive supranuclear palsy or multiple system atrophy). The economic burden of PD includes direct costs, such as Parkinsonism secondary to external causes, such as manganese for medication, physicians, hospitals, and chronic care facilities. poisoning or carbon monoxide poisoning, although now rare, Estimated indirect costs resulting from the loss of labor of both is referred to as secondary parkinsonism. Because the burden patients and caregivers typically exceed direct costs. The quality of these diseases to the patient is similar to or greater than that of life of both patients and caregivers is adversely affected. for PD and there is no evidence for addressing these disorders separately, they will not be distinguished here. Interventions Treatment of PD is based on symptomatic relief, except for pre- Prevalence, Incidence Rate, and Mortality venting secondary parkinsonism caused by neurotoxins. Prevalence estimates vary widely across populations (Tanner and Goldman 1996; Zhang and Roman 1993). Recent reports, Population-Based Interventions. No determinants of PD contrary to previous reports, suggest that the prevalence in amenable to population-based interventions have been developing and developed countries may be similar (Marras identified. and Tanner 2002). Few incidence studies have been performed, and none in developing countries. Van Den Eeden and others Personal Interventions. Specific curative or neuroprotective (2003) report the incidence rate of PD in the United States as treatments for PD have not been established. Interventions approximately 13 per 100,000 person-years. Men are affected are primarily directed at palliation of symptoms and include more commonly than women (Tanner and Goldman 1996). pharmaceuticals, surgery, physical therapy, and--in some Lower PD incidence in African Americans--and by extension countries--traditional medicines. Africans--has been suggested but is controversial (Van Den Levo-dopa (l-dopa), l-dopa/decarboxylase inhibitor is the Eeden and others 2003). Most mortality estimates available for most widely used therapy for PD. It provides partial relief of all developed countries show about a twofold overall increased PD symptoms. Despite its benefits, chronic side effects after mortality, independent of age, in those with PD (Berger and long-term use can cause significant morbidity. others 2000). Researchers in developing countries have studied the use of traditional medicines for PD. Clinical trials have shown that the seeds of Mucuna pruriens, which contain l-dopa, are a safe and Causes and Risk Factors effective treatment for PD (Parkinson's Disease Group 1995), The cause of PD is unknown. A specific environmental risk and in animal studies, they are two to three times more effec- factor has not been identified. Pure genetic forms account for tive than synthetic l-dopa dose per dose (Hussain and Manyam 10 to 15 percent of cases or fewer. Increasing age and male gen- 1997). This substance is available in ayurvedic formulations in der are risk factors worldwide (Marras and Tanner 2002). India at a much lower cost than that of synthetic antiparkin- Exposure to toxins, head trauma, frequent infections, diets high sonian drugs. Another traditional medicine is derived from in animal fat, and midlife adiposity have been reported to Banisteriopsis caapi, which tribal societies of the Amazonian increase PD risk, but none do so consistently (Tanner and jungle use to make a potent hallucinogenic brew. It reportedly Goldman 1996). The most consistent association is an inverse showed dramatic positive effects on rigidity and akinesia in association with cigarette smoking and caffeine consumption, 15 patients with postencephalitic parkinsonism (Lewin and suggesting a protective effect (Ascherio and others 2001). Schuster 1929). A third traditional option is tai chi, a basic exer- cise in traditional Chinese medicine that may help with some of the motor deficits of PD. Burden of Disease Surgical treatment for PD by deep brain stimulation is gen- The BOD estimates for PD include Parkinson's disease and sec- erally recommended to address the loss of efficacy of dopami- ondary parkinsonism. Mathers and others (2003) estimate the nergic drugs. For most patients, it is not effective independent DALYs for PD as 2,325,000, with the burden being slightly of drugs. Although a few will have dramatic improvement and 632 | Disease Control Priorities in Developing Countries | Vijay Chandra, Rajesh Pandav, Ramanan Laxminarayan, and others may be able to reduce or stop drugs, this effect is generally tem- He and others (1995) report the age-adjusted stroke porary. Criteria for selection of patients for deep brain stimula- incidence of 117 per 100,000 population in China. The annual tion include those with advanced disease who are responsive to incidence of stroke in China is reported to have increased in l-dopa, not demented, and in good general health. Additional both men and women, with an average annual percentage considerations are the high cost of the equipment, the need change of 4.5 and 4.2 percent, respectively (Wang, Zhao, and for trained personnel to program the device, and--in most Wu 2001). In Japan, the age-adjusted annual incidence of stroke cases--the need for several visits to a medical center to pro- was 105 per 100,000 (Fukiyama and others 2000). Wide varia- gram the stimulator correctly, with periodic returns to adjust tion within these countries and a high risk of death after the first the settings. stroke in the first year in Japan have been reported. Investigators believe that those observations are due to variations in the prevalence of hypertension and the consequent larger propor- STROKE tion of hemorrhagic stroke (Kiyohara and others 2003). Walker and others (2000) report the yearly age-adjusted Stroke, also known as cerebrovascular accident or brain attack, is mortality rate per 100,000 for age group 15 to 64 ranged from a syndrome caused by an interruption in the flow of blood to 35 to 65 in men and 27 to 88 in women in Tanzania. When part of the brain caused either by occlusion of a blood vessel compared with the rates in England and Wales--11 for men (ischemic stroke) or rupture of a blood vessel (hemorrhagic and 9 for women--these rates are extremely high. The authors stroke). The interruption in blood flow deprives the brain of postulate that the high rates in Tanzania are due to untreated nutrients and oxygen, resulting in injury to cells in the affected hypertension. Many developed countries have experienced a vascular territory of the brain. The occlusion of a blood vessel steep decline in stroke mortality in recent decades, but the rate can sometimes be temporary and present as a reversible neuro- of decline has fallen substantially in recent years (Liu, Ikeda, logical deficit, which is termed a transient ischemic attack. Even and Yamori 2001; Sarti and others 2000). Mortality from stroke though stroke is a clinical diagnosis, brain imaging is required has increased in some Eastern European countries (Sarti and to distinguish ischemic stroke from hemorrhagic stroke. When others 2000). imaging is unavailable, clinical scores can be useful to identify Approximately 15 percent of patients die shortly after a patients with intracerebral hemorrhage (Allen 1983; Poung- stroke. Of the remaining 85 percent, approximately 10 percent varin, Viriyavejakul, and Komontri 1991). recover almost completely, and 25 percent recover with minor impairments (National Stroke Association 2002). Thus, approximately 40 percent experience moderate to severe Frequency of Types of Strokes, Prevalence, Incidence Rate, impairments that require special rehabilitative care. About Mortality, and Disability after Stroke 10 percent will require care in a nursing home or other long- In most parts of the world, about 70 percent of strokes are due to term facility. ischemia, 27 percent are due to hemorrhage, and 3 percent are of unknown cause (Gunatilake, Jayasekera, and Premawardene 2001). Approximately 25 percent of all ischemic strokes are due Risk Factors to cardioembolic causes, with the proportion being higher Risk factors for stroke in general are similar to those for car- among younger individuals. In some parts of the world--for diovascular disease. Moreover, risk factors for first stroke and instance, China and Japan--hemorrhagic strokes account for a recurrence of stroke are also similar if they remain uncon- greater proportion of all strokes, ranging from 17.1 to 39.4 per- trolled after the first attack (see chapter 33). cent in China (Zhang and others 2003) to 38.7 percent in Japan Increasing age, particularly after 55, is one of the most (Fukiyama and others 2000). important risk factors for stroke (Thorvaldsen and others Comparable data do not exist for all parts of the world. Most 1995). Although stroke is more prevalent among men, stroke- morbidity data from Southeast Asian countries, for example, related fatality rates are higher among women (Goldstein and are hospital based and are, thus, likely to be underestimates, others 2001). Hypertension is the most important modifiable because many stroke patients die before they are brought to the determinant of both first and recurrent stroke (Eastern Stroke hospital. Mortality data are also likely to be underestimates, and Coronary Heart Disease Collaborative Research Group because verifying the cause of death is usually difficult. 1998). The association between blood pressure and stroke In India, the prevalence of stroke has been estimated at in East Asian populations seems stronger than in Western 203 per 100,000 population older than 20 (Anand and others populations (Eastern Stroke and Coronary Heart Disease 2001). The male-to-female ratio was one to seven. In Taiwan, Collaborative Research Group 1998). Other risk factors include China, the crude point prevalence was 592 per 100,000 (Huang, smoking, environmental exposure to tobacco smoke, dyslipi- Chiang, and Lee 1997). demia, atrial fibrillation, diabetes and impaired glucose Neurological Disorders | 633 tolerance, generalized and abdominal obesity, physical inactiv- which include not only recognized risk factors in developed ity, excess alcohol consumption, increased homocysteine levels, countries but also locally relevant risk factors, such as rheu- drug abuse, hemostatic factors, and existing cerebrovascular matic heart disease and puerperal stroke. disease (Goldstein and others 2001). Treatment strategies for acute ischemic stroke include the In developing countries, rheumatic heart disease leading to following: embolic stroke is also a major cause. This risk factor is declin- ing in importance with the control of rheumatic fever. · General management. Overall medical care of patients with Dehydration in postpartum women can lead to a stroke, par- an acute stroke is important. Attention to complications ticularly in remote areas where deliveries are conducted at such as bronchoaspiration, fluid and electrolyte imbalance, home. and control of blood sugar, as well as prevention of deep vein thrombosis, is crucial. Experience in developed coun- tries suggests that specialized stroke units provide the best Burden of Disease care for acute stroke patients (Smaha 2004), but in develop- The BOD estimates for stroke include subarachnoid hemor- ing countries, particularly in rural areas, where hospital beds rhage, intracerebral hemorrhage, cerebral infarction, and are scarce and most patients are attended by general physi- sequelae of cerebrovascular disease. Mathers and others (2003) cians, such units are impractical. estimate the DALYs for cerebrovascular disease as 72,024,000, · Platelet antiaggregants. Aspirin can prevent early stroke with the burden being almost similar for females (36,542,000) recurrence if given during the acute phase of stroke (within and males (35,482,000). The burden is highest in East Asia and 48 hours) (Chinese Acute Stroke Trial Collaborative Group the Pacific, followed by South Asia and by Europe and Central 1997; International Stroke Trial Collaborative Group 1997). Asia (table 32.1). The burden in Sub-Saharan Africa is higher The adverse effects of aspirin (cerebral hemorrhage and gas- than in the Middle East and North Africa, which may suggest trointestinal complications) appear to be dose related, and an etiology for stroke other than atherosclerotic disease. most agree that using a low dose of aspirin is prudent Health experts anticipate that the number of stroke cases (Antithrombotic Trialists' Collaboration 2002). Since will increase, particularly in developing countries, because of aspirin can aggravate a hemorrhagic stroke, simple guide- aging populations and increased exposure to major risk factors. lines for the use of platelet antiaggregants should be devel- Corresponding to this increase in the number of stroke cases oped and could be based on scales such as the Siriraj score will be an increase in the number of people with disabilities to rule out hemorrhage (Poungvarin, Viriyavejakul, and surviving after stroke. Komontri 1991). · Thrombolytic therapy. Tissue plasminogen activator and recombinant tissue plasminogen activator (rt-PA) can be Interventions used to halt a stroke by dissolving the blood clot that is Several intervention strategies are available for stroke, but only blocking blood flow to the brain (National Institute of a few can be applied in developing countries. Neurological Disorders and Stroke rt-PA Stroke Study Group 1995). Thrombolytic therapy can increase bleeding Population-Based Interventions. Public health policies to and must be used only after careful patient screening, with a address risk factors for stroke include tobacco and alcohol con- CT scan of the brain within three hours of stroke symptom trol, laws to provide labels showing the fat content of foods, onset, to exclude an intracranial bleed. It also requires and public education about the harm caused by high-fat foods. appropriately trained physicians to administer the medica- Public health programs to control rheumatic fever will reduce tion. These prerequisites for the administration of throm- rheumatic heart disease and the subsequent risk of embolic bolytic agents restrict its use to selected centers in develop- strokes. Better training of birth attendants will reduce the risk ing countries. of peripartum hemorrhage, which leads to puerperal strokes. Strategies for prevention of recurrence of stroke apply Personal Interventions. Modification of adverse lifestyle and equally to individuals who have experienced a transient major risk factors such as hypertension, diabetes, high lipid ischemic attack and to those who have experienced a complete levels, smoking, and alcohol abuse is beneficial both for pri- stroke. These strategies include the following: mary prevention and recurrence of stroke. Some evidence indi- cates that the decline in the incidence of stroke observed in · Platelet antiaggregants. Aspirin therapy is effective in pre- many countries is due to better management of hypertension venting recurrence of stroke, with low daily doses being at (MacWalter and Shirley 2002). Special consideration should be least as effective as higher daily doses (Antithrombotic given to the profile of risk factors in developing countries, Trialists' Collaboration 2002). When compared with aspirin, 634 | Disease Control Priorities in Developing Countries | Vijay Chandra, Rajesh Pandav, Ramanan Laxminarayan, and others clopidogrel has a slight benefit among those who have had · occupational therapy, which helps patients relearn everyday a previous stroke, myocardial infarction, or symptomatic activities, such as eating and drinking peripheral arterial disease. Clopidogrel is an effective and · speech therapy, which helps patients relearn language and safe alternative for patients who do not tolerate aspirin. speaking skills Although clopidogrel may be slightly more effective than · counseling, which can help alleviate some of the mental and aspirin, it is also more expensive. Antiplatelet combination emotional problems that result from stroke. therapy using agents with different mechanisms of action, such as the combination of extended release dipyridamole Comprehensive rehabilitation in a multidisciplinary stroke and aspirin, has been shown to reduce the risk of stroke over unit reduces deaths, disability, and the need for long-term insti- aspirin alone (Sacco, Sivenius, and Diener 2005). In con- tutional care (Smaha 2004), but such facilities are extremely trast, combination therapy with aspirin and clopidogrel limited in developing countries. Home-based rehabilitation offers no advantage over aspirin alone and also increases the services can prevent long-term deterioration in activities of risk of hemorrhage (Diener and others 2004). daily living, although the absolute impact is relatively modest · Anticoagulant therapy. Anticoagulation with warfarin should (Outpatient Service Trialists 2002). However, in developing be considered in stroke patients with atrial fibrillation, countries, the vast majority of patients will be treated either at because of its clear efficacy in preventing embolic strokes, home by a general physician or in a small community hospital provided that patients are appropriately monitored where no skilled rehabilitation therapist is available. (European Atrial Fibrillation Trial Study Group 1993; Mohr and others 2001). Anticoagulant therapy also reduces the risk of embolic stroke in patients with rheumatic heart disease. However, anticoagulation can be hazardous in devel- COST-EFFECTIVENESS OF INTERVENTIONS oping countries because of the lack of monitoring facilities. IN DEVELOPING COUNTRIES · Surgical treatment. In patients with symptomatic carotid disease with stenosis of 70 percent and in asymptomatic We determined incremental cost-effectiveness ratios (ICERs) patients with high-grade stenosis, carotid endarterectomy for selected interventions for each condition by calculating has been shown to be more beneficial than medical total DALYs lost by a population because of the condition with care alone (Asymptomatic Carotid Atherosclerosis Study and without treatment and then dividing the difference by 1995; Asymptomatic Carotid Surgery Trial Collaborative the treatment cost. The disability weights used are presented Group 2004; North American Symptomatic Carotid in table 32.2. All analyses in this section followed the volume Endarterectomy Trial Collaborators 1991). However, editors' standardized guidelines for economic analysis, region- inappropriate selection of patients or high intraoperative specific age structures, and underlying mortality rates. We complications could obviate such benefits. Carotid angio- converted nontradable inputs into U.S. dollars at the market plasty has been suggested as an alternative to carotid exchange rate. We assumed that the costs of tradable inputs endarterectomy in management of severe internal carotid were internationally consistent, as were costs associated with artery disease, but its advantages and disadvantages have yet surgical treatments. Table 32.3 presents the costs of drugs and to be clearly established (Naylor, London, and Bell 1997). medical services. No fixed costs were assumed; therefore, our Carotid endarterectomy for stroke prevention is available at results are not linked with the extent of treatment coverage. only a few centers in developing countries, which makes its widespread use impractical. Table 32.2 Disability Weights Used in ICER Analysis The goal of rehabilitation after a stroke is to enable individ- uals who have experienced a stroke to reach the highest feasible AD and other Acute Recurrent level of independence as soon as possible. Successful rehabilita- Weight dementias Epilepsy PD stroke stroke tion depends on the extent of brain damage, skill of the reha- Untreated 0.627 0.15 0.392a 0.278b 0.556 bilitation team, length of time before rehabilitation is started, Treated 0.627c 0 0.316 0.235b n.a.d and support provided by caregivers. Because each stroke Source: Mathers and others 2006. patient has specific rehabilitation needs, customizing the n.a. not applicable. rehabilitation program is important. Rehabilitation therapies a. Treatment for PD is assumed to be effective for a maximum of 10 years. We also assume that a patient reverts to the untreated disability weight after 10 years. include several complementary approaches: b. Disability is assumed to last a maximum of 10 years; then we assume the patient recovers fully. c. The patient is assumed to experience no benefit from treatment. Benefits are in the form of · physical therapy, which helps stroke patients relearn simple reduced caregiver hours. d. Treatment does not change the disability weight following a recurrent stroke; only the likelihood motor activities, such as walking of experiencing a second stroke is reduced. Neurological Disorders | 635 636 | Disease Table 32.3 Input Requirements for Interventions by Condition Control Primary health care worker Priorities visits to patient in home or Visits to primary health care patient visits to see the worker Specialist care in in doctor in outpatient department in outpatient department outpatient department Inpatient care Developing Patients using the Visits per Patients using the Visits per Patients using the Visits per Patients using the Length Annual drug Condition servicea (percent) year service (percent) year service (percent) year service (percent) of stay costs (US$) Countries AD and other dementias Acetylcholinesterase inhibitors 100 4 100 12 100 2 5 7 638 | Antipsychotics 100 12 100 12 25 6 5 7 10 Vijay Epilepsy Chandra, Phenobarbital 100 2 100 6 10 2 1 3 1 Lamotrigineb 100 2 100 6 10 2 1 3 144 Rajesh Surgery n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 2,600c Pandav PD Levodopa/carbidopa 100 3 100 6 100 2 1 5 71 ,Ramanan Ayurvedic preparations 100 3 100 6 100 2 1 5 19 Deep brain stimulation n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 37,000 Laxminarayan, Stroke (acute attack) Aspirin n.a. n.a. 100 1 100 1 100 14 3 Heparin n.a. n.a. 100 1 100 1 100 14 691 and rt-PA n.a. n.a. 100 1 100 1 1 7 1,777d others Stroke (prevention of recurrence) Aspirin 100 4 100 6 100 1 n.a. n.a. 3 Dipyridamole and aspirin 100 4 100 6 100 1 n.a. n.a. 64 Carotid endarterectomy n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 6,216 Source: Authors. n.a. not applicable. a. Percentage of patients receiving the specified treatment. b. Nondrug treatment costs for lamotrigine are not included in the cost-effectiveness analyses because they are accounted for in the phenobarbital treatment costs. Lamotrigine is taken in addition to phenobarbital. c. Epilepsy surgery also requires screening at a cost of US$600 per screened patient. Because only half of screened patients are eligible for surgery, the cost amounts to US$1,200 per treated patient. d. This treatment requires testing for eligible patients. The costs of screening ineligible patients include all the same hospital and doctor costs as treatment, as well as 80 percent of the drug cost to account for the diagnostic CT. AD and Other Dementias Franky, and Wieser 2000). We assumed that roughly half of We analyzed the use of acetylcholinesterase inhibitors in the surgery recipients experience no more seizures and that the treatment of AD on the basis of the following assumptions: remaining half continue to take phenobarbital despite under- first, only patients who were older than 60 at the time of onset going surgery. Our evaluation of the surgical option included were considered; second, the treatment has no long-term the costs of diagnostic services and the costs associated with benefits--that is, it does not reduce patient disability and has screening patients who ultimately may not be eligible for sur- no effect on mortality. gery. For patients in LMICs who are refractory to phenobarbi- We computed the benefits of reduced caregiver hours on the tal, the ICER of the add-on drug lamotrigine was US$3,000, basis of reports that the improvement in cognitive function in and the ICER of the surgical option plus phenobarbital was AD patients associated with treatment using acetylcholines- US$3,100. The difference between phenobarbital and the other terase inhibitors was a 1.2 point change in the global assessment two options was significant in all regions. scale for cognitive function, as measured by the Mini Mental Among refractory epilepsy patients eligible for surgery State Examination. A 1 point improvement in the score was and according to postoperative outcome studies conducted associated with a 0.56 hour per day reduction in caregiver in developed countries, surgery may be of comparable cost- hours, or roughly 205 hours per year (Marin and others 2003). effectiveness to treatment with a combination of phenobarbital The cost of using acetylcholinesterase inhibitors per hour of and lamotrigine. Because effectiveness data for developing caregiver time saved averaged US$13 across low- and middle- countries are not available, this calculation is based on cost esti- income countries (LMICs) and was at least US$11 in specific mates from a study in Colombia and estimates of the effective- regions (the regions are the same as those in table 32.1). This ness of surgery from developed countries. If the surgical out- amount is substantially more than the wage rate in these come in developing countries were worse than in developed regions, which would generally not exceed US$1 to US$1.50 countries, the cost-effectiveness of surgery would be lower. per hour, even for hired caregivers specifically trained to care Furthermore, we note a number of limitations to the use of for AD patients. We, therefore, conclude that the use of acetyl- surgery in refractory epilepsy, particularly in developing coun- cholinesterase inhibitors in developing countries is not efficient tries, along with the lack of long-term follow-up data on the from an economic perspective. Calculating the cost per DALY outcome of surgery. We stress that the primary treatment of averted for acetylcholinesterase inhibitors would not be mean- epilepsy is with phenobarbital, and effective treatment of epi- ingful, because we assume no benefit to the patient. Finally, the lepsy lies in more efficient use of this highly cost-effective med- use of acetylcholinesterase inhibitors is uncommon in develop- ication to close the treatment gap. ing countries; therefore, reducing its use is not an important concern. Parkinson's Disease We evaluated three interventions for PD: a combination of Epilepsy l-dopa and carbidopa, traditional medicines such as the We analyzed the cost-effectiveness of phenobarbital in the ayurvedic treatment used in India, and deep brain stimulation. treatment of epilepsy, and the results are shown in table 32.4. We assumed that treatment for all three modalities was effec- We assumed that phenobarbital was provided to all patients. tive for 10 years from the onset of treatment. The ICERs in The cost of using phenobarbital per DALY gained in LMICs LMICs for these three modalities were US$1500, US$750 and was US$89. Table 32.4 shows that the benefits of phenobarbital US$31,000, respectively (table 32.4). On the basis of the cost of are large relative to its cost. medication and evidence from clinical trials of effectiveness We did not look at other AEDs, such as phenytoin or carba- (Parkinson's Disease Group 1995) and from animal studies mazepine, because the costs of those medications are much (Hussain and Manyam 1997), we found that ayurvedic treat- greater than that of phenobarbital, but their effectiveness is ment was the most cost-effective option. The relatively favor- essentially the same (Aldenkamp, De Krom, and Reijs 2003). able ICER for ayurvedic treatment is due to the extremely low Although their use may be justified for specific medical rea- medication cost of this intervention. The relatively high ICER sons, phenobarbital is much more cost-effective. for deep brain stimulation was largely attributable to the We analyzed treatment options for patients who are refrac- extremely high cost of surgery. Table 32.4 shows DALYs gained tory to treatment with phenobarbital. We assumed that such for US$1 million of health expenditure. cases were treated either with a combination of phenobarbital and lamotrigine or with a combination of phenobarbital and surgery. We used the cost for epilepsy surgery of US$2,600, Stroke in accordance with a study from Colombia, and applied it to We evaluated two sets of interventions for stroke: treatment of all regions (Malmgren and others 1996; Tureczek, Fandino- acute stroke and prevention of secondary stroke. We assumed Neurological Disorders | 637 638 Table 32.4 Results from Cost-Effectiveness Analysis of Interventions for Alzheimer's Disease, Epilepsy, Parkinson's Disease, and Stroke, by World Bank Region | Disease Low- and middle- East Asia and Europe and Latin America and Middle East and Sub-Saharan Condition income countries the Pacific Central Asia the Caribbean North Africa South Asia Africa Control AD Priorities Cost per care hour reduced using 11 11 12 13 12 11 11 acetylcholinesterase inhibitors (US$) in Epilepsy Developing Incremental costs of DALYs gained per year of treatment compared with no treatment (US$) Countries Phenobarbital 89 78 122 261 165 54 25 Phenobarbital and lamotrigine 2,994 3,306 2,945 4,301 3,344 2,872 1,490 | Phenobarbital and surgery 3,060 3,411 3,049 3,477 2,904 3,097 1,788 Vijay Number of DALYs gained per US$1 million per year Chandra, Phenobarbital 11,262 12,799 8,185 3,828 6,072 18,581 39,632 Phenobarbital and lamotragine 334 302 340 232 299 348 671 Rajesh Phenobarbital and surgery 327 293 328 288 344 323 559 Pandav PD Incremental costs of DALYs gained per year of ,Ramanan treatment compared with no treatment (US$) Levodopa/carbidopa 1,512 1,398 1,760 2,254 1,944 1,311 1,281 Laxminarayan, Ayurvedic preparation 751 638 1,000 1,494 1,184 551 520 Levodopa/carbidopa and deep brain stimulation 31,114 26,941 29,310 29,444 30,770 31,347 34,069 Number of DALYs gained per US$1 million per year and Levodopa/carbidopa 662 715 568 444 514 763 781 others Ayurvedic preparation 1,331 1,568 1,000 669 845 1,815 1,922 Levodopa/carbidopa and deep brain stimulation 32 37 34 34 32 32 29 Stroke (treatment of acute attack) Incremental costs of DALYs gained per year of treatment compared with no treatment (US$) Aspirin 149 109 104 574 534 118 112 Heparin 2,675 2,185 1,318 4,952 5,443 2,967 2,940 rt-PA 1,278 1,169 648 2,158 2,516 1,630 1,623 Number of DALYs gained per US$1 million per year Aspirin 6,691 9,209 9,633 1,742 1,873 8,463 8,942 Heparin 374 458 759 202 184 337 340 rt-PA 783 856 1,543 463 398 613 616 Stroke (prevention of recurrence) Incremental costs of percent recurrence risk averted after 2 years of treatment (US$) Aspirin 4 3 6 9 7 2 2 Dipyridamole and aspirin 5 5 6 8 7 4 4 Carotid endarterectomy 87 87 87 87 87 87 87 Incremental costs of DALYs gained per 2 years of treatment compared with no treatment (US$) Aspirin 70 60 59 233 196 52 34 Dipyridamole and aspirin 93 95 63 194 186 96 69 Carotid endarterectomy 1,458 1,614 836 2,001 2,234 1,759 1,284 Number of DALYs gained per US$1 million per 2 years of treatment Aspirin 14,313 16,569 16,866 4,285 5,093 19,348 29,373 Dipyridamole and aspirin 10,752 10,555 15,969 5,150 5,384 10,369 14,572 Carotid endarterectomy 686 620 1,197 500 448 568 779 Source: Authors. Neurological Disorders | 639 that stroke sufferers have fully recovered 10 years after their last inhibitors, suggests that they should not be widely used in stroke. developing countries. Instead, giving low doses of antipsychotic We evaluated aspirin, heparin, and rt-PA for the treatment medication to patients with any form of dementia who also of acute stroke. The International Stroke Trial Collaborative have behavioral problems may be a better option for reducing Group (1997) reports that, within 14 days of the onset of caregiver stress, although this possibility has not been system- stroke, mortality with heparin treatment is less than with a atically evaluated. placebo; however, after six months, mortality is actually greater Phenobarbital is by far the most cost-effective intervention for patients treated with heparin than with a placebo--that is, for managing epilepsy and should be recommended for wide- there is a negative cost per DALY gained if this effect is incor- spread use in public health campaigns against epilepsy in porated. The estimates presented here are based on the change LMICs. For those patients who do not respond to phenobarbi- in the short-term mortality risk. For LMICs, the cost per DALY tal, the addition of lamotrigine is advisable rather than surgery, averted using aspirin was US$150 (table 32.4). The equivalent because of the resource-intensive evaluation and infrastructure costs of interventions using rt-PA and heparin were US$1,300 required for epilepsy surgery. and US$2,700, respectively. The costs of heparin are higher than Indigenous systems of medicine, such as the ayurvedic med- the costs of rt-PA, despite the expensive equipment required for icines used in India, are much more cost-effective than Western rt-PA, because of the lower effectiveness of heparin. medications or surgical procedures for managing patients with Table 32.4 presents DALYs averted for US$1 million of PD. Other countries may wish to test and standardize such health expenditure for the three treatments. The cost per DALY medications for their own use. gained using aspirin is a conservative estimate, because the use Aspirin is by far the most cost-effective intervention both for of aspirin has additional benefits in terms of preventing a treating acute stroke and for preventing a recurrence of stroke. recurrence of stroke. It is easily available in developing countries, even in rural areas. Table 32.4 shows the costs of preventing a second stroke within two years of the first stroke. For LMICs, aspirin was the least expensive option at US$3.80 per single percentage point RESEARCH AND DEVELOPMENT AGENDA decrease in the risk of a second stroke within two years of the first. This rate translates to roughly US$70 per DALY gained The populations of most developing countries are aging (table 32.4). Combining dipyridamole with aspirin, because of rapidly. Many neurological disorders frequently occur in the higher cost, was slightly more expensive at roughly US$5.20 per elderly, posing an emerging public health problem. As a result, single percentage point decrease in recurrent stroke risk for a developing countries should begin or expand their research single individual, or about US$93 per DALY. In contrast, and development agendas to address issues related to the pre- carotid endarterectomy was US$87 for an equivalent decrease vention, identification, and management of neurological in individual recurrence risk or almost US$1,500 per DALY. disorders. In the short term, they should focus on early identi- The aspirin monotherapy option for preventing a recurrence of fication, optimum treatment, and amelioration of distress and stroke was the most cost-effective approach only in South Asia handicaps and on reduction of the social and economic burden and Sub-Saharan Africa, largely because of the relatively low on patients and their families. In the long term, they should costs of nontradable inputs, such as hospital and doctors' fees, develop and implement strategies for primary prevention of in those regions. Low input costs of nontradables increase the neurological disorders. Specific areas for research and develop- relative importance of drug costs in determining the most cost- ment include the following: effective intervention; therefore, the cheaper drug, aspirin, was most cost-effective. Table 32.4 shows that, though US$1 million · Conducting population-based epidemiological studies in would be most effectively spent on aspirin alone in South Asia developing countries. Population-based data from develop- and Sub-Saharan Africa, investment in aspirin and dipyri- ing countries are insufficient, which limits evidence-based damole treatment would result in a greater DALY gain in the planning. In addition, such data may also suggest important other regions. hypotheses for research if they identify genuine differences across regions (for example, the reported difference in the incidence of AD in developed and developing countries). In RECOMMENDATIONS addition, the identification of risk or protective factors would be useful in the primary prevention of such diseases. The use of acetylcholinesterase inhibitors for treating patients · Enhancing existing health care delivery systems. In most with AD, as assessed by the number of caregiver hours saved, developing countries, approximately 70 to 80 percent of suggests that this intervention is not cost-effective. This find- patients live in rural areas, where medical care is frequently ing, combined with the limited efficacy of acetylcholinesterase provided by nonphysician health care providers or, at best, 640 | Disease Control Priorities in Developing Countries | Vijay Chandra, Rajesh Pandav, Ramanan Laxminarayan, and others by a general physician. Limitations in the availability of countries are directed at treatment of stroke. This approach health care have resulted in a huge treatment gap for many not only is more expensive but also is less beneficial to the neurological disorders. For such situations, a simple model patient. for the management of neurological disorders by existing community-based health care providers, trained to provide such services, would be helpful. Research is needed on ACKNOWLEDGMENTS optimum referral systems for more difficult cases that local communities will accept and can afford. Strategies for home- The lead author would like to acknowledge with gratitude the based care of patients need to be systematically evaluated. support provided by the regional director and director of pro- · Developing cheaper and more efficacious medicines. Many gram management of the South-East Asia Regional Office of currently available medications have significant side effects the World Health Organization. Special mention must be made and are too expensive for many patients in developing coun- of Dr. Daniel Chisholm, who provided input into the cost- tries. Newer medications need to be developed with lower effectiveness analysis, particularly that dealing with epilepsy, costs, fewer side effects, better efficacy, and less frequent and of Dr. Donald Silberberg, who served as the senior adviser dose schedules. to the chapter. The authors wish to thank the many reviewers · Promoting the use of indigenous systems of medicine. Many for their valuable suggestions, which have been incorporated people in developing countries use local indigenous medi- into the chapter. cines. More research needs to be done on the pharmacolog- ical properties of those medications (see chapter 69). · Launching stigma removal campaigns. The stigmatization of REFERENCES patients with neurological disorders and of their families is still prevalent, particularly in rural and remote areas, and it Aldenkamp, A. P., M. De Krom, and R. Reijs. 2003. "Newer Antiepileptic often prevents patients from seeking and obtaining appro- Drugs and Cognitive Issues." Epilepsia 44 (Suppl. 4): 21­29. priate medical care. Effective strategies to address this issue Allen, C. M. C. 1983. "Clinical Diagnosis of Acute Stroke Syndrome." Quarterly Journal of Medicine 42: 515­23. need to be developed and implemented on a large scale. Anand, K., D. Chowdhury, K. B. Singh, C. S. Pandav, and S. K. Kapoor. 2001."Estimation of Mortality and Morbidity Due to Strokes in India." Neuroepidemiology 20 (3): 208­11. MISSED OPPORTUNITIES Antithrombotic Trialists' Collaboration. 2002. "Collaborative Meta- Analysis of Randomised Trials of Antiplatelet Therapy for Prevention of Death, Myocardial Infarction, and Stroke in High Risk Patients." Many research studies have reported that the incidence of AD British Medical Journal 324 (7329): 71­86. is lower in developing countries than in Western countries. Asawavichienjinda, T., C. Sitthi-Amorn, and W. Tanyanont. 2003. Migration studies, such as those looking at the migration of "Compliance with Treatment of Adult Epileptics in a Rural District of Africans to the United States, have shown a change in the risk Thailand." Journal of Medical Association Thailand 86 (1): 46­51. for AD within one or two generations. This finding suggests Ascherio, A., S. M. Zhang, M. A. Hernan, I. Kawachi, G. A. Colditz, F. E. that developing countries may have some protective factors that Speizer, and others. 2001."Prospective Study of Caffeine Consumption and Risk of Parkinson's Disease in Men and Women." Annals of rapidly change on migration to developed countries. Despite Neurology 50 (1): 56­63. this information being available for more than 25 years, no Asymptomatic Carotid Atherosclerosis Study. 1995. "Endarterectomy for systematic efforts have been made to identify these protective Asymptomatic Carotid Artery Stenosis." Journal of the American factors. Given the rapid adaptation of Western lifestyles in Medical Association 273 (18): 1421­28. developing countries, identifying these factors is important Asymptomatic Carotid Surgery Trial Collaborative Group. 2004. "Prevention of Disabling and Fatal Strokes by Successful Carotid before the opportunity is permanently lost. Endarterectomy in Patients without Recent Neurological Symptoms: The successful use of phenobarbital for treating epilepsy was Randomised Controlled Trial." Lancet 363: 1491­502. first described in 1912. Not only is it effective for many types Berger, K., M. M. Breteler, C. Helmer, D. Inzitari, L. Fratiglioni, of epilepsy, but it is also inexpensive. Nevertheless, despite its C. Trenkwalder, and others. 2000. "Prognosis with Parkinson's Disease in Europe: A Collaborative Study of Population-Based Cohorts: availability for more than 90 years and its modest cost, the Neurologic Diseases in the Elderly Research Group." Neurology 54 treatment gap for epilepsy still exceeds 90 percent in many (11 Suppl. 5): S24­27. developing countries. Brodaty, H., and M. Gresham. 1989. "Effect of a Training Programme to Indigenous systems of medicine, such as for the treatment Reduce Stress in Carers of Patients with Dementia." British Medical of PD, have been used for centuries in developing countries. Journal 299 (6712): 1375­79. However, their utility has not been fully exploited. Buck, D., B. A. Gregson, C. H. Bamford, P. McNamee, G. N. Farrow, J. Bond, and others. 1997. "Psychological Distress among Informal Despite evidence of the benefit of control of hypertension Supporters of Frail Older People at Home and in Institutions." in the primary prevention of stroke, most efforts in developing International Journal of Geriatric Psychiatry 12 (7): 737­44. Neurological Disorders | 641 Casetta, I., V. C. Monetti, S. Malagu, E. Paolino, V. Govoni, E. Fainardi, and He, J., M. J. Klag, Z. Wu, and P. K. Whelton. 1995. "Stroke in the People's others. 2002. "Risk Factors for Cryptogenic and Idiopathic Partial Republic of China: II. Meta-Analysis of Hypertension and Risk of Epilepsy: A Community-Based Case-Control Study in Copparo, Italy." Stroke." Stroke 26 (12): 2228­32. Neuroepidemiology 21 (5): 251­54. Henderson, A. S., and A. F. Jorm. 2000. "Definition of Epidemiology of Chandra, V., M. Ganguli, R. Pandav, J. Johnston, S. Belle, and S. T. Dementia: A Review." In Dementia, ed. M. Mario and N. Sartorius, DeKosky. 1998. "Prevalence of Alzheimer's Disease and Other 1­34. West Sussex, U.K.: John Wiley. Dementias in Rural India: The Indo-U.S. Study." Neurology 51 (4): Henderson, V. W. 1997. "The Epidemiology of Estrogen Replacement 1000­8. Therapy and Alzheimer's Disease." Neurology 48 (5 Suppl. 7): S27­35. Chandra, V., R. Pandav, H. H. Dodge, J. M. Johnston, S. H. Belle, S. T. Hendrie, H. C., A. Ogunniyi, K. S. Hall, O. Baiyewu, F. W. Unverzagt, DeKosky, and others. 2001. "Incidence of Alzheimer's Disease in a O. Gureje, and others. 2001. "Incidence of Dementia and Alzheimer Rural Community in India: The Indo-U.S. Study." Neurology 57 (6): Disease in 2 Communities: Yoruba Residing in Ibadan, Nigeria, and 985­89. African Americans Residing in Indianapolis, Indiana." Journal of the Chinese Acute Stroke Trial Collaborative Group. 1997. "CAST: American Medical Association 285 (6): 739­47. Randomised Placebo-Controlled Trial of Early Aspirin Use in 20,000 Hendrie, H. C., B. O. Osuntokun, K. S. Hall, A. O. Ogunniyi, S. L. Hui, F. W. Patients with Acute Ischaemic Stroke: CAST (Chinese Acute Stroke Unverzagt, and others. 1995. "Prevalence of Alzheimer's Disease and Trial) Collaborative Group." Lancet 349 (9066): 1641­49. Dementia in Two Communities: Nigerian Africans and African Cockerell, O. C., A. L. Johnson, J. W. Sander, and S. D. Shorvon. 1997. Americans." American Journal of Psychiatry 152 (10): 1485­92. "Prognosis of Epilepsy: A Review and Further Analysis of the First Huang, Z. S., T. L. Chiang, and T. K. Lee. 1997. "Stroke Prevalence in Nine Years of the British National General Practice Study of Epilepsy: Taiwan: Findings from the 1994 National Health Interview Survey." A Prospective Population-Based Study." Epilepsia 38 (1): 31­46. Stroke 28 (8): 1579­84. Desai, P., M. V. Padma, S. Jain, and M. C. Maheshwari. 1998. "Knowledge, Hussain, G., and B. V. Manyam. 1997. "Mucuna Pruriens Proves More Attitudes, and Practice of Epilepsy: Experience at a Comprehensive Effective than L-DOPA in Parkinson's Disease Animal Model." Rural Health Services Project." Seizure 7 (2): 133­38. Phytotherapy Research 11: 419­23. Diener, H. C., J. Bogousslavsky, L. M. Brass, C. Cimminiello, L. Csiba, International Stroke Trial Collaborative Group. 1997. "The International M. Kaste, and others. 2004. "Aspirin and Clopidogrel Compared with Stroke Trial (IST): A Randomised Trial of Aspirin, Subcutaneous Clopidogrel Alone after Recent Ischaemic Stroke or Transient Heparin, Both, or Neither among 19,435 Patients with Acute Ischaemic Ischaemic Attack in High-Risk Patients (MATCH): Randomised, Stroke: International Stroke Trial Collaborative Group." Lancet Double-Blind, Placebo-Controlled Trial." Lancet 364 (9431): 301­7. 349 (9065): 1569­81. Eastern Stroke and Coronary Heart Disease Collaborative Research Jorm, A. F., and D. Jolley. 1998. "The Incidence of Dementia: A Meta- Group. 1998. "Blood Pressure, Cholesterol, and Stroke in Eastern Asia: Analysis." Neurology 51 (3): 728­33. Eastern Stroke and Coronary Heart Disease Collaborative Research Kaiser, C., G. Asaba, C. Mugisa, W. Kipp, S. Kasoro, T. Rubaale, and others. Group." Lancet 352 (9143): 1801­7. 1998. "Antiepileptic Drug Treatment in Rural Africa: Involving the Engel, J. Jr., S. Wiebe, J. French , M. Sperling, P. Williamson, D. Spencer, Community." Tropical Doctor 28 (2): 73­77. and others. 2003. "Practice Parameter: Temporal Lobe and Localized Karaagac, N., S. N. Yeni, M. Senocak, M. Bozluolcay, F. K. Savrun, H. Neocortical Resections for Epilepsy." Epilepsia 44 (6): 741­51. Ozdemir, and others. 1999. "Prevalence of Epilepsy in Silivri, a Rural European Atrial Fibrillation Trial Study Group. 1993. "Secondary Area of Turkey." Epilepsia 40 (5): 637­42. Prevention in Non-Rheumatic Atrial Fibrillation after Transient Kiyohara, Y., M. Kubo, I. Kato, Y. Tanizaki, K. Tanaka, K. Okubo, and oth- Ischaemic Attack or Minor Stroke." Lancet 342 (8882): 1255­62. ers. 2003. "Ten-Year Prognosis of Stroke and Risk Factors for Death in Foster, N. L., R. C. Petersen, S. I. Gracon, and K. Lewis. 1996."An Enriched- a Japanese Community: The Hisayama Study." Stroke 34 (10): 2343­47. Population, Double-Blind, Placebo-Controlled, Crossover Study of Kotsopoulos, I. A., T. van Merode, F. G. Kessels, M. C. de Krom , and J. A. Tacrine and Lecithin in Alzheimer's Disease: The Tacrine 970-6 Study Knottnerus. 2002. "Systematic Review and Meta-Analysis of Incidence Group." Dementia 7 (5): 260­66. Studies of Epilepsy and Unprovoked Seizures." Epilepsia 43 (11): Fukiyama, K., Y. Kimura, K. Wakugami, and H. Muratani. 2000."Incidence 1402­9. and Long-Term Prognosis of Initial Stroke and Acute Myocardial Leonardi, M., and T. B. Ustun. 2002. "The Global Burden of Epilepsy." Infarction in Okinawa, Japan." Hypertension Research 23 (2): 127­35. Epilepsia 43 (Suppl. 6): 21­25. Gentleman, S. M., D. I. Graham, and G. W. Roberts. 1993. "Molecular Leone, M., E. Bottacchi, E. Beghi, E. Morgando, R. Mutani, R. Cremo, and Pathology of Head Trauma: Altered Beta APP Metabolism and the others. 2002."Risk Factors for a First Generalized Tonic-Clonic Seizure Aetiology of Alzheimer's Disease." Progress in Brain Research 96: 237­46. in Adult Life." Neurological Sciences 23 (3): 99­106. Goldstein, L. B., R. Adams, K. Becker, C. D. Furberg, P. B. Gorelick, G. Levin, E., J. Moriarty, and P. Gorbach. 1994."Better for the Break." London: Hademenos, and others. 2001. "Primary Prevention of Ischemic Her Majesty's Stationery Office, National Institute of Social Work Stroke: A Statement for Healthcare Professionals from the Stroke Research Unit. Council of the American Heart Association." Stroke 32 (1): 280­99. Lewin, L., and P. Schuster. 1929. "Ergebnisse von Banisterinversuchen an Gopinath, B., K. Radhakrishnan, P. S. Sarma, D. Jayachandran, and A. Kranken." Deutsche Medizinische Wochenschrift 55: 419. Alexander. 2000. "A Questionnaire Survey about Doctor-Patient Communication, Compliance, and Locus of Control among South Li, G., Y. C. Shen, C. H. Chen, Y. W. Zhau, S. R. Li, and M. Lu. 1991. "A Indian People with Epilepsy." Epilepsy Research 39 (1): 73­82. Three-Year Follow-up Study of Age-Related Dementia in an Urban Gunatilake, S. B., B. A. Jayasekera, and A. P. Premawardene. 2001. "Stroke Area of Beijing." Acta Psychiatrica Scandanavica 83 (2): 99­104. Subtypes in Sri Lanka: A Hospital-Based Study." Ceylon Medical Liu, L., K. Ikeda, and Y. Yamori. 2001. "Changes in Stroke Mortality Rates Journal 46 (1): 19­20. for 1950 to 1997: A Great Slowdown of Decline Trend in Japan." Stroke Haupt, M., A. Karger, and M. Janner. 2000. "Improvement of Agitation 32 (8): 1745­49. and Anxiety in Demented Patients after Psychoeducative Group MacWalter, R. S., and C. P. Shirley. 2002. "A Benefit-Risk Assessment of Intervention with Their Caregivers." International Journal of Geriatric Agents Used in the Secondary Prevention of Stroke." Drug Safety Psychiatry 15 (12): 1125­29. 25 (13): 943­63. 642 | Disease Control Priorities in Developing Countries | Vijay Chandra, Rajesh Pandav, Ramanan Laxminarayan, and others Malmgren, K., A. Hedstrom, R. Granqvist, H. Malmgren, and E. Ben- Intracerebral Haemorrhage from Infarction." British Medical Journal Menachem. 1996. "Cost Analysis of Epilepsy Surgery and of Vigabatrin 302: 1565­67. Treatment in Patients with Refractory Partial Epilepsy." Epilepsy Prince, M. 2000. "Dementia in Developing Countries: A Consensus Research 25 (3): 199­207. Statement from the 10/66 Dementia Research Group." International Mani, K., G. Rangan, H. V. Srinivas, and S. Narendran. 1993. "Natural Journal of Geriatric Psychiatry 15 (1): 14­20. History of Untreated Epilepsy: A Community-Based Study in Rural Rajkumar, S., S. Kumar, and R. Thara. 1997. "Prevalence of Dementia in a South India." Epilepsia 34 (Suppl. 2): 166. Rural Setting: A Report from India." International Journal of Geriatric Marin, D., K. Amaya, R. Casciano, K. L. Puder, J. Casciano, S. Chang, and Psychiatry 12 (7): 702­27. others. 2003. "Impact of Rivastigmine on Costs and on Time Spent Sacco, R., J. Sivenius, and H. C. Diener. 2005. "Efficacy of Aspirin Plus in Caregiving for Families of Patients with Alzheimer's Disease." Extended-Release Dipyridamole in Preventing Recurrent Stroke in International Psychogeriatics 15 (4): 385­98. High-Risk Populations." Archives of Neurology 62: 403­8. Marras, C., and C. Tanner. 2002. "The Epidemiology of Parkinson's Sarti, C., D. Rastenyte, Z. Cepaitis, and J. Tuomilehto. 2000. "International Disease." In Movement Disorders Neurologic Principles and Practice, Trends in Mortality from Stroke, 1968 to 1994." Stroke 31 (7): ed. R. L. Watts and W. C. Koller, 177­96. New York: McGraw-Hill. 1588­601. Marriott, A., C. Donaldson, N. Tarrier, and A. Burns. 2000. "Effective- Schapel, G. J., R .G. Beran, F. J. Vajda, S. F. Berkovic, M. L. Mashford, F. M. ness of Cognitive-Behavioural Family Intervention in Reducing the Dunagan, and others. 1993. "Double-Blind, Placebo Controlled, Burden of Care in Carers of Patients with Alzheimer's Disease." British Crossover study of Lamotrigine in Treatment Resistant Partial Journal of Psychiatry 176 (1): 557­62. Seizures." Journal of Neurology Neurosurgery Psychiatry 56 (5): 448­53. Mathers, C. D., A. D. Lopez, and C. J. L. Murray. "The Burden of Disease Shumaker, S. A., C. Legault, S. R. Rapp, L. Thal, R. B. Wallace, J. K. Ockene, and Mortality by Condition: Data, Methods, and Results for 2001." In and others. 2003. "Estrogen Plus Progestin and the Incidence of Global Burden of Disease and Risk Factors, eds. A. D. Lopez, C. D. Dementia and Mild Cognitive Impairment in Postmenopausal Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. New York: Women: The Women's Health Initiative Memory Study--A Oxford University Press. Randomized Controlled Trial." Journal of the American Medical Meinardi, H., R. A. Scott, R. Reis, J. W. Sander, and ILAE Commission on Association 289 (20): 2651­62. the Developing World. 2001. "The Treatment Gap in Epilepsy: The Smaha, L. A. 2004. "The American Heart Association Get with the Current Situation and Ways Forward." Epilepsia 42 (1): 136­49. Guidelines Program." American Heart Journal 148 (Suppl. 5): S46­48. Melzer, D., K. Pearce, B. Cooper, and C. Brayne. 2004. "Alzheimer's Disease Tanner, C., and S. Goldman. 1996. "Epidemiology of Parkinson's Disease." and Other Dementias." Department of Public Health and Epide- Neurology Clinics 14: 317­35. miology, University of Birmingham, U.K. http://hcna.radcliffe- oxford.com/dementiaframe.htm. Thorvaldsen, P., K. Asplund, K. Kuulasmaa, A. M. Rajakangas, and M. Schroll. 1995. "Stroke Incidence, Case Fatality, and Mortality in the Mohr, J. P., J. L. P. Thompson, R. M. Lazar, B. Levin, R. L. Sacco, K. L. Furie, WHO MONICA Project: World Health Organization Monitoring and others. 2001. "A Comparison of Warfarin and Aspirin for the Trends and Determinants in Cardiovascular Disease." Stroke 26 (3): Prevention of Recurrent Ischemic Stroke." New England Journal of 361­67. Medicine 345 (20): 1444­51. Tureczek, I. E., J. Fandino-Franky, and H. G. Wieser. 2000. "Comparison of National Institute of Neurological Disorders and Stroke rt-PA Stroke Study the Epilepsy Surgery Programs in Cartagena, Colombia, and Zurich, Group. 1995. "Tissue Plasminogen Activator for Acute Ischemic Stroke: Switzerland." Epilepsia. 41 (Suppl. 4): S35­40. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group." New England Journal of Medicine 333 (24): 1581­87. Van Den Eeden, S. K., C. M. Tanner, A. L. Bernstein, R. D. Fross, A. Leimpeter, D. A. Bloch, and others. 2003. "Incidence of Parkinson's National Stroke Association. 2002. "Recovery and Rehabilitation." Disease: Variation by Age, Gender, and Race/Ethnicity." American National Stroke Association, Englewood, CO. http://www.stroke.org/ Journal of Epidemiology 157 (11): 1015­22. HomePage.aspx?P=435435784753465. Walker, R. W., D. G. McLarty, H. M. Kitange, D. Whiting, G. Masuki, D. M. Naylor, A. R., N. J. M. London, and P. R. Bell. 1997. "Carotid Endarterec- Mtasiwa, and others. 2000. "Stroke Mortality in Urban and Rural tomy versus Carotid Angioplasty." Lancet 349 (9046): 203­24. Tanzania: Adult Morbidity and Mortality Project." Lancet 355 (9216): North American Symptomatic Carotid Endarterectomy Trial Collab- 1684­87. orators. 1991. "Beneficial Effect of Carotid Endarterectomy in Walsh, J. S., H. G. Welch, and E. B. Larson. 1990. "Survival of Outpatients Symptomatic Patients with High-Grade Carotid Stenosis: North with Alzheimer-Type Dementia." Annals of Internal Medicine 113 (6): American Symptomatic Carotid Endarterectomy Trial Collaborators." 429­34. New England Journal of Medicine 325 (7): 445­53. Wang, W. Z., J. Z. Wu, D. S. Wang, X. Y. Dai, B. Yang, T. P. Wang, and others. Outpatient Service Trialists. 2002. "Therapy-Based Rehabilitation Services 2003. "The Prevalence and Treatment Gap in Epilepsy in China: for Stroke Patients at Home." Cochrane Database of Systematic An ILAE/IBE/WHO Study." Neurology 60 (9): 1544­45. Reviews (2) CD002925. Wang, W., D. Zhao, and G. Wu. 2001. "The Trend of Incidence Rate of Pal, D. K., A. Carpio, and J. W. Sander. 2000. "Neurocysticercosis and Acute Stroke Event in Urban Areas, Beijing from 1984 to 1999" (in Epilepsy in Developing Countries." Journal of Neurology Neurosurgery Chinese). Zhonghua Liu Xing Bing Xue Za Zhi 22 (4): 269­72. Psychiatry 68 (2): 137­43. WHO (World Health Organization). 2000. The Global Campaign against Pal, D. K., T. Das, S. Sengupta, and G. Chaudhury. 2002. Help-Seeking Epilepsy (information pack). Geneva: WHO. Patterns for Children with Epilepsy in Rural India: Implications for Service Delivery. Epilepsia 43 (8): 904­11. Zhang, L. F., J. Yang, Z. Hong, G. G. Yuan, B. F. Zhou, L. C. Zhao, and Parkinson's Disease Group. 1995. "An Alternative Medicine Treatment for others. 2003. "Proportion of Different Subtypes of Strokes in China." Parkinson's Disease: Results of a Multicenter Clinical Trial: HP-200 Stroke 34 (9): 2091­96. in Parkinson's Disease Study Group." Journal of Alternative Zhang, Z. X., and G. C. Roman. 1993. "Worldwide Occurrence of Complementary Medicine 1 (3): 249­55. Parkinson's Disease: An Updated Review." Neuroepidemiology 12 (4): Poungvarin, N., A. Viriyavejakul, and C. Komontri. 1991. "Siriraj 195­208. Stroke Score and Validation Study to Distinguish Supratentorial Neurological Disorders | 643 Chapter 33 Cardiovascular Disease Thomas A. Gaziano, K. Srinath Reddy, Fred Paccaud, Sue Horton, and Vivek Chaturvedi Cardiovascular disease (CVD) is the number one cause of Predominant Cardiovascular Diseases death worldwide (Mathers and others 2006; Murray and Lopez This chapter focuses on the most common causes of CVD 1996; WHO 2002b). CVD covers a wide array of disorders, morbidity and mortality: including diseases of the cardiac muscle and of the vascular sys- tem supplying the heart, brain, and other vital organs. This · ischemic heart disease (IHD) chapter reviews the epidemiological transition that has made · stroke CVD the world's leading cause of death, assesses the status of · congestive heart failure (CHF). the transition by region, and indicates regional differences in the burden of CVD. It also reviews the cost-effectiveness of var- These diseases account for at least 80 percent of the burden ious interventions directed at the most relevant causes of CVD of CVD in all income regions, which share many of the same morbidity and mortality. common risk factors; accordingly, similar interventions are appropriate. A fourth manifestation, rheumatic heart disease (RHD), which accounts for 3 percent of all disability-adjusted EPIDEMIOLOGY OF CVD life years (DALYs) lost as a result of CVD, does not contribute significantly to the overall global burden of CVD. The burden At the beginning of the 20th century, CVD was responsible for of RHD will likely continue to diminish, but it is still an impor- less than 10 percent of all deaths worldwide, but by 2001 that tant inflammatory cause of heart disease in developing coun- figure was 30 percent. About 80 percent of the global burden of tries and accordingly is addressed in this chapter. We do not CVD death occurs in low- and middle-income countries. address many other forms of CVD because of the scope of this Murray and Lopez (1996) predicted that CVD will be the lead- volume; the regional rather than global nature of some inflam- ing cause of death and disability worldwide by 2020 mainly matory diseases, such as Chagas disease; or the congenital because it will increase in low- and middle-income countries. abnormalities or genetically based cardiomyopathies for which By 2001, CVD had become the leading cause of death in the prevention and treatment options remain limited. developing world, as it has been in the developed world since the mid 1900s (Mathers and others 2006; WHO 2002a). Nearly Ischemic Heart Disease. IHD is the single largest cause of death 50 percent of all deaths in high-income countries and about 28 in the developed countries and is one of the main contributors percent of deaths in low- and middle-income countries are the to the disease burden in developing countries. The two leading result of CVD (Mathers and others 2006). Other causes of manifestations of IHD are angina and acute myocardial infarc- death, such as injuries, respiratory infections, nutritional defi- tion. In 2001, IHD was responsible for 7.3 million deaths and 58 ciencies, and HIV/AIDS, collectively still play a predominant million DALYs lost worldwide (WHO 2002b). Seventy-five per- role in certain regions, but even in those areas CVD is now a cent of global deaths and 82 percent of the total DALYs resulting significant cause of mortality. from IHD occurred in the low- and middle-income countries. 645 Glossary ACE inhibitors (angiotensin-converting enzyme Dyslipidemia: a condition marked by abnormal concen- inhibitors): a group of antihypertensive drugs that exert trations of lipids or lipoproteins in the blood. their influence through the renin-angiotensin-aldosterone Embolus: a blood clot that moves through the blood- system. stream until it lodges in a narrowed vessel and blocks Antiplatelets: drugs that interfere with the blood's ability circulation. to clot. Endocarditis: inflammation of the lining of the heart and Atheroschlerosis: a chronic disease characterized by its valves. thickening and hardening of the arterial walls. Hypertension: abnormally high arterial blood pressure. Atrial fibrillation: an abnormal rhythm of the heart that Reperfusion: restoration of the flow of blood to a previ- can result in an increased risk of stroke because of the for- ously ischemic tissue or organ. mation of emboli (blood clots) in the heart. Statins: a group of drugs that inhibit the synthesis of cho- Beta-blockers: a group of drugs that decrease the heart lesterol and promote the production of low-density rate and force of contractions and lower blood pressure. lipoprotein (LDL)­binding receptors in the liver, resulting Cardiogenic shock: poor tissue perfusion resulting in a decrease in the level of LDL and a smaller increase in from failure of the heart to pump an adequate amount of the level of high-density lipoprotein (HDL). blood. Thrombolysis: the breaking up of a blood clot. Cardiomyopathy: a disorder of the muscle limiting the heart's function. Thrombus: a blood clot that forms inside a blood vessel or cavity of the heart. Chagas disease: a tropical American disease caused by a parasitic infection. Chronic symptoms include cardiac Transient ischemic attack: transient reduced blood flow problems, such as an enlarged heart, altered heart rate or to the brain that produces strokelike symptoms but no rhythm, heart failure, or cardiac arrest. lasting damage. Angina is the characteristic pain of IHD. It is caused by Stroke. Stroke is caused by a disruption in the flow of blood to atherosclerosis leading to stenosis (partial occlusion) of one or part of the brain either because of the occlusion of a blood more coronary arteries. Patients with chronic stable angina vessel (ischemic stroke) or the rupture of a blood vessel (hem- have an average annual mortality of 2 percent or less. Acute orrhagic stroke). Many of the same risk factors for IHD apply myocardial infarction (AMI) is the total occlusion of a major to stroke; in addition, atrial fibrillation is an important risk fac- coronary artery with a complete lack of oxygen and nutrients tor for stroke. The annual risk of stroke in patients with non- leading to cardiac muscle necrosis. AMI is usually diagnosed valvular atrial fibrillation is 3 to 5 percent, with 50 percent of by changes in the electrocardiogram; by elevated serum thromboembolic stroke being attributable to atrial fibrillation enzymes, such as creatine phosphokinase and troponin T or I; (Wolf, Abbott, and Kannel 1991). Chapter 32 discusses the and by pain similar to that of angina. Thirty-day mortality diagnosis and management of the clinical syndromes in greater after an AMI is high: even with best medical therapy it detail. remains at about 33 percent, with half the deaths occurring before the individual reaches the hospital. Even in a hospital Congestive Heart Failure. CHF is the end stage of many heart with a coronary care unit where advanced care options are diseases. It is characterized by abnormalities in myocardial func- available, mortality is still 7 percent. In a hospital without tion and neurohormonal regulation resulting in fatigue, fluid such facilities or therapies, the mortality rate is closer to 30 retention, and reduced longevity. CHF is caused by pathological percent. Even though mortality among patients who have processes that affect the heart; IHD and hypertension-related recovered from an AMI has declined in recent decades, heart disease are the most common etiologies. The risk of approximately 4 percent of patients who survive initial hospi- developing CHF is two times more in hypertensive men and three talization die in the first year following the event (Antman times more in hypertensive women compared with those who are and others 2004). normotensive. CHF is five times more common in those who 646 | Disease Control Priorities in Developing Countries | Thomas A. Gaziano, K. Srinath Reddy, Fred Paccaud, and others have had an AMI than in those who have not. The prognosis for illness and death. Before 1900, infectious diseases and malnu- those with established CHF is generally poor and worse than for trition were the most common causes of death; however, those with most malignancies (McMurray and Stewart 2000) or primarily because of improved nutrition and public health AIDS, with a one-year mortality rate as high as 40 percent and a measures, they have gradually been supplanted in most high- five-year mortality between 26 and 75 percent. income countries by CVD and cancer. As improvements con- The worldwide burden of CHF is substantial and continues tinue to spread to developing countries, CVD mortality rates to rise. Throughout the developed world the prevalence is are increasing. about 2 to 3 percent, with an annual incidence rate of 0.1 to 0.2 Known as the epidemiological transition, this shift is highly percent (McMurray and Stewart 2000). However, the incidence correlated with changes in personal and collective wealth (the and prevalence of CHF rise dramatically with age. Prevalence is economic transition), social structure (the social transition), 27 per 1,000 population for those older than 65, compared with and demographics (the demographic transition). Omran 0.7 per 1,000 for those younger than 50 (McKelvie 2003). CHF (1971) provides an excellent model of the epidemiological occurs more frequently in men, and incidence and mortality transition that divides it into three basic ages: pestilence and differ substantially according to gender and socioeconomic sta- famine, receding pandemics, and degenerative and human- tus. CHF causes 53,000 deaths in the United States each year created diseases (table 33.1). Olshansky and Ault (1986) add a and contributes to another 213,000, and the death rate attrib- fourth stage: delayed degenerative diseases. uted to CHF rose by 155 percent from 1979 to 2001 in the The consistent pattern for most high-income countries going United States (American Heart Association 2002). CHF is the through the epidemiological transition has been initially high first-listed diagnosis in 1 million hospitalizations. rates of stroke, mostly hemorrhagic. Only in the third phase, with the presence of increased resources, but coupled with Rheumatic Heart Disease. RHD is the consequence of an increased diabetes and smoking rates and adverse lipid profiles, acute rheumatic fever (ARF)--that is, a poorly adapted do rates of IHD climb. This phase is also accompanied by better autoimmune response to group A -hemolytic streptococci. It control of severe hypertension, reducing the rates of hemor- affects the connective tissue, mainly the joints and the heart rhagic stroke, which is then replaced by ischemic stroke. Most valves. The most serious complications are valvular stenosis, regions appear to be following this pattern and have a predomi- regurgitation following the valvulitis, or both (Ephrem, nance of IHD. The two exceptions are East Asia and the Pacific Abegaz, and Muhe 1990). RHD is also a predisposing factor for and Sub-Saharan Africa. The pattern in East Asia and the Pacific infective endocarditis, a disease of younger adults, predomi- is dominated by China and appears to be a result of China's stage nantly males (Koegelenberg and others 2003). in the transition but may also be following a pattern similar to According to 2001 estimates, RHD accounts for 338,000 Japan's--that is, dominated by more strokes and fewer IHD deaths per year worldwide, two-thirds of them in Southeast deaths--whereas Sub-Saharan Africa is in an earlier phase of the Asia and the Western Pacific (WHO 2002b). About 12 million epidemiological transition. people in developing countries, most of them children, suffer Even though countries tend to enter these stages at different from RHD (WHO 1995). Steer and others' (2002) review of times, the progression from one stage to the next tends to pro- developing countries suggests that RHD prevalence in children ceed in a predictable manner. The six World Bank regions are is between 0.7 and 14 per 1,000, with the highest rates in Asia. at various phases of the epidemiological transition (table 33.1), RHD and ARF are the most common causes of cardiac disease and where development has occurred, it has often been at a among children in developing countries (Ephrem, Abegaz, and more compressed rate than in the high-income countries. Muhe 1990; Schneider and Bezabih 2001; Steer and others Although rates of IHD and stroke fell 2 to 3 percent per year in 2002) and account for almost 10 percent of sudden cardiac the high-income countries during the 1970s and 1980s, the rate deaths (Kaplan 1985). of decline has since slowed. Overweight and obesity are esca- Until the 1950s, ARF accounted for a substantial portion of lating at an alarming pace, while rates of type 2 diabetes, hyper- cardiovascular problems among schoolchildren in developed tension, and lipid abnormalities associated with obesity are on countries, and even though it is now far less common, out- the rise. This trend is not unique to the developed countries, breaks still occur (Carapetis, Currie, and Kaplan 1999), however. According to the World Health Organization, world- suggesting that neither antibiotics nor other public health mea- wide more than 1 billion adults are overweight and 300 million sures have been totally effective in controlling ARF. are clinically obese. Even more disturbing are increases in childhood obesity that have led to large increases in diabetes and hypertension. If these trends continue, age-adjusted CVD The Epidemiological Transition mortality rates could increase in the high-income countries in Over the past two centuries, the industrial and technological the coming years. These trends are discussed in greater detail in revolutions have resulted in a dramatic shift in the causes of chapter 45. Cardiovascular Disease | 647 648 Table 33.1 Stages of the Epidemiological Transition and Its Global Status, by Region | Disease Percentage Percentage Life of deaths of the world's Control expectancy attributable population Stage Description (years) Dominant form of CVD to CVD in this stage Regions affected Priorities Pestilence and famine Predominance of 35 RHD, cardiomyopathy caused 5­10 11 Sub-Saharan Africa, parts of all regions excluding malnutrition and infectious diseases by infection and malnutrition high-income regions in Developing Receding pandemics Improved nutrition and public 50 Rheumatic valvular disease, 15­35 38 South Asia, southern East Asia and the Pacific, health leads to increase in IHD, hemorrhagic stroke parts of Latin America and the Caribbean chronic diseases, hypertension Countries Degenerative and Increased fat and caloric intake, 60 IHD, stroke (ischemic and 50 35 Europe and Central Asia, northern East Asia and human-created widespread tobacco use, chronic hemorrhagic) the Pacific, Latin America and the Caribbean, diseases disease deaths exceed mortality Middle East and North Africa, and urban parts of | Thomas from infections and malnutrition most low-income regions (especially India) Delayed degenerative CVD and cancer are leading 70 IHD, stroke (ischemic and 50 15 High-income countries, parts of Latin America and A. diseases causes of morbidity and mortality, hemorrhagic), CHF the Caribbean Gaziano, prevention and treatment avoids death and delays onset; K. age-adjusted CVD declines Srinath Source: Adapted from Olshanksy and Ault 1986; Omran 1971; WHO 2003b. Reddy ,Fred Paccaud, and others Risk Factors portion of the world's population lives in East Asia and the The risk of developing CVD depends to a large extent on the Pacific and South Asia and the incidence of IHD is high in presence of several risk factors. The major risk factors for CVD Europe and Central Asia. include tobacco use, high blood pressure, high blood glucose, lipid abnormalities, obesity, and physical inactivity. The global East Asia and the Pacific. The status and character of the epi- variations in CVD rates are related to temporal and regional demiological transition across the region reflects the diversity of variations in these known risk factors. Discussions of the economic circumstances in East Asia and the Pacific. Since the strength of the associations of the various factors with CVD are 1950s, life expectancy in China has nearly doubled from 37 years found elsewhere (chapters 30, 44, and 45). Although some risk to 71 years (WHO 2003b). Approximately 60 percent of the factors, such as age, ethnicity, and gender, obviously cannot be population still lives outside urban centers, and as is the case in modified, most of the risk is attributable to lifestyle and behav- most developing countries, rates of IHD, stroke, and hyperten- ioral patterns, which can be changed. sion are higher in urban centers. China appears to be straddling the second and third stages of a Japanese-style epidemiological transition, with CVD rates higher than 35 percent, though dom- BURDEN OF DISEASE inated by stroke, not IHD. However, in urban China, the death CVD is the leading cause of death in all World Bank regions with rate from IHD rose by 53 percent from 1988 to 1996. the exception of Sub-Saharan Africa (figure 33.1), where HIV/AIDS has emerged as the leading cause of mortality Europe and Central Asia. The emerging market economies, (Mathers and others 2006). Between 1990 and 2020, IHD is which consist of the former socialist states of Europe, are anticipated to increase by 120 percent for women and 137 per- largely in the third phase of the epidemiological transition. As cent for men in developing countries,compared with age-related a group, they have the highest rates of CVD mortality in the increases of 30 to 60 percent in developed countries (Leeder and world, similar to those seen in the United States in the 1960s others 2004). Even though 80 percent of CVD deaths occur in when CVD was at its peak. Belarus, Croatia, Kazakhstan, low- and middle-income countries, the death rates for most Romania, and Ukraine have seen significant increases in IHD regions are still below the rate for high-income countries, which death rates (figure 33.2). In the Russian Federation, life is 320 per 100,000 population annually. The marked exception is expectancy for men has dropped precipitously since 1986 from Europe and Central Asia, which has a rate of 690 CVD deaths per 71.6 years to about 59 years in 2004, in large part because of 100,000 population. CVD. In the Czech Republic, Hungary, Poland, and Slovenia, age-adjusted CVD rates have been declining. Nevertheless, Regional Burdens CVD rates generally remain higher than in Western Europe. The majority of the burden occurs in East Asia and the Pacific, Europe and Central Asia, and South Asia because a large pro- Croatia 62% 61% Kazakhstan 56% 36% Percentage of total deaths Belarus 53% 30% 70 Cardiovascular diseases Respiratory infections Ukraine 49% 38% 60 Malignant neoplasms Chronic lung diseases Romania 26% 26% Injuries HIV/AIDS 50 Japan 10% 8% 40 Hungary 2% 2% Greece 15% 11% 30 Portugal 29% 19% 20 United States 29% 30% 10 Netherlands 29% 39% 0 Sweden 40% 43% and 20% 43% and Asia and and Luxembourg Asia East Africa Africa Australia 52% 46% Europe South AsiaPacific America Caribbean Sub-Saharan Central East the MiddleNorth Denmark 46% 49% Males Females Latin the Source: Mathers and others 2006. Source: Mackay and Mensah 2004. Figure 33.1 Major Causes of Death in Persons of All Ages in Low- Figure 33.2 Percentage Change in Ischemic Heart Disease Death and Middle-Income Regions Rates in People Age 35 to 74, 1988­98, Selected Countries Cardiovascular Disease | 649 Latin America and the Caribbean. In 2001, CVD accounted disease accounts for the dominance of stroke (Bertrand 1999). for about 31 percent of all deaths in Latin America and the RHD and cardiomyopathies, the latter caused mostly by mal- Caribbean, but that figure is expected to rise to 38 percent by nutrition, various viral illnesses, and parasitic organisms, are 2020 (Murray and Lopez 1996). In recent decades, average life also important causes of CVD mortality and morbidity. expectancy in Latin America and the Caribbean has risen from 51 to 71 years, and the quality of nutrition has improved steadily. Social and Economic Impact At the same time, the region has seen a switch from vegetables as Leeder and others' (2004) report highlights the economic a source of protein to animal protein and an increase in fat impact of cardiovascular diseases in developing economies, intake as a percentage of energy. As a whole, the region seems to which arises largely because working-age adults account for a be in the third phase, but in South America, some areas are still high proportion of the CVD burden. Conservative estimates in in the first phase of the transition. Brazil, China, India, Mexico, and South Africa indicate that each year at least 21 million years of future productive life are Middle East and North Africa. Increasing economic wealth in lost because of CVD. In South Africa, for example, costs for the the Middle East and North Africa has been characteristically direct treatment of CVD were equivalent to 2 to 3 percent of accompanied by urbanization. The rate of CVD has been gross domestic product, or roughly 25 percent of all health care increasing rapidly and is now the leading cause of death, expenditures (Pestana and others 1996). accounting for 25 to 45 percent of total deaths. Over the past Current expenditures in developed countries are indicators few decades, daily per capita fat consumption has increased in of possible future expenditure in developing countries. For most countries in the region, ranging from a 13.6 percent example, Hodgson and others (2001) estimated that in 2003 increase in Sudan to a 143.3 percent increase in Saudi Arabia the direct and indirect costs of CVD in the United States would (Musaiger 2002). IHD is the predominant cause of CVD, with amount to US$350 billion. They also estimated that in 1998 about three IHD deaths for every stroke death. RHD remains a Americans spent US$109 billion on hypertension, equivalent to major cause of morbidity and mortality, but the number of about 13 percent of the health care budget. Studies are limited hospitalizations for RHD is declining rapidly. but suggest that obesity-related diseases are responsible for 2 to 8 percent of all health care expenditures in developed South Asia. Some regions of India appear to be in the first countries. phase of the transition, whereas others are in the second or even the third phase. Nonetheless, India is experiencing an alarming increase in heart disease, which seems to be linked to changes in lifestyle and diet, rapid urbanization, and possibly COST-EFFECTIVENESS OF INTERVENTIONS an underlying genetic component. Diabetes is also a major CVD remains one of the most studied and written about sub- health issue. India has 31.6 million diabetics, and the number jects in medicine. As a result, many interventions exist with is expected to reach 57.2 million by 2025 (Ghaffar, Reddy, and strong evidence for significant reductions in morbidity and Singhi 2004). The World Health Organization estimates that, mortality associated with CVD. by 2010, 60 percent of the world's cardiac patients will be in India. About 50 percent of CVD-related deaths occur among Intervention Effectiveness by Disease people younger than 70, compared with about 22 percent in the West. Between 2000 and 2030, about 35 percent of all CVD This chapter addresses those interventions believed to have the deaths in India will occur among those age 35 to 64, compared largest effect because they result in large reductions in CVD with only 12 percent in the United States and 22 percent in events, are inexpensive, or the prevalence or incidence of the dis- China (Leeder and others 2004). eases to which they are directed is significant.The omission of an intervention does not imply that it is not cost-effective but rather Sub-Saharan Africa. In Sub-Saharan Africa, deaths attributa- that either it had an effect on a smaller percentage of people or ble to CVD are projected to more than double in between the the chapter was unable to encompass all such interventions. years 1990 and 2020. Although HIV/AIDS is the leading over- all cause of death in this region, CVD is the second-leading Acute Myocardial Infarction. Treatment of AMI involves killer and is the first among those over the age of 30. Stroke is medical therapies that reduce myocardial oxygen demand and the dominant form, in keeping with patterns characteristic of fatal arrhythmias (beta-blockers), that restore blood flow by earlier phases of the epidemiological transition. With increas- inhibiting platelet aggregation (aspirin), or that dissolve ing urbanization, levels of average daily physical activity are the thrombus occluding the arterial lumen (thrombolytics) or falling and smoking rates are increasing. Hypertension has an invasive intervention with cardiac catheterization and emerged as a major public health concern, and hypertensive angioplasty. 650 | Disease Control Priorities in Developing Countries | Thomas A. Gaziano, K. Srinath Reddy, Fred Paccaud, and others Beta-blockers are used both during and after an AMI. luminal coronary angioplasty (PTCA), and PTCA with stents. Benefits persist for at least 6 years and up to 15 years after the CABG is the placement of grafts, usually from the saphenous first AMI. The second Thrombolysis in Myocardial Infarction vein or internal mammary artery, to bypass stenosed coronary trial showed significant benefits when beta-blockers were used arteries while maintaining cerebral and peripheral circulation within two hours of symptoms (Roberts and others 1991). by cardiopulmonary bypass. CABG is a major operative proce- Aspirin, an antiplatelet agent, and thrombolytic agents, the dure requiring appropriate surgical and anesthetic environ- standard treatments for reopening the artery in AMI, have ments and has a perioperative mortality of 1 to 3 percent, with demonstrated an additive effect in reducing mortality (GISSI later complication rates of 15 to 20 percent. 1986), with a benefit irrespective of age, sex, blood pressure, Almost 1 million CABGs per year are performed worldwide, heart rate, or previous history of AMI or diabetes (Fibrinolytic with about 519,000 interventions in the United States alone in Therapy Trialists' Collaborative Group 1994). The benefits are 2000 (American Heart Association 2002). The main indication greater the closer the thrombolytics are given to the time of for CABG is for those with left main coronary artery stenosis or onset, and the risk of bleeding is greater the later they are given. those with involvement of multiple coronary arteries with The risk of adverse events following administration of throm- reduced left ventricular function, particularly among diabetics. bolytics is low during the first 24 hours; trials with throm- The prevalence estimates of those with left main coronary bolytics show that the benefits are greatest when they are artery stenosis or involvement of three coronary arteries has administered less than 12 hours after an AMI and preferably varied over time, but current estimates range from 7 to 20 per- less than 6 hours (Antman and others 2004). cent of survivors of myocardial infarction (Kuntz and others The invasive alternative to immediate medical reperfusion 1996; Rogers and others 1991; Topol, Holmes, and Rogers of an occluded coronary artery is angioplasty or percutaneous 1991) For these cases, investigators have shown that CABG is coronary intervention. Its superiority over thrombolysis in more beneficial than medical treatment, both in terms of developed countries remains a matter of debate. Issues that symptoms and of mortality (Eagle and others 1999). remain important in relation to the choice of strategy are over- Both developed and developing countries are increasingly all severity or location of the AMI and the time from symptom using PTCA (Denbow and others 1997). The main indications onset to initiation of treatment. In patients presenting late or for its use are low-risk patients with single- or double-vessel with a high risk of mortality, such as those in cardiogenic disease and poor response to medical treatment. The success shock, percutaneous coronary intervention may be beneficial rate of PTCA is more than 95 percent; however, because it has (Hochman and others 1999). However, as with thrombolytic no mortality benefit when compared with medical therapy agents, the benefits of percutaneous coronary intervention or CABG, we did not evaluate new analyses of the cost- diminish significantly with time between the onset of symp- effectiveness of this intervention, but instead provided infor- toms and the opening of the artery (De Luca and others 2004; mation from experience in developed countries. The addition D. O. Williams 2004). of stents to PTCA has lead to a decrease in restenosis rates and The invasive strategy requires a facility and individual physi- readmissions to hospitals but shows no change in mortality cians who conduct enough of the procedures annually to compared with medical therapy. remain proficient. In the absence of these conditions, the American Heart Association recommends that treatment focus Pharmacological Interventions The pharmacological inter- on thrombolytics (Antman and others 2004). Given either a ventions either prevent thrombosis, as does aspirin, or target lack of facilities and operators for percutaneous interventions the individual risk factors, as do the antihypertensives (diuret- or long distances to such facilities in many developing coun- ics, beta-blockers, and ACE inhibitors) or statins targeting tries, we did not evaluate this procedure. cholesterol. Furthermore, these agents may possibly have addi- tional properties of reducing the risk of fatal arrhythmias, Long-Term Management of Existing Vascular Disease. The improving repair after AMI (remodeling), or stabilizing the management of individuals with chronic vascular disease con- atherosclerotic plaque. sists of invasive techniques, pharmacotherapy, lifestyle and Overall, the long-term administration of antiplatelet agents behavioral changes, and rehabilitative measures. It also involves in those with vascular disease leads to a 25 percent reduction addressing such issues as adherence to treatment,regular follow- in the risk of major vascular events: 33 percent for nonfatal ups to determine compliance and assess risk, and treatment of AMI, 25 percent for nonfatal stroke, and 16 percent for any comorbidities that are likely to have an impact on the progres- vascular death. The use of aspirin has produced similar sion of vascular disease (for instance, renal disease). benefits in individuals with IHD or prior stroke. Antiplatelet treatment in individuals with a previous AMI has been Invasive Interventions The three most common procedures shown to prevent 18 nonfatal myocardial infarctions, 5 nonfa- are coronary artery bypass graft (CABG), percutaneous trans- tal strokes, and 14 vascular deaths for every 1,000 patients Cardiovascular Disease | 651 treated for two years (Antithrombotic Trialists' Collaboration brain), the long-term benefits of lowering blood pressure have 2002). been clearly established. Lowering blood pressure reduces the The benefits of antiplatelet agents for those with vascular overall risk of future stroke by 28 percent and of other vascular disease far outweigh the risks. The risk of intracranial bleeding events and CHF by 26 percent in patients with a history of increases by nearly 25 percent with the use of antiplatelet stroke disease, irrespective of their hypertension status. The agents, but in absolute terms this risk comes to only one or two benefits are even more pronounced for individuals with a his- intracranial bleeds per 1,000 patients treated per year. The risk tory of hemorrhagic stroke. Larger reductions in blood pres- of major extracranial bleeding, mostly gastrointestinal, also sure confer greater benefits, and benefits are present across dif- increases by 60 percent, or one or two excess events per 1,000 ferent age groups, genders, and ethnicities and with varying patients per year. comorbid status. The most established and commonly used agent is aspirin, Beta-blockers are one of the cornerstones of long-term although other agents (for example, clopidogrel or ticlopidine) treatment of individuals with IHD, especially those with a his- with similar efficacy but much greater cost are available. Low tory of AMI. Long-term use of beta-blockers has been associ- doses of aspirin--75 to 100 milligrams (mg) per day--are as ated with 23 percent relative risk reduction in mortality beneficial as higher doses. (Freemantle and others 1999), 25 percent relative risk reduc- Lowering LDL and elevating HDL cholesterol levels is one tion in nonfatal myocardial infarction, and 30 percent relative of the cornerstones of treatment of cardiovascular disease, risk reduction in sudden cardiac death (Yusuf and others and investigators have suggested that suboptimal levels of 1985). The benefits are larger for those at highest risk of sus- cholesterol contribute to almost two-thirds of the global car- taining a vascular event in the future and are present across all diovascular risk (WHO 2002b). Although the usual target of age groups and sexes. Furthermore, beta-blockers provide clear lipid-lowering therapy has been lowering total or LDL choles- benefits in patients with chronic stable angina, where they pro- terol, medical experts are increasingly recognizing the impor- vide symptom relief as well as reductions in vascular events tance of increasing HDL cholesterol and lowering triglyceride (Heidenreich and others 1999). levels, especially in high-risk individuals, such as those with ACE inhibitors have proved invaluable in preventing cardio- diabetes or metabolic syndrome, as well as in ethnic popula- vascular events and CHF in those with IHD. The extent to tions like Southeast Asians. which the benefits conferred by their use are caused by their Recent evidence has demonstrated that the relationship ability to lower blood pressure or by their other properties, between cholesterol levels and vascular events is continuous such as cardiac remodeling and neurohormonal modulation, is and occurs at much lower cholesterol thresholds than previ- not clear. Long-term use of ACE inhibitors in those with a his- ously believed. The clinical trials have consistently demon- tory of myocardial infarction and in other individuals at high strated a 25 to 30 percent reduction in the risk of cardiovascu- risk of vascular disease reduces vascular mortality by 25 percent lar morbidity and mortality. Furthermore, the evidence and other nonfatal events, such as recurrent myocardial infarc- suggests that more aggressive reductions in cholesterol have tion, revascularization, hospitalization, progression or new higher benefits than mild or moderate reductions (Cannon and onset of CHF, and stroke (Teo and others 2002). In those with others 2004; Knatterud and others 2000). No increased risk of asymptomatic or symptomatic left ventricular dysfunction cancers appears to exist, as was previously believed, although a after myocardial infarction, ACE inhibitors reduce the risk of a small increase exists in the risk of inflammation of noncardiac variety of vascular endpoints by 20 to 26 percent. Similarly, the muscle (myopathy) (Pfeffer and others 2002). use of ACE inhibitors even in those with no evident left ven- As with cholesterol, the relationship between blood pressure tricular dysfunction confers a 21 percent reduction in risk for and vascular events is continuous and is discussed further in major coronary events (Dagenais and others 2001), 32 percent chapter 45. Even patients with presumed "normal" blood pres- for stroke (Bosch and others 2002), and 20 to 22 percent for sure and prior vascular disease benefit from lowering blood composite vascular outcomes (Fox 2003). pressure (Nissen and others 2004), confirming earlier evidence that individuals with a history of AMI who have lower blood Nonpharmacological Interventions Cessation of smoking pressure are less likely to have future vascular events. and dietary modifications are important goals of secondary Furthermore, investigators have established mortality and mor- prevention of CVD. Cardiac rehabilitation, including exercise, bidity benefits for several specific classes of drugs to reduce is useful for a wide range of patients with IHD and reduces blood pressure in patients with vascular disease, namely, future vascular events by about 15 percent. Exercise alone beta-blockers, calcium-channel blockers, and ACE inhibitors reduces vascular mortality by 24 percent and vascular end- (Fox 2003). points by 15 percent (Jolliffe and others 2000). Results of trials In patients with a prior history of stroke or transient for psychological interventions targeted at stress, depression, ischemic attack (transient occlusion of artery supplying the low social support, and so on have been conflicting. 652 | Disease Control Priorities in Developing Countries | Thomas A. Gaziano, K. Srinath Reddy, Fred Paccaud, and others Congestive Heart Failure. Diuretics are standard therapy for Linking Costs and Effectiveness in Developing Countries CHF, with the loop and thiazide diuretics most commonly Few intervention trials have been carried out solely in develop- used. Diuretics provide relief of symptoms more rapidly than ing countries, but investigators have extrapolated estimates of any other CHF medication because they are the only drugs cost-effectiveness ratios for the developing world in general that can adequately control the fluid retention associated with based on changes in key input prices (Goldman and others CHF. Using spironolactone, a neurohormonal antagonist, 1991); however, this process is limited by the fact that both the together with a diuretic decreased the risk of mortality by 30 underlying epidemiology and the costs can differ significantly percent and of hospitalization by 35 percent, compared with across and within countries and regions. Thus, our results a placebo in patients with severely advanced heart failure reflect models that used prices and epidemiological data for (Pitt and others 1999); however, this combination requires World Bank regions where applicable. Intervention effects were, intensive monitoring of electrolytes and testing to follow however, based on systematic reviews of randomized trials or patients and thus was not included in our cost-effectiveness meta-analyses in developed countries. Until intervention trials analyses. are conducted in developing countries, this option remains the Investigators have shown that ACE inhibitors reduce risks best for evaluating the cost-effectiveness of various interven- related to a variety of endpoints, including mortality, hospital- tions in the developing regions. In cases in which models for dis- ization, major coronary events, deterioration of symptoms, and eases in selected regions were not developed, we present results progression from asymptomatic to symptomatic left ventricu- of cost-effectiveness analyses from high-income countries. lar dysfunction, by 25 to 33 percent. The benefit is conferred We used estimates of life expectancy for the model from irrespective of the etiology of systolic failure; begins soon after data supplied by the volume editors. The model includes only the start of treatment; persists over the long term; and is inde- the costs related to the intervention itself and to CVD events pendent of age, sex, and baseline use of other medications. and their sequelae. Costs include personnel salaries, health care Furthermore, the use of ACE inhibitors has proved to be highly visits, diagnostic tests, and hospital stays as provided by the vol- cost-effective in developed countries. ume editors. Our analysis does not include indirect costs, such Beta-blockers improve symptoms, decrease hospitalization as those arising from lost work time or family assistance. Drug and deterioration of heart function, and improve mortality. costs are from McFayden (2003). All are in U.S. dollars unless They should be used even when the patient becomes asympto- otherwise specified. Disability weights were taken from matic. Beta-blockers are beneficial at all stages of CHF, reduc- Mathers and others (2006). ing the morbidity and mortality associated with CHF by 25 to 33 percent. Because most patients with CHF die of sudden car- Ischemic Heart Disease. diac death, the protective effects of beta-blockers are probably related to their antiarrhythmic properties. Acute Myocardial Infarction We evaluated four incremental Digitalis decreases hospitalization rates in individuals with strategies for the treatment of AMI and compared them with CHF but has no effect on vascular or total mortality (Digitalis a strategy of no treatment as a base case. The four treatment Investigation Group 1997). Given that it also has a narrow strategies were aspirin (162.5 mg per day for 30 days); aspirin therapeutic-toxic window and requires careful monitoring, its and atenolol (100 mg per day for 30 days); aspirin, atenolol, and role in standard treatment for CHF has diminished and has not streptokinase (1.5 million units); and aspirin, atenolol, and tis- been included in our cost-effectiveness analyses. sue plasminogen activator (100 mg accelerated regimen). Doses for the aspirin and streptokinase were those used by the Second Rheumatic Heart Disease. The management of patients International Study of Infarct Survival Collaborative Group with ARF includes providing antistreptococcal treatment, (ISIS-2 Collaborative Group 1988), the atenolol regimen was managing clinical manifestations, and screening children. In that of the First International Study of Infarct Survival (ISIS-1 the acute stage, all patients with ARF should be treated as if Collaborative Group 1986), and the tissue plasminogen activa- they have a group A streptococcal infection--that is, with a tor dosing was that used in the Global Use of Strategies to Open 10-day course of penicillin. Anti-inflammatory agents Occluded Coronary Arteries (GUSTO)­I trial (GUSTO provide symptomatic relief during ARF but do not prevent Investigators 1993). The relative risk of dying from AMI was RHD. Secondary prophylaxis prevents colonization of the reduced for all patients receiving the medications. Patients upper respiratory tract and consists of penicillin or sulfadiazine receiving the thrombolytics faced increased risks of major for the first five years (and for life for patients with valvular bleeds and hemorrhagic strokes. Because the effectiveness of heart disease). Noncompliance is frequent, reaching rates streptokinase diminishes over time, we carried out two further as high as one-third of patients (Bassili and others 2000). sensitivity analyses to compare its use for patients over and Tertiary treatment entails surgery for valve replacement or under the age of 75 and for patients who receive the intervention valvuloplasty. sooner or later than six hours after the onset of symptoms. Cardiovascular Disease | 653 Table 33.2 presents incremental cost-effectiveness ratios able in developed countries; however, given that hospital facil- (ICERs) for each therapy by region. The incremental cost per ities may not be available to most patients in many developing DALY averted was less than US$25 for all six regions for the regions, we undertook two separate analyses, one with hospital aspirin and aspirin plus atenolol interventions; US$634 to costs and one without. US$734 for aspirin, atenolol, and streptokinase; and slightly In a setting where hospitals are available, a combination of less than US$16,000 for aspirin, atenolol, and tissue plasmino- aspirin and atenolol dominated no therapy and was cost saving gen activator. Minor variations occurred between regions in all regions (table 33.2). The ICERs for the addition of because of small differences in follow-up care costs. The results enalapril ranged from US$660 per DALY in Sub-Saharan Africa for an analysis that evaluated ICERs as cost per life year saved to US$866 per DALY in Europe and Central Asia. The combi- showed no significant differences. nation of all four medications ranged from US$1,720 per Table 33.3 displays the results of the secondary analysis for DALY to US$2,026 per DALY. For CABG the costs per DALY streptokinase and tissue plasminogen activator. Giving the ranged from about US$24,000 to more than US$72,000. streptokinase sooner than six hours following onset reduces the Despite having similar benefits as aspirin and atenolol in rela- incremental cost per DALY to less than US$440 compared with tion to mortality, enalapril and lovastatin demonstrated higher more than US$1,300 if given after six hours. Similar effects are per DALY costs because of the added costs of monitoring renal seen when streptokinase is given to those under 75 compared and liver function, respectively, as is required for these two with those 75 years or older. medications. According to meta-analyses, nitroglycerin has a modest When we assumed that hospitals were not readily available effect on mortality in AMI: a 3 percent reduction. However, (table 33.2), no therapy combination was cost saving compared given that it can have profound effects on blood pressure that with no therapy. The combination of aspirin and atenolol was could limit the use of beta-blockers that confer more signifi- the next best strategy, with ICERs ranging from US$386 per cant benefits, its use should be limited to patients with ongoing DALY in South Asia to US$545 per DALY in Latin America and ischemic pain and systolic blood pressures greater than 90 mil- the Caribbean. The addition of enalapril increased the range of limeters of mercury who do not have ongoing right ventricu- ICERs to US$783 per DALY to US$1,111 per DALY, and the lar infarction. When modeled, it had a reasonable cost- addition of lovastatin increased them still further. CABG was effectiveness ratio of US$70 per life year saved, but we did not not evaluated because of the underlying assumption that hos- include the analysis in the incremental analysis because of the pitals were not available. blood pressure effects of the multiple agents. Table 33.4 shows the number of events prevented with the four-drug combination medical therapy compared with no Secondary Prevention Four medical therapies--aspirin, beta- therapy and the additional number of events averted blockers, statins, and ACE inhibitors--have been the mainstay with CABG compared with the four-drug combination. The of treatment for those with IHD in the developed world. To medical regimen alone would prevent some 2,000 CVD deaths, evaluate the best medical intervention, we used incremental about 4,000 myocardial infarctions, and approximately 200 cost-effectiveness analysis to examine the 15 different possible strokes per million persons treated in each region. The use of combinations of the four standard medical therapies. The four CABG in addition to the medical regimen would prevent an therapies were 75 to 100 mg per day of aspirin, 100 mg per day additional 65­70 deaths, nearly 300 myocardial infarctions, and of atenolol, 10 mg per day of enalapril, and 40 mg per day of up to 30 strokes per million population. lovastatin. In addition, CABG surgery provides an invasive option that gives added mortality benefit when compared with Congestive Heart Failure. The interventions examined for conventional medical therapy in patients with certain anatom- CHF were the addition of the ACE inhibitor enalapril, the beta- ical obstructions in coronary circulation. Thus, we evaluated blocker metoprolol, or both to a baseline of diuretic treatment. CABG in addition to all four medications for those with left As for the IHD interventions, we performed separate analyses main coronary artery disease or with three-vessel coronary for each assumption of whether or not hospital facilities would artery disease and reduced left ventricular function. Because be available. For the model of treatment for CHF assuming these therapies also have significant effects on the incidence of hospitalization (table 33.2), the addition of enalapril is cost sav- stroke, we included the effect on DALYs and costs for stroke in ing and the ICER for the addition of metoprolol ranges from the analyses. US$124 to US$219 per DALY depending on the region. When In addition to the mortality benefits demonstrated by trials the availability of hospitals is limited (table 33.2), the enalapril of the individual medications or surgery, they also resulted in plus diuretics strategy is no longer cost saving, but it costs only significant reductions in hospitalizations in developed coun- US$31 per DALY or less, and the ICER for enalapril, metopro- tries. The cost savings from these reduced hospitalizations lol, and diuretics increases only to about US$275 per DALY. make the cost-effectiveness of such interventions quite favor- These figures are probably underestimates of the cost per 654 | Disease Control Priorities in Developing Countries | Thomas A. Gaziano, K. Srinath Reddy, Fred Paccaud, and others Table 33.2 ICERs for Treatment Compared with No Treatment, by Region US$/DALY ACE inhibitors and beta- blockers for CHF ACE inhibitors compared with and beta-blockers baseline of Medical therapy and for CHF compared diuretics, Medical therapy for Medical therapy and CABG for IHD CABG for IHD compared with baseline limited AMI compared with compared with baseline of no treatment, with baseline of no treatment, of diuretics, hospital baseline of no treatment hospital access limited hospital access hospital access access ASA ASA, BB, ASA, BB, ASA, BB, ASA, BB, ASA, BB, ACEI, ACEI, Region ASA ASA, BB BB, SK TPA ASA, BB ACEI ACEI, Statin CABG ASA, BB ACEI Statin ACEI MET ACEI MET East Asia and 13 15 672 15,867 Cost saving 781 1,914 33,846 461 942 2,220 Cost saving 189 27 274 the Pacific Europe and 19 21 722 15,878 Cost saving 866 2,026 47,942 530 1,097 2,470 Cost saving 144 30 275 Central Asia Latin America and 20 22 734 15,887 Cost saving 821 1,942 62,426 545 1,111 2,497 Cost saving 124 31 275 the Caribbean Middle East and 17 20 715 15,893 Cost saving 672 1,686 72,345 527 996 2,305 Cost saving 128 29 275 North Africa South Asia 9 11 638 15,860 Cost saving 715 1,819 24,040 386 828 2,034 Cost saving 219 25 273 Sub-Saharan 9 11 634 15,862 Cost saving 660 1,720 26,813 389 783 1,955 Cost saving 218 25 273 Africa Source: Authors' calculations. Cardiovascular ASA aspirin, BB atenolol, SK streptokinase, TPA tissue plasminogen activator, ACEI enalapril, Statin lovastatin, MET metoprolol. Note: The intervention in the first column of each set of strategies is compared with the baseline; each successive intervention for each set of strategies is compared with the intervention immediately to its left. Disease | 655 Table 33.3 Sensitivity Analyses: Effect of Time to Treatment may be cost-effective. Secondary prevention using benzathine and Age on Use of Thrombolytics in AMI (All Regions penicillin injections is cost-saving according to Strasser (1985) Combined) and should be considered for all developing countries with the SKa (US$/DALY) TPAa (US$/DALY) infrastructure to perform the required follow-up. Time to thrombolysis 6 hours 374­437 15,800 Cost-Effectiveness Analyses in High-Income Countries 6­12 hours 1,300­1,440 15,700 Table 33.5 summarizes the results of cost-effectiveness analyses Age at treatment for CVD interventions in high-income countries. These results 75 559­650 14,800 include analyses that are similar to ours. The differences are 75 or older 1,260­1,350 21,000 that they reflect costs and treatment patterns in the high- income countries studied, mostly the United States. Costs in Source: Authors' calculations. SK streptokinase; TPA tissue plasminogen activator. developing countries are roughly one-fifth of those in devel- a. In addition to aspirin and atenolol. oped countries (but closer to one-third in Latin America and approaching one-half in South Africa). However, where DALY, given some loss in the mortality benefit for the hospital- patented drugs are involved and patent laws are enforced, the ization that the model does not capture. costs may be much closer to U.S. levels. Because the cost-effectiveness studies have been undertaken Rheumatic Heart Disease. For RHD, except in epidemics, sec- largely in the United States, the results do not always readily ondary prevention is more effective than primary prevention. transfer to developing countries. In some U.S. studies, the Primary prevention by means of antibiotic treatment of strep- alternative procedure considered is medical management; such tococcus infections of the pharynx is not highly cost-effective in facilities simply may not exist in developing countries. endemic situations, given that only 10 to 20 percent of such Similarly, interventions that are cost saving in the United States infections are from streptococcus, less than 3 percent of these may not be cost saving in developing countries but may well be will evolve into rheumatic fever, and only a proportion of these cost-effective in terms of cost per DALY saved. Furthermore, continue on to RHD (Strasser 1985). The development of a the cost-effectiveness analyses reflect morbidity and mortality rapid antigen test for diagnosing group A streptococcal pharyn- rates in developed countries. gitis may make primary prevention more cost-effective (Majeed Interventions that Kupersmith and others (1995) classify as and others 1993). Similarly, in an epidemic in which the pro- highly cost-effective in the United States (less than US$20,000 portion of infections from streptococcus is higher or the rate of per life year saved or quality-adjusted life year saved) may be progression to rheumatic fever is higher, primary prevention cost-effective in many developing countries. Interventions that Table 33.4 Number of Deaths and CVD Events Prevented by the Use of a Four-Component Medical Regimen and CABG per 100,000 Myocardial Infarction Survivors over 10 Years, by Region Number of events prevented with four-component Number of incremental events prevented with medical regimen compared with no therapya CABG compared with medical therapy IHD Stroke Myocardial Myocardial deaths deaths infarctions Strokes IHD deaths Stroke deaths infarctions Strokes Region averted averted prevented prevented averted averted prevented prevented East Asia and 1,900 104 4,077 209 79 11 248 22 the Pacific Europe and 1,990 89 3,964 179 83 1 294 7 Central Asia Latin America and 1,913 83 4,040 118 62 4 258 18 the Caribbean Middle East and 1,908 95 4,294 118 62 1 296 22 North Africa South Asia 1,930 97 4,043 122 34 2 275 30 Sub-Saharan Africa 1,909 91 4,233 173 69 12 254 1 Source: Authors' calculations. a. Aspirin, atenolol, enalapril, and lovastatin. 656 | Disease Control Priorities in Developing Countries | Thomas A. Gaziano, K. Srinath Reddy, Fred Paccaud, and others Table 33.5 Cost-Effectiveness Analyses for CVD Interventions in High-Income Countries Intervention Alternative Cost-effectiveness Source IHD Lovastatin, 20 mg/day Diet Cost saving (males age 45­54); US$4,700/life year saved (females age 45­54) Goldman and others 1991a Defibrillators in emergency vehicles No defibrillators US$47 to US$551/life year saved; up to US$2,600 in rural areas Jermyn 2000; Ornato and others 1988; Rowley, Garner, and Hampton 1990b Propranolol for postmyocardial No beta-blockers US$2,400 for high-risk patients; US$23,400 for low-risk patients Goldman and others 1988 a infarction (beta-blocker) CABG for left main disease Medical management US$2,700 to US$6,700/life year saved Weinstein and Stason 1982;b A. Williams 1985a PTCA (men age 55 with Medical management US$6,400 to US$8,800/life year saved (US$28,000 to US$132,000 for Wong and others 1990b severe angina) mild angina) Primary angioplasty No intervention after AMI US$12,000/quality-adjusted life year Parmley 1999 Three-vessel CABG Medical management US$14,000/life year saved Weinstein and Stason 1982 Streptokinase (reperfusion), with No intervention after AMI US$15,000/quality-adjusted life year Parmley 1999 PTCA available Tissue plasminogen activator (AMI) Steptokinase US$33,500/life year saved Lorenzoni and others 1998 Primary stenting, one-vessel, men PTCA US$32,000/life year saved Cohen and others 1993 over age 55 Three-vessel CABG for severe angina PTCA US$41,000/life year saved Wong and others 1990b Two-vessel CABG Medical management US$33,000 to US$90,000/life year saved Weinstein and Stason 1982; A. Williams 1985 Angiography for coronary artery disease CABG US$45,000/quality-adjusted life year Doubilet, McNeil, and Weinstein 1985a Stroke Anticoagulants (warfarin) for chronic Aspirin Warfarin dominates for high-risk patients; US$10,000/quality-adjusted life Gage, Cardinalli, and Owens 1998c nonvascular atrial fibrillation year for medium-risk patients; US$462,000/quality-adjusted life year for low-risk patients Anticoagulants for mitral stenosis No anticoagulants US$5,500/quality-adjusted life year Eckman, Levine, and Pauker 1992c and atrial fibrillation Carotid endarterectomy Aspirin US$5,100 to US$51,000/life year saved Kuntz and Kent 1996; Matchar, Pauk, and (symptomatic patients) Lipscomb 1996c Cardiac transplant No transplant US$54,000/life year saved Evans 1986a Arrhythmias Implantable cardioverter-defibrillator Medical management US$28,000/life year saved King, Aubert, and Herman 1998; Kuppermann and Cardiovascular for cardiac arrest (long term) others 1990 RHD Benzathine penicillin injections No injections Cost saving Strasser 1985 Disease Source: Authors. Note: All costs have been converted to 2001 U.S. dollars. | a. Surveyed in Kupersmith and others 1995. 657 b. Surveyed in Tengs and others 1995. c. Surveyed in Holloway and others 1999. Kupersmith and others (1995) classify as cost-effective in Primary Prevention the United States (US$20,000 to US$40,000 per life year saved Because the control of many cardiovascular risk factors is or quality-adjusted life year saved) are probably borderline strongly related to the legislative environment--for example, cost-effective for developing countries. Interventions that that pertaining to tobacco use or nutrition--the design and Kupersmith and others (1995) classify as borderline, expensive, implementation of appropriate laws and regulations is likely to or very expensive in the United States are unlikely to merit increase in developing countries. However, any such initiatives public funding in developing countries. need to be monitored and systematically evaluated, especially Thus, medical interventions that are likely to be cost- to estimate the magnitude of the reduction achieved. effective in developing countries include benzathine penicillin Another area of research is the assessment of chemoprophy- injections as secondary prevention for those who have had laxis in primary prevention. Multidrug combinations such rheumatic fever (usually for five years); ACE inhibitors for as the hypothetical "polypill" are likely to be the first practical CHF; and various drugs (beta-blockers, off-patent statins) for initiative of a long list of important innovations. Both the long-term care following a myocardial infarction, confirming efficacy and the effectiveness of new interventions in primary our earlier analyses. Other therapies that are probably cost- prevention should be evaluated as a matter of urgency, because effective but that we did not analyze include antithrombotic no results of large-scale clinical trials in developing countries agents (aspirin, heparin) to prevent venous thromboembolism; are as yet available. anticoagulants for medium- and high-risk nonvalvular atrial fibrillation (stroke); and anticoagulants for mitral stenosis and atrial fibrillation (stroke). Health Services Selected invasive interventions that might possibly be cost- Capacity building--more specifically, education and training-- effective for CVD in certain developing countries include pace- of health care workers in developing countries, is a major issue maker implants for atrioventricular heart block, primary for the future, along with critical evaluations of the perform- angioplasty for acute myocardial infarction, and reperfusion ance of health workers. Such evaluations should compare vari- with streptokinase. Of course, the ability to undertake these ous capacity-building strategies; for instance, they could com- interventions assumes a cost-effective infrastructure for diag- pare the delivery of simplified regimens of care by community nosis and referral and an adequate volume of cases. For exam- health workers versus delivery of care by trained health profes- ple, the American Heart Association recommends acute angio- sionals. plasty in centers where the physician conducts at least 75 such The dissemination of innovations deserves special attention procedures each year and the hospital conducts at least 200 per in a context of scarce resources (Berwick 2003). The transfer of year. For stroke, carotid endarterectomy is potentially cost- technologies to developing countries should be made on cost- effective for symptomatic patients compared with aspirin effectiveness criteria, which implies analysis conducted in the alone, again in an environment with an adequate volume of specific situation of developing countries--for example, cost- cases. Cost-effectiveness is much lower for asymptomatic cases. effectiveness for thrombolytics in a developing country might Interventions that rank as cost-effective for heart disease in be much worse than in the United States if getting to a hospi- the U.S. context and that are borderline cost-effective in devel- tal on time is a problem. Sensitivity analysis of the cost- oping countries include implantable cardioverter-defibrillator effectiveness of surgical and medical interventions in develop- for cardiac arrest, primary stenting for single-vessel disease (the ing countries is also needed. study was for men over age 55), CABG for two-vessel disease, Furthermore, the appropriate incentives for technological and angiography for patients with a high probability of coro- changes in health care should be investigated (McClellan and nary artery disease. Kessler 1999). This line of research includes analyses of the pricing of technologies (including drugs) or of new designs for RESEARCH AND DEVELOPMENT services, such as point-of-care devices for use by community health workers. Even though most of the interventions currently available The long period of incubation of CVD opens up opportu- appear to be expensive and complex for developing countries, nities for extensive screening based on preclinical signs and the demand for effective care for cardiovascular diseases will biomarkers. However, strong lines of research are needed to exert major pressure on health systems in coming decades. secure effective and safe screening programs and should Increased use of these procedures is already documented in include opportunistic screening for places where visits to China and India (Murray and Lopez 1994, 1997; Unger 1999). health centers are limited. In this context, cardiovascular research should be concentrated Finally, all assessments made in relation to health services in the fields of primary prevention, health services, clinical research should take into account the costs related to scaling up guidelines, clinical research, and epidemiology. any procedure evaluated. 658 | Disease Control Priorities in Developing Countries | Thomas A. Gaziano, K. Srinath Reddy, Fred Paccaud, and others Clinical Guidelines Epidemiological Research The diffusion of health technologies usually leads to a widen- A basic task of epidemiological research is to assess geographic ing of the clinical indication beyond the evidence-based scope and secular trends in the distribution of risk factors. Of special of the intervention (PTCA is a classic example) (Dravik 1998), relevance is the movement from regional to country levels and corresponding to a decrease not only in the procedure's the trend within a country. The impact of poor health status in efficacy, but also in its effectiveness (Anderson and Lomas early life should be assessed from the impact of poor fetal 1988; Blustein 1993). Several studies suggest that overuse and health to the consequence of multiple childhood infections on underuse tend to coexist in the same community and that the risk for CVD. Because of the scarce availability of resources, even severe scarcity of resources does not protect against the development and maintenance of health care should be overuse of cardiological interventions, at least among certain supported by a comprehensive information system. Simple, segments of the population (Joorabchi 1979; Soumerai and affordable health information systems are preferable along the others 1997). lines of the framework developed by the World Health The consequences of such trends are more dramatic in Organization. developing than developed countries. Therefore, the introduc- tion of costly care should be accompanied by a corresponding CONCLUSIONS: PITFALLS AND PROMISES effort in relation to the provision of formal education to providers and prescribers, complemented by the development A global CVD epidemic is rapidly evolving, and the burden of of clinical guidelines aimed at avoiding both the overuse and disease is shifting. Twice as many deaths from CVD now occur in the underuse of procedures. developing as in developed countries. The vast majority of CVD Clinical guidelines are already numerous, but all have been can be attributed to conventional risk factors. Even in Sub- established in affluent countries. A new, specific effort should Saharan Africa, high blood pressure, high cholesterol, extensive be made in developing countries to address local issues, such as tobacco and alcohol use, and low vegetable and fruit consump- problems related to the availability of procedures or drugs or to tion are already among the top risk factors for disease.Because of accessibility of services, and the development and maintenance the time lag associated with CVD risk factors, especially in chil- of these guidelines should follow best available standards. dren, the full effect of exposure to these factors will be seen only in the future. Information from more than 100 countries shows Clinical Research that more 13- to 15-year-olds smoke than ever before, and stud- ies show that obesity levels in children are increasing markedly in In most situations, health care innovations should be intro- countries as diverse as Brazil, China, India, and almost all island duced as experimental interventions to permit proper moni- states (Leeder and others 2004). Populationwide efforts now to toring and evaluation. These experiments do not have to reduce risk factors through multiple economic and educational address the efficacy of the procedure (many innovations will policies and programs will reap savings later in medical and already have been tested), but rather issues pertaining to their other direct costs as well as indirectly in terms of improved qual- effectiveness and efficiency in the specific context of developing ity of life and economic productivity. countries. Another reason for the experimental approach is the rapid- ity with which the field of CVD is evolving. It is not reasonable, REFERENCES at the local level, to wait until the publication of trial results and American Heart Association. 2002. Heart Disease and Stroke Statistics-- meta-analyses, which often takes place years after changes have 2003 Update. http://www.americanheart.org/downloadable/heart/ occurred in everyday practice. For this reason, a new culture of 10461207852142003HDSStatsBook.pdf. clinical research should be developed in which every innova- Anderson, G. M., and J. Lomas. 1988. "Monitoring the Diffusion of a Technology: Coronary Artery Bypass Surgery in Ontario." American tion should be taken as an opportunity for systematic experi- Journal of Public Health 78 (3): 251­54. mental evaluation. Antithrombotic Trialists' Collaboration. 2002. "Collaborative Meta- Among various topics in clinical research, adherence analysis of Randomised Trials of Antiplatelet Therapy for Prevention deserves special mention. On average, 50 percent of patients in of Death, Myocardial Infarction, and Stroke in High Risk Patients." developed countries do not take their prescribed medicines British Medical Journal 324 (7329): 71­86. after one year, despite having full access to medicines. In devel- Antman, E. M., D. T. Anbe, P. W. Armstrong, E. R. Bates, L. A. Green, M. Hand, and others. 2004."ACC/AHA Guidelines for the Management of oping countries, this poor adherence is made worse by poor Patients with ST-Elevation Myocardial Infarction--Executive access to health services and drugs, to lack of education, and to Summary: A Report of the American College of Cardiology/American other factors (Bovet and others 2002; WHO 2003a). Options Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 1999 Guidelines for the Management of for improving adherence should be designed and experimented Patients with Acute Myocardial Infarction)." Circulation 110 (5): with. 588­636. Cardiovascular Disease | 659 Bassili, A., S. R. Zaher, A. Zaki, M. Abdel-Fattah, and G. Tognoni. 2000. Fibrinolytic Therapy Trialists' Collaborative Group. 1994. "Indications for "Profile of Secondary Prophylaxis among Children with Rheumatic Fibrinolytic Therapy in Suspected Acute Myocardial Infarction: Heart Disease in Alexandria, Egypt." Eastern Mediterranean Health Collaborative Overview of Early Mortality and Major Morbidity Journal 6 (2­3): 437­46. Results from All Randomised Trials of More Than 1,000 Patients. Bertrand, E. 1999. "Cardiovascular Disease in Developing Countries." In Fibrinolytic Therapy Trialists' (FTT) Collaborative Group." Lancet 343 Cardiology, ed. S. Dalla Volta. New York: McGraw-Hill. (8893): 311­22. Berwick, D. M. 2003. "Disseminating Innovations in Health Care." Journal Fox, K. M. 2003. "Efficacy of Perindopril in Reduction of Cardiovascular of the American Medical Association 289 (15): 1969­75. Events among Patients with Stable Coronary Artery Disease: Randomised, Double-Blind, Placebo-Controlled, Multicentre Trial Blustein, J. 1993. "High-Technology Cardiac Procedures. The Impact of (the EUROPA Study)." Lancet 362 (9386): 782­88. Service Availability on Service Use in New York State." Journal of the American Medical Association 270 (3): 344­49. Freemantle, N., J. Cleland, P. Young, J. Mason, and J. Harrison. 1999. "Beta Blockade after Myocardial Infarction: Systematic Review and Meta Bosch, J., S. Yusuf, J. Pogue, P. Sleight, E. Lonn, B. Rangoonwala, and oth- Regression Analysis." British Medical Journal 318 (7200): 1730­37. ers. 2002. "Use of Ramipril in Preventing Stroke: Double Blind Randomised Trial." British Medical Journal 324 (7339): 699­702. Gage, B. F., A. B. Cardinalli, and D. K. Owens. 1998. "Cost-Effectiveness of Preference-Based Antithrombotic Therapy for Patients with Bovet, P., M. Burnier, G. Madeleine, B. Waeber, and F. Paccaud. 2002. Nonvalvular Atrial Fibrillation." Stroke 29 (6): 1083­91. "Monitoring One-Year Compliance to Antihypertension Medication in the Seychelles." Bulletin of the World Health Organization 80 (1): 33­39. Ghaffar, A., K. S. Reddy, and M. Singhi. 2004. "Burden of Non- communicable Diseases in South Asia." British Medical Journal 328 Cannon, C. P., E. Braunwald, C. H. McCabe, D. J. Rader, J. L. Rouleau, (7443): 807­10. R. Belder, and others. 2004. "Intensive versus Moderate Lipid Lowering GISSI (Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto with Statins after Acute Coronary Syndromes." New England Journal of Miocardico). 1986. "Effectiveness of Intravenous Thrombolytic Medicine 350 (15): 1495­504. Treatment in Acute Myocardial Infarction. Gruppo Italiano per lo Carapetis, J. R., B. J. Currie, and E. L. Kaplan. 1999. "Epidemiology and Studio della Streptochinasi nell'Infarto Miocardico (GISSI)." Lancet 1 Prevention of Group A Streptococcal Infections: Acute Respiratory (8478): 397­402. Tract Infections, Skin Infections, and Their Sequelae at the Close of the Goldman, L., S. T. Sia, E. F. Cook, J. D. Rutherford, and M. C. Weinstein. Twentieth Century." Clinical Infectious Diseases 28 (2): 205­10. 1988. "Costs and Effectiveness of Routine Therapy with Long-Term Cohen, D. J., J. A. Breall, K. K. Ho, R. M. Weintraub, R. E. Kuntz, M. C. Beta-Adrenergic Antagonists after Acute Myocardial Infarction." New Weinstein, and others. 1993. "Economics of Elective Coronary England Journal of Medicine 319 (3): 152­57. Revascularization. Comparison of Costs and Charges for Conventional Goldman, L., M. C. Weinstein, P. A. Goldman, and L. W. Williams. 1991. Angioplasty, Directional Atherectomy, Stenting, and Bypass Surgery." "Cost-Effectiveness of HMG-CoA Reductase Inhibition for Primary Journal of the American College of Cardiology 22 (4): 1052­59. and Secondary Prevention of Coronary Heart Disease." Journal of the Dagenais, G. R., S. Yusuf, M. G. Bourassa, Q. Yi, J. Bosch, E. M. Lonn, and American Medical Association 265 (9): 1145­51. others. 2001. "Effects of Ramipril on Coronary Events in High-Risk GUSTO (Global Use of Strategies to Open Occluded Coronary Arteries) Persons: Results of the Heart Outcomes Prevention Evaluation Study." Investigators. 1993. "An International Randomized Trial Comparing Circulation 104 (5): 522­26. Four Thrombolytic Strategies for Acute Myocardial Infarction." New De Luca, G., H. Suryapranata, J. P. Ottervanger, and E. M. Antman. 2004. England Journal of Medicine 329 (10): 673­82. "Time Delay to Treatment and Mortality in Primary Angioplasty for Heidenreich, P. A., K. M. McDonald, T. Hastie, B. Fadel, V. Hagan, B. K. Lee, Acute Myocardial Infarction: Every Minute of Delay Counts." and others. 1999. "Meta-analysis of Trials Comparing Beta-Blockers, Circulation 109 (10): 1223­25. Calcium Antagonists, and Nitrates for Stable Angina." Journal of the Denbow, C. E., E. E. Chung, W. Foster, H. Gist, and R. E. Vlietstra. 1997. American Medical Association 281 (20): 1927­36. "Percutaneous Transluminal Coronary Angioplasty (PTCA) in Jamaica. Hochman, J. S., L. A. Sleeper, J. G. Webb, T. A. Sanborn, H. D. White, J. D. Preliminary Results." West Indian Medical Journal 46 (4): 115­19. Talley, and others. 1999. "Early Revascularization in Acute Myocardial Digitalis Investigation Group. 1997. "The Effect of Digoxin on Mortality Infarction Complicated by Cardiogenic Shock." New England Journal of and Morbidity in Patients with Heart Failure." New England Journal of Medicine 341 (9): 625­34. Medicine 336 (8): 525­33. Hodgson, T. A., and L. Cai. 2001. "Medical Care Expenditures for Doubilet, P., B. J. McNeil, and M. C. Weinstein. 1985. "The Decision Hypertension, its Complications, and its Comorbidities." Medical Care Concerning Coronary Angiography in Patients with Chest Pain: A 39 (6): 599­615. Cost-Effectiveness Analysis." Medical Decision Making 5 (3): 293­309. Holloway, R. G., C. G. Benesch, C. R. Rahilly, and C. E. Courtright. 1999. Dravik, V. 1998. "PTCA Increase." Canadian Journal of Cardiology 14 "A Systematic Review of Cost-Effectiveness Research of Stroke (Suppl. A): 27A­31A. Evaluation and Treatment." Stroke 30 (7): 1340­49. Eagle, K. A., R. A. Guyton, R. Davidoff, G. A. Ewy, J. Fonger, T. J. Gardner, ISIS-1 (First International Study of Infarct Survival) Collaborative Group. and others. 1999. "ACC/AHA Guidelines for Coronary Artery Bypass 1986. "Randomised Trial of Intravenous Atenolol among 16,027 Cases Graft Surgery: Executive Summary and Recommendations--A Report of Suspected Acute Myocardial Infarction: ISIS-1 (First International of the American College of Cardiology/American Heart Association Study of Infarct Survival Collaborative Group)."Lancet 2 (8498): 57­66. Task Force on Practice Guidelines (Committee to Revise the 1991 ISIS-2 (Second International Study of Infarct Survival) Collaborative Guidelines for Coronary Artery Bypass Graft Surgery)." Circulation Group. 1988. "Randomised Trial of Intravenous Streptokinase, Oral 100 (13): 1464­80. Aspirin, Both, or Neither among 17,187 Cases of Suspected Acute Eckman, M. H., H. J. Levine, and S. G. Pauker. 1992. "Decision Analytic Myocardial Infarction: ISIS-2. ISIS-2 (Second International Study of and Cost-Effectiveness Issues Concerning Anticoagulant Prophylaxis Infarct Survival) Collaborative Group." Lancet 2 (8607): 349­60. in Heart Disease." Chest 102 (4 Suppl.): 538S­549S. Jermyn, B. D. 2000. "Cost-Effectiveness Analysis of a Rural/Urban First- Ephrem, D., B. Abegaz, and L. Muhe. 1990. "Profile of Cardiac Diseases in Responder Defibrillation Program." Prehospital Emergency Care 4 (1): Ethiopian Children." East African Medical Journal 67 (2): 113­17. 43­47. Evans, R. W. 1986. "Cost-Effectiveness Analysis of Transplantation." Jolliffe, J. A., K. Rees, R. S. Taylor, D. Thompson, N. Oldridge, and Surgical Clinics of North America 66 (3): 603­16. S. Ebrahim. 2000. "Exercise-Based Rehabilitation for Coronary 660 | Disease Control Priorities in Developing Countries | Thomas A. Gaziano, K. Srinath Reddy, Fred Paccaud, and others Heart Disease." Cochrane Database of Systematic Reviews (4) McKelvie, R. 2003. "Heart Failure." Clinical Evidence 9: 95­118. CD001800. McMurray, J. J., and S. Stewart. 2000. "Heart Failure: Epidemiology, Joorabchi, B. 1979. "The Emergence of Cardiac Nondisease among Aetiology, and Prognosis of Heart Failure." Heart 83 (5): 596­602. Children in Iran." Israel Journal of Medical Sciences 15 (3): 202­6. Murray, C. J., and A. D. Lopez. 1994. Global Comparative Assessments in the Kaplan, E. L. 1985. "Epidemiological Approaches to Understanding the Health Sector: Disease Burden, Expenditures, and Intervention Packages. Pathogenesis of Rheumatic Fever." International Journal of Geneva: World Health Organization. Epidemiology 14 (4): 499­501. ------. 1996. Global Burden of Disease and Injury Series, Vols. I and II, King, H., R. E. Aubert, and W. H. Herman. 1998. "Global Burden of Global Health Statistics. Boston: Harvard School of Public Health. Diabetes, 1995­2025: Prevalence, Numerical Estimates, and ------. 1997. "Mortality by Cause for Eight Regions of the World: Global Projections." Diabetes Care 21 (9): 1414­31. Burden of Disease Study." Lancet 349 (9061): 1269­76. Knatterud, G. L., Y. Rosenberg, L. Campeau, N. L. Geller, D. B. Musaiger, A. O. 2002. "Diet and Prevention of Coronary Heart Disease in Hunninghake, S. A. Forman, and others. 2000. "Long-Term Effects on the Arab Middle East Countries." Medical Principles and Practice 11 Clinical Outcomes of Aggressive Lowering of Low-Density (Suppl. 2): 9­16. Lipoprotein Cholesterol Levels and Low-Dose Anticoagulation in the Post Coronary Artery Bypass Graft Trial: Post CABG Investigators." Nissen, S. E., E. M. Tuzcu, P. Libby, P. D. Thompson, M. Ghali, D. Garza, Circulation 102 (2): 157­65. and others. 2004."Effect of Antihypertensive Agents on Cardiovascular Events in Patients with Coronary Disease and Normal Blood Pressure: Koegelenberg, C. F., A. F. Doubell, H. Orth, and H. Reuter. 2003. "Infective The CAMELOT Study: A Randomized Controlled Trial." Journal of the Endocarditis in the Western Cape Province of South Africa: A Three- American Medical Association 292 (18): 2217­25. Year Prospective Study." QJM 96 (3): 217­25. Olshansky, S. J., and A. B. Ault. 1986. "The Fourth Stage of the Kuntz, K. M., and K. C. Kent. 1996. "Is Carotid Endarterectomy Cost- Epidemiologic Transition: The Age of Delayed Degenerative Diseases." Effective? An Analysis of Symptomatic and Asymptomatic Patients." Milbank Memorial Fund Quarterly 64: 355­91. Circulation 94 (9 Suppl.): II194­98. Omran, A. R. 1971. "The Epidemiologic Transition: A Theory of the Kuntz, K. M., J. Tsevat, L. Goldman, and M. C. Weinstein. 1996. "Cost- Epidemiology of Population Change." Milbank Memorial Fund Effectiveness of Routine Coronary Angiography after Acute Quarterly 49: 509. Myocardial Infarction." Circulation 94 (5): 957­65. Ornato, J. P., E. J. Craren, E. R. Gonzalez, A. R. Garnett, B. K. McClung, and Kupersmith, J., M. Holmes-Rovner, A. Hogan, D. Rovner, and J. Gardiner. M. M. Newman. 1988. "Cost-Effectiveness of Defibrillation by 1995. "Cost-Effectiveness Analysis in Heart Disease, Part III: Ischemia, Emergency Medical Technicians." American Journal of Emergency Congestive Heart Failure, and Arrhythmias." Progress in Cardiovascular Medicine 6 (2): 108­12. Diseases 37 (5): 30­46. Parmley, W. W. 1999. "Cost-Effectiveness of Reperfusion Strategies." Kuppermann, M., B. R. Luce, B. McGovern, P. J. Podrid, J. T. Bigger Jr., and American Heart Journal 138 (2, part 2): S142­52. J. N. Ruskin. 1990. "An Analysis of the Cost Effectiveness of the Pestana, J. A., K. Steyn, A. Leiman, and G. M. Hartzenberg. 1996. "The Implantable Defibrillator." Circulation 81 (1): 91­100. Direct and Indirect Costs of Cardiovascular Disease in South Africa in Leeder, S., S. Raymond, H. Greenberg, H. Liu, and K. Esson. 2004. A Race 1991." South African Medical Journal 86 (6): 679­84. against Time: The Challenge of Cardiovascular Disease in Developing Pfeffer, M. A., A. Keech, F. M. Sacks, S. M. Cobbe, A. Tonkin, R. P. Byington, Countries. New York: Trustees of Columbia University. and others. 2002. "Safety and Tolerability of Pravastatin in Long-Term Lorenzoni, R., D. Pagano, G. Mazzotta, S. D. Rosen, G. Fattore, R. De Clinical Trials: Prospective Pravastatin Pooling (PPP) Project." Caterina, and others. 1998. "Pitfalls in the Economic Ealuation of Circulation 105 (20): 2341­46. Thrombolysis in Myocardial Infarction: The Impact of National Diff- Pitt, B., F. Zannad, W. J. Remme, R. Cody, A. Castaigne, A. Perez, and erences in the Cost of Thrombolytics and of Differences in the Efficacy others. 1999. "The Effect of Spironolactone on Morbidity and across Patient Subgroups." European Heart Journal 19 (10): 1518­24. Mortality in Patients with Severe Heart Failure." New England Journal Mackay, J., and G. A. Manesh. 2004. The Atlas of Heart Disease and Stroke. of Medicine 341 (10): 709­17. Geneva: WHO. Roberts, R., W. J. Rogers, H. S. Mueller, C. T. Lambrew, D. J. Diver, H. C. Majeed, H. A., L. al-Doussary, M. M. Moussa, A. R. Yusuf, and A. H. Smith, and others. 1991. "Immediate versus Deferred Beta-Blockade Suliman. 1993. "Office Diagnosis and Management of Group A Following Thrombolytic Therapy in Patients with Acute Myocardial Streptococcal Pharyngitis Employing the Rapid Antigen Detecting Infarction: Results of the Thrombolysis in Myocardial Infarction Test: A 1-Year Prospective Study of Reliability and Cost in Primary (TIMI) II-B Study." Circulation 83 (2): 422­37. Care Centres." Annals of Tropical Paediatrics 13 (1): 65­72. Rogers, W. J., J. D. Babb, D. S. Baim, J. H. Chesebro, J. M. Gore, R. Roberts, Matchar, D., J. Pauk, and J. Lipscomb. 1996. "A Health Policy Perspective and others. 1991. "Selective versus Routine Predischarge Coronary on Carotid Endarterectomy: Cost, Effectiveness, and Cost- Arteriography after Therapy with Recombinant Tissue-Type Effectiveness." In Surgery for Cerebrovascular Disease, 2nd ed., ed. Plasminogen Activator, Heparin, and Aspirin for Acute Myocardial W. Moore. Philadelphia: W. B. Saunders. Infarction: TIMI II Investigators." Journal of the American College of Mathers, C. D., A. D. Lopez, and C. J. L. Murray. "The Burden of Disease Cardiology 17 (5): 1007­16. and Mortality by Condition: Data, Methods, and Results for 2001." In Rowley, J. M., C. Garner, and J. R. Hampton. 1990. "The Limited Potential Global Burden of Disease and Risk Factors, eds. A. D. Lopez, C. D. of Special Ambulance Services in the Management of Cardiac Arrest." Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. New York: British Heart Journal 64 (5): 309­12. Oxford University Press. Schneider, J., and K. Bezabih. 2001. "Causes of Sudden Death in Addis McClellan, M., and D. Kessler. 1999. "A Global Analysis of Technological Ababa, Ethiopia." Ethiopian Medical Journal 39 (4): 323­40. Change in Health Care: The Case of Heart Attacks--The TECH Soumerai, S. B., T. J. McLaughlin, D. Spiegelman, E. Hertzmark, Investigators." Health Affairs 18 (3): 250­55. G. Thibault, and L. Goldman. 1997."Adverse Outcomes of Underuse of McFayden, J. E., ed. 2003. International Drug Price Indicator Reference Beta-Blockers in Elderly Survivors of Acute Myocardial Infarction." Guide. Boston: Management Sciences for Health. Journal of the American Medical Association 277 (2): 115­21. Cardiovascular Disease | 661 Steer, A. C., J. R. Carapetis, T. M. Nolan, and F. Shann. 2002. "Systematic ------. 2002a. Integrated Management of Cardiovascular Risk. Geneva: Review of Rheumatic Heart Disease Prevalence in Children in WHO CVD Program. Developing Countries: The Role of Environmental Factors." Journal of ------. 2002b. The World Health Report 2002: Reducing Risks, Promoting Paediatrics and Child Health 38 (3): 229­34. Healthy Life. Geneva: WHO. Strasser, T. 1985. "Cost-Effective Control of Rheumatic Fever in the ------. 2003a. "Adherence to Long-Term Therapies: Evidence for Action." Community." Health Policy 5 (2): 159­64. WHO, Geneva. http://www.who.int/chronic_conditions/adherence_ Tengs, T. O., M. E. Adams, J. S. Pliskin, D. G. Safran, J. E. Siegel, M. C. report.pdf. Weinstein, and others. 1995. "Five-Hundred Life-Saving Interventions ------. 2003b. World Health Report 2003: Shaping the Future. Geneva: and Their Cost-Effectiveness." Risk Analysis 15 (3): 369­90. WHO. Teo, K. K., S. Yusuf, M. Pfeffer, C. Torp-Pedersen, L. Kober, A. Hall, and Williams, A. 1985. "Economics of Coronary Artery Bypass Grafting." others. 2002. "Effects of Long-Term Treatment with Angiotensin- British Medical Journal 291 (6491): 326­29. Converting-Enzyme Inhibitors in the Presence or Absence of Aspirin: Williams, D. O. 2004. "Treatment Delayed Is Treatment Denied." A Systematic Review." Lancet 360 (9339): 1037­43. Circulation 109 (15): 1806­8. Topol, E. J., D. R. Holmes, and W. J. Rogers. 1991. "Coronary Angiography Wolf, P. A., R. D. Abbott, and W. B. Kannel. 1991. "Atrial Fibrillation as an after Thrombolytic Therapy for Acute Myocardial Infarction." Annals Independent Risk Factor for Stroke: The Framingham Study." Stroke 22 of Internal Medicine 114 (10): 877­85. (8): 983­88. Unger, F. 1999. "Cardiac Interventions in Europe 1997: Coronary Wong, J. B., F. A. Sonnenberg, D. N. Salem, and S. G. Pauker. 1990. Revascularization Procedures and Open Heart Surgery." Cor "Myocardial Revascularization for Chronic Stable Angina. Analysis of Europaeum 7: 177­89. the Role of Percutaneous Transluminal Coronary Angioplasty Based Weinstein, M. C., and W. B. Stason. 1982. "Cost-Effectiveness of Coronary on Data Available in 1989." Annals of Internal Medicine 113 (11): Artery Bypass Surgery." Circulation 66 (5, part 2): III56­66. 852­71. WHO (World Health Organization). 1995. "Strategy for Controlling Yusuf, S., R. Peto, J. Lewis, R. Collins, and P. Sleight. 1985. "Beta Blockade Rheumatic Fever/Rheumatic Heart Disease, with Emphasis on during and after Myocardial Infarction: An Overview of the Primary Prevention." Bulletin of the World Health Organization 73 (5): Randomized Trials." Progress in Cardiovascular Diseases 27 (5): 335­71. 583­87. 662 | Disease Control Priorities in Developing Countries | Thomas A. Gaziano, K. Srinath Reddy, Fred Paccaud, and others Chapter 34 Inherited Disorders of Hemoglobin David Weatherall, Olu Akinyanju, Suthat Fucharoen, Nancy Olivieri, and Philip Musgrove As a result of carrier protection against malaria, the inherited human development. In adults two components exist: a major hemoglobin disorders are the commonest diseases attributable hemoglobin, Hb A, and a minor hemoglobin, Hb A2. The bulk to single defective genes. Approximately 7 percent of the of the hemoglobin during later fetal life is Hb F. These hemo- world's population is a carrier, and 300,000 to 500,000 babies globins each consist of two pairs of unlike globin chains. The with severe forms of such disorders are born each year (WHO adult hemoglobins and fetal hemoglobin have chains com- 1989). Although these disorders are most frequent in tropical bined with (Hb A, ), ), or chains (Hb F, 2 2 (Hb A2, 2 2 regions, they are now encountered in most countries because of ). Each of the different globin chains is controlled by dis- 2 2 migrations of populations. tinct genes; two genes exist for the and chains and one for each of the other chains. Their structure and the regions of the genes that control the production of the different globin chains INTRODUCTION have been determined (Steinberg and others 2001; Weatherall and Clegg 2001b). If untreated, many of the inherited hemoglobin disorders result in death during the first few years of life. Their effect on the burden of disease has only recently been recognized, following Spectrum of Inherited Hemoglobin Disorders an epidemiological transition caused by improvements in Inherited hemoglobin disorders fall into two main groups: the hygiene, nutrition, and control of infection that has reduced structural hemoglobin variants and the thalassemias, which are childhood mortality. Babies with severe hemoglobin disorders caused by defective globin production. They all follow a reces- are now able to survive long enough to present for diagnosis sive form of inheritance. Those with a single defective globin and treatment. The impact of these diseases is being felt gene--carriers or heterozygotes--are symptomless. If two car- throughout the Indian subcontinent and much of Asia. riers marry, a one in four chance exists that each child they Although the situation will worsen in Sub-Saharan Africa as produce will receive defective genes from each parent--that is, it undergoes a similar transition, such diseases are already they are homozygous for the particular disorder. responsible for a major health burden. International health The structural variants result mostly from single amino acid agencies and the governments of affected countries need to substitutions in the or chains. Often these are innocuous, understand the future extent of the problem and to develop but in some cases they may alter the stability or functional programs to control and manage these diseases. properties of the hemoglobin and lead to a clinical disorder. They are designated by letters of the alphabet or by the place Normal Hemoglobin names where the condition was first discovered. Even though Hemoglobin (Hb), the pigment in the red blood cells that researchers have identified more than 700 structural hemoglo- transfers oxygen to the tissues, changes structure during bin variants, only three (Hb S, Hb C, and Hb E) are widespread. 663 The homozygous state for the sickle cell gene results in sickle Platt 1998). Even though Hb SC disease is milder than sickle cell anemia, whereas the compound heterozygous state for the cell anemia, it is associated with many complications, including sickle cell and Hb C genes results in Hb SC disease. Hb SC dis- a higher frequency of proliferative retinopathy. ease, although milder, also has important public health impli- cations. Hb E, the commonest variant globally, is innocuous in Thalassemias. The homozygous or compound heterozygous its heterozygous and homozygous states, but because it is syn- states for thalassemia also run a variable course, although thesized less effectively than Hb A, it interacts with tha- without transfusion, death usually occurs in the first few years lassemia to produce an extremely common condition called Hb (Weatherall and Clegg 2001b). With adequate transfusions and E thalassemia, which is becoming an increasingly important the administration of drugs to remove iron, children may health burden in many parts of Asia. develop well and survive to adulthood. However, these drugs The thalassemias are classified according to the ineffectively are expensive, and even when they are available in poorer coun- synthesized globin chains. From a public health viewpoint, tries, many children receive inadequate dosages and die in only the and thalassemias are sufficiently common to be childhood or adolescence from iron overload. The situation is important. further complicated because the common thalassemias of intermediate severity--notably Hb E thalassemia--exhibit a Clinical Features clinical spectrum ranging from transfusion-dependent disease The inherited hemoglobin disorders are characterized by an to a condition compatible with normal survival and growth extremely diverse series of clinical syndromes of varying into adult life without treatment. severity. The thalassemias are equally heterogeneous. The extremely common milder forms (termed thalassemias Sickle Cell Anemia and Related Disorders. The clinical fea- because some chains are produced) produce only a mild o tures of sickle cell disorders reflect the red blood cells' propen- hypochromic anemia in homozygotes. In contrast, the tha- sity to assume a sickle shape in deoxygenated blood, leading to lassemias, so called because of the absence of chain synthesis, shortened red cell survival and a tendency to block small blood result in stillbirth in their homozygous states following preg- vessels (Bunn 1997; Serjeant 1992). Even though patients may nancies with toxemic and postpartum complications. The o adapt to their anemia, their illness is interspersed with acute compound heterozygous states for and thalassemias episodes, including: attacks of bone pain; sequestration of result in Hb H disease, which varies in severity and may be blood into the lungs, liver, or spleen; or thrombosis of cerebral transfusion dependent. vessels, which may cause a stroke. They are extremely prone to The thalassemias are extremely heterogeneous at the molec- infection, particularly during early childhood, and to a wide ular level: more than 200 different mutations of the globin range of chronic complications. For reasons not yet under- genes have been found, and the thalassemias are almost as stood, the severity of the disease varies extensively. Even in varied. Every severely affected population in the world has a populations in eastern Saudi Arabia and parts of India, which few common mutations unique to a particular region, together have a high frequency of thalassemia and an unusual ability with varying numbers of rare ones. to produce Hb F in adult life, both of which, when inherited with sickle cell disease, result in a milder form of the illness, Population Genetics and Dynamics morbidity is still high. The high gene frequencies for the hemoglobin disorders are Although little is known about mortality from "sickling" attributable to the effects of natural selection. Although disorders in developing countries, in Sub-Saharan Africa many severely affected homozygotes would, in the absence of medical children die early because of these conditions (Akinyanju 2001; interventions, have died early in life, asymptomatic heterozy- Fleming and others 1979). Fleming and others, working in gotes for Hb S, Hb C, and probably thalassemia and Hb E, as rural Nigeria, found that even though more than 2 percent of well as those with mild forms of thalassemia, are more resist- all newborns had sickle cell anemia, it was absent in the ado- ant to severe malarial infection than normal persons. Hence, in lescent and adult populations. At the same time, they found environments in which malaria was common, carriers were that urban centers in Nigeria, where medical care was available, protected and survived to have more children, and the gene fre- had an increasing number of affected adults, and by the late quencies rose until they were balanced by loss of severely 1970s, a significant improvement in survival had clearly fol- affected homozygotes from the population. Although some lowed the introduction of antimalarial measures (Molineaux decline in frequency among immigrant populations may occur and others 1979). Both in Jamaica and in the United States, because of lack of exposure to malaria and outbreeding, this death appears to peak between one and three years of age, usu- decline will occur over many generations, and even if malaria ally from infection. Recent U.S. data suggest that the median were completely eradicated, an equally long time would pass age of adult death is 42 for men and 48 for women (Dover and before any significant fall occurred in the global frequency. 664 | Disease Control Priorities in Developing Countries | David Weatherall, Olu Akinyanju, Suthat Fucharoen, and others Changes resulting from variation in selection or in popula- The gene for Hb S is distributed throughout Sub-Saharan tion dynamics will, however, be small compared with the effect Africa, the Indian subcontinent, and the Middle East, where car- of the demographic and epidemiological transitions that many rier frequencies range from 5 to 40 percent or more. Hb E countries have recently undergone. For example, thalassemia is found in the eastern half of the Indian subcontinent and was not identified in Cyprus until 1944, when major improve- throughout Southeast Asia, where carrier rates may exceed ments in public health revealed that the disease was common. By 60 percent. Thalassemia is frequent in a broad band from the the early 1970s, estimates indicated that, in the absence of steps Mediterranean basin and parts of Africa, throughout the Middle to control the disease, in about 40 years approximately 78,000 East, the Indian subcontinent, Southeast Asia, and Melanesia units of blood would be required each year to treat all the and into the Pacific islands. The thalassemias occur right severely affected children, 40 percent of the population would be across the tropical zone, reaching extremely high frequencies in carriers, and the cost to the health system would equal or exceed some populations, whereas the o thalassemias are restricted to the island's health budget (Weatherall and Clegg 2001b). parts of SoutheastAsia and the Mediterranean basin (table 34.1). Several World Health Organization (WHO) workshops Global Distribution and Frequency have attempted to estimate the global burden of the thal- of the Hemoglobinopathies assemias and important structural hemoglobin variants Figures 34.1a and 34.1b show the global distributions of the (Angastiniotis and Modell 1998; Weatherall and Clegg 2001b; hemoglobinopathies. Table 34.1 shows approximate carrier fre- WHO 1989, 1994). There are perhaps 270 million carriers and quencies by region. 300,000 to 500,000 annual births of infants with sickle cell Hb E Hb S Source: Weatherall and Clegg 2001b. Note: Structural hemoglobin variants are Hb E (innocuous unless interacting with or thalassemia) and Hb S (causing sickle-cell disease in the homozygous state). Figure 34.1a Global Distribution of Hemoglobinopathies Hb E and Hb S Inherited Disorders of Hemoglobin | 665 and thalassemias Source: Weatherall and Clegg 2001b. Figure 34.1b Global Distribution of Hemoglobinopathies and Thalassemias Table 34.1 Carrier Frequencies for Common Hemoglobin Disorders, by World Health Organization Region, 2001 (percent) Region Hb S Hb C Hb E thalassemia othalassemia thalassemia Americas 1­20 0­10 0­20 0­3 0­5 0­40 Eastern Mediterranean 0­60 0­3 0­2 2­18 0­2 1­60 Europe 0­30 0­5 0­20 0­19 1­2 0­12 Southeast Asia 0­40 0 0­70 0­11 1­30 3­40 Sub-Saharan Africa 1­38 0­21 0 0­12 0 10­50 Western Pacific 0 0 0 0­13 0 2­60 Sources: Livingstone 1985; Weatherall and Clegg 2001a, 2001b. Note: Many of these data are derived from small population samples. anemia or serious forms of thalassemia. Southeast Asia, where summarize information about the different thalassemia muta- the thalassemias and Hb E predominate, is most severely affect- tions in those regions. ed. Sub-Saharan Africa has the second-highest burden, reflect- These data only approximate the problems for health care ing the high incidence of Hb S. Weatherall and Clegg (2001b) services that the hemoglobin disorders will pose in the future. 666 | Disease Control Priorities in Developing Countries | David Weatherall, Olu Akinyanju, Suthat Fucharoen, and others 1,600 5,000 3,000 5,000 5,500 3,000 2,000 8,000 Sickle cell 216,000 6,000 4,000 Source: Weatherall and Clegg 2001b. Note: These figures give only a broad approximation of the numbers of new births with the hemoglobinopathies, based as they are in many cases on data from single centers. Figures not identified as sickle cell represent births of babies with thalassemia. Figure 34.2 Approximate Annual Births of Babies with Sickle Cell Anemia and Thalassemia Unfortunately, few of the data are based on micromapping assessing the burden that the disorders will impose on health of incidence in different populations. Weatherall and Clegg's services is difficult. As more babies survive and present for (2001b) review of studies in Indonesia, Sri Lanka, and Thailand treatment, the population on long-term therapy will steadily reveals the extent of variability of incidence within relatively expand; the more effective the treatment, the greater the bur- short geographic distances, suggesting that the number of den will be on health services. For example, from 2005 to annual births of babies with thalassemia major or Hb E 2025, an estimated 100,000 cases of Hb E thalassemia will thalassemia may be underestimated. Similarly, published data be added to the Thai population, and 20,000 thalassemia for the annual births of babies with sickle cell anemia in India homozygotes will be born each year in southern China and the Middle East are almost certainly too low, because esti- (Weatherall and Clegg 2001b). If these children all survive to mates based on gene frequency suggest that the figure may be adulthood, they will account for a large proportion of health close to 100,000. The data in table 34.1 and figure 34.2, there- service expenditure. fore, represent a minimal estimate of the future likely health burden resulting from inherited hemoglobin disorders. BURDEN OF DISEASE Furthermore, in many cases, the data are not based on pro- jected increases in birth rates. WHO disease burden estimates do not include the incidence or Because of these uncertainties, including how long coun- prevalence of the hemoglobin disorders, nor the deaths or tries will take to pass through the epidemiological transition, disability-adjusted life year (DALY) losses from sickle cell disease Inherited Disorders of Hemoglobin | 667 or thalassemia. Neither do they treat these disorders as risk per year and possibly as many as 173,000. These translate into factors for anemia, infection, stroke, and other conditions or 0.5 million to 4.5 million DALYs, accounting for less than estimate the prevalence (frequency) of the underlying genetic 1 percent, but perhaps as much as 2 percent, of the burden for factors. Thus, the estimates provided here are necessarily incom- children under five. Life expectancies and the extent of disabil- plete and speculative. ities among survivors in Africa are unknown, so the low DALY For severe thalassemia, figure 34.2 suggests 43,100 births number is no doubt underestimated. per year, nearly all in low- and middle-income countries, where Outside Africa, Weatherall and Clegg (2001b) estimate affected babies are likely to die before reaching two years of age. 60,000 sickle cell births per year concentrated in India and At least 41,500 deaths probably occur each year, or 0.3 percent of the Middle East and among descendants of Africans in the all deaths of children under five. This estimate may be too low, Americas. The actual figure may be as high as 100,000. Without because it does not include the estimated 20,000 births per year treatment, deaths peak in the first 2 years of life, and half of all in China. Thus, the severe thalassemias probably account for deaths occur in the first 20 years. If 25 percent of sufferers die 50,000 to 100,000 deaths per year, or 0.5 to 0.9 percent of all at age 1 and 25 percent at age 10, those deaths would contribute deaths of children under five in low- and middle-income coun- almost 14,000 DALYs for every 1,000 births in a low- or middle- tries. Each death accounts for 29.2 DALYs if it occurs before the income country. Including deaths after age 20 and disability child reaches the age of one. Taken together, all the deaths con- might double the estimate. tribute 1.46 million to 2.92 million DALYs to the world burden. Survival elsewhere is greater than in Africa, because of lower Treated thalassemia victims who survive to age 40 or older risks of infection and greater access to treatment. The United contribute much less to the disease burden because they are Kingdom has about 10,000 survivors (Davies and others 2000), fewer and their residual disability weight is only 0.02 to 0.10 and the United States has some 50,000 (Ashley-Koch, Yang, and (chapter 15 provides an explanation of disability weights). Olney 2000). No good estimates are available of the numbers or Living with poorly treated thalassemia has a weight equal to or age distribution of survivors in most of the rest of the world, greater than 0.1. No global estimates of the number of treated but Hambleton's (2004b) cohort study in Jamaica shows how survivors are available, but estimates indicate that 500,000 may treatment increased survival: 70 percent of those enrolled exist in Thailand alone, of which perhaps 55,000 are transfu- starting in 1973 survived to age 20, as did 80 percent of those sion dependent with severe disabilities. Their total DALY loss, enrolled three to six years later. including disability for those with milder Hb H disease, would Of 1,000 babies born with sickle cell disease, Jamaican clinic be only some 15,000 per year, trivial relative to the DALYs records and follow-up show how many would die at each age in resulting from premature mortality. Deaths by age 10 from each year, allowing an estimate of the burden from premature homozygous thalassemia or by age 30 from Hb E tha- mortality (Hambleton 2004a). Table 34.2 presents those results: lassemia would add 53,600 DALYs in Thailand. 560 deaths per year represent almost 14,000 DALYs. Deaths o thalassemia contributes to the burden of disease primar- after age 50 contribute less because they are fewer, and life ily through stillbirths or deaths shortly following birth and sec- expectancy and DALYs per death decline with age. Thus, 18,000 ondarily through mothers' disability during pregnancy. WHO to 22,000 DALYs per year for deaths at all ages is a reasonable does not count stillbirths, and no data on affected births are estimate of the mortality burden from 1,000 sickle cell births available except for an estimate of 1,250 per year in Thailand, per year at Jamaican levels of treatment coverage and effective- which adds 37,242 DALYs. Assuming that mothers suffer a dis- ness. Applied to the estimated 60,000 to 100,000 births per year ability weight of 0.3 during the last trimester would add only outside Africa, this figure implies at least 1.08 to 2.20 million 100 DALYs. Every 1,000 homozygous o thalassemia pregnan- DALYs, or 0.1 percent of the total burden in low- and middle- cies contribute about 30,000 DALYs, but insufficient informa- income countries. tion is available on incidence elsewhere to use the Thai estimate Three sources of disability also contribute to the burden: ane- to project global or regional levels. mia without painful crises or other complications; disability For sickle cell disease, the burden is harder to estimate from mild or severely painful crises; and other clinical events, because of the higher survival rate and the disability during both acute and chronic (for example,leg ulcers and retinopathy). crises. Figure 34.2 shows an estimated 216,000 births per year For the first source, the disability weight is assumed to aver- in Africa alone, but reliable data on survival are not available. age 0.04. This source adds a constant 0.04 DALYs for every year Early studies suggested a mortality rate greater than 80 percent a sickle cell patient survives. The loss per year per 1,000 births by age five, but more recent estimates indicate that the figure is in Jamaica is multiplied by 2.5 for deaths during each five-year probably greater than 50 percent, with the improvement result- interval (because deaths are assumed to occur at the midpoint ing from treatment and from control of the infections that of the interval) and by 5.0 for survivors, who suffer disability cause most early sickle cell deaths. Mortality of 50 to 80 percent for the entire five years. This loss adds about 10 percent to the at ages one to five implies at least 21,600 to 34,500 deaths loss from premature mortality. 668 | Disease Control Priorities in Developing Countries | David Weatherall, Olu Akinyanju, Suthat Fucharoen, and others Table 34.2 Burden of Sickle Cell Disease by Age Group, Assuming 1,000 Births per Year and Survival to Various Ages, Jamaica, Starting in 1973 Age group (years) Total or Category 0­4 5­9 10­14 15­19 20­24 25­29 30­34 35­39 40­44 45­49 average Number of survivors 876 834 807 777 727 680 627 564 491 440 682.3 Number of deaths 124 42 27 30 50 47 53 63 73 51 560 Death rate (percent/year) 2.61 0.98 0.66 0.75 1.32 1.33 1.61 2.10 2.73 2.17 1.63 Number of DALYs lost/death 28.90 28.59 27.77 26.84 25.82 24.69 23.43 22.00 20.39 18.58 24.70 Total DALY losses from deaths 3,584 1,201 750 805 1,291 1,161 1,242 1,386 1,488 948 13,856 Number of DALYs lost from 188 171 164 158 150 141 130 119 106 93 1,420 background (chronic) anemia Total DALYs lost from deaths and 3,772 1,372 914 963 1,441 1,302 1,372 1,505 1,594 1,041 15,276 chronic anemia Number of pain crises/year 242.7 381.0 383.8 584.4 866.7 600.5 523.6 473.4 309.6 182.2 4,548 Number of other acute clinical 77.5 22.2 182.2 281.9 events Number of other chronic clinical 49.8 14.8 12.8 10.9 88.3 events Source: Authors' calculations based on Hambleton 2004a, 2004b. Note: The number of DALYs lost per death is calculated assuming all deaths occur at the midpoint of the age interval. Life expectancies by age are those for all low- and middle-income countries together. The number of pain crises is the total during one year for all the individuals in an age group. Those who die during the interval are assumed to die in equal numbers at the midpoint of each of the five years and, therefore, to suffer half as much disability from chronic anemia and half as many pain crises in that year as those of the same age who survive the year. Blanks in the table indicate that clinical events are rare in the corresponding age groups and their numbers are not well recorded. The totals of acute and chronic clinical events are therefore probably slight under-estimates. For the second source of disability, even mild crises should be hemoglobin. Even though research directed at their correction weighted considerably worse than background anemia, and by means of somatic cell gene therapy is ongoing, this technol- severe crises requiring hospitalization should be weighted worse ogy will probably not be generally applicable for some time, and still: values of 0.2 and 0.5, respectively, are assumed. The num- when it is developed, it may be extremely expensive. Thus, for ber of crises and the share that are severe vary with age,with peak the moment, the major approaches to the control and manage- severity at ages 21 to 45 for a mean disability weight of 0.35. ment of these diseases are population screening, genetic coun- Because each crisis lasts only 7 to 10 days,or about 0.023 year,the seling and prenatal diagnosis, and management of symptoms. loss per 1,000 births in each five-year age group never exceeds 10 DALYs and makes no difference to the total burden. The third source of disability may carry disability weights of Prevention 0.135 for a leg ulcer, 0.276 for stroke survivors, 0.279 for acute Programs to reduce the number of seriously affected individu- chest syndrome, and 0.567 for retinopathy, but disability als follow two approaches. First, population screening and weights for a number of other conditions are unknown. Even if counseling programs can be established to educate populations acute conditions last one month with an average disability about the risks of having children with similarly affected weight of 0.5, they add less than 10 DALYs in any five-year partners. Data about the effectiveness of this approach are interval. These conditions occur much less frequently than extremely limited. In an early study in Greece, people's knowl- painful crises, but chronic ones may last much longer, con- edge of their genetic makeup had no effect on marriage pat- tributing more or less to the burden than pain crises but adding terns (Stamatoyannopoulos 1973); however, a recent study in little to mortality losses. Table 34.2 therefore includes only the the Islamic Republic of Iran found that about 50 percent of estimated losses from background anemia and the frequencies, affected couples decided to separate, and births with severe but not the DALY losses, of painful crises and other events that thalassemia fell to about 30 percent of those expected (Samavat add negligibly to the burden. and Modell 2004). The reasons for this remarkable discrepancy require further investigation. CONTROL AND TREATMENT The second preventive approach also involves population screening or screening in prenatal clinics. If women are found With the exception of the few patients who can obtain a bone to be carriers, their partners are screened, and following coun- marrow transplant, no cure exists for the inherited disorders of seling they are offered a prenatal diagnosis and termination of Inherited Disorders of Hemoglobin | 669 affected fetuses. This method has been used widely in the transfusions are required. If so, babies and children require Mediterranean region and elsewhere, resulting in a major transfusion at monthly intervals using washed red cells reduction in newborns with serious forms of thalassemia (Cao, rather than whole blood. In addition, blood must be screened Galanello, and Rosatelli 1998; Weatherall and Clegg 2001b). for hepatitis B and C, for HIV, and--in some countries--for Prenatal diagnosis programs are available in China, India, the malaria. Because patients accumulate iron from transfusions, Islamic Republic of Iran, Lebanon, Pakistan, Singapore, they also require lifelong treatment with a chelating agent, the Thailand, the United Arab Emirates, the United States, and most effective being desferrioxamine delivered subcutaneously many European countries; several other countries are estab- overnight using a pump. Oral chelating agents, which would lishing similar programs. undoubtedly improve compliance, are available, but their Because severe thalassemia is incompatible with survival efficacy and safety have yet to be verified. Some children with the without regular blood transfusions, prenatal diagnosis is a logi- major form of thalassemia--and many with the intermediate cal approach where acceptable until more definitive treatments varieties--will at some point require splenectomy, to be pre- become available. The situation with sickle cell anemia is differ- ceded by appropriate vaccinations and followed by prophylactic ent. First, it is not uniformly fatal in early life, and morbidity and penicillin. They also require regular assessments of their iron mortality during this period can be controlled. Second, the clin- status by measurements of serum ferritin or, better, by hepatic ical course of the condition is unpredictable: some patients' iron concentrations. Various complications occur, particularly symptoms are relatively mild, whereas others develop life- for those not adequately transfused, including endocrine defi- threatening complications. Hence, even though some countries ciencies, bone disease, and infection. Bloodborne infections, are practicing prenatal diagnosis, in others neither communities particularly hepatitis C and HIV/AIDS, are an increasing haz- nor doctors consider that it should be applied widely. This com- ard. Most children with thalassemia also require regular folate plex issue would be clarified if the reasons for the phenotypic supplementation and vitamin C with their chelation therapy. variability of the sickling disorders were better defined. The serious forms of thalassemia, o thalassemia, cause Whether or not screening programs are backed up with an stillbirth late in pregnancy and several maternal complications. offer of prenatal diagnosis, they require an intense period of Although some infants have been given exchange transfusion or education of the population about the nature of inherited transfusion in the immediate neonatal period and survived,they hemoglobin disorders. This education requires input from have gone on to a transfusion-dependent life. Because of the many sectors of society, including the media, public health increased risk of congenital malformations as a result of the workers, local volunteer societies, and the medical community. disease--and particularly because of maternal complications-- Programs of this type require careful planning as well as this course of action is not recommended, and this disease is an availability of facilities for screening and counseling when the important indication for prenatal diagnosis. Those who inherit program is initiated. Their development also requires prior dis- o thalassemia from one parent and thalassemia from the cussion between the government, health care workers, and other have a moderately severe form called Hb H disease that members of the community--including religious leaders. is usually compatible with a life independent of transfusions except for periods of stress, such as infection. The tha- lassemias cause no clinical problems,either in their homozygous Treatment or heterozygous states. The treatment of the hemoglobinopathies varies widely Patients with sickle cell anemia are at high risk from infec- depending on the disease. The severe forms of thalassemia tion early in life; therefore, diagnosis as early as possible is vital. require lifelong blood transfusions. The family of thalassemia Excellent evidence, at least in developed countries, indicates intermediate diseases ranges from transfusion-dependent that prophylactic penicillin significantly reduces early morbid- forms to symptomless carrier traits. Hb E thalassemia, the ity and mortality. commonest hemoglobinopathy in Asia, varies in severity from Even though many children adapt well to their anemia, forms that require regular or intermittent transfusions to many eventually develop sickle cell crises (Ballas 1998). The milder anemia that does not require lifelong transfusions most common form, the so-called painful crisis, is character- unless complications arise, particularly hypersplenism. ized by severe bone pain that often requires hospital admission Weatherall and Clegg's (2001b) review of studies in Asia and treatment with analgesics, oxygen, and infection control. indicates that the medical community does not always appreci- More life-threatening crises, including stroke, marrow aplasia ate these subtleties and that many patients who receive regular associated with viral illness, and pulmonary crises with severe transfusions might well have survived without transfusions had hypoxia, require urgent hospital treatment. Regular Doppler their early management been more effective. testing of the cerebral blood flow can anticipate neurological Thus, those managing patients with severe forms of tha- complications that can be prevented by regular transfusions lassemia must make absolutely sure in infancy that regular (Adams and others 1998), which can be continued indefinitely. 670 | Disease Control Priorities in Developing Countries | David Weatherall, Olu Akinyanju, Suthat Fucharoen, and others Because most aplastic crises result from human parvovirus Several simple, cheap tests are available for diagnosing iron infection, the development of a vaccine would be a great deficiency, but thalassemia presents more of a problem. advantage. The other acute complication, splenic sequestration Screening tests will identify those heterozygous for o thal- causing rapid enlargement of the spleen, is associated with pro- assemia or homozygous for thalassemia but will miss most found anemia. It necessitates urgent hospital admission and cases of heterozygosity for thalassemia. However, given the blood transfusion, plus sometimes splenectomy. A variety of restricted distribution of o thalassemia, these distinctions are other complications require hospital treatment, including pri- clinically important only in areas where o thalassemia is com- apism, aseptic necrosis of the femoral or humoral heads, renal mon. Further diagnosis of the thalassemias requires DNA failure, and recurrent hematuria. At every age patients with analysis. sickling disorders seem to be more prone to infection that often Several simple and cheap screening tests are available for sick- requires hospital admission. In most sickling disorders, crises ling disorders, but all of them require confirmation of the are more frequent and anemia worsens during pregnancy. A genotype by hemoglobin electrophoresis. Neonatal screening review of extensive clinical trials in the United States shows requires electrophoresis because the solubility test, which is used that the long-term administration of hydroxyurea reduces widely for adult screening,is unreliable in the first months of life. the frequency of crises and prolongs life in adult sufferers Throughout Asia, screening for Hb E is also necessary, either (Weatherall 2003). with a one-tube dye test or by hemoglobin electrophoresis. The Although milder, Hb SC disease is clinically important, par- Hb E trait is missed by measuring cell size or osmotic fragility. ticularly because of the relatively high incidence of ocular complications. Initial Diagnosis of More Severe Hemoglobinopathies. The initial diagnosis of thalassemia is usually clinical. It can be confirmed by finding typical thalassemic changes of the Requirements peripheral blood, together with an elevated level of Hb F. A Screening and diagnosis for the hemoglobin disorders requires variety of cheap tests for measuring Hb F levels are available, or relatively simple laboratory techniques combined with a well- HPLC analysis can be used. Hemoglobin electrophoresis or o organized program for their application in the community. HPLC is used to diagnose thalassemia homozygotes and children with Hb H disease. Screening and Carrier Detection. Unlike many genetic dis- Sickle cell anemia, Hb SC disease, or combinations of the eases, carrier screening for the main hemoglobin disorders is sickle cell gene with forms of thalassemia can all be identified well established, accurate, and inexpensive. The initial screen- by hemoglobin electrophoresis and can be confirmed by a fam- ing for thalassemia usually measures the mean cell volume and ily study. the mean cell hemoglobin. Thresholds below which the likeli- Further Analysis. More detailed confirmatory analysis, hood of some form of thalassemia is great are well established. including identification of the underlying mutation, is required The diagnosis of thalassemia is confirmed by finding a raised for the thalassemias as a prerequisite for prenatal diagnosis. level of Hb A2 using high-performance liquid chromatography A variety of approaches to mutation analysis based on the poly- (HPLC) or cheaper forms of chromatography or quantitative merase chain reaction, which amplifies particular regions of hemoglobin electrophoresis. Ideally, the initial blood count DNA, are available, but because every population has a number should use an electronic cell counter, and the Hb A2 should be of less common mutations, a central reference laboratory in measured using HPLC. However, the equipment for HPLC each country must be able to sequence the globin genes. analysis is expensive, and the cost per sample is approximately Rapid DNA-based techniques for identifying the different US$2. For this reason, several more economical approaches deletion and nondeletion forms of thalassemia are also avail- have been developed (Fucharoen and others 2004). The initial able (Weatherall and Clegg 2001b). screening can be done using a single-tube osmotic fragility test, which, even though it may result in a relatively high number Facilities and Organization. To provide an adequate labora- of false positive results, usually gives fewer false negatives. tory service, each country with a high incidence of thal- Commercial kits for osmotic fragility testing have recently been assemia or sickle cell anemia (a carrier rate equal to or greater produced, and at least one variety has been validated in than 1 to 2 percent) requires at least one central reference lab- Thailand. Further validation of this approach is required before oratory to carry out accurate hemoglobin and DNA analyses. it can be recommended. Various cheaper methods for measur- Peripheral hospital laboratories with expertise in screening ing the Hb A2 level are available. tests and their quality control are also required. When the red cell indices suggest thalassemia or an osmotic General pediatricians, pediatricians with a special interest in fragility test is positive but the Hb A2 level is normal, it is vital blood diseases, or pediatricians or other clinicians who devote to distinguish between iron deficiency and thalassemia. their entire time to the management of such diseases may care Inherited Disorders of Hemoglobin | 671 for children with severe hemoglobin disorders. A problem arises routinely immunize patients with polyvalent pneumococ- for older patients, who must often change their doctors during cal, meningococcal, and Haemophilus influenzae vac- adolescence. There is a serious dearth of physicians trained to cines. However, the effectiveness of regimens of this type care for older patients. Ideally, every country with a high inci- has not been clearly established even for splenectomized dence of inherited hemoglobin disorders should have centers patients. specially designated for treating patients of all ages. Such centers · Splenectomy. Each year, 2 to 3 percent of patients with severe require outpatient transfusion facilities, space for parents to wait thalassemia require splenectomy. When patients who have while their children are being transfused, inpatient facilities, and received suboptimal treatment first receive standard treat- access to basic laboratory diagnostic services. Centers involved ment, the proportion needing splenectomy can be much in prenatal diagnosis require access to appropriate obstetric higher--up to 30 percent in the first one or two years services. The advantage of specialist centers is continuity of care. (WHO 1994). Because of the spleen's natural tendency to Patients with chronic disease must have confidence in their atrophy, splenectomy is required only in patients with sickle medical advisers. Such confidence can be achieved only if they cell anemia who are having repeated splenic sequestration see the same staff over the entire course of their illness. episodes or who develop hypersplenism. As concerns personnel, a WHO working group has recom- · Bone marrow transplantation.Given the continuing mended one doctor, three nurses, one laboratory technician, improvement in results, bone marrow transplantation is one counselor, and one administrative assistant for every 50 to now a realistic option for patients with severe forms of 100 patients (WHO 1994). Overall, the workload is higher in thalassemia and sickle cell anemia who have histocompati- centers for thalassemia than for sickle cell anemia, largely ble related donors and, in particular, are relatively free of because of the lesser need for regular blood transfusions for complications, particularly chronic hepatitis and severe iron those with the latter condition. However, sickle cell anemia is loading (Giardini 1997). Marrow transplantation requires associated with more acute inpatient episodes per year. specialized facilities and a trained staff, and the initial capital The other major role of centers of this type is education, expenditure is high, but it should be considered wherever including training other clinicians, medical students, coun- the serious hemoglobinopathies are a major problem if the selors, nurses, and others needed to provide information to alternative is lifetime treatment. local communities. · Prenatal diagnosis programs. If prenatal diagnosis programs A number of publications describe treatment protocols are part of an existing genetic and diagnosis service, fetal to use in managing patients (Weatherall and Clegg 2001b; sampling will involve only marginal costs. If they must be WHO 1985, 1987, 1989, 1994). Treatments include the set up from scratch, they require at least two obstetricians following: trained in fetal medicine, access to ultrasound, and a spe- cialist nurse. Disposable or reusable sampling equipment · Blood transfusion. Although regular transfusion is required is also required, together with suitable sterile facilities for most frequently for managing the thalassemias, it is increas- amniocentesis, facilities for termination of pregnancy in ingly being applied for the prophylaxis of serious complica- the first or second trimesters, and access to experienced tions of sickle cell anemia. Treatment centers need to bereavement counselors. Access to a laboratory able to carry cooperate closely with national blood transfusion programs. out mutation analysis is also required. Washed or otherwise leukocyte-depleted blood should · Other treatments. Patients with hemoglobinopathies require always be used. Blood has to be screened for hepatitis B, folate supplements. Those infected with hepatitis C require hepatitis C, HIV, and--in some populations--malarial treatment with antiviral agents. Those with endocrine dam- parasites. Blood requirements for patients with thalassemia age caused by iron loading may require hormone replace- range from 500 to 1,500 liters per 100 patients per year, ment therapy. depending on their age distribution. · Iron chelation. Transfusion-dependent thalassemic patients In addition, centers for hemoglobin disorders require a and patients with sickle cell anemia maintained on transfu- trained counselor or clinical psychologist to handle both sion for prophylactic purposes require 30 to 50 milligrams genetic and social issues. They should interact with local pub- per kilogram per day of desferrioxamine infused subcuta- lic health organizations to disseminate information about neously by pump. One pump is required per patient. inherited hemoglobin diseases to schools and to the commu- Regular assessment of body iron and assessment for com- nity at large. Many countries have parent support groups. The plications are essential. Thalassemia International Federation, an international body · Immunization and other prophylactic measures. Every child run largely by parents and lay members, provides advice and with sickle cell anemia should receive oral prophylactic support for national parent associations and hemoglobinopa- penicillin from the time of diagnosis. Some centers now thy programs. 672 | Disease Control Priorities in Developing Countries | David Weatherall, Olu Akinyanju, Suthat Fucharoen, and others COSTS AND EFFECTIVENESS OF DIAGNOSIS cost of treatment, or 2 percent of the discounted lifetime cost. AND MANAGEMENT Similarly, Scriver and others (1984) estimate the costs of pre- venting an affected birth in Canada as less than annual treat- Defining the full costs of treating patients with inherited disor- ment costs and about 4 percent of lifetime costs. ders of hemoglobin is difficult, and comparing them between countries is even harder. The variables that confound such esti- Treatment. Investigators have made several estimates of the mates include different health care systems, varying methods of costs of treating the thalassemias. Approximate annual costs obtaining donated blood, widely varying practices in screening of care in Thailand in 2003 were as follows (authors' estimates blood for pathogens, and differing costs of drugs and equip- updated from unpublished Thai sources): ment. WHO working parties and others have produced approximate data (Alwan and Modell 1997; WHO 1985, 1987, 1989, 1994). · homozygous thalassemia: US$19.84 million for patients in their first year of treatment and US$17.7 million per year for patients in subsequent years of treatment Thalassemias · Hb E thalassemia: US$100.3 million for patients in the Information about the economic aspects of the thalassemias is first year of treatment and US$92.4 million for patients in sparse, including the costs and effectiveness of different inter- subsequent years of treatment o ventions, three of which are discussed here: screening and · homozygous thalassemia: US$727,000 counseling, treatment using transfusions and chelation, and · Hb H disease: US$7.49 million. cure by means of bone marrow transplant. The life expectancy of a child in Thailand homozygous for Screening and Counseling. Screening and counseling are thalassemia is only 10 years, reflecting an inability to provide cost-effective to the extent that they avert an affected birth or the bulk of patients with expensive chelating drugs. These costs ensure early and adequate treatment of an affected child. are therefore not fully comparable to those from more devel- Several studies, beginning with community screening pro- oped countries, where better care means that children live grams in Montreal (Scriver and others 1984), have provided longer. Han, Han, and Myint (1992) provide data for Myanmar strong evidence that screening and prenatal diagnosis are highly that are comparable to those for Thailand. effective in relation to control of the thalassemias. Alwan and Table 34.3 shows the annual costs for a program in Toronto Modell (1997), Davies and others (2000), and Modell and that offers a high level of symptomatic care, and table 34.4 Kuliev (1991, 1993) provide detailed discussions of the issues presents treatment costs for the eastern Mediterranean, taking involved, together with estimates of the costs of screening and into account the increasing expenditure required as children counseling programs compared with the treatment of estab- grow and require higher doses of maintenance drugs and more lished disease, and even though the estimates are based mainly units of blood each year. Despite the difficulties of making on studies in developed countries, the findings are probably comparisons, tables 34.3 and 34.4 indicate that the costs for more generally applicable. managing thalassemia in the eastern Mediterranean are The effectiveness of prenatal diagnosis programs can be roughly similar to those in Toronto, exclusive of transfusion, quite high if more than 2 percent of the population is a carrier and are therefore probably reasonable estimates of the costs of and the public education programs that precede their establ- managing thalassemias in developed countries. ishment are well designed. In Europe, 80 to 90 percent of coun- Transfusion is required to keep a child with severe thal- seled at-risk couples now request prenatal diagnosis, and a assemia alive beyond age one or two, but by itself prolongs life rapid reduction in affected births has been observed. The thal- only to age 10 to 15. The gain is the added 9 to 14 years lived assemia birth rate fell almost 100 percent between the late with disability weights of 0.1 to 0.5, implying 0.50 DALY gained 1970s and the late 1980s in Cyprus and Sardinia and about for the last years and 0.75 to 0.90 per year earlier. Table 34.4 80 percent in mainland Greece and Italy (Modell and Kuliev shows the costs of transfusion at ages 7 to 11, and figures in 1991). Between 1974 and 1986, such births fell by 40 percent in parentheses in the table show the costs at age 2 and for adults. the United Kingdom (Modell and Kuliev 1991). Costs and health gains run parallel, so it makes little difference In terms of cost-effectiveness, terminating a pregnancy can- whether they are discounted when summed over an interval. not be compared with improving the health of a child carried Table 34.5 shows discounted total costs and DALYs gained for to term, but it is clearly cost saving compared with treatment. death at age 10 or 15 and using disability values of 0.1 and 0.25. Modell and Kuliev (1991) estimate the cost of replacement For both ages, the cost per DALY is about US$2,000 for low (when a couple terminates an affected pregnancy and subse- disability and about US$3,000 for high disability for transfu- quently has a normal child) at only 30 percent of the annual sion only. Inherited Disorders of Hemoglobin | 673 Table 34.3 Annual Costs of Hemoglobinopathy per Outpatient, Excluding Transfusion, Toronto (2001 US$) Thalassemia Sickle cell disease Category Chelated Nonchelated Chelated Nonchelated Clinic staff salaries 1,011.95 183.68 1,011.95 252.99 Clinic supplies 930.19 25.15 930.19 34.65 Medical and surgical outpatient unit 2,069.57 n.a. 2,069.57 n.a. Consultations 92.58 88.39 92.58 11.94 Diagnostic tests 742.58 281.44 905.89 210.99 Laboratory costs 413.96 31.04 414.01 42.74 Laboratory costs (medical dayunit visits) 665.81 n.a. 665.81 n.a. Total 5,926.64 609.70 6,090.00 553.31 Source: Estimated costs provided by Nancy Oliveri of the University of Toronto. Note: n.a. not applicable. Table 34.4 Costs of Treatment of Thalassemia for One Patient Age 7 to 11, Eastern Mediterranean (2001 US$) Category Minimum treatment Full treatment Costs other than iron chelation Day transfusion: hotel and nursing 375 375 12 transfusions/year 1,088 (600­1,575) 2,250 (1,390­3,150) Investigations 135 (135­435) 278 (278­870) Occasional costs (such as operations) 150 645 Staff salaries 300 620 Total if no desferrioxamine therapy 2,048 (1,560­2,835) n.a. n.a. Desferrioxamine therapy (iron chelation) 3,080 (1,440­4,725) 6,165 (2,880­9,450) Total with desferrioxamine therapy 5,128 (3,000­7,560) 10,333 (6,190­15,110) Source: Alwan and Modell 1997. Note: n.a. not applicable. The figures in parentheses show the range of costs for a two-year-old and an adult; where no range is shown, the cost is independent of age. Both minimum and full treatment include transfusion and chelation. Full treatment means more frequent transfusion and consultations, more laboratory work, and more surgery. A thalassemic patient can be kept alive beyond adolescence, Comparing full treatment to none, annual costs rise with possibly for a normal life span, if also chelated. This treatment age from US$6,190 to US$15,110 (table 34.4). As table 34.5 means added annual costs of about US$3,000, or about shows, up to age 15, 10.8 to 11.8 discounted DALYs are gained US$6,000 (table 34.4) for full treatment, plus higher costs of at a total cost of US$121,284 and a cost per DALY of approxi- transfusion and other components. Besides prolonging life, full mately US$10,300 to US $11,200, comparable to lifetime incre- treatment is assumed to reduce disability from 0.10 to 0.02 and mental costs. Beyond age 15, the cost per DALY is some from 0.25 to 0.10. Table 34.5 also shows discounted incremen- US$16,000 to US$17,000, and over the lifetime it is between tal costs, DALY gains, and cost per DALY of full treatment com- US$13,000 and US$15,000. Because costs and gains run pared with those of minimal treatment. The incremental cost parallel--and also because of discounting--the lifetime cost per DALY is high up to age 15 because of the modest gain com- per DALY is not sensitive to age at death of 45 or older, so dif- pared with prolonging life with greater disability. From age 15, ferences in regional life expectancies do not matter. the cost drops because the child would otherwise die. If a life These cost estimates come from the eastern Mediterranean. expectancy of 80 years is assumed, lifetime costs are some In Thailand, the first-year costs are much higher than the US$9,000 to US$11,000 per DALY at a disability weight of 0.1 costs shown in table 34.5 and include the costs of delivering and roughly US$10,000 to US$12,000 at a disability weight the child and protecting it from infection. Thereafter, the cost of 0.25. per year of treating the 6,250 survivors of each birth cohort 674 | Disease Control Priorities in Developing Countries | David Weatherall, Olu Akinyanju, Suthat Fucharoen, and others Table 34.5 Cost-Effectiveness of Treatment for Homozygous and Transfusion-Dependent Hb E Thalassemia DALYs gained/patient Cost/DALY (US$) Cost/patient Disability Disability Disability Disability Category (US$) weight 0.1 weight 0.25 weight 0.1 weight 0.25 Minimal treatment, transfusion only Until death at age 10 17,368 6.96­7.60 6.00­6.39 2,285­2,495 2,718­2,896 Until death at age 15 23,840 10.25­10.81 7.52­7.87 2,206­2,325 3,029­3,170 Full treatment with chelation: incremental compared with minimal treatment Until age 15 60,467 1.03­3.80 0.55­2.03 15,912­58,706 29,787­109,940 Beyond age 15 to maximum age 80 132,901 17.25 16.35 7,704 8,129 Total lifetime 193,368 18.28­21.05 16.90­18.38 9,186­10,578 10,520­11,442 Full treatment with chelation: total compared with no treatment Until age 15 121,284 11.81 10.84 10,273 11,186 Beyond age 15 to maximum age 80 274,662 17.25 16.35 15,922 16,799 Total lifetime 395,946 29.06 27.19 13,625 14,578 Source: Authors' calculations. All costs and DALYs gained are discounted at 3 percent annually, starting at birth. Note: Differences in DALYs gained for a given age range and disability weight depend on how rapidly health is assumed to deteriorate in the years immediately preceding death. with transfusion alone (assuming they live only 10 years) Sickle Cell Disease drops, implying a cost per DALY of US$3,146 to US$3,776. Extensive data on the costs of sickle cell anemia come from Half the 97,500 people with Hb E thalassemia are assumed Davis, Moore, and Gergen's (1997) analysis in the United to require treatment, leading to a cost per DALY of US$2,100 States. An estimated 75,000 hospitalizations of both children to US$2,500, consistent with costs in the eastern and adults occurred each year from 1989 through 1993. The Mediterranean. For chelated patients in Toronto, the estimate average cost of hospitalization in 1996 was estimated at of US$5,927 per year (table 34.3) implies a cost per DALY of US$6,300, resulting in a total direct cost of US$575 million per US$6,600 to US$8,000, considerably less than in the eastern year. In this and subsequent studies, the bulk of hospitaliza- Mediterranean, but the costs do not include transfusion. Thus, tions was confined to a subset of about 10 percent of the total data from three different regions on the cost-effectiveness of patient population. full therapy for victims of the common treatable forms of In the United States, specialized treatment centers provide a thalassemia are roughly comparable. cost-saving approach to the care of sickle cell anemia. Patients De Silva and others' (2000) study provides approximate enrolled in these centers used emergency rooms and inpatient costs for treating thalassemia in Sri Lanka. If we assume that a units significantly less frequently than those cared for in the prevention program will probably not be developed in the near general hospital community, resulting in significantly lower future, the data suggest that management of the disease will health care costs (Nietert, Silverstein, and Abboud 2002; Yang consume approximately 5 to 8 percent of the country's health and others 1995). Currently, only a small proportion of budget based on 1999 figures. Those estimates have not been patients are treated in centers of this kind, even in developed used to derive cost-effectiveness results. countries. A pilot study in Benin found that the development of a comprehensive clinical care program reduced the fre- Bone Marrow Transplantation. Angelucci and Lucarelli quency and severity of acute complications related to sickle cell (2001) discuss the economic aspects of bone marrow trans- anemia (Rahimy and others 2003). The annual cost per family plantation. The 1991 cost was US$73,250, excluding follow-up using the program was US$40, and the annual cost for each but including the expense of setting up a program. This hospitalization was US$100. amount is almost certainly cost saving compared with lifelong transfusion and chelation, and it is also more cost-effective, Neonatal Screening and Prophylaxis. Extensive controlled given the reduction in disability from curing the disease. trials in several developed countries have demonstrated the Inherited Disorders of Hemoglobin | 675 Table 34.6 Cost-Effectiveness of Penicillin Prophylaxis for Sickle Cell Disease Detected by Newborn Screening, Jamaica Category Monthly injection Daily oral dose Total/1,000 children Monthly cost of penicillin (J$) 22 250 26,560 Nurse's time, 10 minutes/month (J$) 90 n.a. 88,200 Clinician's time, 20 minutes 4­6 times/year (J$) 152.67­229.00 152.67­229.00 152,670­229,000 Total year 1, 8 treatments (J$) 2,117­2,728 3,221­3,832 2,140,000­2,750,000 Total, each of years 2­4, 12 treatments (J$) 3,176­4,092 4,832­5,748 3,210,000­4,130,000 Discounted total (discount rate of 3 percent), first 4 years 11,101­14,303 16,889­20,091 11,220,000­14,420,000 Equivalent in U.S. dollarsa US$1 J$49.8 223­287 339­407 220,000­290,000 US$1 J$59.8 186­239 282­336 190,000­240,000 Number of deaths averted by prophylaxis 0.024/child 0.024/child 24 deaths Costs per death averted (US$) 7,750­11,958 11,750­16,958 7,830­12,058 Costs per DALY gained (US$) 267­412 405­585 270­416 Source: Authors' calculations based on data from Hambleton 2004a, 2004b. Note: n.a. not applicable. The results are based on a cohort study of 315 cases. a. Two exchange rates are shown because the exchange rate changed during the course of the study. life-saving effect of giving prophylactic penicillin from birth. A In the Jamaican cohort, up to 2 percent of children take pro- randomized trial suggested that this approach would save sig- phylactic penicillin orally; the rest receive monthly injections nificantly more lives than starting prophylactic penicillin when requiring 10 minutes of a nurse's time. A clinician sees each an infant presents with symptoms of the disease. Without child for 20 minutes four to six times a year. Table 34.6 shows treatment, pneumococcal infection is the leading cause of death the costs of personnel and penicillin. Other recurrent costs, before age five. Penicillin by monthly injection or daily oral dose such as those for syringes, will be low in comparison. No from about age four months to four years reduces bacteremia allowance is made for capital costs. The marginal cost of detect- by 83 to 86 percent (D. Bonds, personal communication, ing sickle cell disease in newborns who have already been September 28, 2004; Gill and others 1989, 1995; Panepinto and screened for other conditions such as phenylketonuria is only others 2000). Prophylaxis also reduces the case-fatality rate about US$3.30 (D. Bonds, personal communication, from infection from 27 to 18 percent. September 28, 2004). The cost of screening children who do Data from the United States show that neonatal screening not have sickle cell disease depends on prevalence and is not for sickle cell disease followed by the use of appropriate included in the estimates. It becomes unimportant as more of prophylactic treatment prevents deaths (Tsevat and others the population is at risk. 1991). With 16 percent of the U.S. population being African On average, preventing a death costs about US$8,000 to American, without screening and treatment 13 deaths would US$12,000 by means of injection and about US$5,000 more occur per million infants. Six of these deaths could be pre- when oral penicillin is used. Because death would typically vented by targeting only African Americans, and two more occur between one and two years of age, each death averted could be prevented by universal screening. Ignoring disability saves 29 DALYs and costs about US$270 to US$400 per DALY. and discounting at 3 percent, we find that targeted screening Penicillin prophylaxis is probably more cost-effective than any costs US$6,709 per life year saved, or somewhat more per other intervention. It is standard practice in Jamaica and the DALY. At more than US$30,000 per life year saved, universal United States and can be recommended for middle-income screening would not be cost-effective compared with targeted countries where the prevalence is high enough--in the general testing (Panepinto and others 2000). population or in those of African origin--to justify screening. Data from a Jamaican cohort of 315 children with sickle cell disease allow for a similar cost-effectiveness analysis in a Application to Other Countries. The limited data available middle-income country (Hambleton 2004a, 2004b). They sug- indicate that specialized treatment centers and neonatal screen- gest that treatment averts seven to eight deaths, or that a newly ing programs are effective approaches toward the control of diagnosed child's chances of dying before age four are reduced sickle cell anemia. Although these conclusions should be valid by 2.4 percentage points. Because bacteremia becomes less fre- for developing countries, many uncertainties remain. In partic- quent with age, few cases occur--and little is known about the ular, data on the causes of infection in infants with sickle cell mortality risk--after childhood. disease in Africa are sparse. Because the spectrum of infection 676 | Disease Control Priorities in Developing Countries | David Weatherall, Olu Akinyanju, Suthat Fucharoen, and others may be different, the value of penicillin prophylaxis in Africa is · Option six: based on option four or five but includes the unknown. availability of a bone marrow transplant program. Laboratory diagnosis is well defined, is cheap, and does not differ depending on the level of available technology. The Options one and two, though still commonly practiced in major uncertainty is whether the increasing indications for many countries, offer little prospect of reducing the frequency prophylactic transfusion in developed countries will be mir- of serious forms of thalassemia. Overall, that reduction is rored in other populations. best achieved by option five, which combines maximum possibilities for reducing the frequency of severe disease with Other Treatments. As indicated in the discussion of the bur- the best possible care for affected children. Although thal- den of disease, sickle cell patients suffer various clinical events assemia births have fallen sharply in some developed countries, for which treatment may be life saving, such as transfusion for the effect of prenatal screening is likely to be lower for the large aplastic crises. For painful crises, intervention (analgesics and mainland populations of Asia if this policy is implemented, possibly hospitalization) is only palliative. Because of the lack which is why this option includes best-practice treatment. of information on costs and of consensus on the associated dis- Limited studies also suggest that families that undergo prenatal abilities with and without treatment, we have not assessed the diagnosis tend to settle at the population norm for the number cost-effectiveness of any of these treatments. of children that they subsequently have and that their views on their ability to have unaffected children are extremely positive. For those countries or groups in which termination of preg- OPTIONS FOR CONTROL AND MANAGEMENT nancy is unacceptable for religious or cultural reasons, option OF INHERITED HEMOGLOBIN DISORDERS four is recommended. Option six, which is possible only in countries where bone marrow transplantation is available, Even though much more work is needed on both the scientific should be exploited by any country with a high frequency and the economic aspects of the hemoglobinopathies, certain of the disease, because it offers a potentially cost-effective issues are now clear. Until more definitive ways of treating approach to managing some proportion of affected children. them are available, reliable knowledge exists on how they can For the extremely heterogeneous intermediate forms of best be managed symptomatically. Furthermore, compelling thalassemia, notably Hb E thalassemia, options three and evidence suggests that population screening programs com- four would probably be best, at least until a better understand- bined with prenatal diagnosis can reduce the financial burden ing has been gained of the clinical heterogeneity of these these increasingly common diseases impose on health services. thalassemias. Defining several options for their control and management is therefore possible. These are based, with some modifications, Thalassemias on Alwan and Modell (1997). Few options are available for the o thalassemias. Homozygous babies with this condition are stillborn. However, because of Severe Thalassemias the serious maternal complications of carrying these babies, This list provides most of the possible options for the control this condition should be screened for prenatally, and affected and management of thalassemia in developing countries. babies should be identified by prenatal diagnosis with a view to terminating the pregnancy. · Option one: best possible patient care, together with retro- Hb H disease, the compound heterozygous state between + spective genetic counseling after the first affected child is and o thalassemia, is generally a relatively mild disorder that diagnosed simply requires careful follow-up and treatment of complica- · Option two: best possible patient care, together with retro- tions. Although some have suggested that screening and prena- spective genetic counseling and the option of prenatal diag- tal diagnosis may be relevant for the more severe forms, more nosis for subsequent pregnancies data are required to reach a conclusion. · Option three: best possible patient care, together with retro- spective genetic counseling and prospective (premarital) Sickle Cell Disorders carrier screening and counseling, but no prenatal diagnosis These options for the management of the sickle cell disorders · Option four: best possible patient care with premarital, are directed particularly at the populations of developing coun- family-based, and population-based prospective carrier tries, although, overall, they are relevant to most countries: screening and genetic counseling, but no prenatal diagnosis · Option five: best possible patient care, premarital and prena- · Option one: best possible patient care with the use of pro- tal prospective carrier screening and genetic counseling, and phylactic penicillin following diagnosis, together with retro- the option of prenatal diagnosis spective genetic counseling Inherited Disorders of Hemoglobin | 677 · Option two: best possible patient care, together with a In the case of thalassemia, the Thalassemia International neonatal screening program and the use of penicillin for all Federation acts as an international coordinating body that helps homozygous babies, together with retrospective screening countries develop workshops for training in diagnosis and treat- and counseling ment and organizes international meetings at which experts · Option three: best possible patient care, together with from different countries share their research and clinical experi- neonatal screening and the use of prophylactic penicillin ences. In 1996, a group of doctors formed the Fédération des from birth for homozygotes, together with population Associations de Lutte contre la Drépanocytose en Afrique screening and prospective genetic counseling (Federation of Associations to Control Sickle Cell Anemia in · Option four: as for option three, plus the availability of pre- Africa). The membership has grown to 13 Sub-Saharan coun- natal diagnosis, bone marrow transplantation, or both. tries. The federation represents a major initiative in relation to regional training in both the diagnosis and the treatment of Option three would be required to combine best manage- sickle cell disease in Sub-Saharan Africa. Unfortunately, it has ment with the possibility of reducing the frequency of the dis- been unable to raise sufficient funds to equip and run even a ease, although whether this option would have any effect by modest secretariat. Considering the success of the Thalassemia altering the pattern of marriage is not clear. Whether prenatal International Federation, particularly in countries with limited screening of mothers would be valuable is also not clear: it facilities for managing the hemoglobin disorders, the lack of would reduce the number of neonates who require screening, support for this initiative in Africa is a clear indication of the but because the cost of screening for sickle cell trait is so small, importance of educating nongovernmental organizations and the issue is probably not important. Prenatal diagnosis can be similar bodies about the increasing public health problems developed (option four), but this option does not seem to be a resulting from the inherited hemoglobin disorders. high priority for sickling disorders in many developed coun- tries, at least until more is known about the reasons for their Ethical and Social Issues phenotypic heterogeneity. By contrast, demand is greater in The various options for controlling and managing the hemo- developing countries with limited facilities for the care of these globin disorders raise many ethical and social issues (see patients. Some developed countries are beginning to immunize Weatherall and Clegg 2001b for more details). These issues children with sickling disorders against infections with pneu- arise most often in developing countries, where the level of mococcus, H. influenzae, and meningococcus, but in develop- education is often low and understanding of genetic diseases is ing countries, this treatment would add enormously to the limited. Serious genetic diseases such as thalassemia are associ- burden of management programs. Clinical trials to test the effi- ated with social problems such as patient stigmatization and cacy of prophylactic penicillin with or without these vaccine broken marriages because one partner blames the other for the regimens are urgently needed. Similarly, more information birth of an affected child. In countries where arranged mar- about the pathogens that cause early deaths in developing riages are still common, screening programs for heterozygotes countries is required before the widespread use of prophylactic may make it difficult for female carriers to find husbands. In penicillin can be recommended. addition, cultural and religious objections about interfering with nature arise when pregnancies are terminated because Bone Marrow Transplantation children have serious genetic diseases. At the same time, if Experience in developed countries indicates that bone marrow governments perceive prenatal diagnosis and termination of transplantation may offer a cost-effective approach to manag- pregnancy to be a highly cost-effective way of controlling these ing a subset of patients with inherited disorders of hemoglobin diseases, which they are, the danger arises that governments (Borgna-Pignatti 1985). In developing countries, if this service will pressure women to undergo these procedures. Therefore, is available at all, it is usually confined to the private sector or before any programs are established, extensive discussions to those who can pay the fees teaching hospitals require. Given between governments, the medical profession, and the com- this context, defining the role of bone marrow transplantation munity about how to control these diseases are vital. in the global control of these diseases is difficult. FURTHER RESEARCH International and National Support Groups Despite the progress made toward understanding the molecu- Largely through the efforts of parents with affected children and lar pathology, pathophysiology, and management of the inher- clinicians who have taken an interest in the hemoglobin disor- ited hemoglobin disorders, many gaps persist: ders, many countries have developed national thalassemia or sickle cell anemia societies that provide support for parents, · First, much better data are required about their frequency workshops for doctors, and a variety of other important inputs. and distribution in many developing countries. 678 | Disease Control Priorities in Developing Countries | David Weatherall, Olu Akinyanju, Suthat Fucharoen, and others · Second, more information is required about the reasons, Doppler Ultrasonography." New England Journal of Medicine 339 both genetic and environmental, for the remarkable clinical (1): 5­11. heterogeneity of these conditions. Akinyanju, O. 2001. "Issues in the Management and Control of Sickle Cell Disorder." Archives of Ibadan Medicine 2 (2): 37­41. · Third, much better criteria are required for the management Alwan, A., and B. Modell. 1997. Community Control of Genetic and of the intermediate forms of thalassemia and of sickle cell Congenital Disorders. Eastern Mediterranean Region Office anemia. Technical Publication Series 24. Alexandria, VA: World Health · Fourth, much more work is required on the role of the envi- Organization. ronment, a topic that has been badly neglected compared Angastiniotis, M., and B. Modell. 1998. "Global Epidemiology of Hemoglobin Disorders." Annals of the New York Academy of Sciences with research on the genetics of these conditions. 850: 251­69. · Finally, further studies are required on better methods for Angelucci, E., and G. Lucarelli. 2001. "Bone Marrow Transplantation in their symptomatic management or a definitive cure. Thalassemia." In Disorders of Hemoglobin, ed. M. H. Steinberg, B. G. Forget, D. R. Higgs, and R. L. Nagel, 1052­72. New York: Cambridge One important approach toward progress in controlling University Press. these diseases is further development of North-South partner- Ashley-Koch, A., Q. Yang, and R. S. Olney. 2000. "Sickle Hemoglobin ships (WHO 2002). Arrangements of this kind have been (HbS) Allele and Sickle Cell Disease: A HuGE Review." American Journal of Epidemiology 151 (9): 839­45. extremely successful for thalassemia but have not evolved for Ballas, S. K. 1998. "Sickle Cell Disease: Clinical Management." Clinical sickle cell anemia research. In both cases such partnerships Haematology 11 (1): 185­214. should evolve and lead to local South-South networks allow- Borgna-Pignatti, C. 1985. "Marrow Transplantation for Thalassemia." ing individual countries to share their expertise about these Annual Review of Medicine 36: 329­36. increasingly important conditions. Bunn, H. F. 1997. "Pathogenesis and Treatment of Sickle Cell Disease." Finally, a great deal more work needs to be carried out, par- New England Journal of Medicine 337 (11): 762­69. ticularly in developing countries, to investigate the economic Cao, A., R. Galanello, and M. C. Rosatelli. 1998. "Prenatal Diagnosis and aspects of these diseases, in terms of both their overall health Screening of the Haemoglobinopathies." Clinical Haematology 11 (1): 215­38. burden and their control and management. Davies, S. C., E. Cronin, M. Gill, P. Greengross, M. Hickman, and C. Normand. 2000. "Screening for Sickle Cell Disease and Thalassaemia: A Systematic Review with Supplementary Research." Health Technology Assessment 4 (3): i­99. CONCLUSIONS Davis, H., R. M. Moore Jr., and P. J. Gergen. 1997."Cost of Hospitalizations Associated with Sickle Cell Disease in the United States." Public Health The inherited hemoglobin disorders are posing an increasing Report 112 (1): 40­43. global health problem, killing thousands of children because of De Silva, S., C. A. Fisher, A. Premawardhena, S. P. Lamabadusuriya, T. E. A. the inadequacy or unavailability of treatment. With appropri- Peto, G. Perera, and others (Sri Lanka Thalassaemia Study Group). ate therapy, many children can survive and have an excellent 2000. "Thalassaemia in Sri Lanka: Implications for the Future Health quality of life, despite requiring lifelong treatment. Even Burden of Asian Populations." Lancet 355 (9206): 786­91. though the full economic burden of managing these disorders Dover, G. J., and O. S. Platt. 1998. "Sickle Cell Disease." In Hematology in Infancy and Childhood, ed. D. G. Nathan and S. H. Orkin, 762­801. is currently unknown, in the case of thalassemia, screening Philadelphia: W. B. Saunders. and prenatal diagnosis are cost-effective means of prevention. Fleming, A. F., J. Storey, L. Molineaux, E. A. Iroko, and E. D. Attai. 1979. Penicillin prophylaxis provides cost-effective protection from "Abnormal Haemoglobins in the Sudan Savanna of Nigeria: I. infection for babies with sickle cell disease and should be stan- Prevalence of Haemoglobins and Relationships between Sickle Cell Trait, Malaria, and Survival." Annals of Tropical Medicine and dard practice wherever it is affordable. Parasitology 73 (2): 161­72. Fucharoen, G., K. Sanchaisuriya, N. Sae-Ung, S. Dangwibul, and S. Fucharoen. 2004. "A Simplified Screening Strategy for Thalassaemia ACKNOWLEDGMENTS and Haemoglobin E in Rural Communities in South-East Asia." Bulletin of the World Health Organization 82 (5): 364­72. The authors thank Bernadette Modell for reprints and Giardini, C. 1997. "Treatment of B-Thalassemia." Current Opinion in Hematology 4: 79­87. preprints of her work in this field and Liz Rose for preparing Gill, F., A. Brown, D. Gallagher, S. Diamond, E. Goins, R. Grover, and the manuscript. others. 1989. "Newborn Experience in the Cooperative Study of Sickle Cell Disease." Pediatrics 83 (5, pt. 2): 827­29. Gill, F., L. Sleeper, S. Weiner, A. Brown, R. Bellevue, R. Grover, and others. REFERENCES 1995. "Clinical Events in the First Decade in a Cohort of Infants with Sickle Cell Disease." Blood 86 (11): 776­83. Adams, R. J., V. C. McKie, L. Hsu, B. Files, E. Vichinsky, C. Pegelow, and Hambleton, I. 2004a. "Lifetime Survival Estimates for People with SS others. 1998. "Prevention of a First Stroke by Transfusions in Children Disease in Jamaica." Note prepared for the Disease Control Priorities with Sickle Cell Anemia and Abnormal Results on Transcranial Project, University of the West Indies, Mona, Jamaica. Inherited Disorders of Hemoglobin | 679 ------. 2004b. "Mortality among People with Homozygous Sickle Cell Stamatoyannopoulos, G. 1973. "Problems of Screening and Counseling in Disease in Jamaica." Note prepared for the Disease Control Priorities the Hemoglobinopathies." In Fourth International Congress on Birth Project, University of the West Indies, Mona, Jamaica. Defects, ed. A. G. Motulsky and W. Lenz, 268­76. Amsterdam: Excerpta Han, A. M., K. E. Han, and T. T. Myint. 1992. "Thalassemia in the Medica. Outpatient Department of the Yangon Children's Hospital in Steinberg, M. H., B. G. Forget, D. R. Higgs, and R. L. Nagel, eds. 2001. Myanmar: Cost Analysis of the Day-Care-Room Services for Disorders of Hemoglobin. New York: Cambridge University Press. Thalassemia." Southeast Asian Journal of Tropical Medicine and Public Tsevat, J., J. B. Wong, S. G. Pauker, and M. H. Steinberg. 1991. "Neonatal Health 23 (2): 273­77. Screening for Sickle Cell Disease: A Cost-Effectiveness Analysis." Livingstone, F. B. 1985. Frequencies of Hemoglobin Variants: Thalassemia, Journal of Pediatrics 118 (4, pt. 1): 546­54. the Glucose-6-Phosphate Dehydrogenase Variants, and Ovalocytosis in Weatherall, D. J. 2003. "Pharmacological Treatment of Monogenic Human Populations. Oxford, U.K.: Oxford University Press. Disease." Pharmacogenomics Journal 3 (5): 264­66. Modell, B., and A. M. Kuliev. 1991. "Services for Thalassaemia as a Model Weatherall, D. J., and J. B. Clegg. 2001a. "Inherited Haemoglobin for Cost-Benefit Analysis of Genetics Services." Journal of Inherited Disorders: An Increasing Global Health Problem." Bulletin of the World Metabolic Disorders 14 (4): 640­51. Health Organization 79 (8): 704­12. ------. 1993. "A Scientific Basis for Cost-Benefit Analysis of Genetics ------. 2001b. The Thalassaemia Syndromes. 4th ed. Oxford, U.K.: Services." Trends in Genetics 9 (2): 46­52. Blackwell Science. Molineaux, L., A. F. Fleming, R. Cornille-Brogger, I. Kagan, and J. Storey. WHO (World Health Organization). 1985. Report of the Third and Fourth 1979. "Abnormal Haemoglobins in the Sudan Savanna of Nigeria: III. Annual Meeting of the WHO Working Group for the Community Control Malaria, Immunoglobulins, and Antimalarial Antibodies in Sickle Cell of Hereditary Anaemias. HMG/WG/85.8. Geneva: WHO. Disease." Annals of Tropical Medicine and Parasitology 73 (4): 301­10. Nietert, P. J., M. D. Silverstein, and M. R. Abboud. 2002. "Sickle Cell ------. 1987. Report of the Fifth WHO Working Group on the Feasibility Anaemia: Epidemiology and Cost of Illness." Pharmacoeconomics 20 Study on Hereditary Disease Community Control Programmes, (12): 357­66. Heraklion, Crete, 24­25 October 1987. WHO/HDP/WG/HA/89.2. Geneva: WHO. Panepinto, J. A., D. Magid, M. J. Rewers, and P. A. Lane. 2000. "Universal versus Targeted Screening of Infants for Sickle Cell Disease: A Cost- ------. 1989. Report of the Fifth WHO Working Group on the Feasibility Effectiveness Analysis." Journal of Pediatrics 136 (2): 201­8. Study on Hereditary Disease Community Control Programmes (Hereditary Anaemias) Cagliari, Sardinia. WHO/HDP/WG/HA/89.2. Rahimy, M. C., A. Gangbo, G. Ahouignan, R. Adjou, C. Deguenon, S. Geneva: WHO. Goussanou, and E. Alihonou. 2003 "Effect of a Comprehensive Clinical Care Program on Disease Course in Severely Ill Children with Sickle ------. 1994. Guidelines for the Control of Haemoglobin Disorders. Report Cell Anemia in a Sub-Saharan African Setting." Blood 102 (3): 834­38. of the Sixth Annual Meeting of the WHO Working Group on Haemoglobinopathies, Cagliari, Sardinia, 8­9 April 1989. Geneva: Samavat, A., and B. Modell. 2004. "Iranian National Thalassaemia WHO. Screening Programme." British Medical Journal 329 (7475): 1134­37. ------. 2002. Genomics and World Health. Geneva: Advisory Committee Scriver, C. R., M. Bardanis, L. Cartier, C. L. Clow, G. A. Lancaster, and on Health Research, WHO. J. T. Ostrowsky. 1984. "Beta-Thalassemia Disease Prevention: Genetic Medicine Applied." American Journal of Human Genetics 36 (5): Yang, Y. M., A. K. Shah, M. Watson, and V. N. Mankad. 1995. "Comparison 1024­38. of Costs to the Health Sector of Comprehensive and Episodic Health Care for Sickle Cell Disease Patients." Public Health Report 110 (1): Serjeant, G. R. 1992. Sickle Cell Disease. Oxford, U.K.: Oxford University 80­86. Press. 680 | Disease Control Priorities in Developing Countries | David Weatherall, Olu Akinyanju, Suthat Fucharoen, and others Chapter 35 Respiratory Diseases of Adults Frank E. Speizer, Susan Horton, Jane Batt, and Arthur S. Slutsky Adult respiratory diseases in the developing world are a major for each decade of life, whereas rates of death from tuber- burden in terms of morbidity and mortality and, particularly as culosis remain relatively constant. Notably, acute respiratory related to chronic respiratory disease, are of increasing concern diseases--in addition to tuberculosis--remain major concerns (Murray and Lopez 1996). For many years, the leading cause of among adults with AIDS. adult respiratory disease mortality has been tuberculosis, which The diagnosis of pneumonia varies according to the patient's still kills far more people than it should, given the increased access to medical care. Often the diagnosis is made simply on efficacy of treatment and preventive regimens (see chapter 16). the basis of cough and fever. For patients with access to a hospi- However, the burden of other acute and chronic adult respira- tal, the likelihood of obtaining a chest x-ray increases; generally tory diseases, which is the focus of this chapter, has been rising the infection is bacteriologically confirmed only in the most throughout the world. These diseases fall into four categories: sophisticated medical centers. The natural history of pneumo- acute diseases, such as pneumonia and influenza; chronic nia without antibiotic treatment varies with the etiologic agent diseases, such as chronic obstructive pulmonary disease and the patient's underlying comorbid conditions and age. (COPD) and asthma; occupational lung diseases, such as byssi- Mortality resulting from these lower respiratory diseases is nosis, asbestosis, and coal worker's pneumoconiosis; and other approximately 10-fold higher in people age 60 to 69 than in parenchymal lung diseases, such as immune-related lung people age 15 to 59 (WHO 2000). Comorbid conditions, mal- diseases. Lung cancer, tuberculosis, and AIDS-related lung dis- nutrition, low socioeconomic status, and cigarette smoking eases are dealt with in chapters 29, 16, and 18, respectively. each play a role in increasing the incidence of disease and wors- ening the prognosis, both with and without treatment. ACUTE DISEASES: PNEUMONIA AND INFLUENZA From studies conducted in the developed world, it would be reasonable to conclude that common antibiotics for pneumo- Obtaining figures on the incidence and burden of pneumonia nias that occur outside a hospital setting would effectively and influenza in adults throughout the developing world has reduce days lost from work and, in the absence of other morbid been surprisingly difficult. Much of the research and surveil- conditions, mortality. The few studies in which sputum speci- lance has been directed toward the pediatric age group (see mens have been cultured suggest that Streptococcus pneumoniae chapter 25). In 2000, fatal lower respiratory infections, as a is found in between 40 and 50 percent of the cases. Gram- class that represents serious pneumonia and influenza, were negative organisms or mixed infections are often isolated, and reported as the cause of 120 deaths per million men and thus, the use of broad-spectrum antibiotics is warranted (Hooi, 76 deaths per million women worldwide for the 15 to 59 age Looi, and Ng 2001; Hui and others 1993; Lieberman and others group (WHO 2000). For both sexes in this age group, this sta- 1996). As would be expected, increased use of antibiotics tistic represents approximately one-third of the deaths caused has resulted in increased resistance to common antibiotics. In by tuberculosis. However, for the age groups over 60, rates addition, 10 to 15 percent of these cases may be tuberculosis of death from lower respiratory disease more than double (Dolin, Raviglione, and Kochi 1994). 681 Scott and others (2000) suggested that, despite the similarity the elderly are at greater risk of suffering from complications of the mortality rates for hospital-treated pneumonia in devel- from influenza. Those criteria, along with the adequacy of sup- oping and developed countries, there are important differences ply, form the basis for choosing who should be considered for in the age distributions. The median age at death among vaccination each year. Because the symptoms of influenza can Kenyan adults was 33 years, in contrast to more than 65 years be quite similar to those of bacterial pneumonia, influenza may in more developed countries. Many patients in developing often be misdiagnosed as pneumonia. Generally, influenza is countries present late in the course of the disease. Often they more self-limiting than pneumonia, although the infectivity die before an appropriate diagnostic workup can be completed, and transmission of influenza from person to person can be thus leading to an underestimate of case-fatality rates. substantial. The current threat of H5N1 influenza has resulted Signs and symptoms of influenza can vary from trivial to in increased human and avian surveillance and preparations explosive. Although the disease is usually self-limiting, it can for a possible pandemic (box 35.1). result in both severe incapacity and, when not properly treated, The recent 2003 outbreak of severe acute respiratory syn- potentially fatal secondary pneumonia. Clearly, patients with drome (SARS; see chapter 53) emphasizes the importance of comorbid conditions, the very young, pregnant women, and accurate and open surveillance and a coordinated response in Box 35.1 H5N1 Influenza Clearly, of even greater concern is the potential for a new sources, it is not clear how effective these control measures influenza A pandemic, as occurred in 1918 and more have been. Although these efforts were thought to help recently in 1958 and 1968, from a newly altered strain of control the spread of the virus, permanent ecological avian influenza. With each additional bird-to-human case, reservoirs appear to have become established in wild fowl modest genetic mutation or re-assortment increases the and domestic chickens over a relatively broad region of chance for the avian virus to be altered to become estab- Southeast Asia. WHO authorities have expressed concern lished and virulent in mammalian species. This may result about the finding that migratory birds that are infected in the establishment of sustained transmission among with H5N1 but are relatively asymptomatic have spread humans. While the pandemics of 1958 and 1968 were viable viruses over large regions with subsequent infection together responsible for approximately 3 million deaths, in domestic poultry. Furthermore, more recently there has mostly in the very young, the elderly, and in those with co- been evidence of disease in wild and zoo mammals as well morbid conditions, the 1918 episode is believed to have as isolated cases of infection in domestic cats. Recent caused over 40 million deaths, mostly in the age group 15 reports from Vietnam include two cases in humans to 35 years. This potential for greatly increased mortality infected through the consumption of uncooked duck among such a robust population has fueled recent concern blood. Further investigation of possible person-to-person (WHO 2005a). transmission is underway. Recently, WHO (2005b) stated, This concern has become more immediate with the "The possible spread of H5N1 avian influenza to poultry identification of a sub-strain of influenza A, H5N1, first in additional countries cannot be ruled out. WHO recom- identified in 1997 in Hong Kong when it jumped from mends heightened surveillance for outbreaks in poultry poultry to humans and killed six of 18 infected people. and die-offs in migratory birds, and rapid introduction of Virtually all of the original cases were believed to have containment measures, as recommended by FAO and OIE. been bird-to-human transmission. Since that time there Heightened vigilance for cases of respiratory disease in have been a few hundred serologically confirmed cases in persons with a history of exposure to infected poultry is Cambodia, Indonesia, Thailand, and Vietnam, with high also recommended in countries with known poultry out- case fatality but no sustained evidence of ongoing human- breaks. The provision of clinical specimens and viruses, to-human transmission (WHO 2005b). from humans and animals, to WHO and OIE/FAO refer- The H5N1 strain is highly pathogenic among poultry. ence laboratories allows studies that contribute to the During 2003­2004 it resulted in outbreaks in 8 countries assessment of pandemic risk and helps ensure that work in Asia, with over 100 million birds dying from disease or towards vaccine development stays on course." being culled. More recently, though an additional 150 mil- Humans have little natural immunity to the H5N1 lion birds have been culled, because much of the develop- viruses. Thus, in contrast to the usual influenza epidemics, ing world's poultry economy depends on rural backyard which affect the very young, elderly, and those with 682 | Disease Control Priorities in Developing Countries | Frank E. Speizer, Susan Horton, Jane Batt, and others comorbid conditions, virtually the entire population in an Given the lack of natural immunity, there is considerable exposed community is at risk. In human cases of avian concern that even if adequate vaccines were available, distri- influenza, following the initial respiratory infection, bution on a worldwide basis would be limited by economic mortality results from two distinct processes. One process considerations as well as distribution problems in the devel- begins with relatively rapid onset of respiratory distress oping world. Efforts are underway to identify the genetic from hypoxia associated with ARDS.a The alternative make-up of the strains of H5N1 that will yield the most process results from secondary bacterial infection with a effective vaccines and to produce such vaccines in a cost- variety of organisms. In the documented H5N1 influenza effective manner. Testing H5N1 vaccines based on recently infections in humans, respiratory symptoms are most identified viruses in normal healthy volunteers suggests the prominent. However, in one case of encephalitis in a child immunologic response may be adequate, but several from Vietnam, H5N1 influenza virus was identified in months of production would be necessary to produce ade- cerebrospinal fluid and fecal matter, and in throat and quate supplies for one region, let alone for worldwide distri- serum samples. Isolates from several cases were resistant bution. Stockpiles of effective antiviral medications are to two commonly used antiviral medications (amantadine being generated in some countries. In the interim, WHO and rimantadine), while two other antiviral medications has encouraged the rapid reporting of cases and the estab- (oseltamivir and zanamivir) still appear to be effective. lishment of procedures for better public health intervention There is no way to predict the outcome of these ongoing strategies before and during a pandemic (WHO 2005c). events. What seems evident is that if human-to-human Many countries have developed pandemic influenza pre- transmission becomes established, a pandemic will follow. paredness plans in anticipation of such an event. Source: Authors. a. ARDS is defined as Acute Respiratory Distress Syndrome resulting from multiple causes, the most likely in this situation being an immunological reaction to the virus. Table 35.1 Public Health Measures in the SARS Episode, 2003 Procedure Comment 1. Isolation of patients Isolate rapidly after onset of symptoms. 2. Quarantine of contacts Usually at home, but separate from patients. When in contact with unexposed subjects, wear masks and avoid public transportation and visits to crowded places. 3. Education Reduce delay between onset of symptoms and isolation. In endemic areas, get subject to monitor temperature daily. Use fever hotlines, fever evaluation clinics. 4. Thermal screening Monitor temperature of travelers from endemic areas (not proven effective). 5. Increased social distance Cancel mass gatherings. Close schools, theaters, public facilities. Require use of masks in public settings. 6. Disinfection Practice frequent hand washing. Use aerosol disinfectant agents. 7. Travel advisories Postpone unessential travel. Screen travelers at entry and exit (not proven effective). Distribute health notices to travelers. Source: Data compiled and summarized from Bell 2004. controlling the spread of newly active influenza strains. The part of the reason the epidemic was contained as promptly as it potential for global spread and the occurrence of worldwide was (see table 35.1). However, because of the high case-fatality epidemics of influenza (presumed to be transmitted to humans rate, the disease caused significant disruption throughout the from domesticated or wild animals and then through close world. proximity to humans with symptomatic disease--generally to caregivers) points out the importance of continued surveil- lance for such episodes (Low and McGeer 2004). The lessons Economic Impact of Influenza and Cost-Effectiveness learned from the SARS epidemic reinforce the importance of of Interventions in the Developed World proven traditional public health measures, such as finding and Influenza is common in developed countries. Annually, it isolating cases, quarantining close contacts, and improving affects 10 to 20 percent of the U.S. population (Lee and others infection control practices (Bell 2004). Those methods, along 2002); those affected experience on average a loss of 2.8 with several other, less traditional efforts, were presumed to be workdays per episode. Those over 65 years of age are more Respiratory Diseases of Adults | 683 susceptible to complications, increased costs of hospitaliza- annual burden of influenza in the tropics from data on such tion, and even death. The cost of outbreaks can be large. The occasional and severe outbreaks. Because many areas (for costs of the 1996­97 epidemic in Germany were estimated example, Sub-Saharan Africa) do not have surveillance centers, at US$1,045 million, and the annual costs of outbreaks at not enough is known at this point to make policy recommen- US$11 million to US$18 million (WHO 2002a). dations. There are also no readily available estimates of the For those over age 65, many countries encourage preventive cost-effectiveness of influenza vaccination in those environ- vaccinations annually, on the basis of studies suggesting that ments. (For further discussion of the role of vaccination, see vaccination (either opportunistic or in a campaign) is cost- chapter 20.) effective in elderly populations (for example, see the model of Scuffham and West 2002). Given a good antigenic match, inac- tivated influenza vaccines prevent laboratory-confirmed illness CHRONIC RESPIRATORY DISEASES: in 70 to 90 percent of healthy adult vaccine recipients (WHO NATURE, CAUSES, AND BURDEN 2002a). Vaccination is less costly than chemoprophylaxis (with ion-channel inhibitors such as amantadine and rimantadine, COPD and asthma have very different diagnoses and causes; or with neuraminidase inhibitors such as zanamivir and hence, they are discussed in separate sections. However, the oseltamivir) or early treatment with the same drugs. In both treatments for these different chronic respiratory diseases share the institutionalized and the healthy elderly, vaccination sub- similarities, and that discussion is therefore combined. One of stantially reduces overall mortality from influenza (by 40 to the difficulties in defining COPD on a worldwide basis is that 68 percent). three distinct levels are used, depending on the sophistication The cost-effectiveness of vaccination for healthy working- of the health care system in the country where the patient is age adults, taking into account workdays lost, is a matter of being evaluated: debate. Demicheli and others (2000) concluded that the most cost-effective option for healthy adults age 14 to 60 was to take · Chronic bronchitis with and without obstruction, which may no action. However, these authors include only medical costs in be part of the COPD diagnosis, is defined by the presence of their calculations. Postma and others (2002) reviewed 11 stud- chronic cough and phlegm for three months per year for ies. Only one shows cost savings on the basis of medical costs two or more years and is generally assessed by standardized alone, but nine of them implied cost savings from vaccination questionnaires. if the value of lost work is included. Because of differences in · Obstructive airways disease is often assessed by reduced pul- costs and health care usage patterns, data on cost savings in monary function as measured by simple spirometry and the developed countries cannot be helpfully extrapolated to devel- presence of a reduced ratio of the forced expiratory volume oping countries. in one second (FEV1) divided by the vital capacity (VC). · For emphysema, which is also part of the syndrome of COPD, pulmonary function (changes in lung volume and Economic Impact of Influenza in the Developing World reduced diffusion capacity), x-ray evidence of bullae forma- In Hong Kong, China (where there is a milder year-round pat- tion, hyperinflation of the chest, and (with the use of high- tern of infection, little influenza-related mortality, and low resolution CT scanning) the presence of characteristic reported work losses), a model suggested that vaccination was changes in lung architecture all may contribute to the not cost saving, even if targeted to the elderly (Fitzner and diagnosis. others 2001). The only case for vaccination was if it controlled the emergence of highly virulent strains and prevented trans- What is apparent is that not all these diagnostic procedures mission to the rest of the world. According to the World Health are applied equally, particularly in the developing world; thus, Organization (WHO), much less is known about the impact of COPD may be seriously underreported. The 1998 Workshop influenza in the developing world. However, in the tropics, Report by the WHO and the National Institutes of Health where viral transmission normally continues year-round, (NIH) on "Global Strategy for the Diagnosis, Management, influenza outbreaks tend to have high attack and case-fatality and Prevention of COPD," developed as part of the Global rates. For example, during an influenza outbreak in Initiative for Chronic Obstructive Lung Disease (GOLD 2001), Madagascar in 2002, more than 27,000 cases were reported uses an international standard for defining the level of obstruc- within three months and 800 deaths occurred despite rapid tion from COPD. This strategy should improve worldwide intervention. An investigation of this outbreak, coordinated by estimates. This standard definition will still require the use of WHO, found that health consequences were severe in poorly equipment that measures pulmonary function (Buist 2002). nourished populations with limited access to adequate health Over the next several years, as the price and distribution of this care (WHO 2002b). It is not possible to extrapolate the exact equipment becomes more favorable and as more groups 684 | Disease Control Priorities in Developing Countries | Frank E. Speizer, Susan Horton, Jane Batt, and others a. Males Rates per million people Percent 25,000 16 14 20,000 12 10 15,000 8 6 10,000 4 2 5,000 0 15­29 30­44 45­59 60­69 70­79 80 Age group 0 15­29 30­44 45­59 60­69 70­79 80 b. Females Age group Percent Males Females 12 Source: WHO 2000. 10 8 Figure 35.2 COPD by Age and Sex, Worldwide 6 4 reduced mortality from COPD, presumably through a mecha- 2 nism that results in a modest improvement in pulmonary 0 function that appears to be related primarily to the extent of 15­29 30­44 45­59 60­69 70­79 80 Age group chronic bronchitis and mucus hypersecretion (Scanlon and others 2000; Speizer and others 1989). Within a few years of COPD Asthma Other respiratory diseases Tuberculosis stopping smoking, smokers' rate of decline of pulmonary func- Source: WHO 2000. tion (that is, FEV1) returns to the rate found in nonsmokers, although little of the lost pulmonary function is regained Figure 35.1 Chronic Respiratory Diseases DALYs as a Percentage of (Fletcher and others 1976). Similar effects are seen in the World Totals developing world. However, because smoking is far less prevalent in developing countries, especially among women, other exposures are related to the development of disease (see undertake the training in its use and in the interpretation of also chapter 46). One of the most important exposures, partic- results from the tests, diagnostic uniformity will improve. ularly for women, is to unvented coal-fired cooking stoves, Unfortunately, as pointed out by Aït-Khaled, Enarson, and starting during childhood and continuing into adult life (see Bousquet (2001), the applicability of these guidelines has not chapter 42). been effectively tested in developing countries. Because the interventions and treatments for COPD overlap In adults, COPD dominates all other chronic respiratory with those for asthma, they will be treated together. diseases in accounting for 2 percent to more than 10 percent The diagnosis of asthma has been debated for centuries. of lost disability-adjusted life years (DALYs) on a worldwide Health care providers can generally agree on the diagnosis in basis. Its incidence increases dramatically with age (fig- the individual patient who is wheezing and in whom other eti- ures 35.1a and 35.1b). Of note, mortality from COPD is low ologic factors are ruled out. They would also agree on the def- before age 45. Over age 45, death rates increase from 50 to 200 inition of the disease as an inflammatory response in the per 10,000 individuals and are consistent across age groups in airways that results in variable and generally reversible airflow men and women, with the exception of death rates in women obstruction with or without treatment. However, depending over age 80, which exceed those in men in that age group on the training of health care providers, the nature of surveil- (figure 35.2). lance, the characteristics of a given community, and the par- Much of COPD in the developed world is related to cigarette ticular environment of the community, the accuracy of the smoking, and there is no question that progression of the dis- estimate of the prevalence of asthma in a community may vary ease is related to the number of cigarettes smoked and the years much more. The reported prevalence of the disease may be of smoking. Smoking cessation has been associated with based on no more than an answer to this question: "Has a Respiratory Diseases of Adults | 685 provider ever told you that you (or your child) has had Economic Impact of Asthma in the Developing World asthma?" The response to this question has been validated in a Data for developing countries are much scarcer. For Estonia, number of studies. In contrast, the diagnosis may depend on Kiivet and others (2001, cited in Lee and Weiss 2002) estimated examination of the patient's chest, physiological testing, the direct annual costs of asthma to be US$104 per year per responsiveness to provocative stimuli to the airways, and asthma patient, equivalent to 1.4 percent of direct health care specific response to therapy. Thus, estimates of community costs. In Singapore, medical costs for asthma constitute 1.3 per- burden from asthma may depend on the threshold used in cent of total health care costs (Chew, Goh, and Lee 1999, cited making the diagnosis. in WHO 2001). Despite variations in diagnostic criteria, worldwide esti- One study (Aït-Khaled, Enarson, and Bousquet 2001, cited mates of the asthma burden among adults have generally in Weiss and Sullivan 2001) found that asthma drugs cost come from surveys within selected communities. In contrast between 3.8 and 25 percent of the patient's monthly income in to other adult respiratory diseases, the prevalence of asthma is 24 developing countries in Asia and Africa. K. R. Smith (2000) relatively low (figures 35.1a and 35.1b). In adults, the DALYs estimates the burden of respiratory disease in India that is for asthma are at a peak of about 2 percent of the total world- attributable to indoor air pollution (only a fraction of all respi- wide in people age 15 to 29, and they decline in each older age ratory disease) as 1.6 billion to 2 billion sick days per year. Of group. This pattern is also reflected in mortality rates, with the that total, asthma is responsible for about one-third, acute res- highest rates occurring in young people and equal rates in piratory infection is responsible for about one-third, and the men and women about age 60. After age 60, reported rates of remainder is attributable to COPD, tuberculosis, and ischemic death caused by asthma in men begin to exceed those in heart disease. Asthma and COPD combined account for 44 per- women, and both become substantial. That shift reflects pri- cent of the burden. marily either increasingly questionable diagnostic accuracy or misclassification of other obstructive respiratory diseases such Cost Effectiveness of Interventions for COPD and Asthma as COPD. in Developed Countries Five recent overviews of the economics of chronic respiratory Economic Impact of Asthma and COPD disease, COPD, and asthma (Friedmann and Hilleman 2001; in the Developed World Lee and Weiss 2002; Ruchlin and Dasbach 2001; Sullivan and In the United Kingdom (where asthma rates are particularly Weiss 2001; Weiss and Sullivan 2001), in addition to many indi- high), respiratory disease accounts for 6.5 percent of hospital vidual studies, focus on developed countries.1 Only a limited admissions. Fifteen percent of the working population report number of studies use cost- or quality-adjusted life years work-limiting health problems caused by respiratory disease, (QALYs) saved as the outcome (others use life years saved). and 18.3 million workdays were lost to asthma problems in (Studies focusing on intermediate health outcomes and on cost 1995­96 (Chung and others 2002). minimization are not discussed here.) In general, costs in In the Netherlands, annual costs associated with asthma developing countries would be about 20 percent of those and COPD (direct and indirect) were estimated to exceed reported here, according to detailed unit cost data by region US$500 million for a population of about 14 million (data for from WHO-CHOICE (Choosing Interventions That Are Cost- the 1980s). Asthma or COPD was responsible for 3 percent of Effective) and on comparisons of respiratory drug prices from absenteeism caused by illness, and asthma was also the main online pharmacies in the United States and from the reason for absence from school among children age 4 to 12 International Drug Price Indicator Guide (http://erc.msh.org). (Rutten­van Mölken and others 1992). The exceptions are interventions involving nondiscounted In the United States in the early 1990s, health care costs drugs that are still under strictly enforced patents, for which the attributable to respiratory disease were US$11 billion (about costs in developing countries would be closer to those in the 2 percent of total health care costs), and an estimated 3 million United States. Table 35.2 summarizes the results. workdays and 10 million schooldays were lost to respiratory Inhaled salbutamol (short-acting beta-2 agonist) is the first disease (Stoloff, Poinsett-Holmes, and Dorinsky 2002). line of treatment for both intermittent asthma (daytime symp- Another survey (Weiss and Sullivan 2001) estimated the toms less than once per week, nocturnal symptoms less than costs of asthma in 1991 US dollars for four developed countries twice per month, and normal spirometry between episodes) (Australia, Sweden, the United Kingdom, and the United and COPD (mild to severe) in both developed and developing States) and one state (New South Wales in Australia). Per countries. This treatment became standard practice beginning patient costs of asthma ranged from US$326 (Australia) to in the 1970s, so there are no cost-effectiveness studies of salbu- US$1,315 (Sweden) annually, with direct costs accounting, in tamol compared with placebo. This medical intervention is most cases, for more than half of total costs. likely the most cost-effective one, but it is still likely to cost 686 | Disease Control Priorities in Developing Countries | Frank E. Speizer, Susan Horton, Jane Batt, and others Table 35.2 Cost-Effectiveness of Interventions for Asthma and COPD in 2001 Reference Intervention Alternative Cost-effectiveness Pharmacological Authors' estimates Inhaled ipratropium bromide Placebo US$6,700­US$8,900/QALY for moderate COPD Paltiel and others 2001 Quick relieversa and inhaled Quick relievers only US$13,900/QALY in adults with mild to moderate corticosteroids asthma: US$10,600 for moderate only Van den Boom and others 2001 Inhaled corticosteroid Placebo US$13,400/QALY COPD treatment (fluticasone propionate) Akins and O'Malley 2000 A-1 antitrypsin augmentation therapy Standard careb US$14,400/QALY, severely deficient individuals Hay and Robin 1991, A-1 antitrypsin augmentation therapy Standard careb US$45,000­US$215,000/life year, depends on age, in Ruchlin and Dasbach 2001 efficacy, and so forth Education Toevs, Kaplan, and Atkins 1984, Education and exercise program Exercise program only US$71,500/QALY in Ruchlin and Dasbach 2001 Long-term oxygen Authors' estimate Home oxygen therapy for COPD No oxygen US$19,000/life year (US$26,700­US$38,000/QALY) Mechanical ventilation Schmidt and others 1983, in Mechanical ventilation Standard hospital carec US$6,400­US$23,600/life year (COPD, asthma, cardiac Rutten­van Mölken (excluding ventilation) patients) excluding physician costs Anon and others 1999 in Mechanical ventilation in Standard hospital carec US$35,000­US$60,700/QALY Ruchlin and Dasbach 2001 intensive care unit, asthma, (excluding ventilation) and COPD patients Surgery Al and others 1998, in Lung transplant in end-stage disease No transplant US$464,000/QALY Ruchlin and Dasbach 2001 Ramsey and others 1995, Lung transplant in those eligible No transplant US$238,200/QALY in Ruchlin and Dasbach 2001 a. Quick relievers refer to rapid-acting bronchodilators (for example, salbutamol) that act to relieve bronchoconstriction and accompanying acute symptoms of wheeze, chest tightness, and cough. b. Standard care includes medical management (ipratropium bromide, beta-2 agonist, steroid) and home oxygen as needed. c. Standard care includes medical management and oxygen. some thousands of dollars per life year saved in the United COPD. The cost per QALY is likely to be lower for severe asth- States. ma, but ethical considerations render random controlled trials The next line of treatment currently recommended for unfeasible. developing countries is inhaled corticosteroids (for example, No cost-effectiveness study could be found for ipratropium beclomethasone) for mild to severe persistent asthma (disease bromide compared with placebo. We estimate that the cost per ranging from daytime symptoms greater than once per week, QALY saved would be between one-half and two-thirds of that nocturnal symptoms more than twice a month, and normal for a new-generation inhaled steroid such as fluticasone spirometry between episodes to daily frequent symptoms asso- propionate. This estimate, which is based on the relative cost ciated with severe obstruction) and inhaled ipratropium bro- of the two drugs in the United States and assumes similar mide for COPD. Both first-generation corticosteroids and ipra- effectiveness of the two drugs, would put the cost of ipratro- tropium bromide are off patent. However, as pointed out by pium bromide between US$6,700 and US$8,900 per QALY. Chan-Yeung and others (2004), the use of corticosteroids either Most of the other interventions summarized in table 35.2 intermittently or chronically is commonly recommended in have a higher cost per QALY. For individuals who develop developed countries, where the background level of tuberculo- COPD related to a severe deficiency in alpha-1 antitrypsin, sis among patients is considerably lower. In developing coun- alpha-1 antitrypsin therapy is sometimes considered, at a cost tries with higher tuberculosis rates, corticosteroids must be of between US$45,000 and US$215,000 per life year.2 The use used with greater caution. of long-acting beta-2 agonists and leukotriene modifiers is now Inhaled steroids cost about US$13,900 per QALY for mild an accepted and integrated component of the treatment of to moderate asthma or when used in early treatment of moderate to severe asthma in the developed world. However, Respiratory Diseases of Adults | 687 the cost savings for the developing world are difficult to pressure ventilation, where it is feasible, is less costly than inva- demonstrate because the endpoints of studies using those sive mechanical ventilation for specific indications. Finally, drugs are often changes in spirometric testing, improved costs of lung transplants are at a level scarcely affordable even quality-of-life measures, steroid-sparing effects, or altered in developed countries; Al and others (1998, cited in Ruchlin hospital admission rates. and Dasbach 2001) estimated costs at US$464,000 per QALY, Likewise, oral or intravenous steroids play a crucial role in and Ramsey and others (1995, cited in Ruchlin and Dasbach the treatment of acute exacerbations in both asthma and COPD, 2001) estimated costs at US$238,000 per QALY (in 2000 U.S. but endpoint assessments in studies typically address decreases dollars). in the duration of hospital stays and increases in the use of All those interventions compare unfavorably with the cost- emergency department facilities, which result in decreases in effectiveness of smoking prevention for preventing COPD health costs in the developed world. Oral steroids are inexpen- (discussed in chapter 46). Smoking prevention is one of the sive, even by standards in developing countries, and in the short most cost-effective health interventions that exists, and there is term might appear to be cost-effective, but they are associated a strong case for moving resources from expensive curative with major medium- to long-term consequences and are not interventions to that intervention. Likewise, prevention of recommended as standard therapy. COPD by switching the cooking source from unventilated Educational programs tend to be cost saving in developed stoves that burn biomass to either improved stoves or kerosene countries, where uncontrolled exacerbations are extremely stoves is more cost-effective than treatment (see chapter 42). costly in terms of hospital care (six such programs are surveyed in Van Mölken and others 1992 and one in Ruchlin and Dasbach 2001). Similarly, exercise rehabilitation programs (six Cost-Effectiveness of Interventions for COPD and Asthma surveyed in Ruchlin and Dasbach 2001) can also be cost sav- in Developing Countries ing. WHO (2001) has commented on cost savings achieved by It is difficult to transfer the costs per QALY saved in developed education programs for asthma from four different U.S. stud- countries to developing countries. The cost of patented drugs ies. Only one of these studies addressed cost per well year, in developing countries should be the same as that in devel- which was estimated at US$71,500 in 2001 (Toevs, Kaplan, and oped countries, whereas the costs of education and of the time Atkins 1984, cited in Ruchlin and Dasbach 2001). However, of medical personnel should be substantially lower (on the there are likely to be monetary savings from fewer workdays order of 20 percent of U.S. levels). In practice, the costs of off- lost, which are not factored into this analysis. patent drugs also vary considerably. Beclomethasone dipropi- WHO (2001) surveyed one self-management training pro- onate (one of the older, off-patent inhaled steroids) is available gram for chronic asthma in India (Ghosh and others 1998), for about US$15 per 200-dose inhaler in Canada in online which resulted in improvements in health status, reduced use pharmacies but is quoted at US$1 to US$3 by agencies and of emergency departments and hospitals, and savings on health suppliers on the International Drug Price Indicator Guide costs. Sudre and others (1999) pointed out that studies of edu- (http://erc.msh.org). A similar price difference exists for salbu- cation programs tend not to provide a good description of the tamol inhalers. Hence, the most cost-effective therapies in actual program content and that a more systematic description developed countries (inhaled salbutamol and first-generation of these interventions needs to be promoted to replicate best corticosteroids for asthma and ipratropium bromide for practice. COPD) are also likely to be cost-effective in the wealthier devel- Long-term oxygen therapy is a life-prolonging intervention oping countries--or more broadly if inexpensive drug supplies in advanced stages of COPD. Recent studies do not quantify the are available. Those drugs are likely to be particularly cost- cost per QALY but instead compare different methods of oxy- effective for those with severe and moderately severe asthma or gen delivery (cylinder or concentrator). These authors' crude COPD but less cost-effective for those with mild disease. estimate for long-term oxygen use is US$19,000 per life year Recent practice suggests that a combination of long-acting beta saved.3 If the quality-of-life scores of patients on long-term agonists and inhaled corticosteroids can control moderate to oxygen were 0.8 or 0.6, the cost per QALY would be US$22,750 severe disease more rapidly. However, to make this form of or US$31,700, respectively. (K. J. Smith and Pesce 1994, cited in therapy cost-effective, the patient needs to be reevaluated to the Harvard Catalogue of Preference Scores, assign a median determine whether one or the other drug can be removed. score of 0.4 to quality of life for patients with severe COPD Because of cost considerations, that may not be feasible in the with high supportive care needs and poor functional status.) developing world. In hospitals, mechanical ventilation in the intensive care Once control has been obtained, education alone appears to unit has been estimated to cost US$35,000 to US$60,700 per be ineffective when only respiratory outcomes are considered QALY in 2001 (Anon and others 1999, cited in Ruchlin and (although education on the benefits of exercise has other health Dasbach 2001). Studies suggest that noninvasive positive benefits: see chapter 44 on lifestyles). However, education 688 | Disease Control Priorities in Developing Countries | Frank E. Speizer, Susan Horton, Jane Batt, and others addressing the appropriate use of medication is extremely There are few reliable estimates of the global burden of important, particularly in developing countries, where timely occupation-related respiratory diseases. Because of the lack of emergency care for severe exacerbations may not be readily systematic surveillance in most developing countries, the few available. Although the cost of educational efforts would be published estimates of occupation-related respiratory diseases expected to be considerably lower in developing countries, this have relied on selected studies involving particular industries area requires more systematic research. that investigators have had unique opportunities to explore. Long-term oxygen is also an option for high-income house- For example, Trapido and others (1996) conducted a survey in holds in middle-income developing countries. The costs are a relatively small group of former mineworkers and found that likely to be lower than in developed countries. In Brazil the approximately 55 percent had pneumoconiosis with or without monthly cost for supplemental home oxygen therapy is close tuberculosis. They estimated that about 25 percent of migrant to US$150 (Sant'Anna and others 2003), compared with and former mineworkers in South African gold mines with US$400 per month paid by Medicare, which would bring the 15 to 25 years of exposure had occupational lung diseases. cost-effectiveness to US$7,000 per life year by these authors' Loewenson (1999) pointed out the difficulties in making crude estimates,or between US$8,750 and US$11,700 per QALY. assessments of occupational risk throughout the African coun- Publicly funded systems are unlikely to be able to pay this rate, tries and suggested a series of methodological issues that need although private insurers and wealthy households might pay to be considered. because such therapy prolongs life. Leigh and others (1999) estimated the global burden of dis- The other interventions in table 35.2 are likely to be too eases related to occupational factors at 4.2 million to 10.0 mil- expensive for most developing countries to use at present. lion cases per year. If one subtracts the rates for established market economy countries, the total burden for the rest of the OCCUPATIONAL LUNG DISEASE AND OTHER world is approximately 3.4 million to 9.1 million cases per year. RESPIRATORY DISEASES Using limited data and applying rates from individual nations and regional groups of countries, the authors made an indirect Although occupational lung diseases are often considered dis- calculation for the expected number of cases of occupation- eases of the industrial world, they are occurring with increased related diseases globally. Figure 35.3 summarizes their esti- frequency in the developing world, where guidelines for worker mates for pneumoconiosis and other chronic respiratory dis- safety are generally more lax or nonexistent. In addition, eases by age and gender. Notably, these two categories of disease because of increased migration from rural areas to more account for approximately 30 percent of all occupational dis- urbanized centers and the transfer of major manufacturing eases. The prevalence of these diseases increases with age and is activities from the developed market economy countries to the higher among men. less developed countries, the number of employees with potentially harmful occupational exposures has increased exponentially in the past 30 years. The general discussion of occupation-related diseases is reviewed in chapter 60. We focus Rates per million people here on specific occupation-related lung diseases. 2,500 Occupational lung diseases are, for the most part, charac- terized as related to particular occupational exposures and 2,000 generally fall into two broad pathophysiological types. One type may result in pulmonary fibrosis, which is manifested by 1,500 restricted lung volume and decreased diffusion capacity on pulmonary function testing and increased interstitial pul- 1,000 monary markings on chest x-ray. Certain occupational lung diseases, such as silicosis, are complicated by a substantially 500 increased risk of tuberculosis, which contributes to the overall burden of respiratory disease in the developing world. The 0 second pattern of occupational lung disease is that of obstruc- 0­14 15­44 45­59 60 Age group tive airways disease, which may be reversible (occupational asthma) or irreversible (chronic bronchitis with or without Males Females obstruction or emphysema or COPD), in which the chest x- Source: Leigh and others 1999. ray often is negative and the diagnosis is dependent largely on reported histories of exposures, symptoms, and pulmonary Figure 35.3 Estimated Combined Pneumoconiosis and Other function testing. Occupation-Related Chronic Respiratory Diseases Respiratory Diseases of Adults | 689 Asbestosis and asbestos-related cancers present a particular immunologically related pulmonary diseases most often asso- problem in developing countries. Asbestosis can manifest both ciated with environmental exposures to specific inhaled anti- as other interstitial lung disease, as described above, and as gens or interstitial inflammation and fibrosis, often of obstructive airways disease. In addition, occupational expo- unknown origin. In the developing world, little systematic sure is associated with the occurrence of lung cancer, and work has been done on these diseases to assess incidence or according to studies in developed countries, the rate of occur- prevalence. These conditions probably occur considerably less rence is synergistically associated with smoking. Because the frequently than asthma and COPD. However, they are likely cost of health care compensation in the developed world to have a higher prevalence in developing countries than is exceeds the potential profit from mining and manufacturing reported in the developed world simply because of the pre- of asbestos products, much of the industry has moved to the sumed associations with exposures to organic dusts and the developing world. increased prevalence of malnutrition (see chapter 56), both of LaDou (2004) has recently summarized the status of the which are likely to occur in more rural and less developed areas potential for reducing occupational exposure on a worldwide of the world. basis and suggests that upward of 10 million lives will be lost if See chapters 16, 18, and 29 for interventions for tuberculo- the current lack of controls and continued increases in mining sis, AIDS-related lung disease, and lung cancer, respectively. and manufacturing are not changed. In 2000, more than 2 mil- Managing immunologic and fibrotic respiratory diseases with lion tons of asbestos products were produced, whereas 25 years medication is extremely difficult and expensive. Therapeutic earlier the total production was 350,000 tons each year. Except trials often fail, presumably because the treatments are not for the Russian Federation and Canada, virtually all the larger aimed at a particular antigen. The most effective way of manag- producers are in the developing world, where the recognition ing these respiratory diseases is to reduce exposure to the and reporting of health effects are less well established. The inciting agents, an approach that hinges on two strong prem- likelihood of reversing this trend and developing an interna- ises, which are not always applicable in the developed world. tional ban on asbestos use is small, particularly because it is the First, the disease must be recognized as related to a common nations that produce more asbestos products that are, in fact, environmental contaminant encountered in an occupational or increasing consumption. avocational exposure--for example, exposure to thermophilic The economic burden of occupational lung diseases is sur- actinomycetes in moldy hay or sugarcane results in farmer's prisingly difficult to document. Most developed countries and lung, and exposure to bird feathers or droppings results in bird some developing countries (for example, South Africa) have fancier's disease. Second, community resources must be directed legislation protecting workers from exposure and compensat- toward educating the public about the importance of limiting ing those who have contracted chronic conditions. In the exposure to these agents. United States, compensation payments from the Social Security Administration and the Department of Labor for black lung disease totaled US$1.6 billion in 1996 (NIOSH 1999). Data GENERAL APPROACH TO LOWERING RISK exist on compensation for claims for various occupational lung OF ADULT RESPIRATORY DISEASE diseases for the United States and the European Union coun- tries. However, claims data represent only a small fraction of Although interventions of various sorts are indicated for each the true economic cost (for example, not all workers make of the disease categories discussed, these interventions are often claims; compensation payments lag considerably). For the costly and sometimes ineffective in lowering or preventing pre- United States, the annual costs of complying with the revised mature mortality. Thus, from an operational perspective, it is respirator standards for 1993 were US$111 million for about important to consider preventive and therapeutic strategies that 5 million workers needing to use a respirator (presumably will have greater societal effect than will the management of the these costs of prevention were far lower than the economic cost manifestations of diseases as they arise in individuals. This of unprotected work) (OSHA 1998). The primary treatment approach applies to acute diseases (vaccination schemes to for affected workers is to remove them from the inciting expo- reduce the burden of influenza, in contrast to individual man- sures. (See chapter 60 for discussion of preventive strategies agement of community-acquired pneumonia) and chronic dis- that need to be considered to reduce the risk of occupational eases (smoking prevention and reduction programs, compared disease.) with availability of routine asthma medication). Primary pre- Some of the other major classes of adult respiratory diseases vention strategies should include efforts by multiple agencies of are discussed in other chapters: tuberculosis, in chapter 16; government and the community coming together to establish AIDS-related lung disease, in chapter 18; and lung cancer, in appropriate priorities for action. Four sources of exposure chapter 29. Other diseases that have been studied, particularly stand out: tobacco smoke, indoor smoke, outdoor air pollu- in the developed world, include the hypersensitivity or tants, and occupational exposure (see chapters 46, 42, 43, 690 | Disease Control Priorities in Developing Countries | Frank E. Speizer, Susan Horton, Jane Batt, and others and 60, respectively). Of these, the most pressing and cost- nities for research that go beyond the primary prevention that effective is a cohesive policy to control tobacco smoking. would result from better smoking control policies. Because not In conjunction with the International Union against all smokers are at increased risk, the interaction of smoking Tuberculosis and Lung Disease (IUATLD) and selected univer- with nutritional status (including micronutrient status), with sities and health institutions in various countries, WHO is genetic factors that determine susceptibility, and with respira- developing the Practical Approach to Lung Health (PAL, previ- tory infections may act as a precursor of susceptibility to envi- ously known as the Adult Lung Health Initiative). The program ronmental (ambient or occupational) pollution and personal is focused on improving primary care services, as well as appro- (smoking) pollution. Similarly, the role of immunologic stim- priate referral to secondary health care facilities, for individuals ulation or immunocompetence needs further exploration as it with tuberculosis, acute respiratory infections (especially pneu- relates to the development of asthma. Synergies between the monia), asthma, and COPD. Four countries (Chile, Morocco, conditions discussed in this chapter and other infections (par- Nepal, and South Africa) are serving as the pilot implementa- ticularly tuberculosis, but possibly others) may be especially tion sites (WHO 2003). important in the developing world. Finally, specific environ- In Chile, where respiratory symptoms account for one-third mental conditions--such as altitude, heat and cold stress, and of primary health care visits, a respiratory disease program was increased ambient pollution from rapid urbanization--and initiated in 2001 as part of ongoing efforts to strengthen pri- their effects on asthma and COPD should be explored. mary health care. The pilot program was implemented in 15 Acute respiratory infections, specifically bacterial pneumo- centers. Standard formats are used to devise scores to deter- nia, have not been addressed nearly as well for adults as they mine follow-up for asthma and COPD. Sentinel centers are have been for children. For example, simple data on the preva- used to provide epidemiologic information. Influenza immu- lence of infecting organisms, typical susceptibilities, the ability nization coverage of the elderly and at-risk population has to train ancillary workers in clinical diagnosis, and the correla- reached 85 percent (WHO 2003). tion of clinical assessment with verified disease would be help- In Morocco, survey work done before establishing a PAL ful in establishing the feasibility of assessment and treatment at strategy showed that 31 percent of patients who consult home versus at a clinic or hospital, specifically in the develop- primary health care centers present with respiratory symp- ing world. Common etiologic agents in North America are toms. Of those patients, 85 percent have acute respiratory common elsewhere; therefore, treatment of disease would be infections, 14 percent have chronic conditions, and 1 percent relatively inexpensive. Most community-acquired acute disease have tuberculosis. In Mexico, an IUATLD study implementing responds relatively well. Certainly, penicillin should be recom- asthma control measures was shown to be cost-effective. mended as a first-line drug for community-acquired pneumo- Control of asthma improved, and the majority of patients nia. Educating local healers on the importance of initiating experienced a decrease in the severity of asthma. The cost of treatment earlier in the course of disease would translate into asthma management decreased because of lower costs for savings with respect to decreased days of work lost and reduced emergency services and hospitalizations (WHO 2003). case-fatality rates. Follow-up monitoring and the development of hospital-based bacteriologic testing should be expanded to identify and control for the emergence of resistant bacterial FUTURE RESEARCH NEEDS strains. Studies of asthma in the developed world have been exten- One of the difficulties in quantifying the burden of respiratory sive and of extremely high quality but are directed specifically diseases in adults is the inability to apply uniform methods of toward the health care structures in which they are tested. diagnosis across economies in which sophisticated diagnostic Specific cost-effectiveness studies in the developing world procedures are possible, let alone across less developed should be done to see, for example, whether a focus on disease economies. The problems relate in part to differences in the education and modification of risk factors in addition to med- language describing the same symptoms, levels of registration ications outweighs simple administration of medications (with of census and disease reporting, availability of diagnostic pro- instructions on use). In the developed world, there is no ques- cedures, and reluctance to make accurate estimates because of tion that an approach that is multitiered and involves multiple the cost of intervention strategies. Furthermore, unless controls health care providers is the best, but we still do not have on cigarette smoking are initiated, little progress in stemming concrete evidence of where monies are best spent in the devel- the increasing burden of chronic respiratory diseases can be oping world. Another possible fruitful area of research is on expected. education programs. Most of the literature relates to education There are still a number of unanswered questions related to programs for specific entities (for example, "asthma triggers") COPD, which remains the dominant respiratory disease in and their costs in developed countries. Education programs adults. The developing world provides some unique opportu- in developing countries that are multidimensional (smoking Respiratory Diseases of Adults | 691 cessation, indoor air quality, vaccination) are likely to be rela- Chung, F., N. Barnes, M. Allen, R. Angus, P. Corris, A. Knox, and others. tively inexpensive and cost-effective. Better methods of educat- 2002. "Assessing the Burden of Respiratory Disease in the U.K." Respiratory Medicine 96: 963­75. ing local healers through the use of demonstration projects Crockett, A. J., J. M. Cranston, J. R. Moss, and J. H. Alpers. 2001. "A Review should be tested, as should more efficient distribution systems of Long-Term Oxygen Therapy for Chronic Obstructive Pulmonary to make relatively inexpensive medication available. General Disease." Respiratory Medicine 95: 437­43. increased awareness of the impact of symptoms on adults and Demicheli, V., T. Jefferson, D. Rivetti, and J. Deeks. 2000. "Prevention and of the potential for earlier intervention in a disease should also Early Treatment of Influenza in Healthy Adults." Vaccine 18: 957­1030. be explored and tested for their effects on reducing respiratory Dolin, P. J., M. C. Raviglione, and A. Kochi. 1994. "Global Tuberculosis Incidence and Mortality during 1990­2000." Bulletin of the World disease burdens. Health Organization 72: 212­20. Fitzner, K. A., K. F. Shortridge, S. M. McGhee, and A. J. Hedley. 2001. "Cost-Effectiveness Study on Influenza Prevention in Hong Kong." Health Policy 56: 215­34. NOTES Fletcher, C. M., R. Peto, C. M. Tinker, and F. E. Speizer. 1976. The Natural History of Chronic Bronchitis: An Eight Year Follow-up Study of Working 1. The survey of the cost-effectiveness of interventions below is based Men in London. New York: Oxford University Press. on a review of the University of York database (http://www.york.ac.uk/ inst/crd), combined with a Medline search (focusing mainly on data after Friedmann, M., and D. E. Hilleman. 2001. "Economic Burden of Chronic 1996). Obstructive Pulmonary Disease: Impact of New Treatment Options." 2. According to Hay and Robin (1991), cited in Ruchlin and Dasbach Pharmacoeconomics 19 (3): 245­54. (2001). Akins and O'Malley (2000) have a much lower estimate, which Ghosh, C. S., P. Ravindran, M. Joshi, and S. C. Stearns. 1998. "Reductions probably does not include all the costs of screening and the like. in Hospital Use from Self Management Training for Chronic 3. This estimate was calculated as follows: the MRC (1981) trials sug- Asthmatics." Social Science and Medicine 46 (8): 1087­93. gest that over five years the mortality in a randomized trial for patients GOLD (Global Initiative for Chronic Obstructive Lung Disease). 2001. with severe hypoxemia is 667 per 100,000 for those not treated with long- "Global Strategy for the Diagnosis, Management and Prevention of term oxygen, compared with 548 per 100,000 for those treated with long- GOLD." National Heart, Lung, and Blood Institute­WHO Workshop term oxygen (reviewed in Crockett and others 2001). The cost per month Report, NHLBI Publication 2701, National Institutes of Health, of home oxygen is taken as US$400 (based on U.S. Medicare reimburse- Bethesda, MD. ments in the early 1990s). Hay, J. W., and E. C. Robin. 1991. "Cost-Effectiveness of Alpha-1 Antitrypsin Replacement Therapy in Treatment of Congenital Chronic Obstructive Pulmonary Disease." American Journal of Public Health 81: REFERENCES 427­33. Hooi, L. N., I. Looi, and A. J. Ng. 2001. "A Study on Community Acquired Aït-Khaled, N., D. Enarson, and J. Bousquet. 2001. "Chronic Respiratory Pneumonia in Adults Requiring Hospital Admission in Penang." Diseases in Developing Countries: The Burden and Strategies for Medical Journal of Malaysia 56: 275­84. Prevention and Management." Bulletin of the World Health Hui, K. P., N. K. Chin, K. Chow, A. Brownlee, T. C. Yeo, G. Kumarasinghe, Organization 79 (10): 971­79. and others. 1993. "Prospective Study of the Etiology of Adult Akins, S. A., and P. O'Malley. 2000. "Should Health-Care Systems Pay for Community Acquired Bacterial Pneumonia Needing Hospitalisation Replacement Therapy in Patients with Alpha(1)-Antitrypsin in Singapore." Singapore Medical Journal 34: 329­34. Deficiency? A Critical Review and Cost-Effectiveness Analysis." Chest Kiivet, R. A., I. Kaur, A. Lang, A. Aaviksoo, and L. Nirk. 2001. "Costs of 117 (3): 875­80. Asthma Treatment in Estonia." European Journal of Public Health 11: Al, M. J., M. A. Koopmanschap, P. J. van Enckevort, A. Geertsma, 89­92. W. van der Bij, W. J. de Boer, E. M. TenVergert. 1998."Cost-Effectiveness LaDou, J. 2004. "The Asbestos Cancer Epidemic." Environmental Health of Lung Transplantation in the Netherlands." Chest 113: 124­40. Perspectives 112: 285­90. Anon, J. M., A. Garcia de Lorenzo, A. Zarazaga, V. Gomez-Tello, and Lee, P. Y., D. V. Matchar, D. A. Clements, J. Huber, J. D. Hamilton, and E. D. G. Garrido. 1999. "Mechanical Ventilation of Patients on Long-Term Peterson. 2002. "Economic Analysis of Influenza Vaccination and Oxygen Therapy with Acute Exacerbations of Chronic Obstructive Antiviral Treatment for Healthy Working Adults." Annals of Internal Pulmonary Disease: Prognosis and Cost-Utility Analysis." Intensive Medicine 137: 225­31. Care Medicine 25 (5): 452­57. Lee, T. A., and K. B. Weiss. 2002. "An Update on the Health Economics of Bell, D. M. 2004. "World Health Organization Working Group on Asthma and Allergy." Current Opinion in Allergy and Clinical Prevention of International and Community Transmission of Immunology 2: 195­200. SARS. Public Health Interventions and SARS Spread, 2003." Emerging Infectious Diseases 10 (11). http://www.cdc.gov/ncidod/EID/ Leigh, J., P. Macaskill, E. Kuosma, and J. Mandryk. 1999. "Global Burden of vol10no11/04-0729.htm. Disease and Injury Due to Occupational Factors." Epidemiology 10: 626­31. Buist, A. S. 2002. "Guidelines for the Management of Chronic Obstructive Pulmonary Disease." Respiratory Medicine 96 (Suppl. C): S11­16. Lieberman, D., F. Schlaeffer, I. Boldur, D. Lieberman, S. Horowitz, M. G. Chan-Yeung, M., N. Aït-Khaled, N. White, K. W. Tsang, and W. C. Tan. Friedman, and others. 1996. "Multiple Pathogens in Adult 2004. "Management of Chronic Pulmonary Disease in Asia and Patients Admitted with Community-Acquired Pneumonia: A One Africa." International Journal of Tuberculosis and Lung Disease 8 (2): Year Prospective Study of 346 Consecutive Patients." Thorax 51: 159­70. 179­84. Chew, F. T., D. Y. Goh, and B. W. Lee. 1999."The Economic Cost of Asthma Loewenson, R. 1999. "Assessment of the Health Impact of Occupational in Singapore." Australian and New Zealand Journal of Medicine 29 (2): Risk in Africa: Current Situation and Methodological Issues." 228­33. Epidemiology 10: 632­9. 692 | Disease Control Priorities in Developing Countries | Frank E. Speizer, Susan Horton, Jane Batt, and others Low, D. E., and A. McGeer. 2004. "SARS--One Year Later." New England Smith, K. R. 2000. "National Burden of Disease in India from Indoor Air Journal of Medicine 349: 2381­2. Pollution." Proceedings of the National Academy of Sciences of the United MRC (Medical Research Council Working Group). 1981. "Long-Term States of America 97 (24): 13286­93. Domiciliary Oxygen Therapy in Chronic Hypoxic Cor Pulmonale Speizer, F. E., M. E. Fay, D. W. Dockery, and B. G. Ferris Jr. 1989. "Chronic Complicating Chronic Bronchitis and Emphysema." Lancet 1 (8222): Obstructive Pulmonary Disease Mortality in Six U.S. Cities." American 681­86. Review of Respiratory Disease 140: S49­55. Murray, C. J. L., and A. D. Lopez, eds. 1996. The Global Burden of Disease: Stoloff, S., K. Poinsett-Holmes, and P. M. Dorinsky. 2002. "Combination A Comprehensive Assessment of Mortality and Disability from Diseases, Therapy with Inhaled Long-Acting 2-Agonists and Inhaled Injuries and Risk Factors in 1990 and Projected to 2020. Cambridge, MA: Corticosteroids: A Paradigm Shift in Asthma Management." Harvard University Press. Pharmacotherapy 22 (2): 212­26. NIOSH (National Institute of Occupational Safety and Health). 1999. Sudre, P., S. Jacquemet, C. Uldry, and T. V. Perneger. 1999. "Objectives, Work-Related Lung Disease Surveillance Report 1999. NIOSH Methods, and Content of Patient Education Programmes for Adults Publication 2000-105. http://www.cdc.gov/niosh/docs/2000-105/ with Asthma: Systematic Review of Studies Published between 1979 2000-105.html. and 1998." Thorax 54: 681­7. OSHA (Occupational Safety and Health Administration). 1998. Sullivan, S. D., and K. B. Weiss. 2001. Health Economics of Asthma and Respiratory Protection Regulations. Section 6: Summary of the Final Rhinitis: II. Assessing the Value of Interventions. Journal of Allergy and Economic Analysis. http://www.osha.gov/pls/oshaweb/owadisp. Clinical Immunology 107: 203­10. show_document?p_id=1052&p_table=PREAMBLES. Toevs, C. D., R. M. Kaplan, and C. J. Atkins. 1984. "The Costs and Effects Paltiel, A. D., A. L. Fuhlbrigge, B. T. Kitch, B. Liljas, S. T. Weiss, P. J. of Behavioral Programs in Chronic Obstructive Pulmonary Disease." Neumann, and K. M. Kuntz. 2001. "Cost-Effectiveness of Inhaled Medical Care 22: 1088­100. Corticosteroids in Adults with Mild-to-Moderate Asthma: Results Trapido, A. S., N. P. Mqoqi, C. M. Macheke, B. G. Williams, J. C. Davies, from the Asthma Policy Model." Journal of Allergy and Clinical and C. Panter. 1996. "Occupational Lung Disease in Ex- Immunology 108 (1): 39­49. Mineworkers--Sound a Further Alarm!" South African Medical Journal Postma, M. J., P. Jansema, M. L. L. van Genugten, M.-L. A. Heijnen, J. C. 86 (5): 559. Jager, and L. T. W. de Jong-van den Berg. 2002. "Pharmacoeconomics Van den Boom, G., M. P. Rutten­van Mölken, J. Molema, P. R. Tirimanna, of Influenza Vaccination for Healthy Working Adults." Drugs 62 (7): C. van Weel, and C. P. van Schayck. 2001. "The Cost Effectiveness of 1013­24. Early Treatment with Fluticasone Propionate 250 Microg Twice a Day Ramsey, S. D., D. L. Patrick, R. K. Albert, E. B. Larson, D. E. Wood, and in Subjects with Obstructive Airway Disease. Results of the DIMCA G. Raghu. 1995. "The Cost-Effectiveness of Lung Transplantation: A Program." American Journal of Respiratory and Critical Care Medicine Pilot Study." Chest 108: 1594­601. 164 (11): 2057­66. Ruchlin, H. S., and E. J. Dasbach. 2001. "An Economic Overview of Weiss, K. B., and S. D. Sullivan. 2001. "The Health Economics of Asthma Chronic Obstructive Pulmonary Disease." Pharmacoeconomics 19 (6): and Rhinitis: I. Assessing the Economic Impact." Journal of Allergy and 623­42. Clinical Immunology 107: 3­8. Rutten­van Mölken, M. P., E. K. Van Doorslaer, and F. F. Rutten. 1992. WHO (World Health Organization). 2000. "Global Burden of Disease "Economic Appraisal of Asthma and COPD Care: A Literature Review 2000: Deaths by Age, Sex and Cause for the Year 2000." WHO, Geneva. 1980­1991." Social Science and Medicine 35 (2): 161­75. ------. 2001. "Innovative Care for Chronic Conditions: Building Blocks Sant'Anna, C. A., R. Stelmach, M. I. Zanetti Feltrin, W. J. Filho, T. Chiba, for Action." WHO/MNC/CCH/02.01. WHO, Geneva. and A. Cukier. 2003. "Evaluation of Health-Related Quality of Life in ------. 2002a. "Influenza." Weekly Epidemiological Record 77.28 (July 12): Low-Income Patients with COPD Receiving Long-Term Oxygen 229­40. Therapy." Chest 123 (1): 136­41. ------. 2002b. "Outbreak of Influenza, Madagascar, July­August 2002." Scanlon, P. D., J. E. Connett, L. A. Waller, M. D. Altose, W. C. Bailey, A. S. Weekly Epidemiological Record 77.46 (November 15): 381­88. Buist, and D. P. Tashkin. 2000. "Smoking Cessation and Lung Function ------. 2003. "Report of the First International Review Meeting, Practical in Mild-to-Moderate Chronic Obstructive Pulmonary Disease: The Approach to Lung Health (PAL) Strategy." http://whqlibdoc.who.int/ Lung Health Study." American Journal of Respiratory and Critical Care hq/2003/WHO_CDS_TB_2003.324.pdf. Medicine 161: 381­90. ------. 2005a. "Strengthening Pandemic Influenza Preparedness and Schmidt, C. D., C. G. Elliott, D. Carmelli, R. L. Jensen, M. Gengiz, J. C. Response." Report by the Secretariat. 58th World Health Assembly, Schmit, and others. 1983. "Prolonged Mechanical Ventilation for A58/13, April 7. Respiratory Failure: A Cost-Benefit Analysis." Critical Care Medicine ------. 2005b. "Communicable Disease Surveillance & Response." 11: 407. Confirmed Human Cases of Avian Influenza A (H5N1). Scott, J. A. G., A. J. Hall, C. Muyodi, B. Lowe, M. Ross, B. Chohan, and http://www.who.int/csr/don/en/. (This site provides weekly updates others. 2000. "Aetiology, Outcome, and Risk Factors for Mortality "Disease Outbreak News" of reported cases to WHO by specific coun- among Adults with Acute Pneumonia in Kenya." Lancet 355: 1225­30. tries and latest assistance with regard to potential pandemic status and Scuffham, P. A., and P. A. West. 2002. "Economic Evaluation of Strategies preparedness. (Last accessed August 19, 2005, Updated 282). for the Control and Management of Influenza in Europe." Vaccine 20: ------. 2005c. "WHO Global Influenza Preparedness Plan." The Role of 2562­78. WHO and Recommendations for National Measures before and Smith, K. J., and R. R. Pesce. 1994. "Pulmonary Artery Catheterization in during Pandemics. WHO/CDS/CSR/GIP/2005.5, pp. 1­49. Exacerbations of COPD Requiring Mechanical Ventilation: A Cost- Effectiveness Analysis." Respiratory Care 39: 961­7. Respiratory Diseases of Adults | 693 Chapter 36 Diseases of the Kidney and the Urinary System John Dirks, Giuseppe Remuzzi, Susan Horton, Arrigo Schieppati, and S. Adibul Hasan Rizvi CAUSES AND CHARACTERISTICS OF THE BURDEN worldwide are on RRT, 80 percent of whom live in Japan, OF DISEASES Europe, and North America (Weening 2004). The percentage of patients on regular dialysis varies across Estimates of the global burden of disease indicate that diseases countries as a consequence of the capacity of health care systems of the kidney and urinary tract account for approximately to provide treatment. Europe is an example. Whereas in the 830,000 deaths and 18,467,000 disability-adjusted life years 15 countries of the European Union (before 2004) the prevalence annually, ranking them 12th among causes of death (1.4 per- rate of RRT was approximately 650 patients per 1 million people, cent of all deaths) and 17th among causes of disability (1.0 per- in Central and Eastern Europe it was only 160 patients per 1 mil- cent of all disability-adjusted life years). This ranking is similar lion people, reflecting differences in gross national product. across World Bank regions (table 36.1). Much less is known about the prevalence of earlier stages of Recent research suggests that the data shown in table 36.1 CKD, when symptoms may be mild, ignored, or undiagnosed. underestimate the global prevalence of kidney disease. Chronic A lack of standardization of the stages of CKD has hampered kidney disease (CKD) patients often suffer from cardiovascular assessments of the burden of CKD. In an attempt to carry out or cerebrovascular disease, and their deaths may be attributed such an assessment, the National Center for Health Statistics of to either complication (Hostetter 2004). Altered kidney func- the Centers for Disease Control and Prevention in the United tion is often found in patients with hypertensive and ischemic States conducted a survey from 1988 to 1994. The center ana- heart disease, both of which are associated with increased car- lyzed a sample of 15,625 noninstitutionalized individuals diovascular morbidity and mortality. Approximately 30 per- age 20 and older and defined five stages of renal dysfunction cent of patients with diabetes have diabetic nephropathy, with according to estimates of renal function and urine albumin higher rates found in some ethnic groups (King, Aubert, and level. Coresh and others (2003) found that the estimated preva- Herman 1998). Table 36.2 shows that both genders are similarly lence of CKD in the United States is 11 percent of the adult affected by kidney disease (Coresh and others 2003). population, or 19.8 million people. Nationally representative Generally, renal diseases progress to a final stage as end- data on U.S. adults older than 20 show that 6.3 percent, or stage renal disease (ESRD) and function is substituted by renal 11 million people, have stage 1 CKD, or kidney damage (pro- replacement therapy (RRT), hemodialysis, peritoneal dialysis, teinuria) with normal kidney function (Glomerular Function or transplantation. National and international registries of Rate (GFR) at least 90 milliliters per minute in 1.73 per meter patients on RRT are useful for providing information on the squared) or stage 2 CKD, that is, mildly reduced kidney func- prevalence of renal diseases in a given country. Data combined tion (60 to 89 ml/min/1.73 m2). Furthermore, 4.3 percent, from different sources show that more than 1.5 million people or 7.6 million people, exhibit stage 3 CKD, or moderately 695 Table 36.1 Contribution of Diseases of the Kidney and Urinary System to the Global Burden of Disease by Gender and Region (thousands) Disability-adjusted Years lived Years of Gender and region Population Deaths life years with disability life lost Females 3,056,384 397 8,008 2,546 5,450 Males 3,093,849 433 10,459 4,493 5,960 World 6,150,233 830 18,647 7,039 11,415 East Asia and the Pacific 1,850,775 233 5,400 1,858 3,530 Europe and Central Asia 447,180 53 1,417 623 793 Latin America and the Caribbean 526,138 70 1,667 779 888 Middle East and North Africa 309,762 57 1,283 460 823 South Asia 1,387,873 156 3,991 1,373 2,619 Sub-Saharan Africa 667,663 107 2,623 1,046 1,576 Source: Mathers and others 2006. Table 36.2 Global Deaths Caused by Diseases of the Genitourinary System by Gender and Age Age (years) Gender Birth­4 5­14 15­29 30­44 45­59 60­69 70­79 80+ Male deaths Number (thousands) 11 7 24 43 80 86 110 88 Percent 3 2 5 10 18 19 24 20 Female deaths Number (thousands) 10 6 21 29 61 66 85 98 Percent 3 2 5 8 16 18 23 24 Source: WHO 2002. reduced kidney function (30 to 59 ml/min/1.73 m2), and 0.2 technology. The characterization of inherited kidney diseases percent, or 400,000, have stage 4 CKD, or severely reduced kid- has improved, and novel mutations leading to selective renal ney function (15 to 29 ml/min/1.73 m2) (Coresh and others defects have been described. Inherited kidney diseases are rare, 2003; Coresh, Astor, and Sarnak 2004; National Kidney with the exception of autosomal dominant polycystic kidney Foundation 2002). A sizable proportion (360,000) of these disease, the fourth most common cause of ESRD in developed patients eventually progress toward ESRD (stage 5, or less than countries. This disease has a prevalence of 1 in 1,000 people 15 ml/min/1.73 m2) and require RRT. Early detection of CKD and affects approximately 10 million people worldwide is, therefore, important to retard or arrest the loss of renal func- (Grantham 1997). Autosomal recessive polycystic kidney dis- tion. Late detection of CKD is a lost opportunity for making ease is less frequent, with an incidence of 1 in 40,000, but is an lifestyle changes and initiating therapeutic measures. important hereditary disease of childhood (Guay-Woodford, Jafri, and Bernstein 2000). Many other inherited diseases can lead to ESRD, but together they account for only a small per- CAUSES OF DISEASES OF THE KIDNEY centage of all people with ESRD. AND URINARY SYSTEM Kidney disease leading to ESRD has many causes. The preva- lence varies by country, region, ethnicity, gender, and age. Glomerulonephritis Glomerulonephritides are a group of kidney diseases that affect the glomeruli. They fall into two major categories: glomeru- Genetic Diseases lonephritis refers to an inflammation of the glomeruli and can Knowledge of inherited kidney disease has changed radically be primary or secondary, and glomerulosclerosis refers to scar- with advances in molecular biology and gene-sequencing ring of the glomeruli. Even though glomerulonephritis and 696 | Disease Control Priorities in Developing Countries | John Dirks, Giuseppe Remuzzi, Susan Horton, and others glomerulosclerosis have different causes, both can lead to 300 million are at risk. The disease causes lesions in the bladder ESRD. Glomerulonephritis ranks second after diabetes as the and predisposes those with the condition to secondary infec- foremost cause of ESRD in Europe. (Stengel and others 2003) tions, bladder cancers, and chronic pyelonephritis. and is the second leading cause of ESRD in the United States, Some 15 to 20 million people have tuberculosis (TB) world- according to the United States Renal Data System (http://www. wide, of whom 8 million to 10 million are infectious. ifrr.net/). Approximately 20 to 35 percent of patients requiring Genitourinary TB is a common form of extrapulmonary TB RRT have a glomerular disease. and is always secondary to the primary lesion, which usually Glomerular diseases are more prevalent and severe in tropi- occurs in the lung (Pasternak and Rubin 1997). Lesions cal regions and low-income countries (Seedat 2003).A common referred to as ulcero-cavernous or miliary affect the kidneys. If mode of presentation is the nephrotic syndrome, with the age of left untreated, such lesions may progress to kidney destruction. onset at five to eight years. Estimates indicate that 2 to 3 percent Early recognition of and effective therapy for TB substantially of medical admissions in tropical countries are caused by renal- decrease the consequences in relation to kidney function. related complaints, most resulting from glomerulonephritis. In the industrial countries, kidney stones are a common A number of kidney diseases that result from infectious dis- problem (Morton, Iliescu, and Wilson 2002), affecting 1 person eases, such as malaria, schistosomiasis, leprosy, filariasis, and in 1,000 annually, and the incidence is increasing in tropical hepatitis B virus, are exclusive to the tropics. HIV/AIDS can be developing countries (Robertson 2003). Factors such as age, complicated by several forms of kidney disease; however, sex, and ethnic and geographic distribution determine preva- patient data are sparse (Seedat 2003). lence. The peak age of onset is in the third decade, and preva- Acute poststreptococcal nephritis following a throat or skin lence increases with age until 70. infection caused by Group A streptococcus has almost disap- Although largely idiopathic, the following risk factors are peared in high-income countries because of improved hygiene associated with stone disease: low urine volume, hyperurico- and treatment but remains an important glomerular disease in suria, hyperoxaluria, hypomagnesuria, and hypocitraturia. India and Africa, where epidemics have been reported (Seedat Diarrhea, malabsorption, low protein, low calcium, increased 2003). consumption of oxalate-rich foods, and low fluid intake may The eradication of endemic infections, along with improve- play a role in the genesis of stone disease. In developing coun- ments in socioeconomic status, education, sanitation, and tries, 30 percent of all pediatric urolithiasis cases occur as blad- access to treatment, is a crucial step toward decreasing the inci- der stones in children. The formation of bladder stones in chil- dence of glomerular diseases in developing countries. dren is caused by a poor diet high in cereal content and low in animal protein, calcium, and phosphates. Kidney stones can have different clinical presentations, Infections, Stones, and Obstructive Uropathy ranging from asymptomatic to large obstructing calculi in the Infections of the urinary tract are a common health problem upper urinary tract that can severely impair renal function and worldwide and can be categorized as either uncomplicated or lead to ESRD. Although specific causes of kidney stones should complicated. Uncomplicated infections include bladder infec- be treated appropriately, general treatment includes increased tions such as cystitis, seen almost exclusively in young women fluid intake, limited daily salt intake, moderate animal protein (Hooton 2000). Among sexually active women, the incidence intake, and medical treatment with alkali and thiazides. of cystitis is 0.5 episodes per person annually, and recurrence The Afro-Asian stone-forming belt stretches from Sudan, develops in 27 to 44 percent of cases. Acute, uncomplicated the Arab Republic of Egypt, Saudi Arabia, the United Arab pyelonephritis, involving the kidney, is less frequent in women Emirates, the Islamic Republic of Iran, Pakistan, India, than is cystitis. Males are less susceptible to acute, uncompli- Myanmar, Thailand, and Indonesia to the Philippines. The dis- cated infections of the bladder or the kidney, with an incidence ease affects all age groups from less than 1 year old to more of five to eight episodes per 10,000 men annually. Even though than 70, with a male to female ratio of 2 to 1. The prevalence of uncomplicated urinary tract infections are considered benign, calculi ranges from 4 to 20 percent (Hussain and others 1996). they have significant medical and financial implications esti- Urolithiasis accounts for some 50 percent of the urological mated at approximately US$1.6 billion per year (Foxman 2003). workload and the bulk of urological emergencies. Patients may As for complicated urinary tract infections, hospitalization present with major complications leading to eventual ESRD results in almost 1 million such infections per year in the United and resulting in significant morbidity and mortality. In devel- States. Bladder catheterization is the most important cause. oped countries, only about 1 percent of patients are on dialysis Developing countries exhibit a different pattern of urinary because of obstructive uropathy, whereas in developing coun- tract infection. Obstructive or reflux nephropathy is often tries such as Indonesia and Thailand, obstructive uropathy is attributed to urinary schistosomiasis (Barsoum 2003). often the leading cause of ESRD, accounting for 20 percent or Worldwide, 200 million people are affected and an estimated more of patients on dialysis. The availability of appropriately Diseases of the Kidney and the Urinary System | 697 trained medical and surgical personnel and of equipment hypertension has also been associated with an increased risk of essential for treating stone disease promptly would reduce the diabetic nephropathy. When specific markers of risk are found, incidence of obstructive uropathy and ESRD. Cost analyses high-risk individuals can be identified early and monitored for indicate that the medical prevention of stones saves more than the development of proteinuria and kidney dysfunction. US$2,000 per person annually (Parks and Coe 1996). The earliest sign of diabetic nephropathy is the appearance of small amounts of protein in the urine (proteinuria). As pro- teinuria increases and blood pressure rises, kidney function Benign Prostatic Hypertrophy declines. The complete loss of kidney function occurs at differ- Benign prostatic hypertrophy is a major cause of lower urinary ent rates among type 2 diabetes patients, but it eventually tract symptoms and leads to obstructive renal failure and occurs in 30 percent of proteinuria cases. The latter have a ESRD. By age 80, 80 percent of men have benign prostatic 10-fold increased risk of dying from associated coronary artery hypertrophy. The World Health Organization quotes a mortal- disease, which may obviate the progression of diabetic ity rate of 0.5 to 1.5 per 100,000 (La Vecchia, Levi, and Lucchini nephropathy to ESRD. As therapies and interventions for coro- 1995). The actual incidence of benign prostatic hypertrophy is nary artery disease improve, patients with type 2 diabetes may difficult to assess because of the lack of epidemiological data. In survive long enough to develop kidney failure. the developed world, the incidence varies between 0.24 and 10.90 per 1,000 annually from age 50 to 80, and the probabili- ty of prostate surgery for benign prostatic hypertrophy ranges Hypertension from 1.4 to 6.0 percent (Oishi and others 1998). Hypertension and kidney disease are closely related. Most pri- mary renal diseases eventually produce hypertension. Arterial hypertension accelerates many forms of renal disease and has- Acute Renal Failure tens the progression to ESRD (Luke 1999). Recent studies have Acute renal failure refers to a sudden and usually temporary loss firmly established the importance of continuous blood pres- of kidney function that may be so severe that RRT is needed sure reduction to slow the progression of many forms of renal until kidney function recovers. Even though acute renal failure injury, particularly glomerular disease (Agodoa and others can be a reversible condition, it carries a high mortality rate. 2001; Peterson and others 1995). Over the long term, damage Acute renal failure is a prominent feature of major earthquakes, to the heart and cardiovascular system resulting from hyper- where many suffer from crush syndrome accompanied by tension represents the major cause of morbidity and mortality severe dehydration and rapid release of muscle cell contents, among ESRD patients (Martinez-Maldonado 1998). including potassium. Kidney function shuts down unless body Before the development of effective antihypertensive agents, fluid and blood pressure are rapidly corrected and frequent 40 percent of hypertensive patients developed kidney damage hemodialysis is available. Recent earthquake rescues in the and 18 percent developed renal insufficiency over time Islamic Republic of Iran and Turkey have demonstrated the (Johnson and Feehally 2000). Elevated serum creatinine devel- benefits of rapid hydration and dialysis (Sever and others 2001). ops in 10 to 20 percent of hypertensive patients, with African Americans and Africans at particularly high risk. In 2 to 5 per- cent of hypertensive patients, progression toward ESRD will Diabetes occur in 10 to 15 years. Despite the relatively low rate of pro- Diabetes is one of the most common noncommunicable dis- gression, hypertension remains the most common cause of eases (see chapter 30). With the serious complication of ESRD after diabetes in the United States, is the foremost cause nephropathy, diabetes has become the single most important of death in all developed countries, and is a likely primary cause of ESRD in the United States and Europe, according to cause in developing countries given its high global prevalence Stengel and others (2003) and the United States Renal Data rate. Native Americans and Hispanic Americans are dispropor- System (http://www.ifrr.net/). Diabetes may account for one- tionately affected relative to Caucasian Americans. third of all ESRD cases. Family-based studies and segregation analyses suggest that inherited factors play a major role in people's susceptibility to GLOBAL PERSPECTIVES IN RELATION TO RRT diabetic renal complications (Seaquist and others 1989). In the United States, the burden of ESRD is threefold to fivefold Despite the lack of uniform data worldwide, the medical com- greater among African Americans, Mexican Americans, and munity is aware that the total number of patients requiring Native Americans than other Americans, and Imperatore and RRT is growing in all high- and middle-income countries. In others (2000) find a 200 percent greater possibility of the occur- the United States, for example, 360,000 people with ESRD rence of inherited diabetic nephropathy. A family history of were on RRT in 2003, compared with 150,000 in 1994, and 698 | Disease Control Priorities in Developing Countries | John Dirks, Giuseppe Remuzzi, Susan Horton, and others according to a recent forecast, by 2014 the figure will have ­ dyslipidemia increased to 650,000 (Xue and others 2001). This increase rep- ­ poor glycemic control in diabetic patients resents a linear growth in new cases combined with longer sur- ­ proteinuria vival by existing patients. · biological markers Levels in middle-income countries are lower, but rising. In ­ hemoglobin Eastern Europe between 1990 and 1996, following economic ­ insulin-resistant syndrome changes, the number of hemodialysis and peritoneal dialysis ­ proteinuria centers increased by 56 and 296 percent, respectively ­ serum creatinine. (Rutkowski 2002), and the number of patients rose by 78 and 306 percent, respectively. Growing evidence suggests that fetal exposure to an abnor- Overall, the incidence of ESRD is increasing worldwide at an mal intrauterine environment leads to an increased risk of annual growth rate of 8.0 percent, far in excess of the annual chronic disease later in life. For example, children of diabetic population growth rate of 1.3 percent.Nearly 1.6 million people, mothers are prone to obesity and diabetes at a young age, and or only 15 percent of those affected, are receiving RRT, 80 per- intrauterine growth retardation can lead to ischemic heart dis- cent of them in developed countries. The remaining 20 percent ease, diabetes, hypertension, and kidney disease. Disadvantaged are treated in more than 100 developing countries, whose pop- racial minorities in developed countries and the impoverished ulations account for more than 50 percent of the world's popu- in developing countries are at risk of intrauterine growth lation. A large proportion of people living in the poorest coun- retardation caused by malnutrition (Nelson 2001; Nelson, tries die of uremia because of a complete lack of RRT. Morgenstern, and Bennett 1998). Attention to maternal nutri- tion and other factors that would reduce low birthweight and impaired nephron development may have long-term implica- Risk Factors for Kidney Disease tions for the development of CKD. The identification of risk factors can prevent or limit disease In low-income countries, poverty is associated with through lifestyle modifications or specific therapeutic inter- increased exposure to infectious diseases that increase suscep- ventions (Appel 2003; McClellan and Flanders 2003). For tibility to CKD, including glomerulonephritis and parasitic dis- example, familial predisposition for a disease, which is not eases. Obesity caused by a diet rich in saturated fats and high in amenable to modification, can be used to identify high-risk salt are risk factors for diabetic nephropathy and hypertensive populations for future monitoring. kidney disease. Change in dietary habits and physical activity Low socioeconomic status and limited access to health care can reduce the overall incidence of diabetes (see chapter 44). are strong risk factors for kidney failure but account for only Smoking and excessive alcohol consumption increase the risk part of the excess of ESRD among African Americans of ESRD (McClellan and Flanders 2003), and analgesic abuse (Perneger, Whelton, and Klag 1995), whereas racial and social and exposure to toxic substances such as lead may affect pro- factors account for most ESRD incidence (Pugh and others gressive renal insufficiency (Lin and others 2001). 1988; Rostand 1992). Factors associated with the progression of CKD include the following: Interventions to Delay CKD During the past 20 years, human and animal research has · unmodifiable variables developed our understanding of CKD and led to preventive ­ old age measures. The notion of renoprotection has resulted in a dual ­ gender approach to renal diseases based on effective and sustained ­ genetics pharmacological control of blood pressure and reduction of ­ ethnicity proteinuria. Lowering blood lipids, stopping smoking, and · risk factors susceptible to social and educational maintaining tight glucose control for diabetes form part of the interventions multimodal protocol for managing renal patients monitored ­ low birthweight by specific biological markers (Ruggenenti, Schieppati, and ­ smoking Remuzzi 2001). ­ alcohol abuse Abnormal urinary excretion of protein is strongly associated ­ illicit drug abuse with the progression of CKD in both diabetic and nondiabetic ­ analgesic abuse and exposure to toxic substance such as renal diseases. Clinical studies have established that a reduction lead in proteinuria is associated with a decreased rate of kidney func- ­ sedentary lifestyle tion loss. A specific category of drugs that lower blood pres- · risk factors susceptible to pharmacological interventions sure, the angiotensin-converting enzyme (ACE) inhibitors or ­ hypertension angiotensin receptor blockers, appear to be more effective than Diseases of the Kidney and the Urinary System | 699 other antihypertensive drugs in slowing the progression of both Trained staff members can carry out screening programs diabetic and nondiabetic CKDs (Brenner and Zagrobelny 2003). inexpensively. Economic analysis, however, suggests that large- The administration of an ACE inhibitor (or of an angiotensin scale programs should be restricted to screening and treating receptor blocker) is an important treatment for controlling only specific high-risk populations. Screening programs can be blood pressure and slowing the rate of progression of chronic implemented using simple, cheap, and reliable tests consisting kidney failure. Other drugs to lower blood pressure are added as of measurements of bodyweight, blood pressure, blood glu- necessary to achieve current targets of 120/80 to 130/80 mil- cose, and creatinine. Screening includes testing urine for hemo- limeters of mercury. Concurrent diuretic therapy is often neces- globin, glucose, leukocytes, and protein (repeat tests may be sary in patients with renal insufficiency, because fluid overload necessary on a spot urine sample); calculating albumin to is an important determinant of hypertension in such cases. creatinine ratios; testing positive results for increased serum Dyslipidemia accelerates atherosclerosis and may promote creatinine and fasting glucose (or glycosylated hemoglobin A1c the progression of renal disease. Careful control of the blood test); and reassessing the urine protein excretion rate, a corner- glucose level in diabetic patients can be beneficial and may stone of kidney assessment. Resulting albumin to creatinine limit other complications. Obesity has not been directly linked ratio categories would indicate a scale of severity of glomerular to the progression of CKD but is an important risk factor for disease, with a cardiovascular risk score based on body mass diabetes and cardiovascular morbidity and mortality. Many index, hypertension, fasting glucose level, microalbuminuria or patients and health care professionals do not appreciate the gross albuminuria, and serum creatinine. Patients with positive benefits of smoking cessation, an important measure in pro- markers for kidney disease would receive the best treatment tecting the kidneys from progressive disease resulting from car- available at the screening center. Incorporating screening for diovascular disease (CVD). Additional elements of secondary kidney disease within screening programs developed for CVD prevention measures include the treatment of anemia and of and diabetes is important because proteinuria and renal dys- abnormal calcium and phosphorus metabolism. function are early sensitive markers of vascular dysfunction The International Society of Nephrology is developing a and CVD patients are at significantly higher risk of kidney dis- program that can be implemented according to the specific ease than the general population. needs of a given developing country. The program has two Resultant medical treatment would focus on the use of ACE objectives: (a) to identify the prevalence of renal disease among inhibitors or angiotensin receptor blockers with a target blood seemingly healthy subjects using a communitywide screening pressure of 120/80 to 130/80 millimeters of mercury. The program, especially among populations at risk, and (b) to ini- greater the level of proteinuria, the more treatment is required; tiate interventions to prevent the progression of renal disease thus, the ACE inhibitor dose would be titrated up as proteinuria and affect both renal and CVD outcomes in subjects with or at levels increased. Diuretics and other antihypertensives would be risk of developing renal disease based on the screening pro- added to meet blood pressure targets. Efforts should be made to gram (Weening 2004). The Kidney Help Trust of Chennai, obtain low-cost (off-patent) ACE inhibitors or other low-cost India, has undertaken a screening program for a population of antihypertensives. Such treatment should delay or stop the pro- 25,000. All those who tested positive for high blood pressure, gression of kidney disease and reduce the risk of CVD. Other diabetes, or both (about 15 percent) were further studied and preventive measures include serum glucose and lipid control then treated with inexpensive antihypertensive and antidia- and low-dose aspirin if a risk of CVD exists (see chapter 44). betic drugs. The cost of the one-year program was Rs 300,000 (US$7,500) or a per capita cost of US$0.27, well within the lim- its of the Indian government's per capita annual health expen- ECONOMIC BENEFITS OF INTERVENTION diture of US$7.67 (Mani 2003). A similar program in Bolivia examined a population of 14,000 and also found that 15 per- An abundance of literature is available on the economics of cent were hypertensive, diabetic, or both. ESRD. In the industrial world, treatment is usually readily An extremely successful program of detection and treat- available and is covered by government or private health insur- ment of renal and cardiovascular diseases among Australian ance. Previous restrictions--for example, treatment being lim- Aborigines was conducted from 1995 to 2000. The ESRD rate ited to certain age groups--have been removed (Chugh and among Aborigines is 3 to 10 times that in developed countries. Jha 1995). Dialysis treatment accounts for 0.7 to 1.8 percent of Treatment consisted of long-acting ACE inhibitors to lower health care budgets in European countries, even though dialy- blood pressure. After an average of 3.4 years of follow-up, the sis patients account for only 0.02 to 0.05 percent of the popu- incidence of ESRD was reduced by 63 percent and nonrenal lation (Schiepatti, Perico, and Remuzzi 2003). deaths were reduced by 50 percent. Hoy and others (2003) esti- The most cost-effective option is prevention. Population mate that this two-year program may have saved US$500,000 screening is not particularly cost-effective, given the low inci- to US$2.7 million in avoided or delayed dialysis costs. dence of ESRD--namely, 100 to 200 per million population 700 | Disease Control Priorities in Developing Countries | John Dirks, Giuseppe Remuzzi, Susan Horton, and others worldwide (Kher 2002)--and given that testing is not highly Table 36.3 Cost-Effectiveness of Selected Interventions for accurate. According to Kiberd and Jindal (1998), screening Kidney Disease costs around US$20 per test, but the positive predictive value Intervention Alternative Outcome (2000 US$) for a single test is only 0.3. Even repeat testing does not improve predictive value dramatically. Screening strategies have, there- Center hemodialysisa No RRT 55,000­80,000/life year 79,000­114,000/QALY fore, focused on specific populations at higher risk of ESRD than the general population. Whereas only 2 to 5 percent of Home hemodialysisa No RRT 33,000­50,000/life year more than 1 billion hypertensive patients will ultimately devel- 47,000­71,000/QALY op nephropathy, approximately 30 percent of type 1 and type 2 Kidney transplanta No RRT 10,000/life year diabetic patients will develop overt nephropathy (Satko and 11,000/QALY Freedman 2001). The conclusion is that treating all diabetics in ACE inhibitors for all No RRT 1,100/QALY developed countries with ACE inhibitors is a cost-saving strat- type 1 diabetics with egy. The modest outlay for ACE inhibitors, which amounts to macroproteinuriab US$320 per year in the United States and is likely to come down Screening diabetic No screening Screening potentially as more ACE inhibitor treatments come off patent, offsets the relatives of cost saving much larger future costs of dialysis and transplant (Golan, nephropathy patientsc Birkmeyer, and Welch 1999; Kiberd and Jindal 1998). Treat all type 2 Screening for Incremental cost- We undertook a crude cost-effectiveness calculation for diabetics with ACE microalbuminuria and effectiveness ratio is treating diabetics in developing countries with ACE inhibitors inhibitorsd treating those who test 7,500/QALY for treating in those cases in which no treatment of ESRD is undertaken. If positive all type 2 diabetics we use Clark and others' (2000) assumptions, 82 percent of Treat all insulin- Screening for Treating all diabetic patients not using ACE inhibitors would survive for dependent diabetics microalbuminuria or insulin-dependent 11 years from the onset of macroproteinuria to ESRD, whereas with ACE inhibitorse macroproteinuria and diabetics dominates treating those who test under a plausible range 72 percent of those using ACE inhibitors would survive for positive of parameters 18 years from the onset of macroproteinuria to ESRD (the annualized death rate for both groups is 1.8 percent). If we Sources: Winkelmayer and others 2002 (review); authors' rough estimates; Satko and Freedman a b c 2001; Golan, Birkmeyer, and Welch 1999; Kiberd and Jindal 1998. d e make the gross assumption that all patients with ESRD in poor developing countries die, this assumption suggests that, at a discount rate of 3 percent and an annual cost for ACE relatives of ESRD patients. They did not calculate any formal inhibitors of US$320, the cost per quality-adjusted life year cost-effectiveness results (table 36.3). (QALY) saved would be about US$1,100 for treating diabetic Kidney transplants are the most cost-effective intervention patients with macroproteinuria. Because of the lack of data, for ESRD. Transplant costs in developed countries have these calculations apply survival rates in developed countries to declined steadily from about US$60,000 in 1970 to about developing countries; thus, the rates are likely too high. Using US$10,000 currently (Winkelmayer and others 2002). In addi- survival rates in developing countries would probably increase tion to facing transplant costs, patients face ongoing costs for the cost per QALY saved, but treatment with ACE inhibitors is immunosuppressive drugs, which start at about US$3,000 per nevertheless likely to be an attractive investment (table 36.3). year initially but can decline thereafter to US$300 per year Satko and Freedman (2001) suggest that screening first- and (Kher 2002). Kidney transplants are cheaper in India than in the second-degree relatives of ESRD patients may be cost-effective. United States, ranging from US$1,500 in government hospitals They cite one study that found 38 percent of first-degree rela- to as much as US$7,000 in private hospitals. Such costs, com- tives of African-American patients with hypertensive ESRD bined with a higher quality of life than obtained with dialysis, had some form of renal disease (Bergman and others 1996). make renal transplantation the most cost-effective option (table Satko and Freedman also cite a study by Freedman, Soucie and 36.3). However, the availability of kidneys is a major limiting McClellan (1997) revealing that in 4,365 incident ESRD factor. Developed countries tend to have well-organized organ patients in the southeastern United States, 14 percent of white retrieval programs, and cadaveric donor transplants are more patients and 23 percent of black patients had first- or second- common than they are in developing countries. Japan, with its degree relatives with ESRD (the rates would probably have extremely low transplant rates, is an exception, perhaps because been higher if subclinical nephropathy had been included). of difficulties in obtaining permission for organ donation. Satko and Freedman (2001) recommend annual screening for Developing countries have limited access to cadaveric donor blood pressure, urinalysis, measurement of serum creatinine programs but better living donor programs. Unrelated living and blood urea nitrogen concentration, and testing for diabetes donors are more common than in developed countries because mellitus, when appropriate, for first- and second-degree poverty increases the willingness of donors to offer kidneys in Diseases of the Kidney and the Urinary System | 701 exchange for payment. The Philippines recently restricted emphasis on CVD, all of which contribute to quality-of-life donations to "emotionally related" donors, but that limitation outcomes, but at an increased cost (National Kidney does not prevent abuses, such as men marrying women of the Foundation 2002). appropriate blood type in the hope of obtaining a kidney. The high mortality rate of dialysis approximates 10 percent Developing countries face particular transplantation prob- per year and has changed little over the past decade; however, lems, such as patients' inability to continue paying for new approaches are emerging for dealing with CVD in RRT immunosuppressive drugs and the transmission of hepatitis B facilities. More patients with kidney disease die before they get and C, malaria, and TB through organ transplant (Kher 2002). to the point at which they need treatment for renal failure, Long-term hemodialysis was introduced in 1960 and is the because early kidney disease is a major marker for CVD and most costly treatment option at approximately US$60,000 per reinfarction, congestive heart failure, and stroke. year at a center and US$40,000 at home in developed countries. In middle-income countries such as Thailand and Turkey and It is most cost-effective if used as an interim measure before in middle-income countries in Latin America (Zatz, Romão, and kidney transplant. Peritoneal dialysis--for example, continu- Noronha 2003), extensive dialysis facilities are available, as they ous ambulatory peritoneal dialysis--was developed in the late are in some low-income countries. For example, in 2003, 1970s and is less expensive--approximately US$20,000 per Pakistan had 110 centers with 2,400 patients on hemodialysis; year (Winkelmayer and others 2002). Most economies con- India had 100 centers with 6,000 patients mostly on hemodialy- tinue to rely on hemodialysis for dialysis patients, except for sis; and China had 75,000 patients on dialysis.Those figures show those mandating that continuous ambulatory peritoneal that needs and markets for dialysis are expanding. However, in dialysis be the first choice--that is, Hong Kong (China), poorer countries, such as Nicaragua and Tanzania, options for Mexico, New Zealand, and the United Kingdom. Switching to RRT are limited because of the lack of equipment, trained continuous ambulatory peritoneal dialysis has the potential of staff, and costly consumables. In addition, many low-income reducing costs for developing countries, especially if they man- countries lack health insurance to defray treatment expendi- ufacture the consumables domestically rather than importing tures, keeping dialysis out of reach. In such countries--for them. Nevertheless, dialysis remains costly and is not a viable example, Nigeria--dialysis directed at preparation for renal long-term solution in places where health budgets are limited. transplantation is the best policy. Recent findings concerning More than 120 countries have dialysis programs (Moeller, primary prevention through lifestyle changes and secondary Gioberge, and Brown 2002). The following data from India prevention by means of pharmaceutical treatment should highlight the stark economics of dialysis (Kher 2002). eventually reduce, but not eliminate, the burden of ESRD. Government hospitals will provide hemodialysis only for acute The acknowledgment by the World Bank and the World renal failure or pretransplant stabilization (Li and Chow 2001), Health Organization that chronic conditions, particularly those and with an incidence of 100 per million population, approxi- resulting from diabetes and hypertension, will increase to mately 100,000 patients develop ESRD each year. Of the 10,000 become a leading cause of death by 2028 has intensified the need who consult a nephrologist, RRT is initiated for 9,000. Of the for prevention and RRT programs. The need to increase aware- 8,500 who begin hemodialysis, about 60 percent are lost to ness, launch targeted screening and intervention studies, pro- follow-up within three months, probably because of the costs vide training for staff, maintain education for physicians in involved. Few remain on dialysis after 24 months. Between 17 kidney and urological disease, and assist centers for RRT is and 23 percent of those on dialysis for two to three months urgent. receive transplants. Developed nations have well-established nephrology and urology centers attached to academic medical institutions and regional public and private secondary and tertiary referral hos- IMPLEMENTATION OF CONTROL STRATEGIES: pitals. They have training programs to meet national require- LESSONS OF EXPERIENCE ments for health professionals--including renal physicians, pri- mary care physicians, and nurses--specializing in kidney and Measures for primary and secondary prevention of CKD are urological disorders. Their centers incorporate the results of now well documented and will eventually reduce the number up-to-date research developments pertaining to kidney disease of patients requiring dialysis. Until recently, the focus has been and clinical applications of the latest advances in care and tech- on RRT to save lives, and considerable efforts are being made to nology. Numerous publications arise from academic endeavors, improve the quality of dialysis. In the United States, guidelines and a close association exists between health care delivery and derived from the Kidney Disease Outcomes Quality Initiative pharmaceutical industries. Each country and region has soci- have added greatly to the quality of dialysis in terms of access eties of nephrology and urology for adults and children. (graft or fistula), adequacy, treatment of anemia, treatment of Middle-income countries may have both public academic secondary hyperparathyroidism, and--more recently--greater centers and private hospitals that offer specialized equipment, 702 | Disease Control Priorities in Developing Countries | John Dirks, Giuseppe Remuzzi, Susan Horton, and others such as lithotripters and imaging technology, and dialysis and in developing countries. Training epidemiologists and physi- transplant programs. Although facilities and trained staff for cians to execute screening strategies and clinical trials in their RRT are more limited than in developed countries, some devel- own settings is urgently needed. The cooperation of global oping countries, such as Turkey, have excellent facilities. funding agencies and training centers; the consistent availabil- In lower-income countries, facilities and staff are in short ity of effective, inexpensive pharmaceuticals; and the assess- supply, and assistance is needed. Large countries, such as ment of the efficacy and side effects of multiple drug therapy China, India, and Pakistan, have kidney centers available but must be coordinated. The priority is to make low-cost drugs have considerable unevenness in development of kidney cen- available, using as a model the recent process that allowed uni- ters and health care in general. Some lower-income countries versal access to inexpensive antiretrovirals for HIV infection. possess remarkable institutions; for instance, the Sindh Institute of Urology and Transplantation in Karachi, Pakistan, which is supported mainly by charitable donations, provides Renal Replacement Therapy every patient who presents with ESRD an opportunity for Successful RRT outcomes depend on reducing morbidity and accessing RRT. Overall, however, centers of excellence are mortality among dialysis patients. RRT costs escalate in concert urgently needed in developing countries. All the "players," from with the rising costs of pharmaceuticals--for example, eryth- governments and international organizations to societies and ropoietin compounds to treat anemia and vitamin D metabo- foundations, need to be congregated in conjunction with lites and calcimimetics to treat secondary hyperparathyroidism national institutions to focus on the continued advantages-- and bone disease. Strategies that will result in less expensive through treatment--that can be delivered to those developing dialysis systems and pharmaceuticals are needed (Schieppati, cardiovascular, diabetic, and kidney disease. Perico, and Remuzzi 2003). Costs relating to renal transplanta- tion have reached a steady state, but the lack of availability of donor kidneys is a serious--and perhaps irresolvable-- RESEARCH AND DEVELOPMENT AGENDA limitation. Significant progress in knowledge about the geographic burden of kidney and urological diseases has taken place during Establishment of Teaching and Research Centers the past three or four decades as a result of more accurate Most high-quality training and research centers for kidney and registries. An international kidney disease data center, in urinary diseases are in the developed world, where training is partnership with the World Bank and the World Health expensive. Important centers of clinical care have emerged in Organization, is now required to progressively increase the countries such as Argentina, China, Mexico, South Africa, amount and quality of data collected worldwide. Thailand, and Turkey. The ability to obtain high-quality train- ing at the local level would be advantageous to developing Basic Knowledge of Kidney Disease countries. For example, the International Society of Recent research findings have advanced the understanding and Nephrology has identified and supported a major clinical treatment of kidney disease. A continuing emphasis on under- training center in South Africa that plays a leading role in train- standing the basic mechanisms of glomerulonephritic, vas- ing nephrologists and urologists for South Africa and other culitic, and autoimmune disease and the detailed mechanisms Sub-Saharan African countries to world standards at lower of the progression of kidney disease to kidney failure is costs than in developed countries and with increased retention required, as well as research into improved therapies. Well- of local physicians. Such local centers should be a national pri- developed research centers are best equipped to deal with these ority in developing countries and should be closely linked to requirements, aided by national governments, charitable international centers for cardiovascular and diabetic disease, organizations and foundations, international organizations, meeting approved international standards for training while and centers in the developing world. recognizing national differences in the pattern of kidney dis- ease. Financial assistance is required to enhance the education and training of health professionals, improve baseline infra- Prevention of Kidney Failure structure, and initiate research studies directed at critical clini- Prevention of acute and chronic kidney disease should be a cal questions and at current and new knowledge relating to the global priority. During the past decade, an array of clinical prevention of kidney disease. The centers should have excellent trials has been directed at assessing the benefits of interven- data collection methods and a computer infrastructure that tional therapy, particularly the success of ACE inhibitors. Such would connect them to current knowledge and allow them to trials can play an important role in increasing knowledge and communicate freely on a global scale. Major priority should be improving the implementation of prevention of kidney disease given to developing leading centers in selected regions. Diseases of the Kidney and the Urinary System | 703 Cost-Effectiveness of Treatment develop emergency policies and practices and be linked with More work is needed in the area of screening and treatment in the appropriate international agencies. both developed and developing countries. Work on the cost- · Have the World Bank and the World Health Organization effectiveness of screening and treating particular subpopula- establish a policy advisory group with relevant internation- tions would be useful, as would the development of better pre- al groups, such as the International Society of Nephrology, dictive tests for microalbuminuria. In addition, cohort studies to address and advise national and regional health min- of hypertensive and diabetic populations might help develop istries on kidney and urological strategies as requested. better indicators that predict susceptibility to progression · Make major health and medical education programs avail- toward nephropathy. able on an annual basis through existing societies and agen- cies to train and update physicians, nurses, technicians, and other relevant health professionals. CONCLUSIONS: PROMISES AND PITFALLS · Develop selected centers of excellence for education, train- ing, clinical care, and prevention of kidney and urological Kidney disease and kidney failure, especially as a complication disease and clinical care of renal failure. At least 10 such cen- of type 2 diabetes mellitus and hypertension, are rising globally ters should be developed in the next decade and located in and are rising faster in developing countries. Kidney failure the countries of the former Soviet Union, Africa, Asia, patients account for a small fraction of the disease burden but a Eastern Europe, and Latin America. Funds should be disproportionately high cost. CKD, along with all chronic dis- provided by international and national agencies and eases, is placing long-term demands on health care. On a global national government organizations and be sustained for up scale, RRT is rising sharply in terms of costs and is usually to 10 years. unavailable in developing countries. Hemodialysis and peri- toneal dialysis are life saving, but in the long term they require coupling with newer, proven, interventional pharmacological treatments that frequently delay or stop continuing progression REFERENCES to ESRD. Advances in the past decade have proven that primary Agodoa, L. Y., L. Appel, G. L. Bakris, G. Beck, J. Bourgoignie, J. P. Briggs, and and secondary prevention measures can now reduce the burden others (African American Study of Kidney Disease and Hypertension of ESRD, and if they are not widely disseminated, the need for Study Group). 2001. "Effect of Ramipril vs. Amlodipine on Renal Outcomes in Hypertensive Nephrosclerosis: A Randomized Controlled RRT will increase along with the certainty that the requirements Trial." Journal of the American Medical Association 285: 2719­28. of kidney disease patients cannot be met. Appel, L. J. 2003. "Lifestyle Modification as a Means to Prevent and Treat The following guidelines for diseases of the kidney and uri- High Blood Pressure." Journal of the American Society of Nephrology 14: nary system are recommended: S99­102. Barsoum, R. S. 2003. "End-Stage Renal Disease in North Africa." Kidney · Expand surveillance of the prevalence of various kidney and International 63 (Suppl. 83): S111­14. urological diseases in developing countries. Provide support Bergman, S., B. O. Key, K. Kirk, D. G. Warnock, and S. G. Rostand. 1996. "Kidney Disease in the First-Degree Relatives of African-Americans for further epidemiological studies in selected countries for with Hypertensive End-Stage Renal Disease." American Journal of assessing the prevalence of kidney disease and interventions Kidney Diseases 27: 341­46. to address it and for establishing an international kidney Brenner, B. M., and J. Zagrobelny. 2003. "Clinical Renoprotection Trials disease data center. Involving Angiotensin II­Receptor Antagonists and Angiotensin- Converting-Enzyme Inhibitors." Kidney International 63 (Suppl. 83): · Promote public awareness in developing countries about S77­85. the nature and early signs of kidney disease along with Chugh, K. S., and V. Jha. 1995. "Differences in the Care of ESRD Patients knowledge of prevention measures and therapies. Worldwide: Required Resources and Future Outlook." Kidney Inter- · Focus more attention on the increasing prevalence of dia- national 48: S7­13. betes and hypertension, and develop kidney disease pro- Clark, W. F., D. N. Churchill, L. Forwell, G. Macdonald, and S. Foster. 2000. grams in that context. Measures of kidney function and Canadian Medical Association Journal 162 (2): 195­98. protein excretion should be taken. The implementation of Coresh, J., B. C. Astor, A. T. Greene, G. Eknoyan, and A. S. Levey. 2003. "Prevalence of Chronic Kidney Disease and Decreased Kidney primary and secondary prevention to reduce the prevalence Function in the Adult U.S. Population: Third National Health and of ESRD should be expanded. Nutrition Examination Survey." American Journal of Kidney Diseases · Increase coordination and resources for efficient and timely 41: 1­12. distribution of supplies and equipment, assessment of Coresh, J., B. Astor, and M. Sarnak. 2004. "Evidence for Increased Cardiovascular Disease Risk in Patients with Chronic Kidney Disease." patients, and frequent dialysis for acute renal failure patients Current Opinion in Nephrology and Hypertension 13 (1): 73­81. caused by crush injuries during such major disasters as Foxman, B. 2003. "Epidemiology of Urinary Tract Infections: Incidence, earthquakes. Countries in earthquake-prone regions should Morbidity, and Economic Costs." Disease-a-Month 49: 53­70. 704 | Disease Control Priorities in Developing Countries | John Dirks, Giuseppe Remuzzi, Susan Horton, and others Freedman, B. I., J. M. Soucie, and W. M. McClellan. 1997. "Family History Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. New York: of End-Stage Renal Disease among Incident Dialysis Patients." Journal Oxford University Press. of the American Society of Nephrology 8: 1942­45. McClellan, W. M., and W. D. Flanders. 2003. "Risk Factors for Progressive Golan, L., J. D. Birkmeyer, and H. G. Welch. 1999. "The Cost-Effectiveness Chronic Kidney Disease." Journal of the American Society of Nephrology of Treating All Patients with Type 2 Diabetes with Angiotensin- 14: S65­70. Converting Enzyme Inhibitors." Annals of Internal Medicine 131 (9): Moeller, S., S. Gioberge, and G. Brown. 2002. "ESRD Patients in 2001: 660­67. Global Overview of Patients, Treatment Modalities, and Development Grantham, J. 1997. "Pathogenesis of Autosomal Dominant Polycystic Trends." Nephrology Dialysis Transplantation 17: 2071­76. Kidney Disease: Recent Developments." In Hereditary Kidney Diseases, Morton, A. R., E. A. Iliescu, and J. W. Wilson. 2002. "Nephrology: 1. ed. A. Sessa, F. Conte, M. Meroni, and G. Battini, vol. 122, 1­9, Investigation and Treatment of Recurrent Kidney Stones." Canadian Contributions to Nephrology. Basel, Switzerland: Karger. Medical Association Journal 166: 213­18. Guay-Woodford, L. M., Z. H. Jafri, and J. Bernstein. 2000. "Other Cystic National Kidney Foundation. 2002. "K/DOQI Clinical Practice Guidelines Kidney Diseases." In Comprehensive Clinical Nephrology, ed. R. J. for Chronic Kidney Disease: Evaluation, Classification, and Johnson and J. Feehally, 50.1­12. London: Mosby. Stratification." American Journal of Kidney Diseases 39 (Suppl. 1): Hooton, T. 2000. "Urinary Tract Infections in Adults." In Comprehensive S1­266. Clinical Nephrology, ed. R. J. Johnson and J. Feehally, 56.1­12. London: Nelson, R. G. 2001. "Diabetic Renal Disease in Transitional and Mosby. Disadvantaged Populations." Nephrology 6: 9­17. Hostetter, T. H. 2004. "Chronic Kidney Disease Predicts Cardiovascular Nelson, R. G., H. Morgenstern, and P. H. Bennett. 1998. "Birth Weight and Disease." New England Journal of Medicine 351 (13): 1344­46. Renal Disease in Pima Indians with Type 2 Diabetes Mellitus." Hoy, W. E., Z. Wang, P. R. A. Baker, and A. M. Kelly. 2003. "Secondary American Journal of Epidemiology 148: 650­56. Prevention of Renal and Cardiovascular Disease: Results of a Renal and Oishi, K., P. Boyle, M. J. Barry, R. Farah, F. L. Gu, S. Jacobson, and others. Cardiovascular Treatment Program in an Australian Aboriginal 1998. "Epidemiology and Natural History of Benign Prostatic Community." Journal of the American Society of Nephrology 14: Hyperplasia." In Fourth International Consultation on BPH, S178­85. Proceedings, ed. L. Denis, K. Griffiths, S. Khoury, A. T. K. Cockett, Hussain, M., M. Lai, B. Ali, S. Ahmed, N. Zafar, A. Naqvi, and A. Rizvi. J. McConnell, C. Chatelain, G. Murphy, O. Yoshida (Health Publication 1996. "Management of Urinary Calculi Associated with Renal Failure." Ltd.), 23­59. Plymouth, U.K.: Plymbridge Distributors Ltd. Journal of the Pakistan Medical Association 45 (8): 205­8. Parks, J., and F. L. Coe. 1996. "The Financial Effects of Kidney Stone Imperatore, G., W. C. Knowler, D. J. Pettitt, S. Kobes, P. H. Bennett, and Prevention." Kidney International 50 (5): 1706­12. R. L. Hanson. 2000. "Segregation Analysis of Diabetic Nephropathy Pasternak, M. S., and R. H. Rubin. 1997. "Urinary Tract Tuberculosis." In in Pima Indians." Diabetes 49: 1049­56. Diseases of the Kidney, 6th ed., ed. R. W. Schrier and C. W. Gottschalk, Johnson, R., and J. Feehally. 2000. "Introduction to Glomerular Disease: 989­1009. Boston: Little, Brown. Clinical Presentation." In Comprehensive Clinical Nephrology, ed. R. J. Perneger, T. V., P. K. Whelton, and M. J. Klag. 1995. "Race and End-Stage Johnson and J. Feehally, 20.1­14. London: Mosby. Renal Disease. Socioeconomic Status and Access to Health Care as Kher, V. 2002. "End-Stage Renal Disease in Developing Countries." Kidney Mediating Factors." Archives of Internal Medicine 155: 1201­8. International 62: 350­62. Peterson, J. C., S. Adler, J. M. Burkart, T. Greene, L. A. Hebert, L. G. Kiberd, B. A., and K. K. Jindal. 1998. "Routine Treatment of Insulin- Hunsicker, and others. 1995."Blood Pressure Control, Proteinuria, and Dependent Diabetic patients with ACE Inhibitors to Prevent Renal the Progression of Renal Disease: The Modification of Diet in Renal Failure: An Economic Evaluation." American Journal of Kidney Diseases Disease Study." Annals of Internal Medicine 123: 754­62. 31 (1): 49­54. Pugh, J. A., M. P. Stern, S. M. Haffner, C. W. Eifler, and M. Zapata. 1988. King, H., R. E. Aubert, and W. H. Herman. 1998. "Global Burden of "Excess Incidence of Treatment of End-Stage Renal Disease in Mexican Diabetes, 1995­2025: Prevalence, Numerical Estimates, and Americans." American Journal of Epidemiology 127: 135­44. Projection." Diabetes Care 21: 1414­31. Robertson, W. G. 2003. "Renal Stones in the Tropics." Seminars in La Vecchia, C., F. Levi, and F. Lucchini. 1995. "Mortality from Benign Nephrology 23: 77­87. Prostatic Hyperplasia: Worldwide Trends 1950­92." Journal of Rostand, S. G. 1992. "U. S. Minority Groups and End-Stage Renal Disease: Epidemiology and Community Health 49: 379. A Disproportionate Share." American Journal of Kidney Diseases 19: Li, P. K. T., and K. M. Chow. 2001. "The Cost Barrier to Peritoneal Dialysis 411­13. in the Developing World: An Asian Perspective." Peritoneal Dialysis Ruggenenti, P., A. Schieppati, and G. Remuzzi. 2001. "Progression, International 21: S307­13. Remission, Regression of Chronic Renal Diseases." Lancet 357: 1601­8. Lin, J. L., D. T. Tan, K. H. Hsu, and C. C. Yu. 2001. "Environmental Lead Rutkowski, B. 2002. "Changing Pattern of End-Stage Renal Disease in Exposure and Progressive Renal Insufficiency." Archives of Internal Central and Eastern Europe." Nephrology Dialysis Transplantation 15: Medicine 161: 264­71. 156­60. Luke, R. G. 1999. "Hypertensive Nephrosclerosis: Pathogenesis and Satko, S. G., and B. I. Freedman. 2001. "Screening for Subclinical Prevalence. Essential Hypertension Is an Important Cause of End- Nephropathy in Relatives of Dialysis Patients." Seminars in Dialysis Stage Renal Disease." Nephrology Dialysis Transplantation 14: 2271­78. 14 (5): 311­12. Mani, M. K. 2003."Prevention of Chronic Renal Failure at the Community Schieppati, A., N. Perico, and G. Remuzzi. 2003. "The Potential Impact of Level." Kidney International 63 (Suppl. 83): S86­89. Screening and Intervention for Renal Diseases in Developing Martinez-Maldonado, M. 1998. "Hypertension in End-Stage Renal Countries." Nephrology Dialysis Transplantation 18: 858­59. Disease." Kidney International 54 (68): 67­72. Seaquist, E. R., F. C. Goets, S. Rich, and J. Barbosa. 1989. "Familial Mathers, C. D., A. D. Lopez, and C. J. L. Murray. "The Burden of Disease Clustering of Diabetic Kidney Disease: Evidence for Genetic and Mortality by Condition: Data, Methods, and Results for 2001." In Susceptibility to Diabetic Nephropathy." New England Journal of Global Burden of Disease and Risk Factors, eds. A. D. Lopez, C. D. Medicine 320: 1161­65. Diseases of the Kidney and the Urinary System | 705 Seedat, Y. K. 2003. "Glomerular Disease in the Tropics." Seminars in Winkelmayer, W. C., M. C. Weinstein, M. A. Mittleman, R. J. Glynn, and Nephrology 23: 12­20. J. S. Pliskin. 2002. "Health Economic Evaluations: The Special Case of Sever, M. S., E. Erek, R. Vanholder, E. Akoglu, M. Yavuz, H. Ergin, and End-Stage Renal Disease Treatment." Medical Decision Making 22: others (Marmara Earthquake Study Group). 2001. "The Marmara 417­30. Earthquake: Epidemiological Analysis of the Victims with Xue, J. L., J. Z. Ma, T. A. Louis, and A. J. Collins. 2001. "Forecast of the Nephrological Problems." Kidney International 60: 1114­23. Number of Patients with End-Stage Renal Disease in the United States Stengel, B., S. Billon, P. van Dijk, K. Jager, F. Dekker, K. Simpson, and to the Year 2010." Journal of the American Society of Nephrology 12: others. 2003. "Trends in the Incidence of Renal Replacement Therapy 2753­58. for End-Stage Renal Disease in Europe, 1990­1999." Nephrology Zatz, R., J. E. Romão Jr., and I. L. Noronha. 2003. "Nephrology in Latin Dialysis Transplantation 18: 1824­33. America, with Special Emphasis on Brazil." Kidney International 63 Weening, J. 2004."Advancing Nephrology around the Globe: An Invitation (Suppl. 83): S131­34. to Contribute." Journal of the American Society of Nephrology 15: 2761­62. WHO (World Health Organization). 2002. "Reducing Risks, Promoting Healthy Life." In The World Health Report 2002, ed. WHO. Geneva: WHO. http://www.who.int/whr/en/. 706 | Disease Control Priorities in Developing Countries | John Dirks, Giuseppe Remuzzi, Susan Horton, and others Chapter 37 Skin Diseases Roderick Hay, Sandra E. Bendeck, Suephy Chen, Roberto Estrada, Anne Haddix, Tonya McLeod, and Antoine Mahé In assigning health priorities, skin diseases are sometimes Sub-Saharan Africa, leads to a similar impact on life quality thought of, in planning terms, as small-time players in the compared with non-HIV-related skin problems, although the global league of illness compared with diseases that cause signif- use of antiretroviral therapy significantly improves quality of icant mortality, such as HIV/AIDS, community-acquired pneu- life (Mirmirani and others 2002). Those findings indicate that monias, and tuberculosis. However, skin problems are generally skin diseases have a significant impact on quality of life. among the most common diseases seen in primary care settings Although mortality rates are generally lower than for other in tropical areas, and in some regions where transmissible dis- conditions, people's needs for effective remedies for skin con- eases such as tinea imbricata or onchocerciasis are endemic, ditions should be met for a number of important reasons. they become the dominant presentation. For instance, the World Health Organization's 2001 report (Mathers 2006) on · First, skin diseases are so common and patients present in the global burden of disease indicated that skin diseases were such large numbers in primary care settings that ignoring associated with mortality rates of 20,000 in Sub-Saharan Africa them is not a viable option. Children, in particular, tend to in 2001. This burden was comparable to mortality rates attrib- be affected, adding to the burden of disease among an uted to meningitis, hepatitis B, obstructed labor, and rheumat- already vulnerable group. ic heart disease in the same region. Using a comparative assess- · Second, morbidity is significant through disfigurement, ment of disability-adjusted life years (DALYs) from the same disability, or symptoms such as intractable itch, as is the report, the World Health Organization recorded an estimated reduction in quality of life. For instance, the morbidity from total of 896,000 DALYs for the region in the same year, similar secondary cellulitis in lymphatic filariasis, which may lead to that attributed to gout, endocrine disease, panic disorders, to progressive limb enlargement, is severe, and subsequent and war-related injuries. As noted later, those figures require immobility contributes to social isolation. confirmation by more detailed studies, and their practical · Third, the relative economic cost to families of treating even application to health interventions needs to be tested. trivial skin complaints limits the uptake of therapies. Assessing the impact of skin disease on the quality of life in Generally, families must meet such costs from an over- comparison with that of chronic nondermatological diseases is stretched household budget, and such expenses in turn difficult; however, the study by Mallon and others (1999), reduce the capacity to purchase such items as essential foods which was not carried out in a developing country, compares (Hay and others 1994). the common skin disease acne with chronic disorders such as · Fourth, screening the skin for signs of disease is an impor- asthma, diabetes, and arthritis and finds comparable deficits in tant strategy for a wide range of illnesses, such as leprosy, yet objective measurements of life quality. Skin disease related to a basic knowledge of the simple features of disease whose HIV, which may constitute an important component of the presenting signs occur in the skin is often lacking at the skin disease burden in developing countries, particularly in primary care level. 707 A shortage of elementary skills in the management of skin pediculosis capitis, tinea capitis, or pyoderma (Figueroa and diseases is a further confounding problem. A number of stud- others 1996). Those figures mirror work carried out elsewhere. ies assessing success in the management of skin diseases in pri- For instance, in Tanzania, in a survey of two village communi- mary care settings in the developing world find that treatment ties, Gibbs (1996) found that 27 percent of patients had a treat- failure rates of more than 80 percent are common (Figueroa able skin disease, and once again, infections were the most and others 1998; Hiletework 1998). An additional point, often common diseases. Overcrowding was a major risk factor in that overlooked, is that skin diseases in the developing world are survey. A similar community-based survey in Sumatra, often transmissible and contagious but are readily treatable Indonesia, showed a 28 percent prevalence of skin disease (Saw (Mahé, Thiam N'Diaye, and Bobin 1997). and others 2001). What seems to influence the overall preva- A number of common diseases account for the vast majority lence and pattern of skin conditions in certain areas is the exis- of the skin disease burden; therefore implementing effective tence of a number of common contagious diseases, notably, treatments targeted at those conditions results in significant scabies and tinea capitis. Hot and humid climatic conditions gains for both personal and public health. Even where eradica- may also predispose populations to pyoderma, thereby affect- tion is impossible, control measures may be important in ing the distribution of disease. reducing the burden of illness; yet few systematic attempts have been made to validate control programs for skin diseases as public health interventions. PATTERNS OF SKIN DISEASES AT THE COMMUNITY LEVEL PREVALENCE OF SKIN DISEASES A recent (unpublished) survey by the International Foundation of Dermatology designed to provide information Few studies aimed at estimating the prevalence of skin diseases about community patterns of skin disease in nine different have been carried out in Western societies. However, Rea, countries across the world--Australia (Northwest Territory), Newhouse, and Halil's (1976) study in Lambeth, south Ethiopia, Indonesia, Mali, Mexico, Mozambique, Senegal, London, which used a questionnaire-based, population- Tanzania, and Thailand)--and poor regions in other tropical centered approach backed by random examination, reveals an environments from Mexico to Madagascar indicates that overall 52 percent prevalence of skin disease, of which the the following were the main skin conditions at community investigators judged that just over half the cases required treat- level: ment. Studies from developing countries have generally adopted a more inclusive approach that uses systematic, · Scabies. Although scabies was often the commonest skin community-based surveys backed by examination. Published disease, it was completely absent in some regions. figures for the prevalence of skin diseases in developing coun- · Superficial mycoses. This group of infections was usually tries range from 20 to 80 percent. reported as one of the three commonest diseases. In a study in western Ethiopia, between 47 and 53 percent of · Pyoderma. This disease was often, but not invariably, associ- the members of two rural communities claimed to have a skin ated with scabies. disease (Figueroa and others 1998), but when they were exam- · Pediculosis. This disease was the subject of much variation ined, 67 percent of those who denied having skin problems but is often overlooked in surveys. Firm, community-level were found to have treatable skin conditions, most of which data on the prevalence of pediculosis are deficient; thus, this were infections. However, prevalence alone does not equate disease is not discussed further in this chapter. with disease burden. For instance, most communities recognize · Eczema or dermatitis. Although this disease was usually scabies as a problem because of its intractable itching and sec- unclassified, irritant dermatitis and chronic lichen simplex ondary infection, whereas they may ignore tinea capitis, which were often cited. is equally common among the same populations, because they · HIV-related skin disease. This disease was reported mainly in are aware that it follows a benign and asymptomatic course in Africa. The pruritic papular dermatitis of AIDS is a specific many patients. problem. Researchers agree about the main risk factors associated · Pigmentary anomalies. Three different problems were cited: with skin disease in developing countries, the most important hypopigmentation, often diagnosed as pityriasis alba, a form of which appears to be household overcrowding. In primary of eczema; melasma; and dermatitis caused by cosmetic schools in western Ethiopia, more than 80 percent of randomly bleaching agents (Mahé and others 2003). examined schoolchildren had at least one skin disease, · Acne. This disease was reported as an emerging and which was usually caused by one of four conditions: scabies, common problem. 708 | Disease Control Priorities in Developing Countries | Roderick Hay, Sandra E. Bendeck, Suephy Chen, and others These diseases are the same as those recorded in the litera- Table 37.1 Search Strategy for Therapies ture described previously. Other skin conditions cited by dif- Search term for ferent members of the group surveyed follow: Disease Search term for disease treatment · Tropical ulcer. The incidence was highly variable, but tropi- Scabies ["scabies"] ["treatment of" or cal ulcer can account for a huge workload in primary care "ivermectin" or centers in endemic areas. "permethrin" or "Lindane" or · Nonfilarial lymphoedema. This condition was mainly con- "malathion" or fined to Ethiopia. "benzoyl benzoate" or · Onchodermatitis, filarial lymphoedema, endemic trepone- "crotamiton" or matoses, Buruli ulcers, and leprosy. These conditions are dis- "sulfur"] cussed in detail elsewhere in this book, but note that they Pyoderma or ["skin diseases, bacterial" or ["drug therapy" or often present with skin changes and symptoms. bacterial skin "ecthyma" or "staphylococcal "prevention & control" infections skin infections" or "impetigo" or "therapy"] According to World Bank (2002) figures for low-income or "pyoderma" or "folliculitis"] populations in 2000, the estimated numbers of individuals Tinea capitis: ["tinea capitis"] ["drug therapy" or infected with pyoderma and scabies, based on the highest "therapy" or prevalence figures from community surveys in the developing "prevention & control"] world, are 400 million and 600 million, respectively. Based on Tinea imbricata: ["tinea imbricata.mp"] the lowest prevalence figures, these estimated numbers are 40 million and 50 million, respectively. For tinea capitis, the Tropical ulcer ["tropical ulcer$.ti"] or ["skin ulcer(explode)" and estimated number of cases based on the highest estimates of "tropic$.mp"] prevalence for Sub-Saharan Africa alone is 78 million. Overall, these data suggest that significant changes could be Source: Authors. made in reducing the burden of skin diseases by focusing on Note: Terms in brackets are medical subject heading terms. If no standard medical subject heading terms were available, databases were searched either using the title option (denoted the small group of conditions, particularly infections, that as ".ti") or the keyword option (denoted as ".mp"). account for the bulk of the community case load. This chapter concentrates on those conditions for which such a strategy SKIN DISEASES could be implemented--namely, scabies, pyoderma, fungal infections, tropical ulcers, HIV/AIDS-related dermatoses, and Scabies pigmentary disorders. Scabies is a common ectoparasitic infestation caused by Sarcoptes scabei, a human-specific mite that is highly prevalent EFFECTIVE THERAPIES in some areas of the developing world. Scabies is transmitted by direct contact. In industrial societies, it is usually seen in sexu- In considering the evidence for effective treatment, a subgroup ally active adults, although it may also appear in the form of of the team (Bendeck, Chen, and McLeod) undertook a data clusters of cases among the elderly in residential homes. Peaks search to establish the evidence base for treatment of the com- of infection in communities may be cyclical. The ease of trans- mon conditions. They carried out comprehensive searches of mission appears to depend, in part, on the parasitic load, and the MEDLINE (1966­April 2003) and EMBASE (1980­April some patients, including the elderly, may have large numbers of 2003) databases to identify therapeutic studies on scabies, parasites present. By contrast, in healthy adults, the total para- pyodermas, and superficial mycoses (but note that many of the site load may be low, but they, nonetheless, may suffer from studies were performed in industrial countries). They used highly itchy lesions. The organisms can also reach high densi- foreign-language articles if an English abstract was provided. ties in patients suffering from a severe depression of immuno- Table 37.1 shows search terms for each of the skin diseases logical responses, as in HIV infection. In this crusted or common in the developing world and for treatment. Norwegian form of scabies, lesions may present with atypical The team members reviewed study titles and abstracts to crusted lesions that itch little. select relevant articles and scrutinized the bibliographies of In developing countries, transmission commonly occurs in selected articles to identify pertinent studies not captured in young children and infants and their mothers and is related to the initial literature search. They defined admissible evidence as close contact, overcrowding, and shared sleeping areas. Sexual primary therapeutic studies, based on clinical evaluation, of the contact is less important as a means of transmission. Scabies is treatment of each disease. also a scourge of prisons in developing countries, where it is Skin Diseases | 709 associated with overcrowding (Leppard and Naburi 2000). No · Permethrin 5 percent cream. This effective, nonirritant treat- evidence exists that transfer is related to inadequate hygiene. ment is usually administered as a cream applied all over the The most important complication of scabies is secondary body. A single application washed off after 8 to 12 hours is bacterial infection, usually caused by Group A streptococci. used. The tubes are small, and adequate quantities should Evidence from studies among the indigenous population of be prescribed. This treatment is also the most costly of the northern Australia indicates that this infection is not always topical therapies. benign and that persistent proteinuria is associated with past scabies infestation, suggesting that nephritis related to second- Treatment failures in developing countries may be related ary infection of scabies may cause long-lasting renal damage to the lack of a suitable place in many communities where (White, Hoy, and McCredie 2001). patients can apply treatment effectively over the entire body The disease presents with itchy papules and sinuous linear from the neck down in privacy. tracks in the skin that can be highly pruritic and particularly Oral ivermectin, which is an important drug in the treatment troublesome at night. Often more than one member of a of onchocerciasis, has also been used in patients with scabies, household has the disease. particularly those with the crusted form or in places such as pris- ons, where large numbers of infected individuals live in close Treatment. The treatments used for scabies are mainly applied proximity. It has also been applied as a community-based treat- topically. Treatment is not based on treating just affected ment and is reported to be effective as such (Hegazy and others individuals, both because of the ease with which scabies 1999). It is not licensed for the treatment of scabies, and the lack spreads and because symptoms may develop days or weeks of safety data on the use of ivermectin in infants limits its use. after infection. The advice given to patients always includes a In addition, insufficient evaluations of its efficacy and cost- recommendation to treat the entire household with a similar effectiveness in developing countries have been carried out. medication, a difficult problem when many people live in the Evidence for Effective Therapies. The team identified 56 arti- same dwelling. The treatments commonly available include cles on therapies for scabies and found the following to be the the following: viable ones: oral and topical ivermectin, permethrin, gamma benzene hexachloride, benzyl benzoate, crotamiton, malathion, · Sulfur ointments. There are no controlled clinical studies of and topical sulfur. Table 37.2 summarizes the evidence for the use of this cheap medication, which is usually made up ivermectin versus a placebo or permethrin and for topical in an ointment base. Soap containing sulfur is available in ivermectin, as well as for the less expensive topical sulfur. some areas. Anecdotally, sulfur ointment needs to be applied for at least one week to the entire body. Irritation is a com- Community-Based Treatments for Scabies. Few studies have mon side effect, and lower concentrations, such as 2.5 per- addressed the problem of community-administered treatments cent, are applied to infants. for scabies, despite the argument that without a community · Benzyl benzoate. A 10 to 25 percent benzyl benzoate emul- approach to therapy in many developing countries, the success- sion is applied over the entire body and left on the skin for ful management of scabies in areas where it affects more than up to 24 hours before washing off. Current recommenda- 5 to 6 percent of the population is doomed to failure. Taplin tions suggest that one to three applications may be suffi- and others' (1991) study of the use of 5 percent permethrin cient, but consensus on the optimal treatment regimen cream in the San Blas Islands, Panama, confirms this view. A would be useful. Benzyl benzoate emulsion is an irritant and three-year program of treatments backed by surveillance can lead to secondary eczema in some patients. reduced the prevalence of scabies from 33 percent to less than · Gamma benzene hexachloride (Lindane). This product is 1 percent; however, a three-week break in regular treatment was widely available and is used as a single application washed followed by a rapid increase in prevalence to 3 percent. The off after 12 to 24 hours. Concerns have arisen about the results of treatments involving the application of similar proto- increasing risk of drug resistance and the absorption of the cols, but using other topical agents, are not available. Oral iver- drug through the skin. It is also not used in children because mectin lends itself to a community-based treatment approach of reports of neurotoxicity and fits. This product is not and has been used in this way (Hegazy and others 1999; Usha available in many countries. and Gopalakrishnan Nair 2000), but insufficient follow-up data · Malathion (0.5 percent) in an aqueous base. The highly puri- are currently available to comment further on this approach. fied commercial forms are effective after a single applica- tion, although a second is advised. No data are available on the use of this preparation in developing countries. Bacterial Skin Infections or Pyoderma · Crotamiton cream or monosulfiram 25 percent. These alter- Bacterial skin infections or pyoderma are common in most native therapies have highly variable efficacy rates. developing countries (Mahé, Thiam N'Diaye, and Bobin 1997). 710 | Disease Control Priorities in Developing Countries | Roderick Hay, Sandra E. Bendeck, Suephy Chen, and others Table 37.2 Evidence of the Efficacy of Treatments for Scabies Treatment and Number average wholesale of people price Strongest evidence in study Results Comments Ivermectin oral Randomized clinical trial 55 79.3 percent cure with single dose of · Will treat concomitant strongy- US$5.20 (3 mg), given at (versus placebo) ivermectin 200 g/kg versus 16.0 percent loidiasis and onchocerciasis 200 g/kg, one or (Hegazy and others 1999) cure with placebo (p 0.001) · Not approved for scabies by the two doses U.S. Food and Drug Administration · Safety not established for children under five and pregnant women Randomized clinical trial 85 Single dose: 70.0 percent cure with · A single application of permethrin (versus permethrin) (Taplin ivermectin 200 g/kg versus 97.8 percent is superior to a single dose of and others 1991) cure with permethrin 5 percent ivermectin, which suggests that Second dose (two-week interval): ivermectin may not be effective at 95.0 percent cure with ivermectin all stages in the life cycle of the 200 g/kg (statistically equivalent cure parasite rates with ivermectin and permethrin used as single dose/application) Ivermectin (topical) Open-label, prospective, 32 100 percent cure rate with two doses of · Subjects treated with 1 percent single group (Macotela-Ruiz ivermectin 1 percent solution at six weeks ivermectin in a solution of and Ramos 1996) (no statistics reported) propylene glycol at 400 g/kg repeated once after one week · Well tolerated Sulfur compounds Open-label, nonrandomized, 102 71 percent cure at four weeks using · Typically used as 5 percent to (topical) prospective cohort (Usha and sulfur, 5 percent in children younger than 10 percent in petrolatum Ointment (480 grams) Gopalakrishnan Nair 2000) 12 months, and 10 percent in children · Messy and smelly US$2.32 older than 12 months · Must be applied repetitively for three nights · Mild local irritation may occur Source: Authors. g microgram; kg kilogram; mg milligram; p probability. Generally these infections arise as primary infections of the epidermis, causing a necrotic ulcer--a condition known as skin known as impetigo or as secondary infections of other ecthyma. However, some evidence suggests that streptococcal lesions such as scabies or insect bites. The usual bacterial causes infection may cause additional long-term damage through the are Group A streptococci or Staphylococcus aureus. Bacterial development of prolonged proteinuria, as described earlier in infections are common in communities. In many cases, no bac- relation to scabies. teriological confirmation is available from cultures, but surveys show that Group A streptococci account for a substantial num- Treatment. Treatment with topical antibacterials, such as ber of cases (Carapetis, Currie, and Kaplan 1999; Taplin and fusidic acid or mupirocin, is expensive; thus, the use of cheap- others 1973), which is not often the case in similar infections er agents, such as antiseptics, is an important option but one in temperate climates, where S. aureus dominates. This finding that has been evaluated in only a few instances. Chlorhexidine carries implications for the selection of treatment options. The and povidone iodine have both been used, but potassium per- reasons for this finding are not clear, although humidity and manganate is also said to be clinically effective. Gentian violet heat are associated with increased risk of bacterial skin infec- at concentrations of 0.5 to 1.0 percent is a cheap agent that is tion. In addition to these superficial infections, S. aureus also widely used, with proven in vitro efficacy against agents com- causes folliculitis, or hair follicle infections and abscesses. Rarer monly involved in pyoderma. Most of those compounds have causes of skin infection in developing countries include been used to prevent rather than to treat infections. The most cutaneous diphtheria and anthrax, as well as necrotizing infec- extensively evaluated topical preparations are fusidic acid oint- tion caused by Vibrio vulnificus. ment and mupirocin, which are given daily for up to 10 days. Bacterial infection causes irritation and some discomfort. In Those drugs are effective in eradicating bacterial infections but, some cases, the infection penetrates deep down through the as noted, are not cheap options. Group A streptococci are still Skin Diseases | 711 sensitive to penicillin, which can be used for treatment, with the treatment of pyodermas, a number of studies reported alternatives for staphylococcal infections being cloxacillin, effective topical therapies, namely: povidone-iodine solution, flucloxacillin, and erythromycin. Industrial countries largely hydrogen peroxide cream, electrolyzed strong acid aqueous view methicillin resistance among staphylococci as a nosoco- solution, tea ointment, Soframycin ointment, honey, fusidic mial problem, yet it has now spread to the community, and acid cream, trimethoprim-polymyxin B sulfate cream, rifax- skin infections provide an ideal medium for the spread of imin cream, sulconazole cream, miconazole cream, neomycin/ resistance, even in developing countries. S. aureus strains iso- polymyxin B-bacitracin (Neosporin) cream, terbinafine lated from skin sites, even in remote tropical areas, are now cream, and mupirocin. Systemic agents cited were cephalexin, resistant to beta-lactam penicillins and tetracyclines through erythromycin, penicillin, Augmentin, amoxicillin, sultami- the spread of resistance genes. Tetracycline ointment is still cillin, (di)cloxacillin, azithromycin, cefadroxil, cefpodoxime, available in many rural pharmacies and is widely used to treat cefaclor, ceftizoxime, clindamycin, clarithromycin, tetracy- superficial skin lesions, even though some bacterial infections cline, fluoroquinolones, and fusidic acid. will be unresponsive. Topical neomycin and bacitracin are Table 37.3 presents the evidence for commonly used anti- widely available, are associated with identifiable levels of treat- septics and some of the specific antibacterial agents. In prac- ment failure, and also carry a risk of sensitization or adverse tice, topical treatments such as chlorhexidine, povidone, and in effects. some cases neomycin or mupirocin will provide the most cost- effective control measures. For extensive infection, cloxacillin Evidence for Effective Treatment. The team reviewed 727 or erythromycin provides alternatives. However, current studies of therapies for pyoderma or bacterial skin infections. evaluations are subject to some weaknesses, such as a lack of These studies could be grouped into either prophylactic large, comparative studies, particularly of the topical therapies, regimens or therapeutic trials. For the prevention of pyoderma, including antiseptics, used in developing countries. the studies surveyed included the following effective therapies: Community-applied measures for managing skin infections chlorhexidine solution, hexachlorophene scrubbing, and have not been evaluated, but measures such as early treatment neomycin/polymyxin B-bacitracin (Neosporin) cream. For of scabies or basic wound care of sores might provide Table 37.3 Evidence of the Efficacy of Topical Treatments for Pyoderma Treatment, level of evidence, cost (manufacturer, Number of formulation, average people wholesale price) Evidence in study Results Commentsa Chlorhexidine gluconate Open-label, prospective 3,602 6.3 percent clinical pyoderma on · Neonatal cord pyoderma (4 percent) detergent solution cohort (versus nothing) postdischarge in the chlorhexidine group; · Prophylaxis study (Taplin and others 1973) 24 percent in the nonchlorhexidine group Level of evidence: VI (no statistics reported) Cost: Open-label, prospective 5,220 Hospital A: 15.2 percent of group without · Neonatal cord pyoderma · Clay-Park cohort (versus 70 per- and 2.1 percent with chlorhexidine · Prophylaxis study prevented cord pyoderma; hospital B: · Liquid, topical, 4 percent cent ethanol and versus · Performed and reported at two nothing) (Taplin and 21.0 percent with ethanol and 1.0 percent · 120 ml, US$7.01 different hospitals others 1973) with chlorhexidine prevented pyoderma (no statistics reported) Povidone-iodine solution Double-blind RCT 160 92 percent improvement with fusidic acid · Impetigo (Betadine) (fusidic acid cream plus and 88 percent with placebo · 14 percent of placebo group versus povidone iodine versus Level of evidence: II 4 percent in fusidic acid group may placebo cream plus have received antibiotics in weeks Cost: povidone iodine) 2 and 4, potentially explaining the · Alpharma (Seeberg and others lack of difference in efficacy U.S. Pharmaceutical 1984) Directory Open-label, prospective 25 12/12 Betadine responded; 0/13 salicylic · Disinfection of chronic wounds of acid responded lymphedematous patients · Solution, topical product, cohort (versus salicylic 10 percent, acid) (Linder 1978) · Outcome measure and statistics not clear · 400 ml, US$5.46 712 | Disease Control Priorities in Developing Countries | Roderick Hay, Sandra E. Bendeck, Suephy Chen, and others Table 37.3 Continued Treatment, level of evidence, cost (manufacturer, Number of formulation, average people wholesale price) Evidence in study Results Commentsa Potassium permanganate Level of evidence: none Cost: · A-A Spectrum · Crystal, NA · 500 gm, US$16.10 Mupirocin Double-blind RCT 52 100 percent of mupirocin patients versus · Impetigo/ecthyma (versus placebo vehicle) 85 percent of placebo (difference not Level of evidence: I · Outcome: cure or improvement (Koning and others significant) · 38 in final evaluation; no ITT Summary: Efficacy supported 2002) by two RCTs and several Double-blind, RCT 106 85 percent of mupirocin versus 53 percent · Secondarily infected dermatoses comparison studies; some (versus vehicle) vehicle-treated patients (p 0.007) with S. aureus or S. pyogenes concern about resistance (Daroczy 2002) · Outcome: marked or moderate Cost: improvement · GlaxoSmithKline (GSK) · 92 in final evaluation; no ITT Pharmaceuticals Open-label RCT (versus 97 90 percent of erythromycin and 96 percent · Impetigo contagiosa · Ointment, TP, 2 percent oral erythromycin) for mupirocin (no statistics given); long- · Outcome: cure or clinical (Eells and others 1986) term follow-up: 9 erythromycin versus · 22 gm, US$41.36 improvement 3 mupirocin patients developed new · Also looked at long-term (up to lesions (p 0.05) one month) follow-up Open-label RCT (versus 60 No significant difference in various · Impetigo oral erythromycin) evaluations of clinical efficacy except · Both articles present the same (Barton, Friedman, investigator's global evaluation research and Portilla 1988; (efficiency/safety performance) (p 0.01) · More adverse effects with Breneman 1990) erythromycin Investigator-blinded, 75 93 percent mupirocin versus 96 percent · Impetigo RCT (versus oral erythromycin (no statistical difference) · Also looked at bacterial erythromycin) Recurrence with erythromycin: 10 percent recurrences (McLinn 1988) of patients with S. aureus and 6 percent of · 53 patients clinically and bacterio- patients with S. pyogenes; recurrence with logically assessable; no ITT mupirocin: none. Source: Authors. gm gram; ITT intent to treat; ml milliliter; p probability; RCT randomized clinical trial; TP topical product. a. Comments include type of skin infection; indication of prophylaxis, otherwise therapeutic trial; ITT analysis; and other comments. significant benefits. In this area, carefully designed pilot control and some other common causes of foot infection, such as programs would provide extremely valuable data. Scytalidium. The clinical and social impact of fungal infections on individuals varies with local conditions. For instance, tinea pedis is a treatable condition that causes cracking and inflam- Fungal Infections mation with itching between the toes. It is generally viewed as Fungal infections that affect the skin and adjacent structures a nuisance that only marginally affects the quality of life; how- are common in all environments. They include infections such ever, under certain conditions its significance is far greater. For as ringworm or dermatophytosis; superficial candidosis and example, fungal infections of the web spaces and toenails in infections caused by lipophilic yeasts and Malassezia species; diabetics provide a portal of entry for S. aureus, an event closely Skin Diseases | 713 related to the development of serious foot complications in with T. tonsurans. This factor poses a dilemma in management, patients with peripheral vascular disease and neuropathy. because where the disease is common and endemic, a regular Similarly, foot infections originally caused by dermatophytes source will always exist for new, severe, inflammatory infec- can develop into more serious disabling infections through sec- tions in children. Therefore, addressing this issue by tackling ondary Gram-negative bacterial infection among certain occu- individual cases without addressing the reservoir, albeit illogi- pational groups in the tropics, such as workers in heavy indus- cal, may ultimately be the most practical approach. try, the police, or the armed forces. Wearing heavy footwear is The diagnosis of tinea capitis is difficult to make clinically a risk factor for the emergence of this problem. in mild cases because the main presenting signs are localized Other infections, such as oropharyngeal candidosis, are patches of hair loss with fine scaling. In some children, the hair important complications of HIV. This commonest infectious loss is more diffuse. With the inflammatory forms, circum- complication of AIDS is a potential early marker. Whereas in scribed patches of hair loss with erythema and pustulation also many patients it may simply have nuisance value, in others it occur, and the whole area is raised into a boggy mass. The only has a more serious impact and leads to dysphagia and loss of way to confirm the diagnosis accurately is to take hair samples appetite. Malassezia infections such as pityriasis versicolor are for culture and microscopy, which is not possible in many areas also common in the developing world and often occur in more because they lack laboratory diagnostic facilities. One specific than 50 percent of the population; however, they are generally form of tinea capitis,favus,is clinically recognizable and distinct, asymptomatic but cause patches of depigmentation, and because the scalp is covered with white plaques called scutula. patients seldom seek treatment. The infection is chronic and can develop into permanent, scar- Some fungal infections are extremely widely distributed or ring alopecia.Inhabitants of endemic areas often recognize favus common in defined endemic areas. They include tinea capitis as a distinct condition that causes chronic illness, and as a result, and tinea imbricata. the uptake of consultation for treatment is higher. Highly effective, topically applied treatments for tinea Tinea Capitis. Tinea capitis is a common, contagious disease capitis are unavailable, and even though simple remedies of childhood that can spread extensively in schools. It is caused such as benzoic acid compound (Whitfield's ointment) may by dermatophyte fungi of the genera Trichophyton and lead to clinical improvements, relapse is almost universal. Microsporum (Elewski 2000). Infections can spread from child Nevertheless, the use of topical therapies may limit the spread to child (anthropophilic infections) or from animals to children of tinea capitis. Treatment depends on the use of oral therapies. (zoophilic infections). Anthropophilic infections tend to be The most widely available of these is griseofulvin, which is endemic or epidemic, whereas the zoophilic forms occur spo- given to children in doses of 10 to 20 milligrams per kilogram radically. The commonest sources and causes of zoophilic daily for a minimum of six weeks. Noncontrolled studies show infections are cats and dogs (Microsporum canis), cattle and that a single dose of 1 gram of griseofulvin given under super- camels (Trichophyton verrucosum), and rodents (T. mentagro- vision can eradicate infection in more than 70 percent of indi- phytes). The causes of the anthropophilic form of this infection viduals, but such regimens have not been adequately assessed vary in different areas of the world. Although in areas of the under trial conditions to determine their effect on community developing world this condition is endemic at high levels, levels of infection, nor are follow-up data available. in many parts of Africa it is a common condition affecting Recent years have seen the development of a number of more than 30 percent of children in primary schools. The effective, new, oral antifungals, including terbinafine, itracona- main African species are M. audouinii, T. soudanense, and zole, and fluconazole. Terbinafine is a highly active agent that is T. violaceum. The last is also found in the Middle East and effective in the treatment of dermatophyte infections. It is given India. T. tonsurans, the form of tinea capitis endemic in the in doses of 62.5 milligrams for those under 10 kilograms, United States (Wilmington, Aly, and Frieden 1996) and in parts 125 milligrams for those weighing 10 to 40 kilograms, and of Europe, such as France and the United Kingdom (Hay and 250 milligrams for those over 40 kilograms. Evidence indicates others 1996), is extremely resistant to treatment. No evidence that it is effective after one week of therapy in T. violaceum and indicates that this form has spread to Africa yet, although this T. tonsurans infections, but the best responses are seen when it possibility exists. is used for four weeks. Unfortunately, at these doses it is less Families of children with tinea capitis seldom present for effective for Microsporum infections, although some data sug- treatment. However, in a small proportion of individuals, tinea gest that responses are significant if the doses are doubled. This capitis produces a highly inflammatory lesion with suppura- drug is, therefore, difficult to administer in standardized proto- tion on the scalp along with permanent scarring and local cols when the cause of infection is uncertain. Itraconazole is hair loss. The numbers of infected individuals showing this also effective, but no suitable pediatric formulation is available highly symptomatic change are not known with any accuracy, because it is marketed in a capsule form that is difficult to but it is believed to occur in about 5 percent of cases, more administer to young children. Fluconazole is also effective, 714 | Disease Control Priorities in Developing Countries | Roderick Hay, Sandra E. Bendeck, Suephy Chen, and others although comparative studies of its use are not available. All is endemic, it can be a significant problem occupying much of three drugs are costly, and a community-based program that the time of health aid post staff. uses them would be difficult to fund and implement. Individual treatments have depended on the antifungals The team found a total of 432 articles for the treatment of described earlier, including griseofulvin. Terbinafine and itra- tinea capitis. Table 37.4 presents key references for the oral conazole are highly effective, but their cost has constrained therapies, the mainstay of therapy. The effective treatments their use. As table 37.5 shows, the relapse rates after itracona- included topical therapies (benzoic acid, bifonazole, selenium zole are also higher than after terbinafine (Budimulja and sulfide, ketaconazole shampoo, and miconazole shampoo) as others 1994). Topical agents such as benzoic acid compound well as systemic agents (griseofulvin, terbinafine, itraconazole, (Whitfield's ointment) are helpful, but are seldom curative and fluconazole,and ketoconazole).The results of topical treatments are difficult to apply over such large areas. Some patients may appear inferior to those of oral therapy, although they have not be treated with locally derived treatments, such as the sipoma been directly compared, and some of the topical agents were paint used in Papua New Guinea, which contains salicylic acid, applied to prevent transmission rather than to treat infection. brilliant green, and kerosene. Traditional treatments have also Attempts at community control of tinea capitis have been been used, but never evaluated. The leaves of Cassia alata, for devised but have not been monitored adequately. The methods instance, are widely used in the western Pacific. have been based on surveillance through culture and treatment The team found studies of the use of griseofulvin, of all infected children. Culture-based diagnosis is difficult to terbinafine, and itraconazole for tinea imbricata. Some studies implement regularly in developing countries. The treatment did mention sipoma paint and Cassia alata, but no studies used for community therapy has been griseofulvin in conven- evaluating their efficacy have been performed. The team also tional daily or large single doses, but those approaches have not found case reports supporting the use of griseofulvin. been compared. In addition, control protocols usually advise Different treatments for use on a community basis need to treating carriers with topically applied agents such as selenium be evaluated because the impact of this condition on local sulfide (which is relatively cheap) or a miconazole shampoo health services in areas of high prevalence is heavy in terms of (which is moderately priced). In practice, some "carriers" are both time and staff workload. really patients with extremely localized and hard-to-detect infections, and such patients will not respond to topical treat- ment in the long term. A second problem is the absolute Tropical Ulcer reliance on laboratory confirmation of cultures to direct Tropical ulcer is a common condition found mainly in chil- treatment of carriers. Therefore, other strategies need to be dren and teenagers in well-defined tropical regions. It usually evaluated, such as reducing the community load, perhaps by affects the lower limbs (Bulto, Maskel, and Fisseha 1993), topical therapy or single-dose griseofulvin, to reduce the risk of causing the sudden appearance of regular and deep ulceration. spread. An alternative would be to continue with the existing It is mainly seen in Africa, India, and the western Pacific and practice of treating individual cases while recognizing that this in parts of Indonesia and the Philippines. The disease is process ignores the community reservoir. caused by a combined infection of a number of different bac- teria together with a fusiform bacterium, Fusobacterium Tinea Imbricata (Tokelau Ringworm). In many parts of the ulcerans, and an as yet unidentified spirochete. The disease is developing world, tinea imbricata is an exotic and unusual associated with poor living conditions and exposure to water, infection, with isolated foci occurring in remote areas of particularly flood or stagnant water and mud. In endemic Brazil, India, Indonesia, Malaysia, Mexico, and the western areas, it is a constant drain on resources. Morris and others' Pacific. However, in some specific locations, it is common and (1989) study of aid posts in East Sepik province, Papua New endemic, reaching prevalence rates of more than 30 percent Guinea, shows that management of tropical ulcer was occupy- in some communities in the western Pacific. For example, ing a third of the posts' time and almost half their health care extrapolating from a school survey in Goodenough Island, budgets. Papua New Guinea, Hay and others (1984) estimate that more The lesion usually starts with mild discomfort and overlying than 7,000 people out of a population of about 20,000 were hyperpigmentation on the skin that progresses rapidly over a infected. few days until the skin breaks down and sloughs, revealing an The disease presents in the form of widespread scaling, underlying ulcer. The lesion is often clean on first presentation often arranged in concentric rings or with large sheets of and round with smooth edges. It generally starts on the lower desquamation. The infection may develop early in life and leg or ankle, and in about 10 percent of cases, it progresses to persist into old age without the development of effective become an irregular, enlarged, and chronic ulcer. immunity. Tinea imbricata often affects wide areas of the body, The condition heals well in most patients with simple sparing only body folds and scalp skin. In those areas where it cleansing and treatment with penicillin; however, early grafting Skin Diseases | 715 Table 37.4 Evidence of the Efficacy of Different Regimens for Tinea Capitis Treatment, level of evidence, cost (manufacturer, Number of formulation, average people wholesale price) Evidence in study Results Comments Benzoic acid compound Investigator-blinded RCT 41 Mycological cure: 12/20 using benzoic acid · Neither treatment is fully (Whitfield's ointment) (versus miconazole compound and 10/19 using miconazole efficacious cream) (Wilmington, cream Level of evidence: III Aly, and Frieden 1996) Cost: not found Observational study Prevalence dropped from 7.8 percent to · Prevalence study of dermatophyto- (Hay and others 1996) 5.8 percent (p 0.05) mycoses in rural schools · After institution of treatment by 12 trained community health workers, only prevalence of tinea capitis dropped significantly Griseofulvin Multicenter, single- 200 Effective treatment: 46 50 (92 percent) · ITT analysis performed blinded, RCT (versus griseofulvin, 47/50 (94 percent) terbinafine, Level of evidence: III · Griseofulvin for six weeks similar terbinafine, itracona- 43 50 (86 percent) itraconazole, 42 50 in efficacy to terbinafine, itracona- Cost: zole, and fluconazole) (84 percent) fluconazole (p 0.33) zole, and fluconazole for two to (Wright and Robertson · Pedinol, tablets, 125 mg, three weeks 1986) US$63.00 for 100 tablets Single-cohort retrospec- 474 60.7 percent responded well; 39.3 percent · Martec, tablets, 125 mg, · Observation over a two-year period tive analysis (Schmeller, returned less than eight months later; US$34.10 for 100 tablets · Conclusions: griseofulvin may be Baumgartner, and 10.7 percent had a recurrence later ineffective in one-third or more Dzikus 1997) patients Multicenter, open-label, 210 No statistically significant differences · 147 patients were evaluable; no ITT RCT (four weeks (cure 67 percent in both groups); · Four weeks of treatment with oral terbinafine versus eight however, graphical presentation of data terbinafine had a similar efficacy weeks griseofulvin) demonstrates a slightly higher proportion to eight weeks of treatment with (Gupta and others 2001) of patients in terbinafine group achieved griseofulvin "cure" earlier Parallel-group, multi- 134 Terbinafine for six weeks had a similar · Four oral terbinafine groups (6, 8, center, double-blind RCT efficacy to griseofulvin 10, or 12 weeks) compared with (versus terbinafine) 12 weeks of griseofulvin (Abdel-Rahman, · ITT analysis performed Nahata, and Powell · Six weeks of terbinafine could 1997) represent an alternative to griseofulvin Double-blind RCT 50 Week 8: 76 percent griseofulvin and · Outcome: cure rates at weeks 8 (versus terbinafine) 72 percent terbinafine (not statistically and 12 (Fuller and others 2001) significant); · Terbinafine is a good alternative week 12: 44 percent griseofulvin and for less-frequent recurrences 76 percent terbinafine (p 0.05) Double-blind RCT 35 88 percent itraconazole versus 88 percent · Tinea corporis and tinea capitis (versus itraconazole) griseofulvin evaluated together (Lipozencic and others · Outcome measure: cure 2002) · 34 patients evaluable for efficacy; no ITT · Two griseofulvin patients discon- tinued therapy because of vomiting · Itraconazole has the same efficacy as griseofulvin and fewer side effects Source: Authors. ITT intention to treat; p probability; RCT randomized clinical trial. 716 | Disease Control Priorities in Developing Countries | Roderick Hay, Sandra E. Bendeck, Suephy Chen, and others Table 37.5 Evidence of the Efficacy of Terbinafine for Tinea Imbricata Treatment, level of evidence, cost (manufacturer, brand Number of name, formulation, people average wholesale price) Evidence in study Results Comments Terbinafine Double-blind random- 83 Clinical and mycological cure rate: · Terbinafine has a slightly higher ized clinical trial cure rate and a lower reinfection Level of evidence: II · 37/37 for terbinafine, 31 35 for (terbinafine versus itraconazole (p 0.05) and relapse rate than itraconazole Cost: itraconazole) At week 17 follow-up, reinfection or · 72 patients eligible for follow-up (Lopez-Gomez and · Novartis relapse: others 1994) · Lamisil · 6 37 (16 percent) evaluable terbinafine · Tablets, 250 mg, patients US$260.51 for 30 tablets · 24 31 (75 percent) evaluable (AWP) itraconazole patients (p 0.001) · Tablets, 250 mg, US$868.16 for 100 tablets (AWP) · Cream, TP, 1 percent, 15 gm, US$32.61 (AWP) · Cream, TP, 1 percent, 30 gm, US$58.40 (AWP) Source: Authors. AWP average wholesale price; gm gram; mg milligram; p probability. may be necessary if healing is delayed. Treatment, therefore, HIV-Related Skin Diseases consists of early treatment with penicillin, a strategy that may A wide range of skin conditions may develop as a consequence also fit with a syndromic approach to ulceration, because it will of HIV infection, but most are beyond the scope of this also be effective for yaws. The alternative is oral metronidazole, chapter. They include conditions that are a significant drain but no evidence of the comparative efficacy of these two on scarce resources. These include Kaposi's sarcoma and toxic approaches is available. epidermal necrolysis, a potentially life-threatening form of In searching the literature for effective remedies for tropical skin failure that is often drug induced and requires the level of ulcer, the team found little evidence. The team did find studies care and attention that would be deployed for patients with evaluating metronidazole and topical dressings, and several severe burns. articles mentioned the efficacy of penicillin and split skin graft- The commonest skin-related complication of HIV, particu- ing, but no randomized controlled trials have been performed. larly in Africa, is the itchy papular eruption or papular pruritic A single case report supports the use of co-trimoxazole. The eruption of HIV. It presents with fiercely itchy multiple papules management strategy thereafter depends on keeping the on the face and upper trunk. It is of unknown etiology and wound clean to allow appropriate healing using local antisepsis responds only to symptomatic treatment--for instance, and cleansing, such as potassium permanganate solution, antipruritic preparations such as antihistamines--although chlorhexidine, or even saline, and protecting the area from simple topical preparations, such as calamine or menthol further abrasion or secondary infection with sterile dressings. creams, may alleviate the itching. Recognizing this condition is Clinical experience suggests that if this regimen is not followed, important, because it is seen only in HIV/AIDS cases and is the risk of developing chronic leg ulceration is substantial. often mistakenly treated as acne. It does not respond to treat- No community strategies for preventing tropical ulcer are ments for acne. known, although the process of infection suggests that simple, hygienic measures to disinfect and clean the affected limb, perhaps modified from those used in lymphatic filariasis, Pigmentary Disorders might be effective as a simple preventive regimen. The possible The development of pigmentary change is an important source use of vaccines has been substantially researched for the animal of concern in many communities (Taylor 1999). Disorders counterpart, sheep foot rot, which is caused by a similar com- associated with pigmentary changes are common and range bination of organisms. from hereditary defects such as albinism (Lookingbill, Skin Diseases | 717 Lookingbill, and Leppard 1995) to increased pigmentation, or onchocercal skin disease), as well as a paper on the direct hyperpigmentation, associated with inflammatory skin lesions costs of treating scabies in Italy. These studies are shown in such as acne. Albinism is a significant cause of life-threatening table 37.6. skin cancer in the developing world. Examples of drug costs (tables 37.2 to 37.4) for tinea capitis, For many of these conditions, no effective remedies are scabies, and pyoderma can be estimated as follows: available. For instance, hyperpigmentation secondary to inflammation cannot be removed effectively, although it may · Treatment of a single case of scalp ringworm using griseo- fade with time. Similarly, no effective cure exists for vitiligo, a fulvin purchased from two differently priced U.S. sources to common disease involving loss of pigment, although experi- achieve the published efficacy rates (table 37.4) with a con- mental treatments such as melanocyte grafting do produce ventional therapeutic course of six weeks, assuming a daily localized repigmentation. Therefore, advising patients of the dose of 250 milligrams, would provide between 61 and current comparative ineffectiveness of treatments for these 92 percent efficacy at a drug cost per individual of US$29 or conditions is important. Preventing the use of therapies that do US$53, depending on the drug source. Alternatively, a single not lead to effective outcomes should be an important part of supervised dose of 1 gram would cost US$1.40 or US$2.50. the strategy for treating skin diseases. With supervision of treatment, the total cost per cure using Some forms of increased pigmentation, such as melasma, daily treatment ranges from US$35 to US$88 per patient. which is hyperpigmentation of the cheek and forehead areas · Treatment of 100 people with scabies using sulfur ointment, and is seen mainly in women, respond to the application of assuming 500 grams per individual, would cost US$58 or hydroquinone derivatives. However, because such treatments US$0.58 per person. This regimen would provide a 71 per- are often misused, they would not be used at the community cent cure rate at three months and a cost per cure of $1.30 level and would be used only with advice from a trained prac- per patient. titioner. Depigmenting creams, lotions, and emulsions are · Treatment with povidone of an individual with pyoderma widely available as cosmetic preparations in many local mar- would cost US$0.68, assuming that 400 milliliters would kets and shops, and in a study in Dakar, Senegal, more than treat eight people. This regimen would provide a cure rate of 50 percent of women questioned stated that they were regularly 88 percent at three months and a cost per cure of US$1.10 using bleaching creams ranging from hydroquinones to corti- per patient. costeroids (Mahé and others 2003). Hydroquinones are poten- tially damaging to the skin and with continuous use cause These calculations have taken into account ideal community patchy increased pigmentation and scarring of the facial skin. treatment conditions, where the recurrence rate is negligible. Similarly, misuse of corticosteroids is associated with a range of However, if such a community-based scheme is not effectively secondary effects from skin thinning to increased infection developed, more than 50 percent of those with scabies are likely rates. Warning people about the potential risks of depigment- to be reinfected. The figures are lower for tinea capitis (15 per- ing creams would be a useful health promotion strategy in cent) and pyoderma (10 percent). Table 37.7 shows the costs of many communities. treating large populations. Skin depigmentation is also a feature of leprosy. Thus, Although little information is currently available, in partic- teaching health care workers responsible for leprosy surveil- ular about the effect of local pricing of medications on overall lance to recognize skin patterns is a practical strategy of great effective treatment costs, the studies cited in this chapter indi- potential value in continuing progress toward eliminating this cate that the financial burden of skin diseases within families disease. may well be significant and that producing a series of robust analyses of the cost implications of both treatment and failure to provide adequate management strategies for these common ECONOMIC ASSESSMENTS AND SKIN DISEASES conditions is critical. IN DEVELOPING COUNTRIES The 1990 global burden of disease study estimated that the disability weighting associated with skin disease was at Apart from the studies mentioned here in relation to families' least 0.02. However, the disability weighting for severe scabies costs for treating community-acquired skin diseases in (25 percent of cases) and patients with ecthyma (10 percent of Mexico (Hay and others 1994) and costs to health posts of pyoderma cases) is 0.10. If we take skin cases with the lower managing tropical ulcer in Papua New Guinea (Morris and disability estimates--for example, mild to moderate scabies others 1989), no published studies are available of the eco- and pyoderma--the cost per DALY gained would be about nomic burden of skin disease. An extensive literature search US$1.00 to US$1.50 (table 37.7). For tinea capitis, the cost per did reveal some studies related to diseases that affect the skin DALY gained using daily treatment would be considerably but discussed elsewhere in this work (Buruli ulcer and higher, US$175 at the lower drug cost. 718 | Disease Control Priorities in Developing Countries | Roderick Hay, Sandra E. Bendeck, Suephy Chen, and others Table 37.6 Literature Review on the Economic Impact of Skin Diseases Year of Study type and Disease Author research Country population Cost categories and indicators Results Buruli ulcer Asiedu 1998 Ghana Retrospective study of · Health care costs (inpatient · Total costs: US$783.27 per 102 cases at a district services including medicines, patient hospital in the Ashanti surgery, laboratory) · Health care costs: region · Indirect costs (loss of produc- US$233.78 per patient tivity, food, miscellaneous) · Indirect costs: US$549.49 per patient · Percentage of total health care cost relative to district budget: 40 percent Onchocerciasis Workneh 1993 Ethiopia Males age 18 to 54 · Days of leave · Those with OSD had (OSD) working at a coffee · Income significantly more days of plantation with OSD leave and less income than and without OSD controls Onchocerciasis Oladepo 1997 Nigeria Matched pairs of male · Current cultivated farm size · Those with OSD had signi- (OSD) farmers with OSD and · Personal wealth ficantly smaller farm sizes without OSD and less personal wealth Onchocerciasis Benton 1998 Ethiopia, Communities · Educational impact · Children of OSD heads of (OSD) Nigeria, · Direct costs household had double the and Sudan · Indirect functional capacity risk of dropping out of costs, for example, from school disability · People with severe OSD spend US$20 more per year on health (15 percent of their incomes) · People with severe OSD spend longer time seeking care Scabies Papini 1999 Italy Outbreaks in two · Health care costs (medical · US$151.17 per resident nursing homes consults, treatment, disinfesta- tion procedures, laundry, extra staffing, disposable materials) Source: Asiedu 1998; Workneh 1993; Oladepo 1997; Benton 1998; and Papini 1999. OSD onchocercal skin disease. Table 37.7 Cost of Cure and Impact on DALYs for the Three Most Common Skin Diseases, Using the Cheapest Effective Treatments Cost of cure Number of people cured Disease (US$/million population) for US$1 million Cost per DALY gained (US$) Comment Tinea capitis 5,250,000 285,000 175 (assuming cost per drug of Estimated on the basis of a high- US$29 for course of treatment) prevalence (15 percent) region such as Ethiopia Scabies 58,000 1,700,000 1.00­1.50 Estimated on the basis of a high- prevalence (10 percent) region Pyoderma 55,000 900,000 1.00­1.50 Estimated on the basis of a high- prevalence (5 percent) region. Source: Authors. Skin Diseases | 719 The benefits of devising control measures for treatable skin Benton, B. 1998. "Economic Impact of Onchocerciasis Control through disease are also affected by the high prevalence figures for skin the African Programme for Onchocerciasis Control: An Overview." Annals of Tropical Medicine & Parasitology 92 Suppl 1: S33­39. diseases in low-income countries with total populations of Breneman, D. L. 1990. "Use of Mupirocin Ointment in the Treatment of between 40 million and 600 million affected, depending on Secondarily Infected Dermatoses." Journal of the American Academy of variations in disease prevalence. Dermatology 22: 886­92. Budimulja, U., K. Kuswadji, S. Bramono, J. Basuki, L. S. Jadanarso, S. Untung, and others. 1994. "A Double-Blind, Randomized, Stratified Controlled Study of the Treatment of Tinea Imbricata with Oral CURRENT STATUS OF COMMUNITY CONTROL Terbinafine or Itraconazole." British Journal of Dermatology 130: MEASURES IN DERMATOLOGY 29­31. Bulto, T., F. H. Maskel, and G. Fisseha. 1993. "Skin Lesions in Resettled and Despite the logic of developing community-focused services Indigenous Populations in Gambela, with Special Emphasis on the for dermatology, such services have seldom been achieved Epidemiology of Tropical Ulcer." Ethiopian Medical Journal 31: 75­82. (Hay, Andersson, and Estrada 1991). Perhaps the best current Carapetis, J. R., B. J. Currie, and E. L. Kaplan. 1999. "Epidemiology and Prevention of Group A Streptococcal Infections: Acute Respiratory example of a concerted, community-based approach is the Tract Infections, Skin Infections, and Their Sequelae at the Close of the Regional Training Center for Dermatology in Moshi, Tanzania, 20th Century." Clinical Infectious Diseases 28: 205­10. which focuses on developing a primary care skills base in Daroczy, J. 2002. "Antiseptic Efficacy of Local Disinfecting Povidone- African countries for the care of patients with skin and sexu- Iodine (Betadine) Therapy in Chronic Wounds of Lymphedematous Patients." Dermatology 204: 75­78. ally transmitted diseases (Kopf 1993). The program has now Eells, L. D., P. M. Mertz, Y. Piovanetti, G. M. Pekoe, and W. H. Eaglestein. trained more than 100 medical assistants and nurses, who were 1986. "Topical Antibiotic Treatment of Impetigo with Mupirocin." placed in 15 different countries at the primary care level and Archives of Dermatology 122: 1273­76. who, in many cases, play key roles in developing local health Elewski, B. 2000. "Tinea Capitis: A Current Perspective." Journal of the programs. A key issue is that action proportional to the severity American Academy of Dermatology 42: 1­20. of the problem is needed. For instance, one option would be to Estrada, R., M. Romero, G. Chavez, and G. Estrada. 2000. "Dermatologia help nonspecialized health workers significantly improve their communitaria: diez años de experiencia. Estudio epidemiológico com- parativo entre población urbana y rural del estado de Guerrero." skills in managing common skin diseases. That option would Dermatologia Revista Mexicana 44: 268­73. present a new challenge for the teaching of dermatology. Along Figueroa, J. I., L. C. Fuller, A. Abraha, and R. J. Hay. 1996. "The Prevalence those lines, a recent initiative to effect change through a control of Skin Disease among Schoolchildren in Rural Ethiopia: A Preliminary and education program in Mali targeted at pyoderma, scabies, Assessment of Dermatologic Needs." Pediatric Dermatology 13: 378­81. and tinea capitis is currently being evaluated. Early assessments ------. 1998. "Dermatology in Southwestern Ethiopia: Rationale for a Community Approach." International Journal of Dermatology 37: indicate that the teaching methods have been effective in instil- 752­58. ling recognition skills among primary care health workers. The Fuller, L. C., C. H. Smith, R. Cerio, R. A. Marsden, G. Midgley, A. L. Beard, effect on community levels of skin diseases is not yet known. and others. 2001. "A Randomized Comparison of Four Weeks of Skin diseases remain a low priority for many health author- Terbinafine versus Eight Weeks of Griseofulvin for the Treatment of Tinea Capitis." British Journal of Dermatology 144: 321­27. ities, despite the large demand for services. Addressing the Gibbs, S. 1996. "Skin Disease and Socioeconomic Conditions in Rural potential for controlling skin problems by means of simple and Africa: Tanzania." International Journal of Dermatology 35: 633­39. effective public health measures should be a realistic target for Gupta, A. K., P. Adam, N. Dlova, C. W. Lynde, S. Hofstader, N. Morar, and alleviating a common and solvable source of ill health. An others. 2001. "Therapeutic Options for the Treatment of Tinea Capitis effective plan, team, and basic dermatological formulary can do Caused by Trichophyton Species: Griseofulvin versus the New Oral much to improve matters (Estrada and others 2000). This Antifungal Agents, Terbinafine, Itraconazole, and Fluconazole." Pediatric Dermatology 18: 433­38. chapter outlines some of the challenges for such programs and Hay, R. J., N. Andersson, and R. Estrada. 1991. "Mexico: Community some of the deficiencies of current provision. Dermatology in Guerrero." Lancet 337: 906­7. Hay, R. J., Y. M. Clayton, N. De Silva, G. Midgley, and E. Rossor. 1996. "Tinea Capitis in Southeast London: A New Pattern of Infection with REFERENCES Public Health Implications." British Journal of Dermatology 135: 955­58. Abdel-Rahman, S. M., M. C. Nahata, and D. A. Powell. 1997. "Response to Hay, R. J., R. Estrada, H. Alarcon, G. Chavez, L. F. Lopez, S. Paredes, and Initial Griseofulvin Therapy in Pediatric Patients with Tinea Capitis." N. Andersson. 1994. "Wastage of Family Income on Skin Disease in Annals of Pharmacotherapy 31: 406­10. Mexico." British Medical Journal 309: 848. Asiedu, K., and S. Etuaful. 1998. "Socioeconomic Implications of Buruli Hay, R. J., S. Reid, E. Talwat, and K. MacNamara. 1984. "Endemic Tinea Ulcer in Ghana: A Three-year Review." American Journal of Tropical Imbricata: A Study on Goodenough Island, PNG." Transactions of the Medicine & Hygiene 59: 1015­22. Royal Society of Tropical Medicine and Hygiene 78: 246­51. Barton, L. L., A. D. Friedman, and M. G. Portilla. 1988. "Impetigo Hegazy, A. A., N. M. Darwish, I. A. Abdel-Hamid, and S. M. Hammad. Contagiosa: A Comparison of Erythromycin and Dicloxacillin 1999. "Epidemiology and Control of Scabies in an Egyptian Village." Therapy." Pediatric Dermatology 5: 88­91. International Journal of Dermatology 38: 291­95. 720 | Disease Control Priorities in Developing Countries | Roderick Hay, Sandra E. Bendeck, Suephy Chen, and others Hiletework, M. 1998. "Skin Diseases Seen in Kazanchis Health Center." Morris, G. E., R. J. Hay, A. Srinavasa, and A. Bunat. 1989. "The Diagnosis Ethiopian Medical Journal 36: 245­54. and Management of Tropical Ulcer in East Sepik Province of Papua Koning S., L. W. van Suijlekom-Smit, J. L. Nouwen, C. M. Verduin, R. M. New Guinea." Journal of Tropical Medicine and Hygiene 92: 215­20. Bernsen, A. P. Oranje, and others. 2002. "Fusidic Acid Cream in the Oladepo, O., W. R. Brieger, S. Otusanya, O. O. Kale, S. Offiong, and Treatment of Impetigo in General Practice: Double-Blind Randomised M. Titiloye. 1997. Farm Land Size and Onchocerciasis Status of Placebo Controlled Trial." British Medical Journal 324: 203­6. Peasant Farmers in South-western Nigeria. Tropical Medicine & International Health 2: 334­340. Kopf, A. W. 1993. "International Foundation for Dermatology: A Challenge to Meet the Dermatologic Needs of Developing Countries." Papini, M., R. Maccheroni, and P. L. Bruni. 1999. "O Tempora o Mores: Dermatologic Clinics 11: 311­14. The Cost of Managing Institutional Outbreaks of Scabies." International Journal of Dermatology 38: 638­39. Leppard, B., and A. E. Naburi. 2000. "The Use of Ivermectin in Controlling an Outbreak of Scabies in a Prison." British Journal of Dermatology 143: Rea, J. N., M. L. Newhouse, and T. Halil. 1976. "Skin Disease in Lambeth: 520­23. A Community Study of Prevalence and Use of Medical Care." British Journal of Preventive and Social Medicine 30: 107­14. Linder, C. W. 1978. "Treatment of Impetigo and Ecthyma." Journal of Family Practice 7: 697­700. Saw, S. M., D. Koh, M. R. Adjani, M. L. Wong, C. Y. Hong, J. Lee, and others. 2001. "A Population-Based Prevalence Survey of Skin Diseases in Lipozencic, J., M. Skerlev, R. Orofino-Costa, V. C. Zaitz, A. Horvath, E. Adolescents and Adults in Rural Sumatra, Indonesia, 1999." Transac- Chouela, and others. 2002. "A Randomized, Double-Blind, Parallel- tions of the Royal Society of Tropical Medicine and Hygiene 95: 384­88. Group, Duration-Finding Study of Oral Terbinafine and Open-Label, High-Dose Griseofulvin in Children with Tinea Capitis Due to Schmeller, W., S. Baumgartner, and A. Dzikus. 1997. "Dermatophyto- Microsporum Species." British Journal of Dermatology 146 (5): 816­23. mycoses in Children in Rural Kenya: The Impact of Primary Health Care." Mycoses 40: 55­63. Lookingbill, D. P., G. L. Lookingbill, and B. Leppard. 1995. "Actinic Damage and Skin Cancer in Albinos in Northern Tanzania: Findings in Seeberg, S., B. Brinkhoff, E. John, and I Mer. 1984. "Prevention and 164 Patients Enrolled in an Outreach Skin Care Program." Journal of Control of Neonatal Pyoderma with Chlorhexidine." Acta Paediatrica the American Academy of Dermatology 32: 653­58. Scandinavica 73: 498­504. Lopez-Gomez, S., A. Del Palacio, J. Van Cutsem, M. Soledad Cuetara, L. Taplin D., L. Lansdell, A. A. Allen, R. Rodriguez, and A. Corets. 1973. Iglesias, and A. Rodriguez-Noriega. 1994. "Itraconazole versus "Prevalence of Streptococcal Pyoderma in Relation to Climate and Griseofulvin in the Treatment of Tinea Capitis: A Double-Blind Hygiene." Lancet 1: 501­3. Randomized Study in Children." International Journal of Dermatology Taplin, D., S. L. Porcelain, T. L. Meinking, R. L. Athey, J. A. Chen, P. M. 33: 743­47. Castillero, and R. Sanchez. 1991. "Community Control of Scabies: A Macotela-Ruiz, E. I. C., and Q. F. B. E. N. Ramos. 1996. "Tratamiento de Model Based on Use of Permethrin Cream." Lancet 337: 1016­18. escabiasis con Ivermectina por via oral en una comunidad rural cerrada: Taylor, S. C. 1999. "Cosmetic Problems in Skin of Color." Skin Implicaciones epidemiológicas." Dermatologia Revista Mexicana 40: Pharmacology and Applied Skin Physiology 12: 139­43. 179­84. Usha, V., and T. V. Gopalakrishnan Nair. 2000. "A Comparative Study Mahé, A., F. Ly, G. Aymard, and J. M. Dangou. 2003. "Skin Diseases of Oral Ivermectin and Topical Permethrin Cream in the Treatment Associated with the Cosmetic Use of Bleaching Products in Women of Scabies." Journal of the American Academy of Dermatology 42: from Dakar, Senegal." British Journal of Dermatology 148: 493­500. 236­40. Mahé, A., H. Thiam N'Diaye, and P. Bobin. 1997. "The Proportion of White, A. V., W. E. Hoy, and D. A. McCredie. 2001. "Childhood Post- Medical Consultations Motivated by Skin Diseases in the Health Streptococcal Glomerulonephritis as a Risk Factor for Chronic Renal Centers of Bamako (Republic of Mali)." International Journal of Disease in Later Life." Medical Journal of Australia 174: 492­96. Dermatology 36: 185­86. Wilmington, M., R. Aly, and I. J. Frieden. 1996. "Trichophyton Tonsurans Mallon, E., J. N. Newton, A. Klassen, S. L. Stewart-Brown, T. J. Ryan, and Tinea Capitis in the San Francisco Bay Area: Increased Infection A. Y. Finlay. 1999. "The Quality of Life in Acne: A Comparison with Demonstrated in a 20-Year Survey of Fungal Infections from 1974 to General Medical Conditions Using Generic Questionnaires." British 1994." Journal of Medical and Veterinary Mycology 34: 285­87. Journal of Dermatology 140: 672­76. Workneh, W., M. Fletcher, and G. Olwit. 1993. "Onchocerciasis in Field Mathers, C. D., A. D. Lopez, and C. J. L. Murray. "The Burden of Disease Workers at Baya Farm, Teppi Coffee Plantation Project, Southwestern and Mortality by Condition: Data, Methods, and Results for 2001." In Ethiopia: Prevalence and Impact on Productivity." Acta Tropica 54: Global Burden of Disease and Risk Factors, eds. A. D. Lopez, C. D. 89­97. Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. New York: World Bank. 2002. World Development Indicators. Washington, DC: Oxford University Press. World Bank. McLinn, S. 1988. "Topical Mupirocin versus Systemic Erythromycin Treat- Wright, S., and V. J. Robertson. 1986. "An Institutional Survey of Tinea ment for Pyoderma." Pediatric Infectious Disease Journal 7: 785­90. Capitis in Harare, Zimbabwe, and a Trial of Miconazole Cream versus Mirmirani, P., T. A. Maurer, T. G. Berger, L. P. Sands, and M. M. Chren. Whitfield's Ointment in Its Treatment." Clinical and Experimental 2002. "Skin-Related Quality of Life in HIV-Infected Patients on Highly Dermatology 11: 371­77. Active Antiretroviral Therapy." Journal of Cutaneous Medicine and Surgery 6: 10­15. Skin Diseases | 721 Chapter 38 Oral and Craniofacial Diseases and Disorders Douglas Bratthall, Poul Erik Petersen, Jayanthi Ramanathan Stjernswärd, and L. Jackson Brown The oral cavity is an essential part of the body and contributes prevent formation of a cavity. Thus, the main risk factors to total health and well-being. Recent research indicates that include presence of cariogenic biofilms and frequent consump- poor oral health affects general health and that some systemic tion of fermentable carbohydrates. Exposure to fluorides in diseases can affect oral health. A variety of diseases involve the optimum concentrations reduces the risk, and normal saliva oral cavity; the two main oral diseases present worldwide and flow and saliva protective systems are also important to coun- lead to tooth destruction or tooth loss: teract the cariogenic factors. Untreated caries can give rise to infection of the tooth pulp, · dental caries, the disease that leads to cavities in the teeth which can spread to the supporting tissues and the jaws, culmi- · periodontal disease, which leads to loosening of teeth. nating in advanced disease conditions that are often painful. For example, in Thailand, recent surveys of a sample of 12-year-old Both diseases are preventable, and strong efforts have been children revealed that 53 percent had suffered from pain or dis- made to control them. Other diseases and conditions are much comfort from teeth over the past year (Petersen and others less prevalent, yet serious, and sometimes even life threatening: 2001). The corresponding figures in China were 34 percent for oral precancer and cancer, oral manifestations of HIV and 12-year-olds (Peng, Petersen, Fan, and others 1997) and 74 per- AIDS, noma, developmental disorders, and fluorosis of teeth. cent for adults (Petersen, Peng, and Tai 1997). Tooth decay is a public health problem worldwide. According to the U.S. Surgeon General's report (U.S. Public DENTAL CARIES Health Service 2000), dental caries is the single most common chronic childhood disease in the United States. Epidemiological Dental caries develops by the localized dissolution of the tooth data for almost 200 countries are available in the World Health hard tissues, caused by acids that are produced by bacteria in Organization (WHO) Country/Area Profile Programme the biofilms (dental plaque) on the teeth and eventually lead to (CAPP) oral health database (http://www.whocollab.od.mah. "cavities." The biofilm consists of microorganisms, including se/index.html) (see table 38.1 for examples). Caries prevalence the highly cariogenic mutans streptococci, and a matrix made of permanent teeth is expressed by the decayed, missing, and up mainly of extracellular polysaccharides. The destructive filled teeth (DMFT) index (calculated by counting the number acids are produced when fermentable carbohydrates (sugars) of DMFT of individuals and taking the mean for the group reach these biofilms, each episode resulting in tooth damage examined). One indicator age group used for international (attack). If this process does not occur frequently, then the nat- comparisons is 12-year-old children. The WHO oral health ural capacity of the body (through saliva) to remineralize will goal was to achieve three DMFT or fewer among 12-year-olds 723 Table 38.1 Mean DMFT and SiC Index of 12-Year-Olds for Some Countries, by Ascending Order of DMFT Mean Country DMFT SiC Index Year Sample size Reference Australia 0.8 2.4 1999 29,130 Armfield, Roberts-Thomson, and Spencer 2003 Nepal 0.8 2.5 2000 623 Data from WHO, courtesy P. E. Petersen Sweden 0.9 2.6 2001 71,896 Sundberg 2002 Jamaica 1.0 2.8 1995 362 Data from PAHO, courtesy E. D. Beltran and S. Estupinan-Day China 1.0 3.0 1996 23,452 Data from WHO, courtesy P. E. Petersen Senegal 1.2 2.8 1994 300 Sembene, Kane, and Bourgeois 1999 Sri Lanka 1.4 3.6 1994­95 2,003 Abayaratna and Krishnarasa 1997 England, U.K. (Northwest) 1.4 3.2 2000­1 12,029 Pitts and others 2002 United States 1.4 3.6 1988­91 176 Data from PAHO, courtesy E. D. Beltran and S. Estupinan-Day Portugal 1.5 3.6 1999 800 Data from WHO, courtesy P. E. Petersen Germany 1.7 4.1 1997 1,043 Micheelis and Reich 1999 Israel 1.7 4.1 2002 1,327 Courtesy S. P. Zusman, Division of Dental Health, Israel South Africa 1.7 4.3 1988­89 1,571 van Wyk 1994 Greece (Northeastern 1.8 4.2 2001 2,217 Demertzi and Topitsoglou 2002 province) (11-year-olds) Scotland (U.K.) 1.8 4.3 1996­97 6,165 Data from K. Woods from the study Pitts, Evans, and Nugent 1998 France 2.0 4.7 1998 6,000 Hescot and Roland 2000 Thailand 2.4 4.9 2001 1,116 Data from WHO, courtesy P. E. Petersen Mexico (state of Mexico) 2.5 5.0 1997 1,138 Irigoyen and Sanchez-Hinojosa 2000 Uruguay 2.5 5.3 1999 596 Sector Público 1999 Comoros 2.6 6.1 2000 142 Data from WHO, courtesy P. E. Petersen Belarus 2.7 5.4 1999 2,537 Data from WHO, courtesy P. E. Petersen Romania 2.7 5.8 2001 785 Data from WHO, courtesy P. E. Petersen Nicaragua 2.8 5.7 1997 365 Data from PAHO, courtesy E. D. Beltran and S. Estupinan-Day Greenland 3.5 7.0 2002 236 Data from WHO, courtesy P. E. Petersen Latvia 3.8 7.1 1998 416 Data from WHO, courtesy P. E. Petersen Poland 3.9 7.2 1997 1,732 Data from WHO, courtesy P. E. Petersen Honduras 4.0 7.5 1997 307 Data from PAHO, courtesy E. D. Beltran and S. Estupinan-Day Bolivia 4.7 8.8 1995 389 Data from PAHO, courtesy E. D. Beltran and S. Estupinan-Day Slovak Republic 5.9 14.3 1998 1,589 Data from WHO, courtesy P. E. Petersen Costa Rica 8.5 13.7 1988 1,349 Data from PAHO, courtesy E. D. Beltran and S. Estupinan-Day Source: Authors. PAHO Pan American Health Organization; SiC Significant caries. by 2000. According to the CAPP database, 70 percent of the level in populations with skewed distribution. The Significant countries had achieved three DMFT or fewer by 2001, repre- Caries (SiC) Index was proposed to bring attention to those senting 85 percent of the world population. Several developing hidden high caries groups (Bratthall 2000). The SiC Index is cal- economies, however, have reported a trend toward higher levels culated by simply taking the mean DMFT of the one-third of the of dental caries. group having the highest DMFT in a population (figure 38.1). A detailed analysis of caries data for many countries, both Table 38.1 shows several countries having fewer than three industrial and developing, shows skewed distributions of the mean DMFT but high SiC Index values, thus illustrating the disease--that is, a proportion of a population of children show- hidden caries burden for children (Nishi and others 2002). ing a high or very high number of caries and the rest showing a Dental caries is found not only in children and young adults low number of caries or none. Expressing caries prevalence as but also in all age groups. The elderly, in particular those mean DMFT may, therefore, not accurately describe the disease with exposed tooth root surfaces, constitute a special risk 724 | Disease Control Priorities in Developing Countries | Douglas Bratthall, Poul Erik Petersen, Jayanthi Ramanathan Stjernswärd, and others · Individuals with poor oral hygiene and frequent sugar DMFT SiC intake are at increased risk. 10 · Individuals not exposed to fluorides--for example, from 9 fluoridated water or toothpastes--are at increased risk of 8 caries. 7 · Persons with individual risk factors, such as reduced saliva flow or exposed tooth root surfaces, or with certain general 6 diseases are also at increased risk of caries. 5 4 Caries Intervention Programs 3 Since the discovery of the caries-preventive effect of fluorides 2 in the 1930s, different forms of fluoride administration pro- 1 grams have been implemented, often with remarkable caries- 0 reducing effects. Fluoride has been added to different vehicles, 0 6 12 18 24 30 36 42 48 55 61 67 73 79 85 91 97 Percentage of group such as water, salt, toothpaste, and milk. Fluoride tablets and fluoride mouth rinsing have been used among young children Source: Adyatmaka and others 1998. Note: The mean DMFT is 2.3. The Significant Caries Index is 5.4. Arrow indicates the and in schools, and more recently even among adults at high proportion of individuals who are included in the calculation of the index. West caries risk (Petersen 1989, 1990). For individual use, fluoride Kalimantan is one of the most caries-affected provinces in Indonesia. in high concentrations has been added to various forms of gels Figure 38.1 DMFT for 331 12-Year-Olds, West Kalimantan, Indonesia and varnishes to be applied on the teeth. Furthermore, fluor- ide in chewing gum is available in some countries. When a population (Barmes 2000). A Swedish study reported DMFT group of international experts on cariology were asked in a values of 21.4 and 24.4 for 50- and 70-year-olds, respectively, study to identify the main causes of the caries decline seen in indicating that nearly all teeth were affected in these age groups several Western countries during recent decades, practically (Hugoson and others 1995). Thomson (2004), reviewing longi- all the experts pointed to fluoride dentifrice as the most signif- tudinal studies of older adults (age 50+), found an incidence of icant factor (Bratthall, Hänsel-Petersson, and Sundberg 1996). root surface caries varying from 29 to 59 percent and conclud- According to WHO (1994), community water fluoridation ed that older people are a caries-active group, experiencing new is safe and cost-effective in preventing dental caries in every caries at a rate comparable to that of adolescents. With increas- age group, benefiting all residents served by the community ing numbers of people becoming 50 years of age or older in water regardless of their social or economic status (Burt 2002; some developing countries, root surface caries may become a Petersen and Lennon 2004; White, Antczak-Bouckoms, and significant problem. Weinstein 1989). Examples of countries with fluoridated water When we consider the global epidemiology of dental caries, supplies for significant parts of the populations are Argentina, the main patterns seem to be the following: Brazil, Brunei Darussalam, Canada, Chile, Ireland, New Zealand, the United Kingdom, and the United States. In many · Countries with low mean sugar consumption (less than developing countries, lack of community water supplies makes 10 to 15 kilograms of sugar per person per year) generally water fluoridation impossible. have low mean caries prevalence. Effective fluoride toothpastes have been available for about · Countries with high mean sugar consumption (more than 40 years (WHO 1994). They have been tested in numerous 20 to 25 kilograms of sugar per person per year) and with- studies, in particular in school-based programs. The most out effective preventive programs generally have high mean commonly used concentrations are 1,000 or 1,500 parts per caries prevalence. million (ppm). Because most studies have been conducted in · Countries with high mean sugar consumption (more than developed countries, WHO launched a program testing a so- 20 to 25 kilograms of sugar per person per year) using effec- called "affordable fluoridated toothpaste" in developing coun- tive preventive programs have been able to reduce the caries tries. In the West Kalimantan Province of Indonesia, a super- prevalence. vised school-based toothbrushing program was implemented over a period of three years, resulting in a reduction of 12 to If we consider the prevalence of caries within a population, 40 percent of caries incidence in the study groups when com- the main patterns seem to be as follows: pared to control groups (Adyatmaka and others 1998). · Disadvantaged or poor population groups have higher Domestic salt fluoridation is another method of automatic dental caries experience than advantaged groups. fluoridation. In the early 1950s, Switzerland and Austria Oral and Craniofacial Diseases and Disorders | 725 introduced this approach by offering their populations fluori- 2001). The Chinese health authorities have emphasized pre- dated salt for the table and for cooking. The fluoride concen- ventive oral care and oral health education since the late 1980s. tration in the salt originally was 90 ppm and was later increased The nationwide mass campaign "Love Teeth Day" has been to 250 ppm. Fluoridated salt is now available in several coun- conducted annually since 1989, and the effective transmission tries in Europe and in South and Central America. A compari- of oral health messages to the public has shown improved oral son of caries data for Jamaica in 1984 (before salt fluoridation) health knowledge and behavior in children as well as in adults and 1995 (after salt fluoridation) showed a reduction of caries (Peng, Petersen, Tai, and others 1997). experience of 69 percent, 84 percent, and 87 percent among In addition, various dental organizations (Cohen 1990) and 15-, 12-, and 6-year-olds, respectively (Estupinan-Day and private companies have developed and carried out successful others 2001). oral health programs worldwide. For example, toothpaste man- Milk fluoridation projects are being conducted in several ufacturers have donated toothpastes, toothbrushes, and educa- countries, including Bulgaria, China, the Russian Federation, tional material promoting oral health in several countries. Thailand, and the United Kingdom. In Bulgaria, a milk fluorida- tion project resulted in a 79 percent lower DMFT in those chil- Effectiveness of the Oral Health Programs dren who had participated in the full five years of the program In countries with systematic national oral disease prevention than in the control children (Pakhomov and others 1995). programs, the total cumulative effect of these programs is Fluoride tablets and fluoride mouth-rinsing programs reflected in the epidemiological figures demonstrating caries under supervision in schools have been implemented in several decline (table 38.2) and in the growing proportions of caries- countries, including the Scandinavian countries, the United free individuals. However, singling out the effects of specific Kingdom, and the United States. The requirement that teachers activities or methods of programs is difficult because several and students be motivated has limited such approaches. In program components often operate simultaneously. For exam- recent years, many national fluoride programs have been ple, in industrial countries, practically all individuals use fluor- adjusted as the additional caries-reducing effects of topical idated toothpaste, and removing this preventive measure from applications with daily use of fluoridated toothpaste have been a group of individuals just to evaluate the effect of another questioned (Petersen and Torres 1999). fluoride program would be unethical. In addition, other factors affect caries reduction, such as changing lifestyles, changing patterns of sugar consumption, and improving living Oral Health Education and Promotion Programs conditions. The WHO Global Oral Health Programme has developed a The current trend in clinical health care and public health is manual for integration of oral health with school health pro- to base recommendations on evidence derived from systematic grams (WHO 2003). In many industrial countries, school health reviews of the literature and critical assessment of the quality of education programs have included oral health, and researchers results (U.S. Public Health Service 2000). The office of the U.S. have shown that children's self-care capacity improved in regard Surgeon General (U.S. Public Health Service 2000) and to regular toothbrushing with the use of fluoridated toothpaste the Swedish Council on Technology Assessment in Health Care (Flanders 1987; Honkala, Kannas, and Rise 1990; Petersen and (SBU 2002) are examples of entities that have attempted to Torres 1999; Sogaard and Holst 1988; Wang and others 1998). determine the effectiveness in public health of evidence-based Examples also exist from school oral health education in devel- approaches and technologies. oping countries. Some programs have been organized within Oral Health in America, the U.S. Surgeon General's report the context of the WHO Health Promoting Schools Initiative. In (U.S. Public Health Service 2000), reviewed experiences from Madagascar, the evaluation of program outcomes has shown the administration of fluorides. Primarily based on U.S. remarkably good results in reducing dental caries risk, improv- studies, the report had these conclusions: ing self-care capacity of children and mothers, and introducing higher levels of dental knowledge and attitudes (Razanamihaja · Strong evidence exists supporting the effectiveness of water and Petersen 1999). Other successful examples are available fluoridation in preventing crown and root caries in children from Tanzania (Petersen and others 2002; van Palenstein and adults. Helderman and others 1997), Zimbabwe (Frencken and others · Strong evidence exists of the effectiveness of the school- 2001), and Namibia (Priwe 1998). based fluoride supplement (tablets) program. The program, In China, principles from the WHO Health Promoting with motivated supervising personnel, such as teachers, is Schools Initiative have been applied in certain provinces; posi- recommended for children at high risk for caries. tive effects of programs were obtained regarding health-related · Evidence supports the effectiveness of school-based fluoride knowledge and behavior, but the clinical outcome measures (0.2 percent sodium fluoride) mouth-rinsing programs were less evident (Petersen and others 2004; Tai and others conducted before 1985 (before the introduction of fluoride 726 | Disease Control Priorities in Developing Countries | Douglas Bratthall, Poul Erik Petersen, Jayanthi Ramanathan Stjernswärd, and others Table 38.2 Declining Caries Experience in Some Countries DMFT in Country Year 12-year-olds Reference African region Côte d'Ivoire 1996 1.8 Guinan and others 1999 1993 2.6 Data from Oral Health Programme, WHO Niger 1997 1.3 Petersen and Kaka 1999 1992 1.5 Data from Oral Health Programme, WHO 1988 1.7 Woodward and Walker 1994 American region Colombia 1998 2.3 Data from PAHO 1984 4.8 Woodward and Walker 1994 Costa Rica 1999 2.3 Data from PAHO 1996 4.8 Data from Ministry of Health 1993 4.9 Data from PAHO Guyana 1995 1.3 Beltran-Aguilar, Estupinan-Day, and Baez 1999 1983 2.7 Woodward and Walker 1994 Haiti 2000 1.0 Data from PAHO 1994 2.2 Data from PAHO Honduras 1997 3.7 Beltran-Aguilar, Estupinan-Day, and Baez 1999 1987 5.7 Beltran-Aguilar, Estupinan-Day, and Baez 1999 Jamaica 1995 1.1 Beltran-Aguilar, Estupinan-Day, and Baez 1999 1984 6.7 Beltran-Aguilar, Estupinan-Day, and Baez 1999 Nicaragua 1997 2.8 Beltran-Aguilar, Estupinan-Day, and Baez 1999 1983 6.9 Beltran-Aguilar, Estupinan-Day, and Baez 1999 Panama 1997 3.6 Beltran-Aguilar, Estupinan-Day, and Baez 1999 1989 4.2 Beltran-Aguilar, Estupinan-Day, and Baez 1999 United States 1992­94 1.28 NHANES III, Courtesy D. Bruce 1988­91 1.4 Beltran-Aguilar, Estupinan-Day, and Baez 1999 1986­87 1.8 Beltran-Aguilar, Estupinan-Day, and Baez 1999 Venezuela, R. B. de 1997 2.1 Beltran-Aguilar, Estupinan-Day, and Baez 1999 1986 3.6 Data from PAHO Middle Eastern region Saudi Arabia 1995 1.7 Data from Oral Health Programme, WHO 1991 2.1 Data from Oral Health Programme, WHO United Arab Emirates 1995 1.6 Nithila and others 1998 1993 2.0 Data from Oral Health Programme, WHO European region Belarus 2000 2.7 Leous and Petersen 2002 1994 3.8 Leous and Petersen 2002 Denmark 2002 0.9 Data from National Board of Health, Denmark 1995 1.2 Data from National Board of Health, Denmark 1980 5.0 Data from National Board of Health, Denmark (Continues on the following page.) Oral and Craniofacial Diseases and Disorders | 727 Table 38.2 Continued DMFT in Country Year 12-year-olds Reference France 1998 1.9 Hescot and Roland 2000 1993 2.1 Hescot and Roland 2000 1990 3.0 Hescot and Roland 2000 Hungary 1996 3.8 Szoke and Petersen 2000 1991 4.3 Szoke and Petersen 2000 1985 5.0 Szoke and Petersen 2000 Israel 2002 1.66 Data from Dr S. P. Zusman, Division of Dental Health, Israel 1989 3.0 Zadik, Zusman, and Kelman 1992 Latvia 2000 3.9 Latvia, State Dentistry Centre 2000 1998 4.2 Latvia, State Dentistry Centre 2000 Norway 2000 1.5 Data from Norwegian Board of Health 1992 2.2 von der Fehr 1994 1986 3.1 Haugejorden 1994 Poland 2000 3.8 Wierzbicka and others 2002 1991 5.1 Wierzbicka and others 2002 Portugal 1999 1.5 de Almeida and others 2003 1990 3.2 de Almeida and others 2003 1984 3.7 de Almeida and others 2003 Romania 2000 2.7 Petersen and Rusu 2002 1990 3.9 Petersen and others 1994 Sweden 2001 0.9 Sundberg 2002 1995 1.4 Sundberg 2002 1985 3.1 Sundberg 2002 United Kingdom 1996­97 1.1 Pitts, Evans, and Nugent 1998 1983 3.1 Downer 1994 Asian region Bangladesh 2000 1.0 Ullah 2001 1981 1.5 Data from Oral Health Programme, WHO Sri Lanka 1994­95 1.4 Abayaratna and Krishnarasa 1997 1983­84 1.9 Sri Lanka, Ministry of Health 1985 Western Pacific region Australia 1999 0.8 Armfield, Roberts-Thomson, and Spencer 2003 1990 1.4 Armfield, Roberts-Thomson, and Spencer 2003 1980 3.6 Carr 1988 Hong Kong (China) 2001 0.8 Hong Kong, Department of Health 2003 1986 1.5 Lo, Evans, and Lind 1990 Japan 1999 2.4 Data from Ministry of Health and Welfare 1993 3.6 Miyazaki and Morimoto 1996 1987 4.9 Miyazaki and Morimoto 1996 Malaysia 1997 1.6 Malaysia, Dental Services Division 1997 1988 2.4 Malaysia, Dental Services Division 1997 Source: Authors. PAHO Pan American Health Organization; NHANES III Third U.S. National Health and Nutritional Examination Survey. Note: Numbers in italics indicate that the country did not achieve the WHO global goal of fewer than three DMFT by 2000 but shows caries decline. 728 | Disease Control Priorities in Developing Countries | Douglas Bratthall, Poul Erik Petersen, Jayanthi Ramanathan Stjernswärd, and others toothpastes) in preventing caries in children. The cost- public funds. Moreover, a filling does not affect the disease effectiveness of this intervention is reduced with the current process causing the cavities. Treatment must be directed against decline in prevalence of caries. It is recommended for use in the causative factors (described earlier). For the individual case, high-risk children consistently over a period of time. several options are available in addition to the various fluoride · Strong evidence supports the effectiveness of sealants in programs mentioned: dietary counseling, sugar substitutes, preventing pits and fissure caries. The report recommends antimicrobial agents to reduce plaque and specific bacteria, and that the programs be limited to high-risk children and high- the use of saliva-stimulating products. risk teeth. In many developing countries, the lack of dental manpower · Fluoride varnishes were not approved for use in the United means that carious teeth remain untreated. The ratio of den- States until 1994; hence, investigations are ongoing of the tists to population is particularly unfavorable in the African effectiveness of this intervention. region compared with Western European countries. For instance, according to CAPP, the ratio is 1 to 1.2 million in The Swedish Council on Technology Assessment in Health Ethiopia, 1 to 225,000 in Mali, and 1 to 166,000 in Zambia, Care (SBU 2002) applied strict criteria of evidence of effective- against about 1 to 1,000 in Scandinavian countries and 1 to ness; that is, the study had to be randomized and have a sample 2,100 in the United Kingdom (see http://www.whocollab.od. representing the total population. For permanent teeth, a mah.se/index.html). In India, the ratio is 1 to 27,000 in the three-year follow-up was necessary. The number of studies urban areas but 1 to 300,000 in the rural areas (Shah 2001). meeting all the criteria was not very high. Here are some Such ratios mean that neither dental caries disease nor the conclusions of this review: cavities will receive proper attention. After taking into consideration the high costs for dental · Daily use of fluoridated toothpaste is an effective method treatment and the lack of dentists, atraumatic restorative treat- to reduce caries in permanent teeth among children and ment (ART) was introduced. This approach requires only hand adolescents. Daily, weekly, or biweekly fluoride mouth rins- instruments rather than sophisticated electric dental drills, and ing can reduce caries, but together with daily fluoride trained dental auxiliaries can deliver ART. The public dental toothpaste use, the additional effects are not strong. health services in South Africa adopted the approach as an · Daily fluoride mouth rinsing can reduce root surface caries appropriate and economic means of providing basic restorative in the elderly, and professional application of fluoride care in certain communities. A randomized clinical trial con- varnish twice a year has a caries-reducing effect in perma- ducted in Tanzania showed no statistically significant differ- nent teeth among youth, as does the use of fluoridated ences between the retention of occlusal amalgam (74 percent) toothpaste. and ART occlusal restorations (67 percent) after a six-year · Fissure sealants have a caries-reducing effect. follow-up (1992­98) (Mandari, Frencken, and Van't Hof 2003). A potentially affordable treatment procedure that could pre- According to the SBU report, it was difficult to interpret the vent untreated carious teeth from being extracted, ART may effect of programs aimed at reducing the intake of sugars or the have relevance to some middle-income countries, although the effect of so-called sugar substitutes. Systematic evaluation of method is not realistic for most low-income countries, where community preventive programs should be carried out in the sustainability of such programs would be low. future, particularly to help identify appropriate alternatives for developing countries. PERIODONTAL DISEASES: CHRONIC GINGIVITIS ASPECTS OF TREATMENT OF CAVITIES AND CHRONIC PERIODONTITIS AND OF CARIES DISEASE Gingivitis, the inflammation of gum tissue caused by bacteria One has to differentiate between treatment of cavities and accumulating in the plaque along the gingival margin, pre- treatment of the disease process resulting in cavities. The cedes chronic periodontitis. The more destructive form of normal treatment of a tooth with a cavity is a filling or, if the periodontal disease, which breaks down the supporting tissues cavity is large, a crown. Large cavities may involve "root- of the teeth, progressively leading to loosening of teeth and fillings" or even extraction of the tooth. A variety of materials tooth loss, affects 10 to 15 percent of most adult populations are used globally: composites, amalgam, gold, porcelain, and (Papapanou 1999). Cigarette smoking and diabetes mellitus others. Options for replacing extracted teeth include removable (with poorly controlled diabetes) are two major risk factors prostheses, fixed bridges, or implants. The more complex treat- associated with periodontal disease and appear markedly to ments are costly, and no country has been able to afford to affect the initiation and progression of the disease (Genco introduce systems in which all dental costs are covered by 1996; Papapanou 1999). Oral and Craniofacial Diseases and Disorders | 729 In recent years there has been a growing awareness of the Screening populations and routine examination in dental association between some systemic diseases and oral disease, and medical clinics for oral precancer and early cancer lesions especially periodontal diseases. For example, a national study in would reduce the mortality, morbidity, and cost of treatment the United States found that the prevalence of diabetes mellitus associated with oral cancers. Not all oral premalignancies in patients with periodontitis was significantly greater show malignant transformation, and detection of these (twofold) than in nonperiodontal patients (Soskolne and oral lesions by biopsies are straightforward, not requiring Klinger 2001). Periodontal disease may be considered one of sophisticated equipment. Tobacco cessation programs aimed the complications of diabetes. Effective control of periodontal at younger and older age groups and control of excessive alco- infection in diabetics appears to reduce the levels of advanced hol intake are definitely beneficial in the prevention of oral glycogen end-products in the serum. cancer. Proper oral hygiene practices can prevent both gingivitis and advanced periodontal disease. All intervention programs leading to improved oral hygiene are instrumental in the ORAL MANIFESTATIONS OF HIV/AIDS control of periodontal disease and will reduce risk of future tooth loss. The school-based oral health educational programs The scarce epidemiological data available on oral manifesta- previously discussed are effective in preventing gingivitis, but tions of HIV in developing countries are difficult to interpret no community-based intervention program addresses peri- because these studies are not standardized (Holmes and odontal disease, especially among adults. Tobacco cessation Stephan 2002). In the study groups, the prevalence of oral programs are also important in the prevention of periodontal lesions in Africa ranged from 15 percent to more than 90 per- diseases. Treatment of periodontal diseases consists of plaque cent of infected individuals; in India the prevalence was 72 per- removal, scaling, and sometimes surgery, plus motivation and cent; and in Thailand it was 82 percent. Reviews are available instruction in oral hygiene. Dental hygienists can perform parts on the different studies performed on oral manifestations of the treatment program. of HIV and AIDS (Naidoo and Chikte 1999; Patton and others 2002). ORAL PRECANCER AND CANCER Candida infections, oral hairy leukoplakia, oral ulcers, and Kaposi's sarcoma are some of the common oral manifestations The most frequent form of oral precancerous lesion, leuko- of HIV and AIDS. Notably, Kaposi's sarcomas were never plakia, appears as a white patch that cannot be rubbed off, typ- detected in the Asian populations studied in India, Singapore, ically in the buccal mucosa, lateral borders of the tongue, and and Thailand but were seen in South African, Zambian, and floor of the mouth. The prevalence of leukoplakia among those Zimbabwean studies (Arendorf and others 1998; Hodgson 15 years old and above ranged from 1.1 percent in Cambodia, 1997; Holmes and Stephan 2002; Lim and others 2001; to 1.7 percent in Myanmar, to 3.6 percent in Sweden (Axell Nittayananta and Chungpanich 1997; Ranganathan and others 1976; Ikeda and others 1995). Malignant transformation varies 2000). The presence of oral candidiasis and hairy leukoplakia in different populations; nearly 5 percent of lesions are found alone or at the same time in an apparently healthy individual to be malignant at first biopsy, and 5 percent develop into could be an early indicator of the undetected HIV infection malignancy at a later stage. Erythroplakias appear as red patches progressing to AIDS. Those signs may be used as indicators and are less common but have a higher tendency (90 percent or during clinical examinations in developing countries where more) than leukoplakias to transform into malignancies technology for laboratory tests is not available or is too expen- (Sudbo and Reith 2003). sive (Greenspan and Greenspan 2002; Holmes and Stephan Oral cancers affect about 300,000 people worldwide 2002). annually (Ferlay and others 2001) and often develop from oral precancerous lesions (Sudbo and Reith 2003). Early detection of oral precancerous lesions, notably oral leukoplakia and NOMA (CANCRUM ORIS) erythroplakia, could easily prevent the development of the disfiguring disease oral cancer and premature death. Noma usually begins as a small ulcer of the gingiva and devel- Tobacco use in any form (smoking or chewing) and exces- ops into a rapidly spreading gangrenous condition of the oral sive alcohol consumption remain the primary risk factors in and facial tissues. Seen mainly in debilitated and malnourished the development of these precancerous lesions ("Early children, it is disfiguring and deadly. The condition is reported Diagnosis and Prevention of Oral Cancer and Precancer" 1995; in developing countries in several regions of the world, partic- Reichart 2001). Factors such as local irritation, Candida albi- ularly in Sub-Saharan Africa (Enwonwu, Falkler, and Idigbe cans infection, and nutritional deficiencies are also associated 2000; Naidoo and Chikte 2000; Petersen 2003). Noma disap- with the presence of leukoplakia. peared from the industrial world in the 20th century except 730 | Disease Control Priorities in Developing Countries | Douglas Bratthall, Poul Erik Petersen, Jayanthi Ramanathan Stjernswärd, and others during World War II. In contrast, risk factors such as poverty, FLUOROSIS OF TEETH poor hygiene, and malnutrition, eventually in combination with infectious diseases such as HIV and AIDS, may have Fluorosis of teeth develops during formation of teeth when recently increased the prevalence of this disease in Sub-Saharan children are young. Drinking water having more than 1.5 ppm Africa (Enwonwu 1995; Naidoo and Chikte 2000). Most of fluoride can give rise to enamel defects and discoloration of important, 90 percent of infected children die without having teeth, leading to endemic fluorosis in the population. These received any care. effects may vary from mild to severe. For example, in the Great Although the specific etiologic factors for noma are not Rift Valley area of East Africa, the ground water has high levels known, poverty has been identified as the single most impor- of fluoride, leading to high rates of dental fluorosis--nearly tant risk indicator. Accordingly, improving the overall socio- 90 percent in some parts of Kenya (Chibole 1987). Some indi- economic conditions can prevent noma. Public health viduals in developed countries can acquire fluorosis of teeth as approaches such as providing a high-protein diet, clean water, a result of the widespread use of different forms of fluorides in and sanitation and preventing communicable diseases such the prevention of caries, though the degree of fluorosis often is as diphtheria, dysentery, and tuberculosis would be needed mild compared with endemic fluorosis. for effective prevention of noma in Africa. Prognosis of Defluoridation of the central water supplies is possible noma is considerably better with timely administration of when naturally occurring fluoride is excessive in the drinking antibiotics. water. However, most developing countries do not have central water distribution systems, and the cost of defluoridation equipment and its maintenance can be high. WHO encourages effective and inexpensive methods that are useful for individual DEVELOPMENTAL DISORDERS households or community defluoridation of drinking water Developmental disorders involve teeth and the craniofacial (WHO 1994). Such methods exist, but a number of operational structures. A few of these disorders are congenital diseases of the problems have been identified, requiring further initiatives in enamel or dentin; problems related to the number, size, or shape this field (Kloos and Haimanot 1999). of teeth; and craniofacial birth defects, such as cleft lip and palate (CL/P). Among the most common congenital malforma- tions seen in humans, cardiovascular malformation is ranked as COMMON-RISK-FACTOR INTERVENTION the first and CL/P as the second. Unilateral CL/P occurs six PROGRAMS times more frequently than the bilateral form. Females are more prone to get cleft palates, whereas cleft lip or CL/P is most com- New research is pointing to associations between chronic oral mon in males (U.S. Public Health Service 2000). infections--particularly periodontitis--and heart and lung The incidence of CL/P differs from 0.18 to 3.74 per 1,000 diseases, stroke, osteoporosis, low birthweight, and premature live births, the highest incidence being seen in Native births in addition to diabetes. Such findings strengthen WHO Americans at 3.74 per 1,000, closely followed by the Japanese at health promotion strategies that are based on the common- 3.36 per 1,000 live births. A fairly uniform incidence of 1 per risk-factor approach, which controls essential risk factors that 600 to 700 live births is reported among Europeans. Overall, contribute to a large number of chronic diseases (Petersen the incidence rates appear high among Asians (0.82 to 3.36 per 2003). Risk behaviors such as smoking; alcohol; diets rich in 1,000 live births), intermediate in Caucasians (0.9 to 2.69 per fats and sugars and low in fiber, fruit, and vegetables; stress; 1,000 live births), and often very low in black Africans (0.18 to poor hygiene; and sedentary lifestyle are factors leading to such 1.67 per 1,000 live births) (Hewson and McNamara 2000; major chronic diseases as cardiovascular diseases, cancers, dia- Vanderas 1987; Wantia and Rettinger 2002). betes, obesity, osteoporosis, dental caries, and periodontal dis- The causes of CL/P are complex, involving multiple genetic ease. These principal risk factors for major chronic diseases are and environmental risk factors. Not all cases of CL/P are inher- often seen to cluster in the same individuals. ited. A number of risk factors, such as folic acid deficiencies, The WHO Global Oral Health Programme recommends the maternal smoking, and maternal age, have been implicated in common-risk-factors approach (Petersen 2003), which implies the formation of clefts (Wantia and Rettinger 2002). development of integral activities in health promotion and Advanced surgery, specific prosthetic appliances, and ortho- disease prevention, involving health education, community dontic treatment can improve the quality of life for those born empowerment, and legislative policy development. For exam- with clefts. However, such treatment is not accessible to chil- ple, such programs could aim at reducing the caries levels dren of several developing countries. Tobacco cessation among preschool children and simultaneously improving gen- programs aimed at pregnant mothers are essential in the pre- eral health. Promoting the reduction of sugar consumption vention of CL/P. would improve not only oral health but also general health Oral and Craniofacial Diseases and Disorders | 731 Table 38.3 Prevention Strategies for Oral Health Disease or condition Causes Actions needed and methods Dental caries High or frequent sugar consumption, plaque present, Targeted actions against causative factors on community highly cariogenic microorganisms, nonuse of fluorides, and individual levels reduced saliva flow, systemic diseases, and other Health education toward self-care capacity, fluoride pro- individual risk factors grams, sugar restriction, actions based on risk assessment of individuals and groups Periodontal diseases Plaque present, pathogenic bacteria, influence of systemic Improved oral hygiene, professional cleaning, antibiotics, diseases, tobacco use identification and treatment of systemic diseases Elimination of pockets if present and removal of local dental irritants, such as rough fillings Tobacco cessation Oral precancer and cancer Tobacco and alcohol use; see chapter 29 Tobacco cessation; see chapter 29 Oral manifestations of HIV/AIDS See chapter 18 See chapter 18. Special oral care Noma (cancrum oris) Probably bacterial in connection with severe Antibiotics together with nutritional support; surgery malnourishment sometimes necessary Developmental disorders Various genetic or environmental causes such as Tobacco cessation programs aimed at pregnant mothers tobacco use Fluorosis of teeth Too high concentration of fluoride in drinking waters or Identification of water sources and reduction of fluoride or from other sources recommendation of other water sources Source: Authors; partly based on Bratthall and Barmes 1993. Note: This table is by no means complete. Many other oral diseases or conditions are important and need attention. The listed ones are of special relevance for developing countries. through better quality of children's diet. Some prevention hormones, biomarkers for oral cancer, inflammatory media- strategies for oral health, suitable for developing countries are tors, and more. Future developments may result in other outlined in table 38.3. affordable and effective devices. Continuous attempts are being made to assess the sociobe- havioral factors in oral health and the information on risk fac- RESEARCH AND FUTURE ACTIONS tors. Caries risk assessment models are tested also for the indi- vidual cases. Through present knowledge, individuals in need Several promising actions against factors causing the two major of targeted actions can be identified. Another strong trend is to oral diseases, caries and periodontal disease, are ongoing: use evidence-based reviews. This type of research is, of course, attempts to control the formation of the biofilm with its not restricted to oral health. Several reviews have already been microflora are of high priority. One research line is to identify done, and a frequent conclusion is that the number of ran- pathogenic bacteria and try to replace them with genetically domized clinical trials is limited, in particular for common modified, less pathogenic bacteria or to eradicate them clinical procedures. This trend will change as the quality of by antibiotics or antiseptics. Preventing dental caries by a vac- future oral health research improves, but not all research prob- cine is not a new idea, and efforts continue. Among other ideas lems can be solved by such studies. Community-based partic- is the use of plantibodies (plant-derived therapeutic antibodies) ipatory research is another approach that may be used to or genetically modified bacteria,releasing components targeting improve oral health studies (O'Fallon and Dearry 2002). pathogens. Functional foods, which include various elements in Within the management of dental diseases--in particular, food, may be another future option to control oral diseases. caries--is the "minimally invasive dentistry" approach, which Although pilot or small-scale studies seem promising, it will be promotes the concept that large restorations (crowns, bridges) several years before such methods can possibly be of use in pop- are not as necessary as believed. Risk assessment, preventive ulations because large clinical trials have not even started. measures, and improved dental materials with good adhesion Saliva is believed to be usable as a diagnostic tool, providing capacity are some of the several components in this approach. noninvasive assessment of a number of oral and systemic dis- Research in transfer of knowledge using the Internet or eases. Devices are being designed to identify in saliva various other electronic media is another strongly expanding area, bacteria and their virulence factors, drugs, metabolic products, from which developing countries should be able to benefit. 732 | Disease Control Priorities in Developing Countries | Douglas Bratthall, Poul Erik Petersen, Jayanthi Ramanathan Stjernswärd, and others COST-EFFECTIVENESS OF ORAL HEALTH CARE DMFT 9 Using the evidence available, the U.S. Surgeon General's report 8 (U.S. Public Health Service 2000) and the report of the Swedish Council on Technology Assessment in Health Care (SBU 2002; 7 see also Kallestal and others 2003) have attempted to determine 6 the cost-effectiveness of oral health intervention programs 5 from developed countries. Among the findings in the U.S. report were the following: 4 3 · Water fluoridation costs about a dollar per person per year 2 for water serving most individuals in the United States. Community water fluoridation is believed to be an effective 1 and cost-effective caries preventive method. 0 · Economic analyses of community dental sealant programs 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 suggest that they are cost-effective and may even provide US$ cost savings when used in high-risk populations. Source: For health expenditure: http://www.who.int/en/; for DMFT: http://www. whocollab.od.mah.se/index.html, both for June 2003. Note: Original data for health expenditures were in international dollars and were The Swedish report (SBU 2002), reviewing original studies converted to U.S. dollars using the exchange rate of US$1 0.70681 (period average June 2003). Because the exchange rate varies over time, the data should be taken as on economic evaluation of caries prevention (a total of 17 approximate values. selected from 1966 to 2003 MEDLINE and manual Internet searches), commented that no conclusion could be drawn Figure 38.2 DMFT as Related to Health Care Expenditure per Capita for 12-Year-Olds in 149 Countries owing to the low evidence values and contradictory results. This comment prompted the group to present its own calcula- tion for cost-effectiveness based on Swedish caries prevalence on the effectiveness of traditionally accepted preventive meas- and charges used in Swedish dental care. The group found that ures, that political pressures on health care are motivated by the cost-effectiveness for fluoridated toothpaste is extremely economic pressures." Finally, he stated that traditional cost- good (cost per prevented DMFT very low), which, of course, is benefit and cost-effectiveness analyses have not been able to not surprising, given the significant caries-reducing results in help the decision makers choose wisely and that the time per- combination with low cost for society. spective for the real effects of prevention lies beyond the inter- No clear correlation appears to exist between caries experi- ests of decision makers. ence and health care investment for individual countries. However, without proper prevention, the alternative strategy Some countries with the lowest health care expenditures have is restorative dentistry--that is, to make fillings, crowns, and values for caries experience (DMFT) that are similar to or dentures. Is this a feasible alternative for developing countries? even lower than those countries having the highest expendi- Yee and Sheiham (2003) give some examples: In Nepal, a sim- tures on health (figure 38.2). Those low-income countries ple amalgam filling would cost about US$4, which does not often have low per capita sugar consumption and, therefore, include the many additional expenses for impoverished rural do not need to install expensive measures for treatment or families, who may have to travel by bus or walk for a day or two prevention. to get to the clinic. The total expenses incurred, including den- It may seem surprising that so few studies are available tal fees, meals, and lodging but not including lost wages, would regarding the cost-effectiveness of caries prevention, or of any amount to US$12, an enormous sum considering the average other oral disease. In a critical review article, Schwarz (1998) Nepalese's earning of US$0.75 per day, and it is enough to buy analyzed the issue. He wrote, "Several decades after consider- food for a month. Yee and Sheiham conclude that treating able improvements in the oral disease situation were docu- caries with the traditional method of restorative dentistry is mented in Scandinavia, doubts are still expressed about beyond the financial capabilities of most low-income nations whether preventive measures are cost-effective." In addition, he because three-quarters of these countries do not even have suf- recommended that four elements be considered when a pre- ficient resources to finance an essential package of health care ventive effect was evaluated: the definition of prevention, the services for their children. Yee and Sheiham (2003) estimate practical perception of effective prevention, the appropriate- that treating dental caries by the traditional amalgam restora- ness of traditional cost-effectiveness analysis, and the time fac- tive dentistry in the permanent dentition of the child popula- tor. He pointed out that "caries prevention is not uniformly tion would cost about US$2,000 for 1,000 children of mixed defined by the profession, that dental research is casting doubt ages from 6 to 18 years, which would require financial Oral and Craniofacial Diseases and Disorders | 733 resources beyond the capabilities of low-income nations. Axell, T. 1976. "A Prevalence Study of Oral Mucosal Lesions in an Adult Hence, they propose a public health and health promotion Swedish Population." Odontologisk Revy 27 (Suppl. 36): 1­103. approach to reduce caries burden instead of the restorative Barmes, D. E. 2000. "Public Policy on Oral Health and Old Age: A Global View." Journal of Public Health Dentistry 60: 335­37. approach. Beltran-Aguilar, E. D., S. Estupinan-Day, and R. Baez. 1999. "Analysis of Although several studies evaluating the effectiveness of Prevalence and Trends of Dental Caries in the Americas between the intervention and oral health promotion programs in develop- 1970s and 1990s." International Dental Journal 49: 322­29. ing countries are becoming available (Estupinan-Day and Bratthall, D. 2000. "Introducing the Significant Caries Index Together with others 2001; Pakhomov and others 1995; Petersen and others a Proposal for a New Global Oral Health Goal for 12-Year-Olds." International Dental Journal 50: 378­84. 2004), a definite need exists for further cost-effectiveness analy- Bratthall, D., and D. E. Barmes. 1993. "Oral Health." In Disease Control sis on such programs, which should be addressed in the future. Priorities in Developing Countries, ed. D. T. Jamison, W. H. Mosley, It would also be useful if studies were commenced on inter- A. R. Measham, and J. L. Bobadilla, 647­59. Washington, DC: World vention programs using the common-risk approach suggested Bank. by WHO (Petersen 2003). Bratthall, D., G. Hänsel-Petersson, and H. Sundberg. 1996. "Reasons for the Caries Decline: What Do the Experts Believe?" European Journal of Oral Sciences 104: 416­22. Burt, B. A. 2002. "Fluoridation and Social Equity." Journal of Public Health CONCLUSIONS Dentistry 62: 195­200. Carr, L. M. 1988. "Dental Health of Children in Australia, 1977­1985." Dental caries and periodontal diseases are the most known oral Australian Dental Journal 33: 205­11. diseases, but other conditions can strongly and negatively Chibole, O. 1987. "Epidemiology of Dental Fluorosis in Kenya." Journal of influence the quality of life. Effective programs to reduce the the Royal Society of Health 107: 242­43. burden of oral diseases--in particular, caries--are available in Cohen, L. K. 1990. "Promoting Oral Health: Guidelines for Dental Associations." International Dental Journal 40: 79­102. principle, but to run these programs in developing countries, de Almeida, C. M., P. E. Petersen, S. J. Andre, and A. Toscano. 2003. new approaches are needed. The WHO strategy of identifying "Changing Oral Health Status of 6- and 12-Year-Old Schoolchildren common risk factors seems promising for health promotion. In in Portugal." Community Dental Health 20: 211­16. broad terms, the most important challenges for oral health in Demertzi, A., and V. Topitsoglou. 2002."Caries Prevalence of 11-Year-Olds the 21st century relate to the transfer of knowledge and experi- between 1989­2001." Abstract. Community Dental Health 19: 203. ences in preventive oral care to the poor and disadvantaged Downer, M. C. 1994. "The 1993 National Survey of Children's Dental population groups in both developing and developed Health: A Commentary on the Preliminary Report." British Dental Journal 176: 209­14. countries. "Early Diagnosis and Prevention of Oral Cancer and Precancer: Report of Symposium III." 1995. Advanced Dental Research 9: 134­37. Enwonwu, C. O. 1995. "Noma: A Neglected Scourge of Children in Sub- ACKNOWLEDGMENTS Saharan Africa." Bulletin of the World Health Organization 73: 541­45. Enwonwu, C. O., W. A. Falkler, and E. O. Idigbe. 2000. "Oro-Facial We acknowledge with great appreciation comments and sug- Gangrene (Noma/Cancrum Oris): Pathogenetic Mechanisms." Critical Reviews in Oral Biology and Medicine 11: 159­71. gestions from Dr. Lois Cohen, Dr. Kevin Hardwick, Dr. Jeanne Estupinan-Day, S. R., H. Baez, R. Horowitz, R. Warpeha, B. Sutherland, C. Sinkford, and Thomas Wall. Sir George Alleyne, our editor, and M. Thamer. 2001. "Salt Fluoridation and Dental Caries in is to be congratulated for his constructive ideas and patience. Jamaica." Community Dentistry and Oral Epidemiology 29: 247­52. Ferlay, J., F. Bray, P. Pisani, and D. M. Parkin. 2001. GLOBOCAN 2000: Cancer Incidence, Mortality and Prevalence Worldwide. Version 1.0. IARC CancerBase, International Agency for Research on Cancer, Lyon. REFERENCES http://wwwdep.iarc.fr/globocan/globocan.htm. Flanders, R. A. 1987."Effectiveness of Dental Health Educational Programs Abayaratna, S., and K. Krishnarasa. 1997. National Oral Health Survey in Schools." Journal of the American Dental Association 114: 239­42. 1994­95. Colombo: Ministry of Health. Frencken, J. E., K. Borsum-Andersson, F. Makoni, F. Moyana, S. Adyatmaka, A., U. Sutopo, P. Carlsson, D. Bratthall, and G. Pakhomov. Mwashaenyi, and J. Mulder. 2001. "Effectiveness of an Oral Health 1998. School-Based Primary Preventive Programme for Children: Education Programme in Primary Schools in Zimbabwe after 3.5 Affordable Toothpaste as a Component in Primary Oral Health Care-- Years." Community Dentistry and Oral Epidemiology 29: 253­59. Experiences from a Field Trial in Kalimantan Barat, Indonesia. Geneva: World Health Organization. Genco, R. J. 1996. "Current View of Risk Factors for Periodontal Diseases." Journal of Periodontology 67: 1041­49. Arendorf, T. M., B. Bredekamp, C. A. Cloete, and G. Sauer. 1998. "Oral Manifestations of HIV Infection in 600 South African Patients." Greenspan, J. S., and D. Greenspan. 2002. "The Epidemiology of the Oral Journal of Oral Pathology and Medicine 27: 176­79. Lesions of HIV Infection in the Developed World." Oral Diseases 8 Armfield, J. M., K. F. Roberts-Thomson, and A. J. Spencer. 2003. The Child (Suppl. 2): 34­39. Dental Health Survey, Australia 1999: Trends across the 1990s. Adelaide: Guinan, J. C., R. Bakayoko-Ly, M. Samba, A. L. Kattie, and A. E. Oka. 1999. Australian Institute of Health and Welfare. http://www.cecdo.org/ "Caries Assessment of School Children 12 Years of Age in 1996 in Ivory pages/database%20intro.html. Coast." Tropical Dental Journal 22: 48­54. 734 | Disease Control Priorities in Developing Countries | Douglas Bratthall, Poul Erik Petersen, Jayanthi Ramanathan Stjernswärd, and others Haugejorden, O. 1994. "Changing Time Trend in Caries Prevalence in Nishi, M., J. Stjernsward, P. Carlsson, and D. Bratthall. 2002. "Caries Norwegian Children and Adolescents." Community Dentistry and Oral Experience of Some Countries and Areas Expressed by the Significant Epidemiology 22: 220­25. Caries Index." Community Dentistry and Oral Epidemiology 30: Hescot, P., and E. Roland. 2000. La santé dentaire en France, 1998. L'Union 296­301. Française pour la Santé Bucco-Dentaire, Paris. Nithila, A., D. Bourgeois, D. E. Barmes, and H. Murtomaa. 1998. "WHO Hewson, A. R., and C. M. McNamara. 2000. "Cleft Lip and/or Palate in the Global Oral Data Bank, 1986­96: An Overview of Oral Health Surveys West of Ireland, 1980­1996." Special Care in Dentistry 20: 143­46. at 12 Years of Age." Bulletin of the World Health Organization 76: 237­44. Hodgson, T. A. 1997. "HIV-Associated Oral Lesions: Prevalence in Zambia." Oral Diseases 3 (Suppl. 1): 46­50. Nittayananta, W., and S. Chungpanich. 1997. "Oral Lesions in a Group of Thai People with AIDS." Oral Diseases 3 (Suppl. 1): S41­45. Holmes, H. K., and L. X. G. Stephan. 2002. "Oral Lesions of HIV Infection O'Fallon, L. R., and A. Dearry. 2002. "Community-Based Participatory in Developing Countries." Oral Diseases 8 (Suppl. 2): 40­43. Research as a Tool to Advance Environmental Health Sciences." Hong Kong, Department of Health. 2003. Oral Health Survey 2001. Environmental Health Perspectives 110 (Suppl. 2): 155­59. http://www.info.gov.hk/tooth_club/survey_eng.htm. Pakhomov, G. N., K. Ivanova, I. J. Moller, and M. Vrabcheva. 1995. "Dental Honkala, E., L. Kannas, and J. Rise. 1990. "Oral Health Habits in 11 Caries-Reducing Effects of a Milk Fluoridation Project in Bulgaria." European Countries." International Dental Journal 40: 211­17. Journal of Public Health Dentistry 55: 234­37. Hugoson, A., G. Koch, T. Bergendal, A. L. Hallonsten, C. Slotte, B. Papapanou, P. N. 1999."Epidemiology of Periodontal Diseases: An Update." Thorstensson, and H. Thorstensson. 1995. "Oral Health of Individuals Journal of the International Academy of Periodontology 1: 110­16. Aged 3­80 Years in Jönköping, Sweden, in 1973, 1983 and 1993." Patton, L. L., J. A. Phelan, F. J. Ramos-Gomez, W. Nittayananta, C. H. Swedish Dental Journal 19: 243­60. Shiboski, and T. L. Mbuguye. 2002. "Prevalence and Classification of Ikeda, N., Y. Handa, S. P. Khim, C. Durward, T. Axell, T. Mizuno, and HIV-Associated Oral Lesions." Oral Diseases 8 (Suppl. 2): 98­109. others. 1995. "Prevalence Study of Oral Mucosal Lesions in a Selected Peng, B., P. E. Petersen, M. W. Fan, and B. J. Tai. 1997. "Oral Health Status Cambodian Population." Community Dentistry and Oral Epidemiology and Oral Health Behaviour of 12-Year-Old Urban Schoolchildren in 23: 49­54. the People's Republic of China." Community Dental Health 14: 238­44. Irigoyen, M. E., and G. Sanchez-Hinojosa. 2000."Changes in Dental Caries Peng, B., P. E. Petersen, B. J. Tai, B. Y. Yuan, and M. W. Fan. 1997. "Changes Prevalence in 12-Year-Old Students in the State of Mexico after 9 Years in Oral Health Knowledge and Behaviour 1987­95 among Inhabitants of Salt Fluoridation." Caries Research 34: 303­7. of Wuhan City, PR China." International Dental Journal 47: 142­47. Kallestal, C., A. Norlund, B. Soder, G. Nordenram, H. Dahlgren, L. G. Petersen, P. E. 1989. "Evaluation of a Dental Preventive Program for Petersson, and others. 2003. "Economic Evaluation of Dental Caries Danish Chocolate Workers." Community Dentistry and Oral Prevention: A Systematic Review." Acta odontologica Scandinavica 61: Epidemiology 17: 53­59. 341­46. ------.1990."Self-Administered Use of Fluoride among Danish Chocolate Kloos, H., and R. T. Haimanot. 1999. "Distribution of Fluoride and Workers." Scandinavian Journal of Dental Research 98: 189­91. Fluorosis in Ethiopia and Prospects for Control." Tropical Medicine and ------. 2003. "The World Oral Health Report 2003: Continuous International Health 4: 355­64. Improvement of Oral Health in the 21st Century--The Approach of Latvia, State Dentistry Centre. 2000. Annual Report of Dental Health Care the WHO Global Oral Health Programme." Community Dentistry and in Latvia. Riga: State Dentistry Centre. Oral Epidemiology 31 (Suppl. 1): 1­21. Leous, P., and P. E. Petersen. 2002. Oral Health Status of Schoolchildren in Petersen, P. E., I. Danila, A. Delean, O. Grivu, G. Ionita, M. Pop, and Belarus, 2000. Copenhagen: WHO Regional Office for Europe. A. Samolia. 1994. "Oral Health Status among Schoolchildren in Lim, A. A., Y. S. Leo, C. C. Lee, and A. N. Robinson. 2001. "Oral Romania, 1992." Community Dentistry and Oral Epidemiology 22: 90­3. Manifestations of Human Immunodeficiency Virus (HIV)­Infected Petersen, P. E., N. Hoerup, N. Poomviset, J. Prommajan, and A. Watanapa. Patients in Singapore." Annals of the Academy of Medicine, Singapore 2001. "Oral Health Status and Oral Health Behaviour of Urban and 30: 600­6. Rural Schoolchildren in Southern Thailand." International Dental Journal 51: 95­102. Lo, E. C., R. W. Evans, and O. P. Lind. 1990. "Dental Caries Status and Treatment Needs of the Permanent Dentition of 6­12-Year-Olds in Petersen, P. E., and M. Kaka. 1999. "Oral Health Status of Children and Hong Kong." Community Dentistry and Oral Epidemiology 18: 9­11. Adults in the Republic of Niger, Africa." International Dental Journal 49: 159­64. Malaysia, Dental Services Division. 1997. Dental Services. Kuala Lumpur: Ministry of Health. Petersen, P. E., and M. A. Lennon. 2004. "Effective Use of Fluorides for the Prevention of Dental Caries in the 21st Century: The WHO Mandari, G. J., J. E. Frencken, and M. A. Van't Hof. 2003. "Six-Year Success Approach." Community Dentistry and Oral Epidemiology 32: 319­21. Rates of Occlusal Amalgam and Glass-Ionomer Restorations Placed Using Three Minimal Intervention Approaches." Caries Research 37: Petersen, P. E., U. Nyandindi, E. N. Kikiwilu, L. Mabelya, B. S. Lembariti, 246­53. and V. J. Poulsen. 2002. Oral Health Status and Oral Health Behaviour of Schoolchildren, Teachers, and Adults in Tanzania. Geneva: World Micheelis, W., and E. Reich. 1999. The Third German Oral Health Study Health Organization. (DMS III). Cologne, Germany: Deutscher Aerzte-Verlag. Petersen, P. E., B. Peng, and B. J. Tai. 1997. "Oral Health Status and Oral Miyazaki, H., and M. Morimoto. 1996. "Changes in Caries Prevalence in Health Behaviour of Middle-Aged and Elderly People in PR China." Japan." European Journal of Oral Sciences 104: 452­58. International Dental Journal 47: 305­12. Naidoo, S., and U. M. Chikte. 1999. "HIV/AIDS--The Evolving Pandemic Petersen, P. E., B. Peng, B. Tai, Z. Bian, and M. Fan. 2004. "Effect of a and Its Impact on Oral Health in Sub-Saharan Africa." South African School-Based Oral Health Education Programme in Wuhan City, Dental Journal 54: 616­30. People's Republic of China." International Dental Journal 54: 33­41. ------. 2000. "Noma (Cancrum Oris): Case Report in a 4-Year-Old Petersen, P. E., and M. Rusu. 2002. Oral Health Status of Romanian HIV-Positive South African Child." South African Dental Journal 55: Schoolchildren--National Survey 2000. Copenhagen: WHO Regional 683­86. Office for Europe. Oral and Craniofacial Diseases and Disorders | 735 Petersen, P. E., and A. M. Torres. 1999. "Preventive Oral Health Care and Szoke, J., and P. E. Petersen. 2000. "Evidence for Dental Caries Decline Health Promotion Provided for Children and Adolescents by the among Children in an East European Country (Hungary)." Municipal Dental Health Service in Denmark." International Journal of Community Dentistry and Oral Epidemiology 28: 155­60. Paediatric Dentistry 9: 81­91. Tai, B., M. Du, B. Peng, M. Fan, and Z. Bian. 2001. "Experiences Pitts, N. B., D. J. Evans, and Z. J. Nugent. 1998. "The Dental Caries from a School-Based Oral Health Promotion Programme in Wuhan Experience of 12-Year-Old Children in the United Kingdom: Surveys City, PR China." International Journal of Paediatric Dentistry 11: Coordinated by the British Association for the Study of Community 280­91. Dentistry in 1996/97." Community Dental Health 15: 49­54. Thomson, W. M. 2004. "Dental Caries Experience in Older People over Pitts,N.B.,D.J.Evans,Z.J.Nugent,and C.M.Pine.2002."The Dental Caries Time: What Can the Large Cohort Studies Tell Us?" British Dental Experience of 12-Year-Old Children in England and Wales: Surveys Journal 196: 89­92. Coordinated by the British Association for the Study of Community Ullah, M. S. 2001. An Epidemiological Oral Health Study on 12-Year-Old Dentistry in 2000/2001." Community Dental Health 19: 46­53. Bangladeshi Schoolchildren. Thesis, Faculty of Dentistry, Oslo. Priwe, C. 1998. The Smiling School Project: A School Based Oral Health U.S. Public Health Service. 2000. Oral Health in America: A Report of the Promotion Programme: Mid-Term Progress Report. Windhoek: Ministry Surgeon General. Washington, DC: U.S. Public Health Service. of Health and Social Services. http://www.nidr.nih.gov/sgr/sgrohweb/TOC.htm. Ranganathan, K., B. V. Reddy, N. Kumarasamy, S. Solomon, R. Vanderas, A. P. 1987."Incidence of Cleft Lip, Cleft Palate, and Cleft Lip and Viswanathan, and N. W. Johnson. 2000. "Oral Lesions and Conditions Palate among Races: A Review." Cleft Palate Journal 24: 216­25. Associated with Human Immunodeficiency Virus Infection in 300 South Indian Patients." Oral Diseases 6: 152­57. van Palenstein Helderman, W. H., L. Munck, S. Mushendwa, M. A. van't Hof, and F. G. Mrema. 1997. "Effect Evaluation of an Oral Health Razanamihaja, N., and P. E. Petersen. 1999. "School Based Oral Health Education Programme in Primary Schools in Tanzania." Community Promotion Programmes in Madagascar." In Health Care Systems in Dentistry and Oral Epidemiology 25: 296­300. Africa--Patterns and Perspectives, ed. L. Blegvad, 123­29. Copenhagen: University of Copenhagen Press. van Wyk, P. J. 1994. National Oral Health Survey, South Africa, 1988/89. Pretoria: Department of Health. Reichart, P. A. 2001. "Identification of Risk Groups for Oral Precancer and Cancer and Preventive Measures."Clinical Oral Investigations 5: 207­13. von der Fehr, F. R. 1994. "Caries Prevalence in the Nordic Countries." International Dental Journal 44 (Suppl. 1): 371­78. SBU (Swedish Council on Technology Assessment in Health Care). 2002. Att förebygga karies: En systematisk litteraturöversikt. Gothenburg: Wang, N. J., C. Kalletstal, P. E. Petersen, and I. B. Arnadottir. 1998. "Caries Swedish Council on Technology Assessment in Health Care. Preventive Services for Children and Adolescents in Denmark, Iceland, Norway and Sweden: Strategies and Resource Allocation." Community Schwarz, E. 1998. "Is Caries Prevention Cost-Effective? Does Anybody Dentistry and Oral Epidemiology 26: 263­71. Care?" Acta odontologica Scandinavica 56: 187­92. Wantia, N., and G. Rettinger. 2002. "The Current Understanding of Cleft Sector Público. 1999. Encuesta de salud bucal, En escolares de 11 a 14 años. Lip Malformations." Facial Plastic Surgery 18: 147­53. Montevideo: Ministry of Public Health. Sembene, M., A. W. Kane, and D. Bourgeois. 1999. "Caries Prevalence in White, B. A., A. A. Antczak-Bouckoms, and M. C. Weinstein. 1989. "Issues 12-Year-Old Schoolchildren in Senegal in 1989 and 1994." in the Economic Evaluation of Community Water Fluoridation." International Dental Journal 49: 73­75. Journal of Dental Education 53: 646­57. Shah, N. 2001. "Geriatric Oral Health Issues in India." International Dental WHO (World Health Organization). 1994. Fluorides and Oral Health. Journal 51: 212­18. Technical Report Series No. 846. Geneva: WHO. Sogaard, A. J., and D. Holst. 1988. "The Effect of Different School Based ------. 2003. School Oral Health Promotion. WHO School Health Dental Health Education Programmes in Norway." Community Dental Information Series 11. Geneva: WHO. Health 5: 169­84. Wierzbicka, M., P. E. Petersen, F. Szatko, E. Dybizbanska, and I. Kalo. 2002. Soskolne, W. A., and A. Klinger. 2001. "The Relationship between "Changing Oral Health Status and Oral Health Behaviour of Periodontal Diseases and Diabetes: An Overview." Annals of Schoolchildren in Poland." Community Dental Health 19: 243­50. Periodontology 6: 91­98. Woodward, M., and A. R. Walker. 1994. "Sugar Consumption and Dental Sri Lanka, Ministry of Health. 1985. National Oral Health Survey, 1983­84. Caries: Evidence from 90 Countries." British Dental Journal 176: Colombo: Ministry of Health. 297­302. Sudbo, J., and A. Reith. 2003. "Which Putatively Pre-Malignant Oral Yee, R., and A. Sheiham. 2003. "Is Treating Caries in Children in Lesions Become Oral Cancers? Clinical Relevance of Early Targeting of Developing Countries by the Restorative Approach a Rational High-Risk Individuals." Journal of Oral Pathology and Medicine 32: Objective?" Developing Dentistry 3: 10­17. 63­70. Zadik, D., S. P. Zusman, and A. M. Kelman. 1992. "Caries Prevalence in 5- Sundberg, H. 2002. Tandhälsan hos barn och ungdomar 1985­2001. and 12-Year-Old Children in Israel." Community Dentistry and Oral Newsletter. Swedish National Board of Health and Welfare, Stockholm. Epidemiology 20: 54­55 736 | Disease Control Priorities in Developing Countries | Douglas Bratthall, Poul Erik Petersen, Jayanthi Ramanathan Stjernswärd, and others Chapter 39 Unintentional Injuries Robyn Norton, Adnan A. Hyder, David Bishai, and Margie Peden This chapter examines the issue of unintentional injuries and accounting for around 7 percent of all deaths in those coun- focuses on a selected number of cause-specific unintentional tries. Similarly, more than 90 percent of DALYs that were injuries. Injuries have traditionally been defined as damage attributed to unintentional injuries occurred in LMICs, to a person caused by an acute transfer of energy (mechanical, accounting for about 8 percent of all DALYs in those countries. thermal, electrical, chemical, or radiation) or by the sudden Injury death rates per 100,000 population were higher in absence of heat or oxygen. Unintentional injuries consist of LMICs (62 per 100,000) than globally (57 per 100,000). that subset of injuries for which there is no evidence of pre- Males accounted for almost two-thirds of the deaths attrib- determined intent. The cause-specific unintentional injuries uted to unintentional injuries in LMICs in 2001, with rates of examined here include those that the World Health both injury death and DALY losses higher among males than Organization (WHO) routinely analyzes and publishes data on females (table 39.1). Compared with other age groups, young and that individually account for the greatest unintentional people age 15 to 29 accounted for the largest proportion of injury burden in terms of mortality and disability-adjusted life deaths from unintentional injuries in LMICs (figure 39.2). years (DALYs). These include road traffic injuries (RTIs), poi- RTIs accounted for the greatest burden of deaths from sonings, falls, burns, and drowning (figure 39.1). unintentional injuries in LMICs in 2001, or about 34 percent of the total burden, and the greatest burden of DALYs from unintentional injuries in LMICs in 2001, accounting for BURDEN AND CAUSES OF UNINTENTIONAL 28 percent of the burden (figure 39.1). Whereas young people INJURIES age 15 to 29 years accounted for the highest proportion of all unintentional injuries, those age 45 to 59 accounted for the This section provides a brief outline of the burden of uninten- highest proportion of injuries from poisonings, while those tional injuries and then reviews the available evidence about age 70 to 79 accounted for the highest proportion of injuries known and potential causes of such injuries. from falls (figure 39.2). Economic Burden of Unintentional Injuries. Estimates of the Burden of Unintentional Injuries burden of unintentional injuries as measured in terms of eco- Worldwide, unintentional injuries accounted for more than nomic costs are almost nonexistent. The best estimates avail- 3.5 million deaths in 2001, or about 6 percent of all deaths and able are for RTIs. Using road crash costs from 21 developed 66 percent of all injury deaths. Unintentional injuries were also and developing countries, the Transport Research Laboratory responsible for more than 113 million DALYs in 2001, or about Ltd. finds that the average annual cost of road crashes was 8 percent of all DALYs and some 70 percent of all injury equivalent to about 1.0 percent of gross national product in DALYs. More than 90 percent of unintentional injury deaths developing countries, 1.5 percent in transition countries, and occurred in low- and middle-income countries (LMICs), 2.0 percent in highly motorized countries. The annual burden 737 a. Deaths Table 39.1 Cause-Specific Death Rates and DALYs Lost because of Unintentional Injuries, by Gender, Worldwide Other and in LMICs, 2001 unintentional injuries Road traffic 26% Global LMICs injuries 34% Category Total Males Females Total Males Females Deaths (per 100,000 population) All unintentional 57 75 41 61 80 44 Drowning injuries 11% RTIs 19 28 11 20 30 11 Poisonings Fires Poisonings 6 7 4 6 8 5 10% 9% Falls Falls 6 8 5 6 8 5 10% Fires 5 4 6 6 4 7 b. DALYs Drowning 6 9 4 7 10 5 Other unintentional 15 19 11 16 20 11 Other Road traffic injuries unintentional injuries injuries 28% DALY losses (per 1,000 population) 37% All unintentional 20 25 14 22 28 16 injuries RTIs 6 8 3 6 9 4 Poisonings Poisonings 1 2 1 1 2 1 6% Falls 2 3 2 3 3 2 Drowning Falls Fires 2 1 2 2 2 2 8% Fires 12% Drowning 2 2 1 2 2 1 9% Other unintentional 7 9 5 8 10 6 Source: Authors. injuries Figure 39.1 Distribution of Unintentional Injuries, Low- and Middle- Source: Authors. Income Countries, 2001 Note: All figures are rounded to the nearest 1,000. Percentage 35 30 25 20 15 10 5 0 All unintentional Road traffic injuries Poisonings Falls Fires Drowning Other injuries Type of injury 0­4 years 5­14 years 15­29 years 30­44 years 45­59 years 60­69 years 70­79 years 80 + years Source: Authors. Figure 39.2 Distribution of Unintentional Injuries by Type of Injury and Age Group, LMICs, 2001 738 | Disease Control Priorities in Developing Countries | Robyn Norton, Adnan A. Hyder, David Bishai, and others of road crash costs is about US$518 billion globally and about considers the extent to which information obtained from stud- US$65 billion in LMICs, exceeding the total annual amount ies conducted in HICs may be relevant. these countries receive in development assistance (Jacobs, Aeron-Thomas, and Astrop 2000). Risk Factors for Road Traffic Injuries. The increasing volume of traffic is one of the main factors contributing to the increase in RTIs in LMICs. Motorization rates rise with income (Kopits and Causes of Unintentional Injuries in LMICs Cropper 2005), and a number of LMICs experiencing growth As in the case of most diseases, unintentional injuries are caused have seen a corresponding increase in the number of motor vehi- by multiple factors. The traditional epidemiological paradigm cles (Ghaffar and others 1999). In some LMICs, this growth has of host, vector, and environmental factors that in combination been led by an increase in motorized two-wheeled vehicles, one contribute to the incidence of disease has been adapted and of the least safe forms of travel, which has resulted in concurrent applied in determining the causes of unintentional injury. increases in related injuries (Zhang and others 2004). However, this paradigm has been extended to consider each The rapid growth in motor vehicles in many LMICs has not factor in relation to the time of the injury--that is, factors oper- been accompanied by improvements in facilities for these road ating before, during, and after the injury that might be associ- users or by facilities that respond to the continued predomi- ated with both its incidence and its severity (Haddon 1968). nance of nonmotorized traffic (Khayesi 2003).Many of the tech- Although the matrix, called the Haddon matrix, was initially nical aspects of planning, highway design, traffic engineering, developed to address the problem of RTIs only, it provides a and traffic management that are the hallmarks of transportation comprehensive framework in which researchers can consider systems in many HICs are absent in LMICs, which need to plan the multitude of factors that may play a role in the causal injury for a level of heterogeneity in traffic that HICs do not encounter pathway, as outlined in table 39.2. (Tiwari 2000). In the past two decades, the evidence base for the identifica- Studies undertaken primarily in HICs show a strong rela- tion of risk factors for unintentional injuries in high-income tionship between the increase in vehicle speeds and increased countries (HICs) has increased dramatically as the number risk of crash and injury, both for motor vehicle occupants and of injury researchers and research institutions has increased. for vulnerable road users, particularly pedestrians (European However, because of the paucity of injury researchers and Road Safety Action Program 2003). This relationship is likely to research institutions in LMICs, the evidence base for the iden- be true for LMICs, and indeed, data obtained from routinely tification of risk factors for unintentional injuries in these collected police reports in a number of LMICs show that speed countries is growing more slowly. is listed as the leading cause of road traffic crashes, accounting Although knowledge about risk factors for injuries in HICs for up to 50 percent of all crashes (Afukaar 2003; Odero, may also be relevant for LMICs, the material presented in the Khayesi, and Heda 2003; Wang and others 2003). following section focuses on information that has been Several case-control studies in HICs have confirmed the role obtained from studies in LMICs. However, the section also of alcohol in the increasing risk of road crashes (Peden and Table 39.2 The Haddon Matrix as Applied to Road Traffic Injuries Factors Phase Nature of intervention Human Vehicles and equipment Environment Precrash Crash prevention Information Roadworthiness Road design Attitudes Lighting Road layout Impairment Braking Speed limits Police enforcement Handling Pedestrian facilities Speed management Crash Injury prevention Use of restraints Occupant restraints Forgiving roadside (for during crash Impairment Other safety devices example, crash barriers) Crash-protective design Postcrash Life sustaining First-aid skill Ease of access Rescue facilities Access to medical personnel Fire risk Congestion Source: Authors. Unintentional Injuries | 739 others 2004). Studies conducted in LMICs showed that drivers Studies in HICs suggest that roadside hazards, such as had consumed alcohol in 33 to 69 percent of crashes in which trees, poles, and road signs, may contribute to between 18 and drivers were fatally injured and in 8 to 29 percent of crashes in 42 percent of road crashes and increase injury severity which drivers were not fatally injured (Odero and Zwi 1995). (Kloeden and others 1998), although the extent to which this is Alcohol consumption by pedestrians also increases their risk of also true in LMICs has not been determined. injuries in HICs; moreover, in at least some LMICs, more than 50 percent of fatally injured pedestrians had consumed alcohol Risk Factors for Poisonings. The literature on poisonings in (Peden and others 1996). LMICs includes comprehensive information about intentional Other factors that increase the risks of road crashes in HICs poisonings; significant information about occupation-related include fatigue, use of hand-held mobile telephones, and inad- poisonings, especially pesticide poisonings; and a growing equate visibility of vulnerable road users (Peden and others body of information about lead poisoning. Each of these types 2004), all of which are equally likely to increase risks in LMICs. of poisoning is covered elsewhere in this book. This chapter Indeed, a recent case-control study from China shows that the focuses on risk factors for other types of poisoning in LMICs, risks of a crash doubled with chronic sleepiness on the part of and, in particular, focuses on risk factors for poisonings in the driver (G. F. Liu and others 2003), and surveys of commer- young children. cial and public road transport in a number of African countries The literature's focus on risk factors for childhood poison- have shown that drivers often work long hours and go to work ing probably reflects the fact that child poisoning victims are exhausted (Mock, Amegashi, and Darteh 1999; Nafukho and seen more often than adults in most hospital settings (Ellis and Khayesi 2002). Studies in Malaysia clearly show that motorcy- others 1994; Nhachi and Kasilo 1992). This fact is in stark con- clists who use daytime running lights have a crash risk about trast to the data presented earlier, which clearly show that mid- 10 to 29 percent lower than those who do not because of their dle-aged individuals sustain the vast majority of deaths and greater visibility (Radin Umar, Mackay, and Hills 1996). DALYs from poisonings in LMICs. Those numbers no doubt Road- and vehicle-related factors may also increase the risk of reflect the importance of work-related poisonings. crash involvement. Specific factors related to road planning Young males consistently appear to be at higher risk of poi- include traffic passing through residential areas, conflicts sonings than females (Ellis and others 1994; Fernando and between pedestrians and vehicles, schools located on busy roads, Fernando 1997; Soori 2001). The most common agents lack of median barriers to prevent dangerous passing on two-lane involved in childhood poisonings are paraffin (or kerosene) roads,and lack of barriers to prevent pedestrian access onto high- and other household chemicals; pesticides; and various plants speed roads, although few studies have specifically examined the or animals, including snakes (Fernando and Fernando 1997). risks associated with those factors (Ross and others 1991). Several case-control studies in LMICs indicate the impor- Although the severity of crash injuries is related to in- tance of a number of sociodemographic risk factors, including vehicle crash protection, evidence indicates that many engi- young parents, residential mobility, and limited adult supervi- neering advances found in vehicles in HICs are not present in sion of children (Azizi, Zulkifli, and Kasim 1993; Soori 2001). vehicles in LMICs (Odero, Garner, and Zwi 1997). Perhaps one The studies also suggest that previous poisoning may be a risk of the most important factors contributing to injury severity factor (Soori 2001). Another important factor seems to be stor- relates to crash protection for vulnerable road users. However, age, including the number of storage containers used in the few HICs, let alone LMICs, require the fronts of cars or buses residence; the use of nonstandard containers for storage (for to be designed in a way that would protect vulnerable road example, beverage bottles for storing kerosene); and the storage users (Mohan 2002). of poisons at ground level (Azizi, Zulkifli, and Kasim 1993; A significant risk factor for increased severity of injuries of Chatsantiprapa, Chokkanapitak, and Pinpradit 2001; Soori users of motorized two-wheeled vehicles is riders' failure to use 2001). motorcycle helmets (B. Liu and others 2004). Studies in a num- ber of Asian countries have shown that failure to use helmets, Risk Factors for Fall-Related Injuries. Risk factors for fall- use of nonstandard helmets, and use of improperly secured related injuries in older people are generally considered in helmets are not uncommon, even in countries with mandatory terms of risk factors for falling, risk factors associated with helmet laws (Conrad and others 1996; Kulanthayan and others the severity of the impact following the fall, and risks factors 2000). Failure to wear helmets is also a risk factor for increased associated with low levels of bone mineral density--insofar injury severity among bicyclists (Attewell, Glase, and as almost all fall-related injuries in older people involve bro- McFadden 2001). Although the failure to use seat belts is a sig- ken bones. The risk factors associated with the latter two nificant risk factor associated with injury severity among vehi- categories are generally related to aspects of the aging process cle occupants, many LMICs have no requirements for seat belts and, as a consequence, are considered in more detail in to be fitted or used (Peden and others 2004). chapter 51. 740 | Disease Control Priorities in Developing Countries | Robyn Norton, Adnan A. Hyder, David Bishai, and others Analytical studies conducted in a variety of LMICs have Investigators have undertaken case-control studies aimed at tended to show that risk factors for fall-related injuries, espe- identifying risk and protective factors for burn-related injuries cially hip fractures, are consistent with the risk factors identi- in Africa, Asia, and South America, and all focus on identify- fied in HICs. Those risk factors include low bone density; poor ing risk factors for children. Environmental risk factors that nutritional status and low body mass index; low calcium have been identified include lack of a water supply, storage of intake; comorbid conditions, such as hypertension and dia- a flammable substance in the home, cooking equipment in the betes; poor performance in activities of daily living; low levels kitchen within reach of children, and housing that is located in of engagement in physical activity; poor cognitive function; slums and congested areas. Persons with personal and socio- poor perceived health status; poor vision; environmental fac- economic risk factors included children who were not the first tors affecting balance or gait; family history of hip fracture; and born, who had a pregnant mother, whose mother recently was alcohol consumption (Boonyaratavej and others 2001; Clark dismissed from a job, who had recently moved, who had a pre- and others 1998; Jitapunkul, Yuktananandana, and Parkpian existing impairment, whose sibling died from a burn or had 2001). a history of burn, whose parents lacked alertness to burns, Some studies have identified other factors that may be more whose clothing was made of synthetic fabrics, whose parents relevant in the context of LMICs. For example, studies in were illiterate, and whose parents were of low economic status. Thailand suggest that factors associated with poor socioeco- Protective factors included the presence of a living room, nomic status may increase risk--for example, not having elec- better maternal education, and a history of previous injury tricity in the house and living in Thai-style houses or huts among males who lived in good environmental conditions (Jitapunkul, Yuktananandana, and Parkpian 2001). (Daisy and others 2001; Delgado and others 2002; Forjuoh The literature specifically identifying risk factors for falls in and others 1995; Werneck and Reichenheim 1997). younger people in LMICs is sparse, but the information there is indicates that such falls usually occur in and around the home, Risk Factors for Drowning. Most drowning incidents in with a significant proportion being associated with falls from LMICs are not associated with recreation or leisure, as is com- heights, including rooftops and trees (Adesunkanmi, Oseni, monly the case in HICs, but instead are associated with every- and Badru 1999; Bangdiwala and Anzola-Perez 1990; Kozik and day activities near bodies of water, including rivers, wells, and others 1999; Raja, Vohra, and Ahmed 2001). However, falls buckets (Celis 1997; Hyder and others 2003; Kobusingye, other than from heights predominate and are frequently related Guwatudde, and Lett 2001). to engagement in vigorous levels of physical activity. As noted earlier, men account for a higher proportion of drowning incidents than women, and children age one to four Risk Factors for Burn-Related Injuries. Despite the focus of and young people appear to be at greatest risk, with drowning WHO's data on burn-related injuries sustained as a result of accounting for a high proportion of injury-related deaths in fires, a number of country-specific surveys conducted in med- those age groups (Celis 1997; Kibel and others 1990; Kozik and ical facilities suggest that scalds from hot water may be equally others 1999; Tan, Li, and Bu 1998). Some surveys also suggest important or more important causes of burn-related injuries that older people may be at particularly high risk (Tan, Li, and (Chan and others 2002; Delgado and others 2002; Forjuoh, Bu 1998). Guyer, and Smith 1995; Rossi and others 1998). However, in Descriptive surveys indicate that those living in rural areas some countries, including China and particularly India, fire- are at greater risk than those living in urban areas (Kobusingye, related injuries clearly outweigh scald-related injuries (Ahuja Guwatudde, and Lett 2001), probably indicating greater expo- and Bhattacharya 2002; Jie and Ren 1992). sure to unprotected water surfaces. A number of studies find Overall, women are at greater risk of fire-related burn that most adult drowning incidents appear to be associated injuries than are men; however, data from population-based with positive blood alcohol tests (Carlini-Cotrim and da Matta and medical center surveys suggest that in some settings Chasin 2000; Celis 1997). (excluding India), males may be at greater risk of burns than are Case-control studies of drowning in young children have females (Chan and others 2002; Zhu, Yang, and Meng 1988). In identified both sociodemographic risk factors and risk factors many studies, burn-related injuries account for a much higher associated with proximity to bodies of water. Ahmed, Rahman, proportion of injuries in young children compared with other and van Ginneken's (1999) study in Bangladesh shows that the age groups (Jie and Ren 1992; E. H. Liu and others 1998). risk of drowning increased with the age of the mother and Rural location appears to be a consistent risk factor for increased much more sharply the larger the number of children burn-related injuries (Courtright, Haile, and Kohls 1993; in the family. Celis's (1997) case-control study in Mexico finds Zhu, Yang and Meng 1988), as is the home (Delgado and that the risk of drowning associated with having a well at home others 2002; Forjuoh, Guyer, and Smith 1995; E. H. Liu and was almost seven times that for children in homes without a others 1998). well. Unintentional Injuries | 741 INTERVENTIONS obvious. Other interventions, particularly those that may have modest but important benefits, may require rigorous evalua- Interventions to prevent unintentional injuries have tradition- tion methods. ally been considered in terms of the three E's--education, Evidence of the effectiveness of interventions in LMICs, as enforcement, and engineering--and within the framework of opposed to HICs, is also relatively uncommon. Although the the Haddon matrix. That is, interventions are considered proven efficacy of some interventions in HICs does not require in terms of preventing the occurrence of the injury, minimiz- replication in LMICs--for example, the use of motorcycle ing the severity of injury at the time of the injury, and mini- helmets--strategies that may be effective in increasing the rates mizing the severity of injury following the injury event. of helmet wearing in HICs may not necessarily be appropriate Although randomized controlled trials are the gold standard in LMICs. Thus, WHO and others increasingly endorse for assessing the effectiveness of injury interventions, such tri- tailoring interventions found to be effective in HICs to LMICs, als are still relatively rare in relation to injuries, and in many followed by rigorous evaluation (Peden and others 2004). cases such trials may be impractical to implement. Studies Table 39.3 provides a summary of promising and recom- comparing the incidence of injury before and after the imple- mended interventions, as well as interventions that have mentation of an intervention, sometimes with reference to a specifically been shown to be effective in LMICs. control population in which the intervention has not been introduced, more commonly provide the only evidence of effectiveness. In some areas, findings from observational stud- Road Traffic Injuries ies, such as case-control studies, provide the best available evi- Many working to reduce RTIs use the "safer roads, safer vehi- dence. However, randomized controlled trials are clearly not cles, safer people, and safer systems" motto. A recent augmenta- needed for some interventions because their benefits are tion of this motto derives from the recognition of the important Table 39.3 Promising and Effective Interventions for Injuries in LMICs Interventions shown to be effective in Injury Promising interventions LMICs (references) RTIs Reducing motor vehicle traffic: efficient fuel taxes, changes in land-use policy, safety Increasing the legal age of motorcyclists from impact assessment of transportation and land-use plans, provision of shorter and 16 to 18 years (Norghani and others 1998) safer routes, trip reduction measures Making greater use of safer modes of transport Minimizing exposure to high-risk scenarios: restricting access to different parts of the road network, giving priority to higher occupancy vehicles or to vulnerable road users, restricting the speed and engine performance of motorized two-wheelers, increasing the legal age for operating a motorcycle, using graduated driver's licensing systems Safer roads Safety awareness in planning road networks, safety features in road design, and remedial action in high-risk crash sites: making provisions for slow-moving traffic and vulnerable road users; providing passing lanes, median barriers, and street lighting Traffic calming measures, such as speed bumps Speed bumps in reducing pedestrian injuries (Afukaar, Antwi, and Ofosu-Amaah 2003) Speed cameras Safer vehicles Improving the visibility of vehicles, including requiring automatic daytime Daytime running lights on motorcycles (Radin running lights Umar, Mackay, and Hills 1996; Yuan 2000) Incorporating crash protective design into vehicles, including installing seat belts Mandating vehicle licensing and inspection Safer people Legislating strategies and increasing enforcement of, for example, speed limits, Increases in fines and suspension of driver's alcohol-related limits, hours of driving for commercial drivers, seat belt use, bicycle licenses (Poli de Figueiredo and others 2001) and motorcycle helmet use Legislation and enforcement of motorcycle helmets (Ichikawa, Chadbunchachai, and Marui 2003; Supramaniam, Belle, and Sung 1984). 742 | Disease Control Priorities in Developing Countries | Robyn Norton, Adnan A. Hyder, David Bishai, and others Table 39.3 Continued Interventions shown to be effective in Injury Promising interventions LMICs (references) Poisonings Better storage, including positioning and nature of storage vessels Free distribution of child-resistant containers Use of child-resistant containers (Krug and others 1994) Warning labels First-aid education Poison control centers Fall-related injuries Older people Muscle strengthening and balance retraining, individually prescribed Tai chi group exercise Home hazard assessment and modification for high-risk individuals Multidisciplinary, multifactorial screening for health and environmental risk factors Younger people Multifaceted community programs of the Children Can't Fly type Burn-related injuries Fire-related injuries Introducing programs to install smoke alarms Separating cooking areas from living areas Locating cooking surfaces at heights Reducing the storage of flammable substances in households Supervising children more effectively Introducing, monitoring, and enforcing standards and codes for fire-resistant garments Scald-related injuries Separating cooking areas from play areas Improving the design of cooking vessels Fire- and scald-related injuries Increasing awareness of burns prevention Providing first-aid education Drowning Limiting exposure to bodies of water close to dwellings, such as by fencing Providing learn-to-swim programs Providing education about risks for drowning Increasing supervision and providing lifeguards at recreational facilities Equipping boats with flotation devices and ensuring their use Legislating and enforcing rules about the numbers of individuals carried on boats Having trained and responsive coast guard services Source: Authors. role of appropriate transport and land-use policies in managing Safer Roads. Intervention strategies focusing on safer roads exposure to the risk of an RTI (Peden and others 2004). should incorporate safety awareness in planning road networks, Managing exposure to risk involves strategies aimed at safety features in road design, and remedial action for high-risk reducing motor vehicle traffic, encouraging the use of safer crash sites. HICs have adopted many of these strategies, and modes of travel, and minimizing exposure to high-risk scenar- though they have not been examined in rigorously controlled ios, as outlined in greater detail in table 39.3. Evidence from studies, such strategies form the basis of best-practice guide- Malaysia shows that increasing the legal age of motorcyclists lines and manuals now being used in LMICs (Ross and others from 16 to 18 has been beneficial (Norghani and others 1998), 1991). but evidence of the effectiveness of many of the other strategies Traffic calming measures are among the strategies recom- is not yet available for LMICs. mended for incorporating safety features into road design. Unintentional Injuries | 743 Although evidence from randomized controlled trials is not Poisonings yet available (Bunn and others 2003), a before-and-after study The prevention of unintentional poisonings includes consider- conducted in Ghana suggested that speed bumps were effective ation of both occupational and nonoccupational poisonings; in reducing traffic injuries, especially pedestrian injuries however, chapter 60 provides a fuller discussion of effective (Afukaar, Antwi, and Ofosu-Amaah 2003). A recent summary interventions in relation to work-related poisonings, so these of research findings also suggests that automated speed are not discussed here. enforcement virtually eliminates speeding (ICF Consulting Suggested interventions to reduce exposure to nonoccupa- Ltd. and Imperial College Centre for Transport Studies 2003). tional poisonings include better storage of poisons in terms of both the location and the nature of the storage vessels used. Safer Vehicles. Strategies focusing on safer vehicles include Specific interventions include storing poisons outside the improving the visibility of vehicles, incorporating crash protec- home and above children's head height and reducing the use tive design into vehicles, and promoting further development of secondhand household containers--for example soda of "intelligent" vehicles. However, in LMICs, strategies that bottles--along with introducing and enforcing legislation to simply ensure regular maintenance of older vehicles or removal prohibit the sale of poisons in such containers (Nhachi and of vehicles in poor condition from the roads, as well as vehicle Kasilo 1994). licensing and inspection, have the potential to be cost-effective The efficacy of child-resistant containers in preventing access (Peden and others 2004). to poisons has been demonstrated, and data from a controlled Meta-analyses of the effects of automatic daytime running before-and-after study in South Africa suggest that free distri- lights on cars consistently show that they reduce road crashes bution of child-resistant containers is a highly effective means of (Elvik and Vaa 2004). Studies in both Malaysia and Singapore preventing poisoning in children (Krug and others 1994). show similar positive effects for daytime running lights on motorcycles (Radin Umar, Mackay, and Hills 1996; Yuan 2000). Although the fitting of seat belts--probably the most Fall-Related Injuries well-known and effective safer vehicle strategy--is covered by Interventions proven effective for preventing falls by older peo- technical standards worldwide and is mandatory in most coun- ple in HICs include muscle strengthening and balance retrain- tries, anecdotal evidence suggests that vehicles in many LMICs ing that is individually prescribed at home by a trained health lack functioning seat belts (Forjuoh 2003). professional; tai chi group exercise; home hazard assessment and modification that is professionally prescribed for older Safer People. Intervention strategies aimed at improving road people with a history of falling; and multidisciplinary, multi- user behavior are increasingly focusing on the introduction factorial health and environmental risk factor screening and and enforcement of relevant legislation rather than on educa- intervention programs, both for community-dwelling older tional efforts. For example, Poli de Figueiredo and others' people in general and for older people with known risk factors (2001) research in Brazil indicates that increasing fines and (Chang and others 2004). suspending drivers' licenses immediately reduced RTIs and In relation to fall-related injuries among young children, deaths. other than general recommendations about increased supervi- A large body of research, although little of it conducted in sion of children and the importance of height reductions and LMICs, shows that setting and enforcing speed limits reduces appropriate ground surfacing to prevent playground injuries, RTIs by up to 34 percent. It also shows that RTIs are reduced only one intervention provides evidence of effectiveness that in varying magnitudes by setting and enforcing legal blood may be relevant for LMICs. The Children Can't Fly Program alcohol limits and minimum drinking-age laws, using alcohol has four major components, which include surveillance and checkpoints, and running mass media campaigns aimed at follow-up, media campaigns, community education, and the reducing drinking and driving (Peden and others 2004). provision of free, easily installed window guards to families The introduction and enforcement of mandatory seat belt with young children living in high-risk areas (Spiegel and and child restraint laws reduces occupant deaths and injuries Lindaman 1977). The program has been shown to be effective by up to 25 percent; however, such laws have not been intro- in reducing falls in low-income areas. duced in all LMICs (Peden and others 2004). Both bicycle and motorcycle helmets reduce head injuries among riders by up to 85 percent. Though education may be effective in increasing helmet use, the effect is greater when Burn-Related Injuries combined with legislation and enforcement, as demonstrated Evidence of the effectiveness of interventions to prevent fire- in Malaysia and Thailand (Ichikawa, Chadbunchachai, and related injuries is limited. A randomized controlled trial of a Marui 2003; Supramaniam, Belle, and Sung 1984). smoke detector giveaway program in inner London was unable 744 | Disease Control Priorities in Developing Countries | Robyn Norton, Adnan A. Hyder, David Bishai, and others to show evidence of the program's effectiveness on the incidence (Peden and others 2004). Some data are also starting to emerge of fires and fire-related injuries (DiGuiseppi and others 2002). from HICs with respect to the cost-effectiveness of fall-related However, a more recent study suggests that installation pro- injury prevention programs for older people. However, data on grams may be more effective in increasing the use of these either the costs or the cost-effectiveness of interventions to pre- alarms than giveaway programs alone (Harvey and others 2004). vent poisonings, burns, or drownings are limited. Interventions that have been proposed but whose effective- Cost-effectiveness studies done in HICs can only be sugges- ness has not yet been proven include separating cooking areas tive for LMICs, because the costs of property losses, disability, from living areas (including efforts to reduce the use of indoor and medical care are so vastly different. Furthermore, basic fires for cooking), ensuring that cooking surfaces are at heights, efficacy is not always guaranteed when a control strategy that reducing the storage of flammable substances in households, worked in a modern city is exported to a poor LMIC village. and supervising young children more effectively (Forjuoh 2004). Consequently, the ability to extrapolate from high-income to The introduction, monitoring, and enforcement of standards low-income countries is severely limited. Yet as middle-income and codes for and the wearing of fire-retardant garments have countries progress, they will begin to consider interventions also been proposed (Bawa Bhalla, Kale, and Mohan 2000). that have already been proven. Evidence of the effectiveness of interventions to prevent Despite the methodological challenges, we modeled the scald injuries is minimal but promising, although such inter- cost-effectiveness of five potential interventions to prevent ventions primarily focus on education, legislation, and enforce- unintentional injuries using information presented earlier on ment of efforts to regulate the temperature of water flowing known effective interventions in LMICs. In each case, the evi- from household taps (Macarthur 2003). dence for effectiveness in an LMIC setting was strong. How- Finally, interventions directed at increasing awareness of ever, because so few interventions have been evaluated in burn prevention have been proposed, largely because of the LMICs, we had to make certain assumptions to extrapolate success of safe community interventions involving a multitude findings about costs and effectiveness in one LMIC setting to of strategies (Ytterstad and Sogaard 1995). other settings (for an outline of the assumptions associated with this modeling, see Bishai and Hyder 2004). Our economic analyses are highly generalized and indicative of what might be Drowning achieved with the interventions considered. Evidence for the effectiveness of interventions to prevent For the analyses, we present all cost estimates in local cur- drowning is almost nonexistent. The only available data come rency converted to U.S. dollars (2001 exchange rates). We adopt from case-control studies undertaken in HICs that suggest that a societal perspective for each intervention, but if appropriate, fencing domestic swimming pools reduces the risks of drown- we comment on cost-effectiveness from a government perspec- ing (Thompson and Rivara 2000). Extrapolation of those find- tive. The time horizon for each intervention is one year of sus- ings to a low-income setting suggests that covering wells with taining the intervention. Costs are annualized so that a typical grills, fencing nearby lakes or riverbanks, and building flood year of operating the intervention is known. As with any inter- control embankments might be effective in reducing drowning. vention, annual operating costs may fall as those involved learn ways to carry out their tasks more efficiently. Each year of pro- gram operation prevents an estimated number of deaths and COSTS, COST-EFFECTIVENESS, AND ECONOMIC injuries. In each case, we present estimates of the raw numbers BENEFITS OF INTERVENTION of deaths and the undiscounted numbers of life years they rep- resent. However, from an economic perspective, the life years Data on effective interventions for preventing unintentional and DALYs of those who sustain nonfatal injuries count less injuries in LMICs and on the economic costs of these injuries than the deaths. For comparability with other economic esti- are limited. As a result, published data on the costs, cost- mates, and in accordance with the economic analysis guidelines effectiveness, and economic benefits of interventions to prevent provided to authors, we discount estimates of DALYs using unintentional injuries in LMICs are virtually nonexistent. The both a 3 percent and a 6 percent discount rate. The 3 percent economic evaluation of interventions and the measurement of discount rate is standard for economic evaluations in HICs; the economic costs of injuries therefore remain a high research however, a higher discount rate may be appropriate in LMICs. priority. Some data are available from HICs on the costs and, in par- ticular, the net economic benefits of interventions for RTIs. Increased Penalties for Speeding and Other Effective Road Also, a body of evidence suggests that many of the interven- Safety Regulations tions designed to provide safer roads and vehicles, and to Poli de Figueiredo and others (2001) provide evidence from improve driver behavior, have clear net economic benefits Brazil on the effectiveness of an intervention to publicize and Unintentional Injuries | 745 enforce traffic speed and other road safety regulations. This injury incurs US$929 in similar costs (Bishai and Hyder 2004). intervention required three components--legislation to Thus, if we associate 1 prevented traffic fatality with preventing impose stiffer penalties, media coverage of the new regime, and 8 serious injuries worth US$16,128 (8 US$2,016) and 28 better enforcement--and achieved a 25 percent reduction in slight injuries worth US$26,012 (28 US$929), then total traffic fatalities between 1997 and 1998. additional cost savings would amount to US$42,140. On the basis of a model of the costs of media coverage and If intervention costs in Bangladesh were close to the of better police enforcement, we estimate that implementing US$8,105 shown in table 39.4 for South Asia, then the inter- such an intervention in a population of 1 million people might vention would save society more than it cost if it prevented only range from as low US$8,100 in South Asia to US$196,000 in one death and the expected proportion of nonfatal injuries. If LMICs in Europe and Central Asia (table 39.4). Those interven- an enforcement intervention in Bangladesh were as effective tion costs are incremental costs that assume that the population as the one documented in Brazil, it could lower fatalities by already has 50 percent of the necessary police strength. We 25 percent. With 83 traffic fatalities per 1 million population, define adequacy as one officer for every 5,000 vehicles and use the intervention could prevent 21 deaths and lead to net regional data on vehicles per 1 million people to estimate the savings of US$876,835--or (21 US$42,140) US$8,105-- number of police officers and amount of equipment needed to for every million population receiving this intervention. enforce traffic laws. The assumption is that after the interven- tion the population will have enough officers to issue citations to one-third of their beat's 5,000 vehicles each year. This effort Speed Bumps would require them to write about 7 to 10 citations per work- A study in Ghana (Afukaar, Antwi and Ofosu-Amah 2003) day. Using the estimates of traffic injury burdens in the regions showed that road traffic fatalities fell by more than 50 percent listed in the table and the potential to lower traffic deaths by following the introduction of speed bumps. Because speed 25 percent, as reported in the Brazilian study, we estimate bumps are usually most effective when installed at the most potential DALY reductions and cost per DALY averted (table hazardous junctions or near pedestrian crossings, planners 39.4; for details of the calculations see Bishai and Hyder 2004). need to know which junctions are the most hazardous. We The cost estimates shown in table 39.4 do not include assumed that half of a city's crashes occur at junctions and that potential cost offsets from savings derived by preventing cities have different numbers of treatable junctions. A few expenditures on medical care or vehicle repair. Including those junctions would have multiple fatalities per year, but most potential savings would lower the societal cost and enhance the would have one or zero fatalities per year. We assumed that the estimated cost-effectiveness. Those cost offsets will vary widely number of fatalities per junction would be distributed as a by region. To demonstrate the importance of cost offsets, we negative exponential to calculate how many junctions might use data from Bangladesh, for which the Transport Research lack effective speed control modifications and could thereby Laboratory Ltd. (2003) has estimated the medical and property be targeted as those responsible for 10 percent or 25 percent of costs of traffic crashes. On the basis of these estimates, we cal- a city's preventable fatalities. Assuming a 10-year useful life for culate 1 fatality, 8 serious injuries, and 28 slight injuries occur a speed bump and using regionalized estimates of speed bump for every fatal crash in Bangladesh. Each serious injury is asso- construction costs, we modeled the annualized cost of con- ciated with US$2,016 in costs for property, administration, lost structing speed bumps at junctions responsible for 10 percent output, medical care, and pain and suffering, and each slight or 25 percent of fatalities. As before, we lacked an evidence Table 39.4 Costs, DALYs, and Costs per DALY of an Intervention to Improve and Publicize Traffic Enforcement by Region (2001 US$) Present value of annual DALYs averted Cost per DALY averted Cost to intervene in a population of 1,000,000 Discounted at Discounted at Discounted at Discounted at Region for 1 yeara 3 percent per year 6 percent per year 3 percent per year 6 percent per year East Asia and the Pacific 11,817 1,517 956 8 12 Europe and Central Asia 195,971 1,433 903 137 217 Latin American and the Caribbean 225,513 1,333 840 169 268 Middle East and North Africa 114,915 2,166 1,365 53 84 South Asia 8,105 1,528 963 5 8 Sub-Saharan Africa 24,518 2,003 1,370 12 18 Source: Authors' calculations. a. Costs do not include cost offsets from prevented medical care and prevented vehicle repair. 746 | Disease Control Priorities in Developing Countries | Robyn Norton, Adnan A. Hyder, David Bishai, and others Table 39.5 Annualized Costs and DALYs of an Intervention to Build Speed Bumps for the Top 10 Percent of the Most Lethal Junctions in a City of 1 Million, by Region Present value of annual DALYs averted (2001 US$) Cost per DALY averted (2001 US$) Cost to intervene in population of 1,000,000 Discounted at Discounted at Discounted at Discounted at Region for 1 yeara 3 percent per year 6 percent per year 3 percent per year 6 percent per year East Asia and the Pacific 725 167 105 4.34 6.89 Europe and Central Asia 708 158 99 4.48 7.11 Latin American and the Caribbean 299 147 92 2.04 3.23 Middle East and North Africa 1,070 238 150 4.49 7.12 South Asia 324 168 106 1.93 3.06 Sub-Saharan Africa 498 220 151 2.26 3.30 Source: Authors' calculations. a. Annual costs in local currency converted to US$ around 2001. Costs do not include cost offsets from prevented medical care and prevented vehicle repair. base from which to estimate cost offsets from prevented med- 2,478 DALYs at a 3 percent discount rate and 1,562 DALYs ical care or vehicle repair and could not include those poten- at a 6 percent discount rate. Thus, the cost-effectiveness of tial savings. going from zero to 100 percent helmet use in China would be Table 39.5 presents the costs of DALYs saved and costs per US$107 (US$265,000/2,478) per DALY at a 3 percent discount DALY saved by building speed bumps at the top 10 percent of rate or US$170 (US$265,000/1,562) per DALY at a 6 percent the most lethal junctions in a city of 1 million people. We discount rate. assumed the speed bumps could lower injuries by 50 percent, as observed in the Ghana study. Given the low costs per DALY averted and the typical high expenditures for medical care for Motorcycle Helmet Legislation and Enforcement crash victims, we are confident that the medical savings to soci- As with bicycles, we have epidemiological data for China, where ety would more than offset the intervention's costs, but we lack motorcycle-related deaths kill 16 people per 1 million popula- the data to prepare complete estimates. tion per year (Zhang and others 2004). We assume that a pop- ulation of 1 million in China has 125,000 regular motorcyclists, which will require the equivalent of half the time of a police Bicycle Helmet Legislation and Enforcement officer to cite 1 percent of them for helmet violations. At Thompson, Rivara, and Thompson's (1989) case-control Chinese salary levels, this effort would cost the equivalent of study indicates that the effectiveness of a bicycle helmet for a US$7,500 per year. The helmets for this population would cost single rider is 85 percent. The effect on lives saved and DALYs US$250,000 at US$2 per year of helmet use. Thus, the total cost averted depends on how many people in a population ride of the intervention would be US$257,500. Assuming a mean bicycles and the roadway environment for riders. Although age of injury of 20 years and a disability weight of 0.4 for head the degree of energy transferred to the brain in a crash and the injury, we estimate the DALYs averted by motorcycle helmet clinical efficacy of helmets may be the same worldwide, few legislation as 589 at a 3 percent discount rate and 357 at a 6 per- data are available on exposure and the bicycle crash burden in cent discount rate. This intervention therefore costs US$437 LMICs. Thus, in modeling the effects of bicycle helmet legis- (US$257,500/589) per DALY based on a 3 percent discount rate lation, we were limited to assessing the case of the one coun- or US$721 (US$257,500/357) per DALY based on a 6 percent try with adequate data on bicycle injury epidemiology: China. discount rate. In China, bicycle-related deaths kill 22 people per 1 million population per year (Li and Baker 1997). Given estimates of the annualized cost of helmet acquisition for all the bicyclists Childproof Paraffin Containers in a Chinese population of 1 million and of the enforcement The use of childproof paraffin containers is relevant primarily in costs of penalizing unhelmeted riders, we estimate that pro- Sub-Saharan Africa, where households use paraffin as a cooking tecting bicyclists with helmets would cost US$265,000. fuel and frequently store it in bottles previously used to store Assuming that China could convert from zero to 100 percent beverages. Studies from South Africa have significantly adherence to helmet use, it could achieve an 85 percent reduc- enhanced our understanding of the cost-effectiveness of distrib- tion in head injuries from this intervention and would avert uting child-resistant containers. According to Krug and others' Unintentional Injuries | 747 (1994) findings, a population of 1 million who used paraffin IMPLEMENTATION OF PREVENTION regularly in South Africa experienced 1,040 poisonings a year. AND CONTROL STRATEGIES After child-resistant containers were distributed, the incidence dropped to 540, indicating that 500 poisonings per year had Investments in the health sector to address specific problems are been prevented. We therefore assume that (a) in a population of a critical indicator of political commitment, sectoral efforts, 1 million, child-resistant containers would need to be distrib- and priorities at the national and international levels. In some uted to 200,000 households; (b) each child-resistant container cases, investments are so low that they provide a useful reference costs R 0.85 (US$0.33), including the costs of distribution; and point for assessing the returns on additional investments in the (c) total direct costs would be US$66,000 (200,000 US$0.33). future. Such a situation has been described as a null point in The average cost of treating a poisoned child in a South health systems, and current expenditures on injury prevention African hospital was R 256.13 (US$100). Thus, indirect cost and control in LMICs approximate this concept (Murray and savings would be US$50,000 (500 US$100), which would others 2000). partially offset the US$66,000 in direct costs, leading to a net This concept can be illustrated by considering investments cost of US$16,000 to intervene. The mean age of children who in preventing RTIs, which are responsible for the majority of suffered poisoning in South Africa was 12 to 24 months. the burden of unintentional injuries and about which much Although no deaths were reported among children in the is known regarding effective interventions, even though such South African study, the most common figure in the literature interventions have not been examined in the context of rigor- is a 2 percent case fatality rate (Krug and others 1994), suggest- ously controlled studies in LMICs (see box 39.1). ing that the prevention of 500 poisonings averted 10 deaths Peden and others (2004) recognize that, despite the global among children around two years old. Life tables provided to burden of RTIs, the levels of investment are pitifully small, the authors for Sub-Saharan Africa show life expectancy at age largely because of a lack of awareness of the scale of the prob- 2 is 49 years; thus, the US$16,000 intervention could save lem and a lack of awareness that interventions can prevent and 490 life years. Therefore, a rough estimate of the cost-effective- reduce the levels of harm. As a consequence, the report directs ness of child-resistant containers as a way of preventing a number of recommendations at governments and communi- paraffin poisoning in South Africa would be US$1,600 ties in the hope that these recommendations will enable coun- (US$16,000/10) per death averted. tries, particularly LMICs, to begin a sustainable process that Most survivors of paraffin poisoning do not suffer perma- will eventually lead to the adaptation and implementation of nent disability, and because we lack any objective means for effective preventive strategies. The recommendations include assigning disability weights to those who are disabled, we ne- the following: glect years lived with a disability in calculating DALYs. The investment of US$16,000 thus results in 10 children surviving · Identify a lead government agency to guide the national for 49 more years. Undiscounted, this is 490 (10 49) life road safety effort. years. The impact is 263 DALYs averted at a 3 percent discount · Assess the problems, policies, and institutional settings rate or 166 DALYs averted at a 6 percent discount rate. The relating to RTIs and the capacity for preventing RTIs in each cost-effectiveness is US$61 (US$16,000/263) per DALY at a country. 3 percent discount rate or US$96 (US$16,000/166) per DALY · Prepare a national road safety strategy and plan of action. at a 6 percent discount rate. · Allocate financial and human resources to address the problem. · Implement specific actions to prevent crashes, minimize Summary injuries and their consequences, and evaluate the effect of Estimated costs per DALY from the interventions considered those actions. here range from negative (that is, savings) to a few hundred · Support the development of national capacity and interna- U.S. dollars per DALY. The order of magnitude of the costs per tional cooperation. DALY averted using these injury countermeasures suggests that they could be categorized as highly cost-effective (Murray and Although few data are available to show the levels of invest- others 2000). Our estimates of intervention costs neglect the ment in other areas of unintentional injury prevention, those potential savings from prevented medical spending but still levels are no doubt considerably lower than for RTIs. With appear quite promising. Although our estimates provide some increases in the proportions and numbers of older people in indicative information about the economic properties of coun- many LMICs, the burden of fall-related injuries is likely to terinjury interventions, these findings point primarily to the increase significantly in the coming years. Recognition of the lack of information about the global economic burden of injury changing demographics in countries such as China, Mexico, that could enable more comprehensive estimates. and Thailand plus a growing body of evidence on effective 748 | Disease Control Priorities in Developing Countries | Robyn Norton, Adnan A. Hyder, David Bishai, and others Box 39.1 Implementation: Case Study of RTIs Bishai and others (2003) quantify the magnitude of gov- cient investment to conflicts between government min- ernment investment in road safety and the extent of RTIs istries, inefficient civil services, and corruption rather than in Pakistan and Uganda. They estimate that Pakistan spent to a lack of knowledge about possible road crash counter- $0.07 per capita, or 0.015 percent of gross domestic prod- measures (Assum 1998). uct (GDP) per capita, on road safety in 1998 and that RTIs have an inverted U-shaped relationship to eco- Uganda spent $0.09 per capita, or 0.02 percent of GDP per nomic development--injuries rise early during develop- capita. This type of evidence stands in stark contrast with ment, then plateau with investments in road safety, and the high burden of RTIs in these countries. then fall as appropriate interventions succeed (van Beeck, These findings occur in the context of public expendi- Borsboom, and Mackenbach 2000). This biphasic pattern ture on health of 1.8 percent of GDP by Pakistan and is known as a Kuznets curve. Attempts to estimate a 1.6 percent by Uganda (UNDP 1998). Per capita health Kuznets curve for road fatalities suggest that the inflection spending by households in Uganda was $7.70 in 1995/96, point at which fatalities begin to decline occurs at GDP per and public spending on health at the district level was capita in the range of $5,000 to $10,000 (Bishai and others, $4.84 per capita in 1997/98 (Hutchinson 1999). Public forthcoming; Kopits and Cropper 2005). This relationship, spending on road safety amounts to about 1 percent of although based on historical records from HICs, has an public spending on health in each country. It is equivalent important lesson for LMICs: they do not need to experi- to 0.2 percent of military spending in Pakistan and ence massive death and disability from RTIs provided that 1.1 percent of Uganda's military budget. they undertake safety investments now. Waiting for overall A review of road safety initiatives in Benin, Côte economic development before implementing specific d'Ivoire, Kenya, Tanzania, and Zimbabwe found similar interventions will result in the needless loss of millions of underinvestment in road safety and attributes this insuffi- lives. Note: All dollars in box 39.1 are 1998 international dollars. interventions to prevent falls suggest that investments in this less than 10 percent was spent on addressing problems related area could lead to significant benefits. Similarly, increasing to 90 percent of the world's population. Analyses revealed that recognition of the significance of the burden of drowning in RTIs were a highly neglected area for investment compared children is leading to growing awareness of the need to invest with the burden of disease RTIs represented as measured in in that area. However, the absence of any effective evidence- U.S. dollars per DALY. As a result, increasing the level of invest- based interventions may be a barrier to further investment, ment for research and development (R&D) on RTIs and other suggesting that research into the burden of drowning must be injuries should be a focus of global advocacy efforts, and a priority. investment is critical for promoting an R&D agenda on injuries Investment in prevention and control activities in other in LMICs. areas of unintentional injuries is minimal in most LMICs, in Developing and prioritizing a global R&D agenda for unin- large part because the burden of those injuries is unrecognized tentional injuries, though, is challenging, and such an exercise and because evidence of effective interventions is lacking. may be more useful at national or large subnational levels. Therefore, there is clearly a need to consider the development, However, a number of issues requiring R&D are likely to be implementation, and evaluation of prevention strategies in common across a range of LMICs. combination, so that effective interventions can be identified Epidemiological research to describe the existing burden, and promulgated and so that ineffective interventions can be causes, and distribution of unintentional injuries in LMICs identified and discarded. is still needed. Often the data are most limited for areas with the greatest potential burden of injuries. Assessing the loss of health and life from unintentional injuries--whom they affect, RESEARCH AND DEVELOPMENT AGENDA how, and under what specific circumstances--is thus a continuing research agenda for LMICs. Problems of underre- The Global Forum for Health Research (2002) estimates that of porting and other biases in available data also need to be the US$73.5 billion spent on health research globally in 1998, addressed. Unintentional Injuries | 749 The lack of intervention research in LMICs is a huge gap in a major health problem. They cause preventable loss of life and global health research. For the most part, no scientific trials health, and they have major economic implications. As a result, of injury interventions have been conducted in LMICs, and R&D investments are a health and economic imperative for existing and new interventions need to be modified, adapted, developing countries and donor organizations. and tested in those specific settings. Three broad domains should be the foci of intervention research: CONCLUSIONS: PROMISES AND PITFALLS · R&D to enhance the efficiency of currently available effica- Unintentional injuries are an important contributor to global cious interventions. For example, increasing the use of hel- death and disability burdens, especially in LMICs. However, mets would prevent motorcycle injuries in East Asia. the significance of the burden is not matched by substantive · R&D to enhance the cost-effectiveness of interventions that knowledge about risk factors or effective interventions in are currently not being implemented or that could be used LMICs. Nevertheless, the models outlined in this chapter indi- more widely. For example, seat belts and child restraints are cate that several interventions for preventing unintentional known effective interventions, and reducing the cost of such injuries are highly cost-effective and in some cases could result interventions might enhance their wide-scale implementa- in significant cost savings. tion in LMICs. Recent evidence shows that public efforts in injury control, · R&D to develop new interventions for unintentional such as traffic safety, are poorly funded in developing countries injuries and to respond to that proportion of the burden (Bishai and others 2003). The low expenditure compares unfa- that is not currently being addressed. For example, child- vorably with expenditure on other health conditions and with hood falls from rooftops in South Asia have been recognized expenditures in more developed nations, where government as an issue, but a locally derived intervention is currently efforts in relation to such issues as traffic safety are well funded. unavailable. Even adjusting for the 20-fold to 30-fold difference in gross domestic product per capita between HICs and LMICs, the Although some might argue that intervention research investment disparities suggest that LMICs attach a low priority should be the priority in most LMICs, unless the basic under- to injury prevention. pinning research on the burden and determinants of uninten- Given the current low level of investment, initial invest- tional injuries has been undertaken, the political and financial ments in safety, if chosen with care, could turn out to be support for such research will not be forthcoming. extremely beneficial to public health and welfare. If, in the first The dearth of economic and policy analysis of unintention- instance, investments were to be made only in the interventions al injuries in LMICs is an embarrassment for the global health modeled here, then injury reductions would likely be signifi- research community. A recent review of economic analysis of cant. The next step would be to modify other interventions that road traffic interventions found a complete absence of any have proven effective in HICs and to combine the introduction detailed studies from the developing world (Waters, Hyder and of such interventions with evaluations of their effects. Policy Phillips 2004). This gap in health systems research would need makers will be concerned that many of the cost-effective inter- to be addressed to develop and implement successful injury ventions are not low-cost interventions. They save many lives prevention programs. but require an extensive upfront investment. Using cost- Defining a research agenda is necessary but is not by itself effectiveness analyses of these interventions to document high sufficient to conduct research or to implement the results. Two returns would encourage financing of these interventions and key conditions are required for moving forward: a critical mass widespread replication efforts. of people to conduct research and appropriate funds to support Policy makers would be unwise to wait for advanced stages R&D. Developing human resources for all aspects of injury pre- of economic development to attend to the problem of road vention and control in the developing world should be a high safety or other unintentional injuries. Indeed, given the limited priority. Individuals need to be trained and institutions sup- but growing knowledge that low-cost, effective interventions ported and empowered to conduct quality scientific research in exist, for governments not to intervene would be unethical. their own countries and on issues relevant to their locations, Even though institutional obstacles are formidable in develop- which would then be used within their countries. This ing countries, governments routinely overcome them to approach involves paying attention to the issue of strengthen- address other perceived threats to public well-being--such as ing the capacity for research, a major cross-cutting theme for crime, terrorism, and war--that disrupt fewer lives than unin- the health sector in the developing world. tentional injuries. The real enigma is that such a profound loss Funding is and always will be a limiting factor for research; of life can take place each year in LMICs without an outcry that however, the mismatch between the burden of injuries and would trigger sustained and effective political commitment by R&D investments can be corrected. Unintentional injuries are governments and civil society. 750 | Disease Control Priorities in Developing Countries | Robyn Norton, Adnan A. Hyder, David Bishai, and others ACKNOWLEDGMENTS Carlini-Cotrim, B., and A. A. da Matta Chasin. 2000. "Blood Alcohol Content and Death from Fatal Injury: A Study in Metropolitan Area of We would like to acknowledge the assistance of Kristina São Paulo, Brazil." Journal of Psychoactive Drugs 32 (3): 269­75. McDaid and Kylie Monro of the George Institute for Celis, A. 1997. "Home Drowning among Preschool Age Mexican Children." Injury Prevention 3 (4): 252­56. International Health in preparing this chapter. We would also Chan, K. Y., O. Hairol, H. Imtiaz, M. Zailani, S. Kunar, S. Somasundaram, like to acknowledge useful comments on the initial draft of the and others. 2002. "A Review of Burns Patients Admitted to the Burns chapter provided by David Sleet of the Centers for Disease Unit of Hospital Universiti Kebangsaan Malaysia." Medical Journal of Control and Prevention, National Center for Injury Prevention Malaysia 57 (4): 418­25. and Control, and by Tony Bliss of the World Bank. Chang, J. T., S. C. Morton, L. Z. Rubinstein, W. A. Mojica, M. Maglione, M. J. Suttorp, and others. 2004. "Interventions for the Prevention of Falls in Older Adults: Systematic Review and Meta-Analysis of Randomised Clinical Trials." British Medical Journal 328 (7441): 653­54. REFERENCES Chatsantiprapa, K, J. Chokkanapitak, and N. Pinpradit. 2001. "Host and Environment Factors for Exposure to Poisons: A Case Control Study of Preschool Children in Thailand." Injury Prevention 7 (3): Adesunkanmi, A. R., S. A. Oseni, and O. S. Badru. 1999. "Severity and 214­17. Outcome of Falls in Children." West African Journal of Medicine 18 (4): 281­85. Clark P., F. de la Pena, F. Gomez Garcia, J. A. Orozco, and P. Tugwell. 1998. "Risk Factors for Osteoporotic Hip Fractures in Mexicans." Archives of Afukaar, F. K. 2003. "Speed Control in LMICs: Issues, Challenges, and Medical Research 29 (3): 253­57. Opportunities in Reducing Road Traffic Injuries." Injury Control and Safety Promotion 10 (1­2): 77­81. Conrad, P., Y. S. Bradshaw, R. Lamsudin, N. Kasniyah, and C. Costello. 1996. "Helmets, Injuries, and Cultural Definitions: Motorcycle Injury Afukaar, F. K., P. Antwi, and S. Ofosu-Amaah. 2003. "Pattern of Road in Urban Indonesia." Accident Analysis and Prevention 28: 193­200. Traffic Injuries in Ghana: Implications for Control." Injury Control and Safety Promotion 10 (1­2): 69­76. Courtright, P., D. Haile, and E. Kohls. 1993. "The Epidemiology of Burns in Rural Ethiopia." Journal of Epidemiology and Community Health Ahmed, M. K., M. Rahman, and J. van Ginneken. 1999. "Epidemiology of 47 (1): 19­22. Child Deaths Due to Drowning in Matlab, Bangladesh." International Daisy, S., A. K. Mostaque, T. S. Bari, R. R. Kahn, and Q. Quamruzzaman. Journal of Epidemiology 28 (2): 306­11. 2001. "Socioeconomic and Cultural Influence in the Causation of Ahuja, R. B., and S. Bhattacharya. 2002. "An Analysis of 11,196 Burn Burns in the Urban Children of Bangladesh." Journal of Burn Care and Admissions and Evaluations of Conservative Management Techniques." Rehabilitation 22 (4): 269­73. Burns 28 (6): 555­61. Delgado J., M. E. Ramirez-Cardich, R. H. Gilman, R. Lavarello, N. Assum, T. 1998. "Road Safety in Africa: Appraisal of Road Safety Initiatives Dahodwala, A. Bazan, and others. 2002. "Risk Factors for Burns in in Five African Countries." Sub-Saharan Africa Transport Policy Children: Crowding, Poverty, and Poor Maternal Education." Injury Program. Working Paper 33, World Bank, Washington, DC. Prevention 8 (1): 38­41. Attewell, R. G., K. Glase, and M. McFadden. 2001."Bicycle Helmet Efficacy: DiGuiseppi, C., I. Roberts, A. Wade, M. Sculpher, P. Edwards, C. Godward, A Meta-Analysis." Accident Analysis and Prevention 33 (3): 345­52. and others. 2002. "Incidence of Fires and Related Injuries after Giving Azizi, B. H., H. I. Zulkifli, and M. S. Kasim. 1993. "Risk Factors for Out Free Smoke Alarms: Cluster Randomized Controlled Trial." British Accidental Poisoning in Urban Malaysian Children." Annals of Tropical Medical Journal 325 (7371): 995. Paediatrics 13 (2): 183­88. Ellis, J. B., A. Krug, J. Robertson, I. T. Hay, and U. MacIntyre. 1994."Paraffin Bangdiwala, S. I., and E. Anzola-Perez. 1990. "The Incidence of Injuries in Ingestion--The Problem." South African Medical Journal 84 (11): Young People: II. Log-Linear Multivariable Models for Risk Factors in 727­30. a Collaborative Study in Brazil, Chile, Cuba, and Venezuela." Elvik, R., and T. Vaa. 2004. Handbook of Road Safety Measures. Amsterdam: International Journal of Epidemiology 19 (1): 125­32. Elsevier. Bawa Bhalla, S., S. R. Kale, and D. Mohan. 2000. "Burn Properties of European Road Safety Action Program. 2003. Halving the Number of Road Fabrics and Garments Worn in India." Accident Analysis and Prevention Accident Victims in the European Union by 2010: A Shared Responsibility. 32 (3): 407­20. Brussels: European Commission. Bishai, D., and A. Hyder. 2004. "Modeling the Cost Effectiveness of Injury Fernando, R., and D. N. Fernando. 1997. "Childhood Poisoning in Sri Interventions in Lower and Middle Income Countries." Disease Lanka." Indian Journal of Pediatrics 64 (4): 457­60. Control Priorities Working Paper 29, Johns Hopkins Bloomberg Forjuoh, S. N. 2003. "Traffic-Related Injury Prevention Interventions for School of Public Health, Baltimore. Low Income Countries." Injury Control and Safety Promotion 10 (1­2): Bishai, D., A. A. Hyder, A. Ghaffar, R. H. Morrow, and O. Kobusingye. 2003. 109­18. "Rates of Public Investment for Road Safety in Developing Countries: . 2004. "Preventing Burns in Low and Middle-Income Countries." Case Studies of Uganda and Pakistan." Health Policy and Planning Disease Control Priorities Working Paper. 18 (2): 232­35. Forjuoh S. N., B. Guyer, and G. S. Smith. 1995. "Childhood Burns in Bishai, D., A. Qureshi, P. James, and A. Ghaffar. Forthcoming. "National Ghana: Epidemiological Characteristics and Home-Based Treatment." Road Fatalities and Economic Development." Health Economics. Burns 21 (1): 24­28. Boonyaratavej, N., P. Suriyawongpaisal, A. Takkinsatien, S. Wanvarie, R. Forjuoh S. N., B. Guyer, D. M. Strobino, P. M. Keyl, M. Diener-West, and Rajatanavin, and P. Apiyasawat. 2001. "Physical Activity and Risk G. S. Smith. 1995. "Risk Factors for Childhood Burns: A Case-Control Factors for Hip Fractures in Thai Women." Osteoporosis International Study of Ghanaian Children." Journal of Epidemiology and Community 12 (3): 244­48. Health 49 (2): 189­93. Bunn, F., T. Collier, C. Frost, K. Ker, I. Roberts, and R. Wentz. 2003. "Area- Ghaffar, A., A. A. Hyder, M. I. Mastoor, and I. Shaikh. 1999. "Injuries in Wide Traffic Calming for Preventing Traffic-Related Injuries." Pakistan: Directions for Future Health Policy." Health Policy and Cochrane Database of Systematic Reviews (1) CD003110. Planning 14 (1): 11­17. Unintentional Injuries | 751 Global Forum for Health Research. 2002. The 10/90 Report on Liu, B., R. Ivers, R. Norton, S. Blows, and S. K. Lo. 2004. "Helmets for Health Research 2001­2002. Geneva: Global Forum for Health Preventing Injury in Motorcycle Riders." Cochrane Database of Research. Systematic Reviews (4) CD004333. Haddon, W. Jr. 1968. "The Changing Approach to the Epidemiology, Liu, E. H., B. Khatri, Y. M. Shakya, and B. M. Richard. 1998. "A 3 Year Prevention, and Amelioration of Trauma: The Transition to Prospective Audit of Burns Patients Treated at the Western Regional Approaches Etiologically Rather Than Descriptively Based." American Hospital of Nepal." Burns 24 (2): 129­33. Journal of Public Health and the Nation's Health 58 (8): 1431­38. Liu, G. F., S. Han, D. H. Liang, F. Z. Wang, X. Z. Shi, J. Yu, and others. 2003. Harvey, P. A, M. Aitken, G. W. Ryan, L. A. Demeter, J. Givens, R. "Driver Sleepiness and Risk of Car Crashes in Shenyang, a Chinese Sundararaman, and others 2004. "Strategies to Increase Smoke Alarm Northeastern City: Population-Based Case-Control Study." Biomedical Use in High-Risk Households." Journal of Community Health 29 (5): and Environmental Sciences 16 (3): 219­26. 375­85. Macarthur, C. 2003. "Evaluation of Safe Kids Week 2001: Prevention of Hutchinson, P. 1999. "Health Care in Uganda." Discussion Paper 404, Scald and Burn Injuries in Young Children." Injury Prevention 9 (2): World Bank, Washington, DC. 112­16. Hyder, A. A., S. Arifeen, N. Begum, S. Fishman, S. Wali, and A. H. Baqui. Mock, C., J. Amegashi, and K. Darteh. 1999. "Role of Commercial Drivers 2003. "Death from Drowning: Defining a New Challenge for Child in Motor Vehicle Related Injuries in Ghana." Injury Prevention 5 (4): Survival in Bangladesh." Injury Control and Safety Promotion 10 (4): 268­71. 205­10. Mohan, D. 2002. "Road Safety in Less-Motorized Environments: Future ICF Consulting Ltd. and Imperial College Centre for Transport Studies. Concerns." International Journal of Epidemiology 31 (3): 527­32. 2003. "Cost-Benefit Analysis of Road Safety Improvements." Final Murray, C. J. L., D. B. Evans, A. Acharya, and B. Baltussen. 2000. report. London: ICF Consulting Ltd. and Imperial College Centre for "Development of WHO Guidelines on Generalized Cost-Effectiveness Transport Studies. Analysis." Health Economics 9 (3): 235­51. Ichikawa, M., W. Chadbunchachai, and E. Marui. 2003. "Effect of the Nafukho, F. M., and M. Khayesi. 2002. "Livelihood, Conditions of Work, Helmet Act for Motorcyclists in Thailand." Accident Analysis and Regulation, and Road Safety in the Small-Scale Public Transport Prevention 35 (2): 83­89. Sector: A Case of the Matatu Mode of Transport in Kenya." In Urban Jacobs, G., A. Aeron-Thomas, and A. Astrop. 2000. "Estimating Global Mobility for All: Proceedings of the Tenth International CODATU Road Fatalities." TRL Report 445, Transport Research Laboratory, Conference, Lomé, Togo, 12­15 November 2002, ed. X. Godard and I. Crowthorne, U.K. Fatonzoun, 241­45. Lisse, the Netherlands: AA Balkema. Jie, X., and C. B. Ren. 1992. "Burn Injuries in the Dong Bei Area of China: Nhachi, C. F., and O. M. Kasilo. 1992. "The Pattern of Poisoning in Urban A Study of 12,606 Cases." Burns 18 (3): 228­32. Zimbabwe." Journal of Applied Toxicology 12 (6): 435­38. Jitapunkul, S., P. Yuktananandana, and V. Parkpian. 2001. "Risk Factors of Nhachi, C. F., and O. M. Kasilo. 1994. "Household Chemicals Poisoning Hip Fracture among Thai Female Patients." Journal of the Medical Admissions in Zimbabwe's Main Urban Centres." Human and Association of Thailand 84 (11): 1576­81. Experimental Toxicology 13 (2): 69­72. Khayesi, M. 2003. "Liveable Streets for Pedestrians in Nairobi: The Norghani, M., A. Zainuddin, R. S. Radin Umar, and H. Hussain. 1998. Use Challenge of Road Traffic Accidents." In The Earthscan Reader on of Exposure Control Methods to Tackle Motorcycle Accidents in Malaysia. World Transport Policy and Practice, ed. J. Whitelegg and G. Haq, 35­41. Research Report 3/98. Serdang, Malaysia: Road Safety Research Center, London: Earthscan. University Putra Malaysia. Kibel, S. M., F. O. Nagel, J. Myers, and S. Cywes. 1990. "Childhood Odero, W., P. Garner, and A. B. Zwi. 1997. "Road Traffic Injuries in Near-Drowning--A 12-Year Retrospective Review." South African Developing Countries: A Comprehensive Review of Epidemiological Medical Journal 78 (7): 418­21. Studies." Tropical Medicine and International Health 2 (5): 445­60. Kloeden, C. N., A. J. McLean, M. R. J. Baldock, and A. J. T. Cockington. Odero W., M. Khayesi, and P. M. Heda. 2003. "Road Traffic Injuries in 1998. "Severe and Fatal Car Crashes Due to Roadside Hazards: A Kenya: Magnitude, Causes, and Status of Intervention." Injury Control Report to the Motor Accident Commission." Adelaide, Australia: and Safety Promotion 10 (1­2): 53­61. National Health and Medical Research Council Road Accident Odero, W. O., and A. B. Zwi. 1995. "Alcohol-Related Traffic Injuries and Research Unit, University of Adelaide. Fatalities in LMICs: A Critical Review of Literature." In Proceedings of Kobusingye, O., D. Guwatudde, and R. Lett. 2001."Injury Patterns in Rural the 13th International Conference on Alcohol, Drugs, and Traffic Safety, and Urban Uganda." Injury Prevention 7 (1): 46­50. Adelaide, 13­18 August 1995, ed. C. N. Kloeden and A. J. McLean, Kopits, E., and M. Cropper. 2005. "Traffic Fatalities and Economic 713­20. Adelaide, Australia: Road Accident Research Unit. Growth." Accident Analysis and Prevention 37 (1): 169­78. Peden, M., D. Knottenbelt, J. Van der Spuy, R. Oodit, M. J. Scholtz, and Kozik, C. A., S. Suntayakorn, D. W. Vaughn, C. Suntayakorn, R. Snitbhan, J. M. Stokol. 1996. "Injured Pedestrians in Cape Town: The Role of and B. L. Innis. 1999. "Causes of Death and Unintentional Injury Alcohol." South African Medical Journal 86 (9): 1103­5. among Schoolchildren in Thailand." Southeast Asian Journal of Tropical Peden, M., R. Scurfield, D. Sleet, D. Mohan, A. A. Hyder, E. Jarawan, and Medicine and Public Health 30 (1): 129­35. others, eds. 2004. World Report on Road Traffic Injury Prevention. Krug, A., J. B. Ellis, I. T. Hay, N. F. Mokgabudi, and J. Robertson. 1994."The Geneva: World Health Organization. Impact of Child-Resistant Containers on the Incidence of Paraffin Poli de Figueiredo, L. F., S. Rasslan, V. Bruscagin, R. Cruz, and M. Rocha e (Kerosene) Ingestion in Children." South African Medical Journal 84 Silva. 2001. "Increases in Fines and Driver Licence Withdrawal (11): 730­34. Have Effectively Reduced Immediate Deaths from Trauma on Kulanthayan, S., R. S. Umar, H. A. Hariza, M. T. Nasir, and S. Harwant. Brazilian Roads: First-Year Report on the New Traffic Code." Injury 32 2000. "Compliance of Proper Safety Helmet Usage in Motorcyclists." (2): 91­94. Medical Journal of Malaysia 55 (1): 40­44. Radin Umar, R. S., G. M. Mackay, and B. L. Hills. 1996. "Modelling of Li, G., and S. P. Baker. 1997. "Injuries to Bicyclists in Wuhan, People's Conspicuity-Related Motorcycle Accidents in Seremban and Shah Republic of China." American Journal of Public Health 87 (6): 1049­52. Alam, Malaysia." Accident Analysis and Prevention 28 (3): 325­32. 752 | Disease Control Priorities in Developing Countries | Robyn Norton, Adnan A. Hyder, David Bishai, and others Raja, I. A., A. H. Vohra, and M. Ahmed. 2001. "Neurotrauma in Pakistan." UNDP (United Nations Development Programme). 1998. Human World Journal of Surgery 25 (9): 1230­37. Development Report 1998. New York: Oxford University Press. Ross, A., C. Baguley, V. Hills, M. McDonald, and D. Silcock. 1991. Towards van Beeck, E. F., G. J. Borsboom, and J. P. Mackenbach. 2000. "Economic Safer Roads in Developing Countries: A Guide for Planners and Development and Traffic Accident Mortality in the Industrialized Engineers. Crowthorne, U.K.: Transport Research Laboratory. World, 1962­1990." International Journal of Epidemiology 29 (3): Rossi, L. A., E. C. Braga, R. C. Barruffini, and E. C. Carvalho. 1998. 503­9. "Childhood Burn Injuries: Circumstances of Occurrences and Their Wang, S., G. B. Chi, C. X. Jing, X. M. Dong, C. P. Wu, and L. P. Li. 2003. Prevention in Ribeirão Preto, Brazil." Burns 24 (5): 416­19. "Trends in Road Traffic Crashes and Associated Injury and Fatality in Soori, H. 2001. "Developmental Risk Factors for Unintentional Childhood the People's Republic of China, 1951­1999." Injury Control and Safety Poisoning." Saudi Medical Journal 22 (3): 227­30. Promotion 10 (1­2): 83­87. Spiegel, C. N., and F. C. Lindaman. 1977 "Children Can't Fly: A Program Waters, H. R., A. A. Hyder, and T. L. Phillips. 2004. "Economic Evaluation to Prevent Childhood Morbidity and Mortality from Window Falls." of Interventions for Reducing Road Traffic Injuries--A Review of American Journal of Public Health 67 (12): 1143­47. Literature with Applications to Low and Middle-Income Countries." Asia Pacific Journal of Public Health 16 (1): 23­31. Supramaniam, V., V. Belle, and J. Sung. 1984. "Fatal Motorcycle Accidents and Helmet Laws in Peninsular Malaysia." Accident Analysis and Werneck, G. L., and M. E. Reichenheim. 1997. "Paediatric Burns and Prevention 16 (3): 157­62. Associated Risk Factors in Rio de Janeiro, Brazil." Burns 23 (6): 478­83. Tan, Z., X. Li, and Q. Bu. 1998. "Epidemiological Study on Drowning in Wujin, Jiangsu, 1997." Zhonghua Liu Xing Bing Xue Za Zhi 19 (4): Ytterstad, B., and A. J. Sogaard: 1995. "The Harstad Injury Prevention 208­10. Study: Prevention of Burns in Small Children by a Community-Based Intervention." Burns 21 (4): 259­66. Thompson, D. C., and F. P. Rivara. 2000. "Pool Fencing for Preventing Drowning in Children." Cochrane Database of Systematic Reviews (2) Yuan, W. 2000. "The Effectiveness of the `Ride Bright' Legislation for CD001047. Motorcycles in Singapore." Accident Analysis and Prevention 32 (4): 559­63. Thompson, R. S., F. P. Rivara, and D. C. Thompson. 1989. "A Case-Control Study of the Effectiveness of Bicycle Safety Helmets." New England Zhang, J., R. Norton, K. C. Tang, S. K. Lo, J. Zhuo, and W. Geng. 2004. Journal of Medicine 320 (21): 1361­67. "Motorcycle Ownership and Injury in China." Injury Control and Safety Promotion 11 (3): 159­63. Tiwari, G. 2000. "Traffic Flow and Safety: Need for New Models for Heterogeneous Traffic." In Injury Prevention and Control, ed. D. Mohan Zhu, Z. X., H. Yang, and F. Z. Meng. 1988. "The Epidemiology of and G. Tiwari, 71­88. London: Taylor and Francis. Childhood Burns in Jiamusi, China." Burns, Including Thermal Injury 14 (5): 394­96. Transport Research Laboratory Ltd. 2003. Guidelines for Estimating the Cost of Road Crashes in Developing Countries. Department for International Development, Project R7780. London: Babtie, Ross, Silcock. Unintentional Injuries | 753 Chapter 40 Interpersonal Violence Mark L. Rosenberg, Alexander Butchart, James Mercy, Vasant Narasimhan, Hugh Waters, and Maureen S. Marshall Violence kills more than 1.6 million people each year. The that more than 90 percent of all violence-related deaths occur in impact of nonfatal violence cannot be quantified, but it is even LMIC countries (Dahlberg and Krug 2002). The estimated rate more pernicious given resultant disabilities and long-term of violent death in LMICs was 32.1 per 100,000 people in 2000, physical, psychological, economic, and social consequences. compared with 14.4 per 100,000 in high-income countries. The direct and indirect costs of violence are enormous. This chapter is based on a public health approach to pre- Violence directly affects health care expenditures worldwide. venting interpersonal violence. A public health approach has Indirectly, violence has a negative effect on national and local three overriding characteristics: it applies scientific methodol- economies--stunting economic development, increasing eco- ogy, emphasizes prevention, and encourages collaboration. nomic inequality, eroding human and social capital, and Applying a scientific methodology to a public health appro- increasing law enforcement expenditures (Waters and others ach involves collecting and analyzing data to define the magni- 2004). tude, scope, and characteristics of the problem, examining the The U.S.-based Centers for Disease Control and Prevention factors that increase or decrease the risk for violence, and iden- identified violence as a leading public health problem in the tifying the factors that can be modified through interventions. mid 1980s and early 1990s (Rosenberg 1985; Rosenberg and Interventions are designed, tested, and evaluated. Efficacious Fenley 1991), as did the World Health Assembly in 1996 and promising interventions are implemented, and their effects (Resolution WHA49.25). Contributing to the World Health and cost-effectiveness are evaluated. Ongoing monitoring of Organization (WHO) report on global violence and health, intervention effects on risk factors and target problems builds Dahlberg and Krug (2002) divided violence into the following the database to allow quantitative assessment of successes and categories: clear identification of remaining needs. Fundamentally, public health is focused on prevention of · self-directed violence, or violence in which the perpetrator harm caused by disease or violence. Although criminal justice is the victim (for example, suicide) systems have traditionally focused on capturing perpetrators · interpersonal violence, or violence inflicted by another indi- of violence and punishing them for their actions (typically vidual or a small group of individuals through incarceration), the public health system attempts to · collective violence, or violence committed by larger groups, prevent violence from occurring and concentrates on identify- such as states, organized political groups, militia groups, and ing ways to keep people from committing acts of violence. terrorist organizations. Interventions may eliminate or reduce the underlying risk fac- tors and shore up protective factors. Prevention strategies are This chapter focuses on interpersonal violence, which dis- conceived and implemented with reference to the interaction proportionately affects low- and middle-income countries of risk factors among people at different stages of the life cycle (LMICs).1 The WHO report on violence and health estimates (Mercy and Hammond 1999; additional sources online). 755 A public health approach must be collaborative, drawing tional determinants, and nature and severity of resultant on contributions from different sectors and disciplines. Public injuries and other harm. Data sources include death certifi- health analyses of violence aim to encourage integrated actions cates, vital statistics records, medical examiners' reports, hospi- by diverse sectors such as health, education, social services, and tal and other medical records, police and judiciary records, and justice. Each sector has a role to play, and collectively their self-reported information from victim surveys and special actions have the potential to reduce violence. studies. Multiple data sources, with their inherent strengths and limitations, are essential. THE NATURE, BURDEN, AND CAUSES The most widely encountered sources of information are OF INTERPERSONAL VIOLENCE from the health and criminal justice sectors. Reliable data on violent deaths are not routinely collected in most countries. WHO (WHO Global Consultation on Violence and Health Where data collection systems are in place, coroner and mortu- 1996, 2­3) defines violence as follows: "The intentional use of ary reports, death certificates, and vital statistics records usually physical force or power, threatened or actual, against oneself, provide additional data about the victim. The health sector typ- another person, or against a group or community that either ically documents characteristics of the decedent and the cause, results in or has a high likelihood of resulting in injury, death, location, circumstances, and time of death. The criminal justice psychological harm, mal-development, or deprivation." This sector documents deaths or arrests resulting from interpersonal definition emphasizes that, for the act to be classified as vio- violence, including sometimes recording information about the lence, a person or group must intend to use force or power relationship between the victim and the offender, the circum- against another person. Thus, violence is distinguished from stances surrounding the violence, and the demographics of the unintended incidents that result in injury or harm. perpetrator. The nature or mode of violence may be physical, sexual, or Theoretically, health and criminal justice sector data include psychological, or it may involve deprivation and neglect. Given information about nonfatal violence at all levels of severity, the difficulties of measuring deprivation and neglect,this chapter including threats of violence and instances of psychological vio- concentrates on the physical, sexual, and psychological modes. lence, deprivation, and neglect. In practice, however, only data Acts of interpersonal violence are classified as family vio- about violence-related injuries presenting at hospital emergency lence or community violence. Family violence is further cate- departments are collected. Studies from a variety of coun- gorized by victim: child, intimate partner, or elder. Child abuse, tries show that for every victim reporting violence to the as defined by WHO (1999, 15), is "physical and/or emotional police, at least two more present only at health agencies ill-treatment, sexual abuse, neglect or negligent treatment or (Houry and others 1999; Kruger and others 1998; Sutherland, commercial or other exploitation, resulting in actual or poten- Sivarajasingam, and Shepherd 2002; additional sources online). tial harm to the child's health, survival, development, or dignity Victims of nonfatal violence treated by the health sector may in the context of a relationship of responsibility, trust, or provide information about the perpetrator-victim relationship, power." Behavior within an intimate relationship that causes about the circumstances surrounding the attack, and about physical, psychological, or sexual harm is typically labeled inti- contextual and developmental risk factors. However, the health mate partner violence or domestic violence. Elder abuse is mis- sector is frequently restricted in recording information about treatment of older people, generally those older than age 60 or perpetrators. 65, in the home or in an institutional setting. In LMICs, population-based surveys are a more useful Community violence is categorized by two types of perpe- source of information about violence-related injuries at all trators: acquaintances and strangers. It includes sexual assault severity levels (Sethi, Habibula, and others 2004). Such surveys by strangers and violence in institutional environments, such as have been conducted in Bangladesh (Rahman, Andersson, and residential care facilities, jails, workplaces, and schools. Youth Svanstrom 1998); Colombia (Duque, Klevens, and Ramirez violence, with perpetrators and victims typically 10 to 29 years 2003); Iraq (Roberts and others 2004); Pakistan (Ghaffar 2001); of age, is also a form of community violence. South Africa (Butchart, Kruger, and Lekoba 2000; additional sources online); and Uganda (Kobusingye, Guwatudde, and Outcomes of Interpersonal Violence Lett 2001). Demographic and health surveys with questions Identifying the outcomes of interpersonal violence helps to about violent victimization also collect information about the determine the magnitude of the problem. relationship between violence and other health conditions, but they can provide only limited insight into the perpetrators. Data. As noted earlier, a fundamental aspect of the public Hospital emergency departments have been used in some health approach is the collection of accurate information, such postconflict settings to monitor weapons-related injuries as demographic characteristics of victims and perpetrators, and evaluate the relative contributions of collective and weapon involvement, settings in which violence occurs, situa- interpersonal violence to the caseload (Meddings and 756 | Disease Control Priorities in Developing Countries | Mark L. Rosenberg, Alexander Butchart, James Mercy, and others O'Connor 1999; Michael and others 1999). Some developing Table 40.1 Estimated Violence-Related Deaths, by Type countries, such as Bangladesh, Kenya, and Uganda, also use and Region, 2001 violence and injury surveillance systems based in health facili- Rate Proportion ties to monitor hospitalizations resulting from violence and per 100,000 of total other causes of injury (Kobusingye and Lett 2000; Odero and Category Numbera populationb (percent) Kibosia 1995; Rahman and others 2001). Where emergency Suicide 875,000 15.2 53.3 and forensic medical services are reasonably well developed Homicide 557,000 9.3 34.0 and where access to such services is equitable, violence and War-related fatality 208,000 3.5 12.7 injury surveillance tools have been integrated into hospital emergency departments (Hasbrouck and others 2002; addi- Total 1,640,000 28.0 100.0 tional sources online), prenatal clinics (Dunkle and others LMICs 1,489,000 31.0 90.8 2004), forensic service centers for rape victims (Swart and oth- High-income countries 150,000 14.3 9.2 ers 2000), and mortuaries (Butchart and others 2001). Those Source: Mathers and others 2006. efforts have proven effective in obtaining victim-based, a. Rounded to the nearest thousand. descriptive epidemiological information and insights into the b. Age standardized. relationships between victims and perpetrators. Homicide rates differ markedly by age and sex (table 40.2). Deaths Resulting from Interpersonal Violence. Global bur- Gender differences were least marked for children. For the 15 to den of disease estimates indicate that, in 2001, approximately 29 age group, male rates were nearly six times those for female 1.6 million people died as a result of violence. Of those deaths, rates; for the remaining age groups, male rates were from two 34 percent were due to interpersonal violence (table 40.1). to four times those for females. Female homicide rates doubled Rates and patterns of violent death vary by country and after age 14 and gradually but steadily increased with age, and region (figure 40.1). Homicide rates were highest in developing male rates increased more than 14 times after age 14, peaked in countries in Sub-Saharan Africa and Latin America and the the 15 to 29 age group, and then gradually decreased with age. Caribbean and lowest in East Asia, the western Pacific, and Overall,homicides resulted in the deaths of 3.4 males per female. some countries in northern Africa. Studies show a strong, inverse relationship between homicide rates and both eco- Violence-Related Burden of Disease. The sum of years of nomic development and economic equality (Butchart and potential life lost because of premature mortality and years of Engstrom 2002; Fajnzylber, Lederman, and Loayza 2000). productive life lost because of disability is not a particularly Poorer countries, especially those with large gaps between the useful measure of the burden of violence. Disability-adjusted rich and the poor, tend to have higher rates of homicide than life years rely, in part, on estimates of nonfatal events. In the wealthier countries. case of violence, those estimates are restricted to injuries and Incidence per 100,000 population 30 Homicide Suicide War 25 20 15 10 5 0 East Asia and Europe and Latin America and Middle East and South Asia Sub-Saharan the Pacific Central Asia the Caribbean North Africa Africa Source: Mathers and others 2006. Figure 40.1 Homicide, Suicide, and War-Related Fatality Rates, by Region, 2001 Interpersonal Violence | 757 Table 40.2 Estimated Global Homicide and Suicide Rates, Intimate partner violence results in an increased incidence by Age Group, 2001 of suicide and suicide attempts, as well as in depression, anxi- (number per 100,000 population) ety, and phobias (Heise and Garcia-Moreno 2002). Additional consequences include substance abuse, eating and sleep disor- Homicides Suicides ders, poor self-esteem, posttraumatic stress disorder, psycho- Age Males Females Males Females somatic disorders, and risky sexual behaviors. Sexual assault 0­4 years 2.1 2.0 0.0 0.0 results in consequences that can be long lasting and severe, 5­14 years 1.6 1.5 1.4 1.1 including posttraumatic stress disorder, depression, and con- 15­29 years 23.1 3.9 18.9 13.2 duct disorders, as well as sleep and eating disorders (Jewkes, 30­44 years 20.9 4.7 22.9 13.0 Sen, and Garcia-Moreno 2002). 45­59 years 16.5 5.0 29.0 15.8 According to Jewkes, Sen, and Garcia-Moreno (2002), among adolescents and women age 12 to 45, the frequency of 60 years 12.6 5.4 41.7 20.8 pregnancy as a result of rape varies from 5 to 18 percent. In Total 14.3 3.7 17.7 10.7 addition, younger rape victims often have an increased rate of Source: Mathers and others 2006. later, unintended pregnancies. Rape frequently results in gyne- cological problems, problems of sexual functioning, and sexu- physical disabilities, both markedly underreported. In addition, ally transmitted diseases, including HIV infection. HIV infec- given that psychological and other noninjury health conse- tion and the stigma it carries put both female and male victims quences of violence are substantial, failure to include them in of sexual assault at increased risk of further violence. A similar the measurement of disability-adjusted life years means that range of reproductive health consequences may also follow estimates of the nonfatal burden of violence may be grossly intimate partner violence. underestimated. A strong, graded relationship exists between the breadth of Violence-related morbidity can be analyzed as four distinct, exposure to abuse or household dysfunction during childhood but often co-occurring, outcome clusters: injuries and disabili- and the presence of adult diseases, including ischemic heart ties, mental health and behavioral consequences, reproductive disease, cancer, chronic lung disease, skeletal fractures, and liver health consequences, and other health consequences. disease (Felitti and others 1998). In developed countries, abuse Studies in a number of countries show that, for every homi- and other violent events of childhood have been associated cide among young people age 10 to 24, 20 to 40 other young with a 4- to 12-fold increased risk for alcoholism, drug abuse, people receive hospital treatment for a violent injury (Mercy depression, and suicide attempt; a 2- to 4-fold increased risk for and others 2002). Injuries range from minor, which can be self- smoking, poor self-rated health, 50 or more sexual intercourse treated, to severe. Severe injuries are those that may require partners, and sexually transmitted disease; and a 1.4- to 1.6- resource-intensive emergency medical treatment and inpatient fold increased risk for physical inactivity and severe obesity care and may result in lifelong disabilities, such as amputations, (Anda and others 1999; Dietz and others 1999; Dube and brain damage, or paraplegia. Few countries have information others 2001, 2002; Hillis and others 2000, 2001; Williamson systems for monitoring nonfatal violent injuries, and existing and others 2002). Similar exposures to violence in developing systems typically record only data on violent injuries present- countries may have different, yet equally wide-ranging, impacts ing at hospital emergency departments. Data from those sites beyond direct physical and psychological injuries. cannot be directly compared, given the marked differences between and within countries in the availability and accessibil- Data on Violence in Developing Countries ity of emergency medical services. Studies documenting the human and economic toll of violence The mental health consequences of violence are far reach- in LMICs are strikingly scarce. In addition to disparate levels ing. Child abuse has well-documented sequelae of psychiatric of economic development, other differences between countries disorders and suicidal behaviors (Runyan and others 2002). strongly influence levels and patterns of interpersonal violence Both short- and long-term sequelae have been demonstrated and the toll that such violence takes on society. Countries with (Mercy and others 2002, Heise and Garcia-Moreno 2002), weak governments and institutions are at considerably higher including depression, anxiety disorders, substance abuse disor- risk for interpersonal violence than countries with developed ders, aggression, cognitive problems, sleep disorders, and post- institutions, and countries at war are likewise at higher risk than traumatic stress disorder. The severity and duration of those countries at peace. The same factors that lead to high levels of consequences vary with the child's age and the length of time interpersonal violence--lack of economic development; weak the child suffers the abuse, as well as the duration and intensi- social, political, and judicial institutions; social disturbances; ty of the abuse, the child's relationship to the abuser, and the and warfare--also adversely affect nations' ability to collect data treatment received (Runyan and others 2002). and to address the causes or consequences of this violence. 758 | Disease Control Priorities in Developing Countries | Mark L. Rosenberg, Alexander Butchart, James Mercy, and others Table 40.3 Risk Factors for Becoming a Victim or Perpetrator of Violence Level of the ecological model Risk factors Individual Early developmental experience (biological and personal history factors Demographic characteristics (for example, age, education, family, or personal income) that influence how individuals behave) Victim of child abuse and neglect Psychological and personality disorders Physical health and disabilities Alcohol or substance abuse problems History of violent behavior Youth Male Gun ownership Relationship Marital conflicts around gender roles and resources (with family members, friends, intimate Association with friends who engage in violent or delinquent behavior partners, peers) Poor parenting practices Parental conflict involving use of violence Low socioeconomic status of household Community High residential mobility (neighborhoods, schools, workplaces) High unemployment High population density Social isolation Proximity to drug trade Inadequate victim care services Poverty Weak policies and programs in, for example, workplaces, schools, residential care facilities Societal Rapid social change (broad factors that reduce inhibitions Economic inequality against violence) Gender inequality Policies that create and sustain or increase economic and social inequalities Norms that give priority to parents' rights over child welfare Norms that entrench male dominance over women Poverty Weak economic safety nets Poor rule of law Poor criminal justice system that supports the use of excessive violence by police officers against citizens and leaves perpetrators immune from prosecution Social or cultural norms that support violence Availability of means (for example, firearms) Conflict or postconflict situation Source: Krug and others 2002a. Risk Factors for Understanding Violence in an ecological model that classifies risk factors for violence by Risk factors for violence are conditions that increase the pos- four levels: individual, relationship, community, and societal sibility of becoming a victim or perpetrator of violence. No (Dahlberg and Krug 2002). Although some risk factors may be single factor explains why a person or group is at a high or low unique to a particular type of violence, the various types of risk of violence. Rather, violence is an outcome of a complex violence more commonly share a number of risk factors interaction among many factors. This relationship is captured (table 40.3). Interpersonal Violence | 759 ECONOMIC IMPACT OF VIOLENCE views with crime victims, to estimate the incidence and calcu- late the direct costs of sexual assault. For example, Miller, Violence exacts an extraordinary economic toll. Cohen, and Rossman (1993) calculate average psychological costs of US$66,600 for each rape and total costs of US$85,000 Costs of Violence for sexual assault resulting in physical injury. Psychological Estimates of the costs of violence vary broadly, with many of the costs, also referred to as "pain and suffering," are considered differences resulting from the inclusion or exclusion of different indirect costs. Because many studies do not include those types categories. Cost categories can be broadly grouped into direct of costs, cost estimates vary widely. costs, which result directly from acts of violence or attempts to Violence at the workplace also extracts an economic toll, but prevent them, and indirect costs, which include the opportunity studies of its magnitude are not well developed and are ham- cost of time, lost productivity, and impaired quality of life. pered by measurement difficulties and nonstandardized Those and other methodological issues lead to differing methodologies. Biddle and Hartley (2002) study homicides in estimates of the costs of violence.2 Researchers have calculated the workplace in the United States and calculate an annual cost the value of a human life using lost wages, estimates of the of approximately US$970 million. An international report quality of life, wage premiums for risky jobs, willingness to pay commissioned by the International Labour Organization on for safety measures, and individual behavior related to safety the costs of violence and stress in work environments estimates measures. The value of human life used in U.S. studies ranges that losses from stress and violence at work are equivalent to from US$3.1 million to US$6.8 million (Fisher, Chestnut, and 1.0 to 3.5 percent of GDP over a range of countries (Hoel, Violette 1989; Viscusi 1993; additional sources online). The rate Sparks, and Cooper 2001). All those studies use a broad defini- used to discount future costs and benefits also varies, generally tion of workplace violence, including psychological violence from 2 to 10 percent.3 such as sexual harassment and bullying. Fromm (2001) reviews a variety of sources and calculates an Violence committed by juveniles is particularly costly to aggregate total of US$94 billion in annual costs to the U.S. society. Miller's (2001) analysis of violent crimes committed in economy resulting from child abuse, which is equal to 1 percent Pennsylvania in 1993 finds that juvenile violence accounted of gross domestic product (GDP). The estimate includes direct for 24.7 percent of all violent crimes and 46.6 percent of total medical costs and related costs of legal services, policing, and victim costs from violent crime. Cohen (1988) calculates that incarceration, as well as the value of indirect productivity the total cost to society of a youth engaging in a life of crime losses, psychological costs, and future criminality. Using sec- ranges from US$1.9 million to US$2.6 million. ondary sources, Courtney (1999) calculates direct costs of US$14 billion, including counseling and child welfare services, Proximate Risk Factors resulting from child abuse in the United States. Alcohol, drugs, and guns contribute to the costs of interper- The Centers for Disease Control and Prevention (CDC sonal violence. According to estimates by the Children's Safety 2003) cite an estimated 5.3 million victimizations involving Network Economics and Insurance Resource Center (1997), intimate partner violence each year in the United States among the cost of violent crime committed under the influence of women 18 and older, resulting in nearly 2 million injuries. alcohol equaled US$33.3 million in 1995, or 8.3 percent of the More than 550,000 of those injuries require medical attention. cost of all violent crime in the United States. The National The costs of intimate partner violence, including medical care, Crime Prevention Council (1999) estimates that the cost of all mental health care, and lost productivity, exceed US$5.8 billion drug-related crime, including productivity costs, amounts to annually. US$60 billion to US$100 billion annually in the United States, As a percentage of GDP, estimates of the costs of intimate with violent crime accounting for approximately 10 percent of partner violence are considerably higher in LMICs than in this figure. high-income countries. Morrison and Orlando (1999) calcu- Cook and Ludwig (2000) estimate that the annual costs of late the costs of domestic violence against women on the basis gun violence in the United States are on the order of US$100 bil- of stratified random samples of women. Using only the lost lion. Miller and Cohen (1997) calculate a significantly higher productive capacity of the women, they extrapolate total costs estimate for the toll of gun-related violence in the United of US$1.73 billion in Chile and US$32.7 million in Nicaragua. States: US$155 billion (including psychological costs and the In a subsequent publication, Buvinic and Morrison (1999) cal- value of quality of life). They also calculate that, on a per capita culate that the direct medical costs plus lost productivity are basis, the cost of gun violence in Canada equals one-third of equivalent to 2.0 percent of GDP in Chile and 1.6 percent of the U.S. cost. Peden and van der Spuy's (1998) study at the GDP in Nicaragua. Groote Schurr Hospital in Cape Town, South Africa, finds that Several studies have used the U.S. National Crime direct medical costs averaged R 30,628 (US$10,308) per gun- Victimization Survey, an annual survey based on 100,000 inter- shot victim. 760 | Disease Control Priorities in Developing Countries | Mark L. Rosenberg, Alexander Butchart, James Mercy, and others Effects on Public Finances and the individual attributes brought to these contexts. The public sector (and thus society in general) bears many of Intervention may therefore attempt to influence aspects or risk the costs of interpersonal violence. Several studies (Klein and factors at any or all of the model's four levels (Dahlberg and others 1999; Payne and others 1993; additional sources online) Krug 2002; Mercy and Hammond 1999). find that 56 to 80 percent of U.S. health care costs for stabbing Table 40.4 presents our typology of prevention strategies. and gun injuries are either directly paid by public financing The examples presented are not exhaustive, nor have all the or are not paid at all. In the latter case, government and society strategies proven effective. Rather, they illustrate the breadth absorb the costs in the form of uncompensated care financ- of potential solutions and emphasize the need to consider ing and overall higher payment rates. In LMICs, society likely addressing the problem simultaneously at different stages of also absorbs the costs of violence through direct public human development and through different social contexts. In expenditures and negative effects on investment and economic many cases, an intervention might have an effect on multiple growth. forms of violence. At this time, data to prove or disprove the effectiveness of most of these interventions are insufficient, and in those cases in which sufficient data are available, they are almost always from high-income countries. INTERVENTIONS The evidence base of ways to prevent violence is expanding Strategic Focuses for Prevention rapidly, but huge gaps remain in relation to effective strategies for reducing the health burden associated with interpersonal A simple understanding of the approaches illustrated in table violence. The greatest strides have come in the areas of youth 40.4 is insufficient for developing a comprehensive violence violence and child abuse, and almost all the prevention knowl- prevention strategy. A public health approach to violence pre- edge has been developed in high-income countries. Despite vention concentrates on identifying ways to keep people from those limitations, an understanding of the epidemiology and committing violent acts. Interventions may eliminate or reduce etiology of violence and prevention provides important the underlying risk factors and shore up protective factors. insights into the spectrum of policies and interventions that Interventions are typically classified in terms of three levels of can be drawn on to prevent violence in LMICs. prevention: primary, secondary, and tertiary (Dahlberg and Krug 2002). Violence Prevention Strategies Primary Prevention. Primary prevention interventions focus The many commonalities among the various forms of violence on preventing violence before it occurs. The literature has given in relation to their epidemiology and etiology suggest that rise to several strategic focuses for the primary prevention of common pathways to prevention may be available (Reza, violence that are important considerations in violence preven- Mercy, and Krug 2001). A typology of prevention strategies is tion planning. Some have been successfully implemented at the useful in sorting through the complexities and commonalities community level in LMICs. of this problem to identify the range of strategies that might The cultural context plays an important role in violent be incorporated into effective violence prevention plans. We behavior. Cultural traditions are sometimes used to justify propose a typology of prevention based on two key dimen- such social practices as female genital mutilation and severe sions: the stages of human development and the ecological physical punishment of children (Mercy and others 2003). model mentioned earlier. Conversely, cultural norms can be a source of protection The epidemiology of violence, including its onset, desis- against violence, such as traditions that promote the equality tance, and continuity, is closely related to the stages of human of women or respect for the elderly. Although evidence-based development (Williams, Guerra, and Elliott 1997). Increasing approaches for changing cultural traditions as a violence pre- evidence points to the existence of discrete developmental vention strategy are not yet available, some countries have pathways to violent behavior (Loeber and others 1993; Tolan adopted this strategy. In South Africa, the Soul City health and Gorman-Smith 1998; U.S. Department of Health and promotion campaign makes residents aware of the extent and Human Services 2001; additional sources online). Thus, inter- consequences of violence and encourages better parenting vening at early developmental stages may reduce the likelihood through role models and improved communication among that violence is expressed during later developmental stages. family members. Evaluations have found shifts in attitudes The ecological model is also an important dimension of the and social norms concerning intimate partner violence and typology, because violence is the product of multiple and over- domestic relations. Willingness to change behavior and take lapping levels of influence on behavior. The ecological model action to stop violence has increased in urban and rural areas assumes that violent behavior is influenced by social contexts among both men and women (Krug and others 2002b, Interpersonal Violence | 761 Table 40.4 Prevention Strategies, by Developmental Stage and Ecological Context Developmental stage Level of the Infant and toddler years Childhood Adolescence Adulthood ecological model (age birth­3) (age 4­11) (age 12­19) (age 20 ) Individual · Reduction in unintended · Social development · Social development · Incentives for postsecondary pregnancies traininga traininga education or vocational · Access to prenatal and · Preschool enrichment · Drug-resistance educationb training postnatal services programsa · Educational incentives for · Services for adults abused as · Treatment programs for · Drug-resistance educationb at-risk, disadvantaged children child witnesses of violence · School-based programs to studentsa · Treatment for child and intimate and victims of maltreatment prevent child maltreatment · Individual counselingb partner abuse offenders to reduce consequences · Community-based · Supervised exposure to · Waiting periods for firearm prevention of child sexual prison and morgue (shock purchases abuse or scare high-risk youthb) · Owner liability for damage by · Gun safety training · Residential programs in guns psychiatric or correctional institutionsb · Academic enrichment programs · Gun safety training · Boot campsb · Waivers to try in adult courtb · School-based violence prevention programsa Relationship · Home visitation servicesa · Parenting traininga · Mentoringa · Programs to strengthen ties to · Parenting traininga · Mentoring · Peer mediation and family · Therapeutic foster care · Partnership programs counselingb · Programs to strengthen ties to between homes and · Temporary foster care jobs schools to promote parental programs for serious and · Couples therapy involvement chronic delinquents · Relationship education · Family therapya Community · Lead monitoring and toxin · Safe havens for children on · Recreational programs · Adult recreation programs removal high-risk routes to and from · Multicomponent gang · Shelters and crisis centers for · Screening by health care school prevention programsb battered women and victims providers for maltreatment · After school programs to · Health care professionals of elder abuse extend adult supervision trained in identification and · Criminal justice reforms to · Recreational programs referral of high-risk youth criminalize child maltreatment, and victims of sexual intimate partner violence, and violence elder abuse · Community policing · Mandatory arrest policies for · Improvements in emergency intimate partner violence response, trauma care, and · Public shaming of intimate access to health services partner violence offenders · Programs to buy back gunsb · Services for identifying and · Metal detectors in schools treating elder abuses · Health care professional train- ing in identification and refer- ral of victims of elder abuse and sexual violence 762 | Disease Control Priorities in Developing Countries | Mark L. Rosenberg, Alexander Butchart, James Mercy, and others Table 40.4 Continued Developmental stage Level of the Infant and toddler years Childhood Adolescence Adulthood ecological model (age birth­3) (age 4­11) (age 12­19) (age 20 ) · Community policing · Community policing · Emergency response and trauma care improvements · Emergency response and trauma care improvements · Health care providers trained in the detection and reporting · Programs to buy back gunsb of child maltreatment · Disruption of illegal gun markets · Programs to buy back gunsb · Prohibition of firearm sales to high-risk purchasers · Promotion of safe storage of firearms and other lethal means · Mandatory sentences for gun use in crimes of inflicting violence · Coordinated community interventions for violence prevention · Prevention and education campaigns to increase awareness · Prevention and education campaigns to increase awareness of of child maltreatment youth violence, intimate partner violence, sexual violence, and · Child protection service programs elder abuse Societal · Promote cultural norms to · Reduce violent content of · Reduce violent content of · Establish job creation value and protect life movies, television, video movies, television, video programs for the chronically · Promote strength-based games, and Internet sites games, and Internet sites unemployed cognitive and socio- available to children available to children emotional skills from birth · Launch public information · Enforce laws prohibiting campaigns to promote illegal transfers of guns to pro-social norms youths · Strengthen police and judicial systems · Strengthen police and judicial systems · Deconcentrate poverty · Promote safe storage of firearms · Reduce income inequality · Deconcentrate poverty · Reduce income inequality · Change cultural norms that support violence and abuse of children and adults Source: Authors. a. These programs have been demonstrated to be effective in reducing violence or risk factors for violence. b. These programs have been found to be ineffective in reducing violence. box 9.1). In the Kapchorwa district of Uganda, a community deaths decreased following the 1988 implementation of legisla- health program has enlisted the support of elders in adopting tion that required the registration of all firearms and strength- alternative practices to female genital mutilation that are ened licensing regulations; a mandatory waiting period was consistent with their cultural traditions (United Nations added in 1996 (Ozanne-Smith and others 2004). However, the Population Fund 1998). evidence to determine whether such strategies are effective in The lethality of interpersonal violence is affected by the reducing firearm-related homicides is currently insufficient means people use to carry out this violence. Reducing access to (Hahn and others 2003), although several policies hold prom- lethal means, such as firearms, may help minimize the health ise (Hemenway 2004; Ludwig and Cook 2003). consequences of violence. A wide variety of strategies have Inadequate parental involvement in children's and adoles- been used to restrict access to firearms, such as mandating cents' activities and lack of supervision are well-established risk waiting periods before purchase, promoting safe storage of factors for youth violence (U.S. Department of Health and firearms, and limiting where firearms can and cannot be Human Services 2001). Evidence indicates that a supportive carried. In the mid 1990s, Colombian officials in Bogotá and relationship with parents or other adults is protective against Cali, noting that homicide rates increased during weekends fol- antisocial behavior. Although not widely evaluated, some men- lowing paydays, on national holidays, and near elections, toring programs that match high-risk youths with a positive implemented a ban on carrying handguns during those times, adult role model appear to be effective in reducing youth vio- which resulted in a 13 to 14 percent reduction in homicide lence (Grossman and Garry 1997; Thornton and others 2002); rates (Villaveces and others 2000). In the Australian state of however, negative findings have also been reported for mentor- Victoria, firearm-related suicides, assaults, and unintentional ing, particularly when mentors receive little training and when Interpersonal Violence | 763 the relationships between adults and youths break down. The suggests that these interventions can improve the mental design of mentoring programs varies considerably, and partic- health of victims, less information is available on other benefits ipation by both mentors and youths can be uneven. (Runyan and others 2002; additional sources online). One Programs that target those who influence children are more approach to preventing child sexual abuse in the United States effective than interventions that target all adults. For example, challenges social norms by offering help to those at risk of preschool enrichment, home visitation, and parenting pro- offending and by encouraging adults to watch for and act on grams have been found to have both short- and long-term warnings of child sexual abuse before an offense is committed effects on preventing violence (Farrington 2003; Mercy and (CDC 2001). Under such programs, individuals voluntarily others 2002; Utting 2003; additional sources online). Early turn themselves in for treatment and thereby prevent potential intervention can help shape attitudes, knowledge, and behavior future violence. of children at a time when they are more open to positive influ- The outcome of injury from interpersonal violence depends ences and can affect their behavior over their lifetime (Mercy not only on its severity, but also on the speed and appropriate- and others 1993). ness of treatment (Committee on Trauma Research 1985). Income inequality is a risk factor universally associated with Establishment of trauma systems designed to treat and manage interpersonal and collective violence (Butchart and Engstrom injured victims efficiently and effectively is an important factor 2002; Zwi, Garfield, and Loretti 2002; additional sources in reducing the health burden of violence. Research suggests online). Poverty itself does not appear to be consistently asso- that reductions in criminal assaults resulting in death in the ciated with violence, but the juxtaposition of extreme poverty United States are partly explained by the increased survival of with extreme wealth appears to be a key ingredient in recipes victims. Developments in medical technology and trauma for violence. Economic programs or policies that reduce or services may be reducing the number of interpersonal violence minimize the effects of income inequality may be strategic in fatalities (Harris and others 2002). Hospital emergency depart- violence prevention, although the evidence base for such inter- ments may also provide an opportunity to intervene with ventions has not been established. victims who might otherwise seek revenge against their attack- ers or victims who are at greater risk for revictimization Secondary and Tertiary Prevention. Although an emphasis (Muelleman and others 1996). on primary prevention is essential for reducing the health burden associated with violence, secondary prevention pro- grams and services are necessary for addressing the immediate COST-EFFECTIVENESS OF INTERVENTIONS consequences of violent actions and behaviors, and tertiary Studies show that implementing preventive interventions costs programs focus on long-term care. Efforts targeted at victims less than dealing with the outcomes of violence, in some cases of violence are extremely important for mitigating the physical by several orders of magnitude. and psychological consequences of the various forms of violence and abuse and for reducing victims' risks for future violence (National Center for Injury Prevention and Examples of Cost-Saving Interventions Control 2002). To date, most evaluations of preventive interventions measure Physicians and other health professionals are gatekeepers in cost and effects in high-income countries. Although cost sav- efforts to monitor, identify, treat, and intervene in cases of inter- ings may not be comparable in LMICs, effects may be greater. personal violence. As previously noted, more cases of interper- sonal violence come to the attention of health care providers Legislation and Shelter for Abused Women. The 1994 than of police. The role of health care providers in prevention Violence against Women Act in the United States has resulted efforts is neither widely understood nor embraced, and many in an estimated net benefit of US$16.4 billion, including institutional and educational barriers limit their effectiveness US$14.8 billion in averted victims' costs (Clark, Biddle, and (Cohen, De Vos, and Newberger 1997). Programs to educate Martin 2002). This wide-ranging legislation introduced pro- health care providers are under way worldwide. Many hospital grams aimed at deterring crimes against women and providing emergency departments, doctors' offices, and clinic settings use assistance to female victims of crimes. Interventions include screening programs to identify victims of intimate partner penalties for repeat offenders, use of sexual history in criminal violence, child abuse, or elder abuse, although the effectiveness and civil cases, programs for victims of child abuse, safe homes of those interventions in reducing subsequent violence is not for women, confidentiality of the abused person's address, and well understood (Heise and Garcia-Moreno 2002; Runyan and pretrial detention in sex offense cases. Chanley, Chanley, and others 2002; Wolf, Daichman, and Bennett 2002). Campbell's (2001) analysis shows that providing shelters for Therapeutic approaches have been implemented in many victims of domestic violence results in an estimated cost- parts of the world to reduce child abuse. Though some research benefit ratio of 18.4 to 6.8. 764 | Disease Control Priorities in Developing Countries | Mark L. Rosenberg, Alexander Butchart, James Mercy, and others Parent Training and Home Visitation. Caldwell (1992) esti- violence and the costs versus benefits of prevention efforts is mates that the costs of child abuse and neglect in Michigan are urgently needed. US$1 billion a year, including the costs of crimes committed by the victims of child abuse later in life and the costs of their incarceration. The study estimates that prevention costs, IMPLEMENTATION OF PREVENTION including a home visitor program for every family and a com- STRATEGIES prehensive parent education program, are just one-nineteenth of the cost of child abuse. Armstrong's (1983) cost-benefit Promoting violence prevention involves encouraging and analysis of a child abuse prevention program in Yeardon, supporting the development, implementation, and evaluation Pennsylvania, finds net savings of US$647,000 per year and a of programs explicitly designed to stop the perpetration of cost-benefit ratio of 1.86. violence at local, regional, and national levels. The 2003 World Health Assembly Resolution (WHA56.24) Registering Firearms. Chapdelaine and Maurice (1996) on implementing the recommendations of the World Report on quantify the costs and benefits of a Canadian law that required Violence and Health (Krug and others 2002b; see also Butchart gun owners to register their firearms by January 1, 2001. and others 2004) advocates a five-point strategy. Implementing a universal licensing and registration system cost approximately US$70 million (2001 U.S. dollars), includ- Increasing Capacity for Collecting and Managing Data ing a significant one-time expense, compared with annual direct health care costs of gun-related violence of US$50 mil- Increased capacity for collecting health, criminal justice, and lion. When the indirect costs of gun violence are included, the social service sector data on violence and its consequences is economic benefits of the law are much clearer. Miller (1995) fundamental to building a sustained, high-level policy and estimates the total costs of firearm-related injuries in Canada at intervention programming response in LMICs. Population- US$5.6 billion, including lost productivity and psychological level data are needed to design and evaluate community-level costs, equivalent to 1 percent of Canada's GDP. intervention trials. Health sector data can cover a larger and often different subset of violence-related injuries than police statistics and, with criminal justice and social service data, can Youth Intervention. Greenwood and others (1996) compare strengthen abilities to define the problem, identify causes and interventions to reduce youth crime in the United States and risk factors, design appropriate interventions, and monitor the find that providing high school students with incentives to interventions' effectiveness. graduate, which costs US$14,100 per program participant, is the Information systems play a large role in recent efforts to most cost-effective intervention, resulting in an estimated 258 address infectious diseases such as tuberculosis and HIV, allow- serious crimes prevented per US$1 million spent. Parent train- ing better identification of high-risk populations and appro- ing prevents an estimated 157 serious crimes per US$1 million, priate interventions. A similar role is imminent for violence compared with 72 for delinquent supervision programs and 11 prevention. As previously stated, information systems must for home visits and day care. All those interventions (excluding integrate data from the criminal justice, labor, education, social home visits) are more cost-effective than California's "three services, and health sectors and must be linked with systems strikes" law, which incarcerates for life those individuals con- housed at multilateral agencies or regional joint initiatives. victed of three serious crimes. Sharing information regionally allows countries to identify opportunities for collaboration and to share best practices. Need for LMIC Cost-Benefit Data Though violence disproportionately affects LMICs, studies of Support for Research the economic effects of violence in those countries are scarce. Supporting research on the causes, consequences, and preven- Comparisons with high-income countries are complicated by tion of violence has proven effective in mobilizing prevention the tendency to undervalue economic losses related to produc- responses in developing countries. In South Africa, the 1997 tivity in lower-income countries, because such losses are typi- Essential National Health Research Conference identified cally based on forgone wages and income. Thus, when the costs research for improved violence prevention and control as a top of violent homicides are calculated, the estimates range from priority, and in 2001, the Medical Research Council established US$15,319 per homicide in South Africa, to US$829,000 in a program to give violence prevention research the same New Zealand, to more than US$2 million in the United States. priority as research into HIV/AIDS, tuberculosis, and malaria Given the existing methodological differences and widespread (Jeenah and others 1997; Medical Research Council of South gaps in the literature, systematic research into the costs of Africa 2004). South Africa also has applied research data in Interpersonal Violence | 765 various prevention contexts, including establishment of a through the system. Services in associated sectors--medical, national violence and injury mortality surveillance system, forensic, social, and legal--should be examined, because inte- passage of firearms legislation, assessment of national and grated efforts can improve access to and value of services that municipal-level burden-of-disease estimates, and design of previously existed in isolation. Training of medical and legal prevention programs (Butchart and others 2001; Groenewald service providers in activities aimed at preventing violence may and others 2003). also affect the success of long-term prevention; however, this The currently limited evidence base and understanding training strategy has not been extensively evaluated. of the causes of all types of violence must be expanded through planned, documented, evaluated, and shared research. Some developing countries have opportunities to develop Developing Action Plans and document prevention programs in special settings, such A national plan of action for preventing interpersonal violence as for refugees, orphans, nomadic, displaced, and homeless and improving victim support and care is the blueprint that populations. provides a set of common goals, a shared time frame, a strategy for coordinating activities, and a framework for evaluating the different sectors involved. Such a plan is therefore central to Promoting Primary Prevention organizing national and community-level interventions that An important first step in promoting primary prevention is involve more than one objective and that depend on input systematic documentation of existing prevention programs. from many different sectors. Strong political support from the Records can include information on the types of violence and highest levels of government is important in aligning the vari- risk factors addressed, target populations served, interventions ous players and ensuring that the plan is implemented and that used, and any monitoring and measurement of the effects. associated programs are maintained. Nongovernmental organ- Such information can help make the programs more visible to izations may provide support and continuity in countries policy makers and development partners and can be used to where programs may be interrupted because of unstable, promote increased investment in programs that apply proven changing governments. and promising interventions. Collaboration among national governments and health- WHO has outlined a methodology for systematic documen- related nongovernmental and multilateral organizations can tation (Sethi, Marais, and others 2004, 22­33) and has initiated establish the importance of formally addressing violence a project to evaluate the feasibility and utility of such docu- through public health approaches. Though legal and criminal mentation in selected cities and provinces in Brazil, India, justice approaches provide a deterrent, experience in high- Jamaica, Jordan, the former Yugoslav Republic of Macedonia, income countries suggests that a proactive public health Mozambique, the Russian Federation, and South Africa. approach can reduce the negative health, social, and economic consequences of interpersonal violence. Strengthening Support Services for Victims A situational analysis of the accessibility and organization of CONCLUSIONS: PROMISES AND PITFALLS emergency, acute, long-term, and rehabilitative services can identify needs and help strengthen care and support services Violence prevention may be seen as a luxury rather than a pub- for victims of violence. The establishment and adequate fund- lic health priority in LMICs; however, the magnitude of the ing of first responder systems, such as police and ambulance problem and the associated health burden negate this view. teams, may lower the costs of violence and contribute to pre- Resolutions on violence prevention passed by the World Health vention. Maps showing hospitals and clinics with specialized Assembly and codified in the World Report on Violence and systems for treating victims of violence can help these first Health and reports from the United Nations Crime Prevention responders. Ready access to legal resources empowers victims. Council present frameworks for approaching violence preven- Mock and others (2004) offer guidelines for strengthening tion. The first World Health Assembly resolution was cospon- victim care and support services. sored by a developing country, South Africa, and a developed Claramunt and Cortes's (2003) assessment in Belize, Costa country, the United States. Both recognized the importance of Rica, El Salvador, Guatemala, Honduras, and Nicaragua estab- making violence prevention a global public health priority even lished baseline information that could be used to advocate for before evidence of effectiveness could be collected. Seven years strengthened medical and legal services for victims of sexual after the resolution, both developing and developed countries violence. Their findings included a lack of adequate medical applauded and adopted the World Report on Violence and and legal information systems, the insufficient training of Health, signaling the beginning of an exciting new agenda for medical staff, and a lack of clear protocols for moving patients public health. 766 | Disease Control Priorities in Developing Countries | Mark L. Rosenberg, Alexander Butchart, James Mercy, and others The public health model for violence prevention focuses on Armstrong, K. A. 1983. "Economic Analysis of a Child Abuse Program." primary prevention and intervention for victims and empha- Child Welfare 62 (1): 3­13. sizes the value of integrating efforts across sectors. However, Biddle, E., and D. Hartley. 2002. "The Cost of Workplace Homicides in the USA, 1980­1997." Abstract submitted for the World Health the model is weakened by a paucity of sustained interventions Organization's Sixth World Conference on Injury Prevention and and measured outcomes in LMICs. Control, Montreal, May 12­15. Currently, the best approach may be to take small, incre- Butchart, A., and K. Engstrom. 2002. "Sex- and Age-Specific Effects of mental steps, focusing on relatively discrete and easily imple- Economic Development and Inequality on Homicide Rates in 0 to 24 Year Olds: A Cross-Sectional Analysis." Bulletin of the World Health mented interventions that address a prevalent problem. LMICs Organization 80: 797­805. should build on programs for which some evidence of effec- Butchart, A., J. Kruger, and R. Lekoba. 2000. "Perceptions of Injury Causes tiveness exists in high-income countries and adopt a "learn as and Solutions in a Johannesburg Township: Implications for you go" approach. Prevention." Social Science and Medicine 50: 331­44. Many LMICs face the daunting challenges of the spreading Butchart, A., M. Peden, R. Matzopoulos, S. Burrows, R. Phillips, N. Bhagwandin, and others. 2001. "The South African National Non- HIV/AIDS epidemic and ongoing intergroup conflict or war. natural Mortality Surveillance System: Rationale, Pilot Results, and Those problems can destroy the infrastructure of civil society, Evaluation." South African Medical Journal 91 (5): 408­17. increase stress and economic hardship, and lead to increases in Butchart, A., A. Phinney, P. Check, and A. Villaveces. 2004. Preventing suicide and interpersonal violence of all kinds--in both sexes Violence: A Guide to Implementing the Recommendations of the World Report on Violence and Health. Geneva: Department of Injuries and and at all ages. In prioritizing violence prevention efforts, policy Violence Prevention, World Health Organization. makers and health care professionals may mitigate some of the Buvinic M., and A. Morrison. 1999. Violence as an Obstacle to secondary repercussions of these deadly factors. Development. 1­8. Washington, DC: Inter-American Development A great deal of progress has been made in violence preven- Bank. tion. There is strong reason to believe that the interventions Caldwell, R. A. 1992. The Costs of Child Abuse vs. Child Abuse Prevention: under way and the capacity to implement violence prevention Michigan's Experience. East Lansing: Michigan Children's Trust Fund. will make a difference. The lessons learned to date during the CDC (Centers for Disease Control and Prevention). 2001. "Evaluation of a public health community's short experience with violence pre- Child Sexual Abuse Prevention Program: Vermont, 1995­1997." vention are consistent with the lessons from the community's Morbidity and Mortality Weekly Report 50 (5): 77­78, 87. much longer experience with the prevention of infectious and ------.2003.Costs of Intimate PartnerViolence againstWomen in the United chronic diseases. Violence can be prevented in LMICs if their States. Atlanta: National Center for Injury Prevention and Control. governments, their citizens, and the global community start Chanley, S. A., J. J. Chanley, and H. E. Campbell. 2001. "Providing Refuge: The Value of Domestic Violence Shelter Services." American Review of now, act wisely, and work together. Public Administration 31: 393­413. Chapdelaine, A., and P. Maurice. 1996. "Firearms Injury Prevention and Gun Control in Canada." Canadian Medical Association Journal ACKNOWLEDGMENTS 155 (9): 1285­89. Children's Safety Network Economics and Insurance Resource Center. The editors thank Linda I. Dahlberg for her thoughtful contri- 1997. Cost of Violent Crime and of Alcohol-Involved and Drug-Involved Violent Crime in the USA, 1995. Landover, MD: National Public butions and Angela Browne for her review. Services Research Institute. Claramunt, M. C., and M. V. Cortes. 2003. Situación de los services medico- legales y de salud para victimas de violencia sexual en Centroamérica. NOTES Serie genero y salud publica 14. San José, Costa Rica: Pan-American Health Organization. 1. The World Bank (2004) classifies countries by annual gross national Clark, K. A., A. K. Biddle, and S. L. Martin. 2002. "A Cost-Benefit Analysis income (2001 U.S. dollars) per capita as follows: low-income, US$735 or of the Violence against Women Act of 1994." Violence against Women less; lower-middle-income, US$736 to US$2,935; upper-middle-income, 8 (4): 417­28. US$2,936 to US$9,075; and high-income, US$9,076 or more. 2. Cost estimates have been converted to 2001 U.S. dollars to facilitate Cohen, M. A. 1988. "The Monetary Value of Saving a High-Risk Youth." comparisons. Journal of Quantitative Criminology 14 (1): 5­33. 3. Whenever possible, we have cited results calculated using a 3 percent Cohen, S., E. De Vos, and E. Newberger. 1997. "Barriers to Physician discount rate, as recommended by the U.S. Panel on Cost-Effectiveness in Identification and Treatment of Family Violence: Lessons from Five Medicine (Gold, Siegel, and Weinstein 2001). Communities." Academic Medicine 72 (Suppl. 1): S19­25. Committee on Trauma Research, National Research Council, Institute of Medicine. 1985. Injury in America: A Continuing Public Health Problem. Washington, DC: National Academy Press. REFERENCES Cook, P. J., and J. Ludwig. 2000. Gun Violence: The Real Costs. New York: Anda, R. F., J. B. Croft, V. J. Felitti, D. Nordenberg, W. H. Giles, D. F. Oxford University Press. Williamson, and others. 1999. "Adverse Childhood Experiences and Courtney, M. E. 1999. "National Call to Action: Working toward the Smoking during Adolescence and Adulthood." Journal of the American Elimination of Child Maltreatment. The Economics." Child Abuse and Medical Association 282 (17): 1652­58. Neglect 23 (10): 975­86. Interpersonal Violence | 767 Dahlberg, L. L., and E. G. Krug. 2002. "Violence: A Global Public Health Hahn, R. A., O. O. Bilukha, A. Crosby, M. T. Fullilove, A. Liberman, E. K. Problem." In World Report on Violence and Health, ed. E. G. Krug, L. L. Moscicki, and others. 2003. "First Reports Evaluating the Effectiveness Dahlberg, J. A. Mercy, A. B. Zwi, and R. Lozano, 1­21. Geneva: World of Strategies for Preventing Violence: Firearms Laws." Morbidity and Health Organization. Mortality Weekly Report 52 (RR14): 11­20. Dietz, P. M., A. M. Spitz, R. F. Anda, D. F. Williamson, P. M. McMahon, Harris, A. R., S. H. Thomas, G. A. Fisher, and D. J. Hirsch. 2002. "Murder J. S. Santelli, and others. 1999. "Unintended Pregnancy among Adult and Medicine: The Lethality of Criminal Assault 1960­1999." Homicide Women Exposed to Abuse or Household Dysfunction during Studies 6 (2): 128­66. Their Childhood." Journal of the American Medical Association 282: Hasbrouck, L. M., T. Durant, E. Ward, and G. Gordon. 2002. "Surveillance 1359­64. of Interpersonal Violence in Kingston, Jamaica: An Evaluation." Injury Dube, S. R., R. F. Anda, V. J. Felitti, D. P. Chapman, D. F. Williamson, and Control and Safety Promotion 9 (4): 249­53. W. H. Giles. 2001. "Childhood Abuse, Household Dysfunction, and the Heise, L., and C. Garcia-Moreno. 2002. "Violence by Intimate Partners." In Risk of Attempted Suicide throughout the Life Span: Findings from World Report on Violence and Health, ed. E. G. Krug, L. L. Dahlberg, the Adverse Childhood Experiences Study." Journal of the American J. A. Mercy, A. B. Zwi, and R. Lozano, 87­121. Geneva: World Health Medical Association 286: 3089­96. Organization. Dube, S. R., R. F. Anda, V. J. Felitti, V. J. Edwards, and J. B. Croft. 2002. Hemenway, D. 2004. Private Guns, Public Health. Ann Arbor: University of "Adverse Childhood Experiences and Personal Alcohol Abuse as an Michigan Press. Adult." Addictive Behaviors 27: 713­25. Dunkle, K. L., R. K. Jewkes, H. C. Brown, G. E. Gray, J. A. McIntryre, and Hillis, S. D., R. F. Anda, V. J. Felitti, and P. A. Marchbanks. 2001. "Adverse S. D. Harlow. 2004. "Gender-Based Violence, Relationship Power, and Childhood Experiences and Sexual Risk Behaviors in Women: A Risk of HIV Infection in Women Attending Antenatal Clinics in South Retrospective Cohort Study." Family Planning Perspectives 33 (5): Africa." Lancet 363 (9419): 1415­21. 206­11. Duque, L. F., J. Klevens, and C. Ramirez. 2003. "Cross Sectional Survey of Hillis, S. D., R. F. Anda, V. J. Felitti, D. Nordenberg, and P. A. Marchbanks. Perpetrators, Victims, and Witnesses of Violence in Bogotá, Colombia." 2000. "Adverse Childhood Experiences and Sexually Transmitted Journal of Epidemiology and Community Health 57 (5): 355­60. Diseases in Men and Women:A Retrospective Study."Pediatrics 106: E11. Ezzati, M., S. Vander Hoorn, A. D. Lopez, G. Danaei, A. Rodgers, Hoel, H., K. Sparks, and C. L. Cooper. 2001. "Estimating the Costs." In C. Mathers, and C. J. L. Murray. 2006. "Comparative Quantification of The Cost of Violence/Stress at Work and the Benefits of a Violence/Stress- Mortality and the Burden of Disease Attributable to Selected Major Free Working Environment, 38­51. Geneva: International Labour Risk Factors for 2001." In A. D. Lopez, C. D. Mathers, M. Ezzati, D. T. Organization. http://www.ilo.org/public/english/protection/safework/ Jamison, and C. J. L. Murray, eds. Global Burden of Disease and Risk whpwb/econo/costs.pdf. Factors. New York: Oxford University Press. Houry, D., K. M. Feldhaus, S. R. Nyquist, J. Abbot, and P. T. Pons. 1999. Fajnzylber, P., D. Lederman, and N. Loayza. 2000. "Crime and Violence: "Emergency Department Documentation in Cases of Intentional An Economic Perspective." Economia 1: 219­78. Assault." Annals of Emergency Medicine 34: 715­19. Farrington, D. P. 2003. "Advancing Knowledge about the Early Prevention Jeenah, M. S., Y. Dada, C. Househam, and D. Harrison. 1997. Essential of Adult Antisocial Behaviour." In Early Prevention of Adult Antisocial National Health Research in South Africa. Document 97.1. Geneva: Behaviour, ed. D. P. Farrington and J. W. Coid, 1­31. Cambridge, U.K.: Council for Health Research and Development. Cambridge University Press. Jewkes, R., P. Sen, and C. Garcia-Moreno. 2002."Sexual Violence." In World Felitti, V. J., R. F. Anda, D. Nordenberg, D. F. Williamson, A. M. Spitz, Report on Violence and Health, ed. E. G. Krug, L. L. Dahlberg, J. A. V. Edwards, and others. 1998. "Relationship of Childhood Abuse and Mercy, A. B. Zwi, and R. Lozano, 147­81. Geneva: World Health Household Dysfunction to Many of the Leading Causes of Death in Organization. Adults: The Adverse Childhood Experiences (ACE) Study." American Klein, S. R., I. J. Kanno, D. A. Gilmore, and S. E. Wilson. 1999. "The Journal of Preventive Medicine 14: 245­58. Socioeconomic Impact of Assault Injuries on an Urban Trauma Fisher, A., L. Chestnut, and D. Violette. 1989."The Value of Reducing Risks Center." American Surgeon 57 (12): 793­97. to Death: A Note on New Evidence." Journal of Policy Analysis and Kobusingye, O. C., D. Guwatudde, and R. R. Lett. 2001. "Injury Patterns in Management 8 (1): 88­100. Rural and Urban Uganda." Injury Prevention 7 (1): 46­50. Fromm, S. 2001. "Total Estimated Cost of Child Abuse and Neglect in the Kobusingye, O. C., and R. R. Lett. 2000."Hospital-Based Trauma Registries United States: Statistical Evidence." Prevent Child Abuse America. in Uganda." Journal of Trauma 48 (3): 498­502. http://www.preventchildabuse.org/learn_more/research_docs/cost_ analysis.pdf. Krug, E. G., L. L. Dahlberg, J. A. Mercy, A. B. Zwi, and R. Lozano. 2002a. Ghaffar, A. 2001. National Injury Survey of Pakistan, 1997­1999. World Report on Violence and Health. Geneva: World Health Islamabad: National Injury Research Centre. Organization. Gold, M. R., J. E. Siegel, and M. C. Weinstein. 2001. Cost-Effectiveness in Krug, E. G., L. I. Dahlberg, J. A. Mercy, A. B. Zwi, and A. Wilson. 2002b. Health and Medicine. New York: Oxford University Press. "The Way Forward: Recommendations for Action." In World Report on Violence and Health, ed. E. G. Krug, L. L. Dahlberg, J. A. Mercy, A. B. Greenwood, P. W., K. E. Model, C. P. Rydell, and J. Chiesa. 1996. Diverting Zwi, and R. Lozano, 243­60. Geneva: World Health Organization. Children from a Life of Crime: Measuring Costs and Benefits. Santa Monica, CA: Rand. Kruger, J., A. Butchart, M. Seedat, and A. Gilchrist. 1998. "A Public Health Approach to Violence Prevention in South Africa." In The Dynamics Groenewald, P., D. Bradshaw, B. Nojilana, D. Bourne, J. Nixon, H. of Aggression and Violence in South Africa, ed. R. van Eeden and Mahomed, and J. Daniels. 2003. Cape Town Mortality, 2001, Part I, M. Wentzel, 399­424. Pretoria: Human Sciences Research Council. Cause of Death and Premature Mortality. Cape Town, South Africa: City of Cape Town, South African Medical Research Council, and Loeber, R., P.Wung, K. Keenan, B. Biroux, M. Stouthamer-Loeber,W. B.Van University of Cape Town. Kammen, and others. 1993. "Developmental Pathways in Disruptive Grossman, J. B., and E. M. Garry. 1997."Mentoring: A Proven Delinquency Child Behavior." Development and Psychopathology 5: 103­33. Prevention Strategy." Juvenile Justice Bulletin NCJ 164386. Washington, Ludwig, J., and P. J. Cook, eds. 2003. Evaluating Gun Policy. Washington, DC: United States Department of Justice, Office of Justice Programs. DC: Brookings Institution. 768 | Disease Control Priorities in Developing Countries | Mark L. Rosenberg, Alexander Butchart, James Mercy, and others Mathers, C. D., A. D. Lopez, and C. J. L. Murray. 2006. "The Burden of Peden, M., and J. van der Spuy. 1998. "The Cost of Treating Firearm Disease and Mortality by Condition: Data, Methods, and Results Victims." Trauma Review 6 (2): 4­5. for 2001." In A. D. Lopez, C. D. Mathers, M. Ezzati, D. T. Jamison, and Rahman, F., Y. Ali, R. Andersson, and L. Svanstrom. 2001. "Epidemiology C. J. L. Murray, eds. Global Burden of Disease and Risk Factors. New of Injury: Results from Injury Registration at District Level Hospital York: Oxford University Press. in Bangladesh: Implications for Injury Surveillance in Low-Income Meddings, D. R., and S. M. O'Connor. 1999. "Circumstances around Countries." Injury Control and Safety Promotion 8 (1): 29­36. Weapon Injury in Cambodia after Departure of a Peacekeeping Force: Rahman, F., R. Andersson, and L. Svanstrom. 1998. "Medical Help Seeking Prospective Cohort Study." British Medical Journal 319 (7207): 412­15. Behaviour of Injury Patients in a Community in Bangladesh." Public Medical Research Council of South Africa. 2004. "Crime, Violence, and Health 112 (1): 31­35. Injury Lead Programme." http://www.mrc.ac.za/crime/about.htm. Reza, A., J. A. Mercy, and E. G. Krug. 2001. "The Epidemiology of Violent Mercy, J. A., A. Butchart, D. Farrington, and M. Cerda. 2002. "Youth Deaths in the World." Injury Prevention 7 (2): 104­11. Violence." In World Report on Violence and Health, ed. E. G. Krug, L. L. Roberts, L., R. Lafta, R. Garfield, J. Khudhairi, and G. Burnham. 2004. Dahlberg, J. A. Mercy, A. B. Zwi, and R. Lozano, 25­56. Geneva: World "Mortality before and after the 2003 Invasion of Iraq: Cluster Sample Health Organization. Survey." Lancet 364 (9448): 1857­64. Mercy, J. A., and W. R. Hammond. 1999. "Preventing Homicide: A Public Rosenberg, M. L., ed. 1985. Violence as a Public Health Problem: Health Perspective." In Studying and Preventing Homicide: Issues and Background Papers for the Surgeon General's Workshop on Violence and Challenges, ed. M. D. Smith and M. Zahn, 274­94. Thousand Oaks, CA: Public Health. Leesburg, VA. October. Sage. Rosenberg, M. L., and M. A. Fenley, eds. 1991. Violence in America: A Public Mercy, J. A., E. G. Krug, L. L. Dahlberg, and A. B. Zwi. 2003. "Violence and Health Approach. New York: Oxford University Press. Health: The United States in a Global Perspective." American Journal of Runyan, D., C. Wattam, R. Ikeda, F. Hassan, and L. Ramiro. 2002. Public Health 93 (2): 256­61. "Child Abuse and Neglect by Parents and Other Caregivers." In World Mercy, J. A., M. L. Rosenberg, K. E. Powell, C. V. Broome, and W. L. Roper. Report on Violence and Health, ed. E. G. Krug, L. L. Dahlberg, J. A. 1993. "Public Health Policy for Preventing Violence." Health Affairs Mercy, A. B. Zwi, and R. Lozano, 57­86. Geneva: World Health 12 (4): 7­29. Organization. Michael, M., D. R. Meddings, S. Ramez, and J. L. Gutierrez-Fisac. 1999. Sethi, D., S. Habibula, K. McGee, M. Peden, S. Bennett, A. A. Hyder, and "Incidence of Weapon Injuries Not Related to Interfactional Combat others. 2004. Guidelines for Conducting Community Surveys on Injuries in Afghanistan in 1996: Prospective Cohort Study." British Medical and Violence. Geneva: World Health Organization. Journal 319 (7207): 415­17. Sethi, D., S. Marais, M. Seedat, J. Nurse, and A. Butchart. 2004. Handbook Miller, T. R. 1995. "Costs Associated with Gunshot Wounds in Canada in for the Documentation of Interpersonal Violence Prevention 1991." Canadian Medical Association Journal 153 (9): 1261­68. Programmes. Geneva: World Health Organization, Department of ------. 2001. "Costs of Juvenile Violence: Policy Implications." Pediatrics Injuries and Violence Prevention. 107 (1): 1­7. Sutherland, I., V. Sivarajasingam, and J. P. Shepherd. 2002. "Recording of Community Violence by Medical and Police Services." Injury Miller, T. R., and M. A. Cohen. 1997. "Costs of Gunshot and Cut/Stab Prevention 8: 246­47. Wounds in the United States, with Some Canadian Comparisons." Accident Analysis and Prevention 29 (3): 329­41. Swart, L., A. Gilchrist, A. Butchart, M. Seedat, and L. Martin. 2000. "Rape Surveillance through District Surgeon Offices in Johannesburg, Miller, T. R., M. A. Cohen, and S. B. Rossman. 1993. "Victim Costs of 1996­1998: Findings, Evaluation, and Prevention Implications." South Violent Crime and Resulting Injuries." Health Affairs 12 (4): 186­97. African Journal of Psychology 30 (2): 1­10. Mock, C., J. D. Lormand, J. Goosen, M. Joshipura, and M. Peden. 2004. Thornton, T. A., C. A. Craft, L. L. Dahlberg, B. S. Lynch, and K. Baer. 2002. Guidelines for Essential Trauma Care. Geneva: World Health Best Practices of Youth Violence Prevention: A Sourcebook for Organization. Community Action. Rev. ed. Atlanta: Centers for Disease Control and Morrison, A. R., and M. B. Orlando. 1999. "Social and Economic Costs of Prevention, National Center for Injury Prevention and Control. Domestic Violence: Chile and Nicaragua." In Too Close to Home: Tolan, P. H., and D. Gorman-Smith. 1998. "Development of Serious and Domestic Violence in the Americas, ed. A. R. Morrison and M. L. Biehl, Violent Offending Careers." In Serious and Violent Juvenile Offenders: 51­80. New York: Inter-American Development Bank. Risk Factors and Successful Interventions, ed. R. Loeber and D. P. Muelleman, R. L., J. Reuwer, T. G. Sanson, L. Gerson, R. H. Woolard, A. H. Farrington, 13­29. Thousand Oaks, CA: Sage. Yancy, and others. 1996. "An Emergency Medicine Approach to United Nations Population Fund. 1998. "Reproductive Health Effects Violence throughout the Life Cycle." Academic Emergency Medicine of Gender-Based Violence." UNFPA Annual Report 1998, 20­21. 3 (7): 708­15. http://www.unfpa.org/about/report/report98/ppgenderbased.htm. National Center for Injury Prevention and Control. 2002. CDC Injury U.S. Department of Health and Human Services. 2001. Youth Violence: A Research Agenda. Atlanta: Centers for Disease Control and Prevention. Report of the Surgeon General.Rockville,MD: U.S.Department of Health National Crime Prevention Council. 1999. Saving Money while Stopping and Human Services. http://www.mentalhealth.org/youthviolence/ Crime. Washington, DC: National Crime Prevention Council. surgeongeneral/SG_Site/home.asp. Odero, W. O., and J. C. Kibosia. 1995. "Incidence and Characteristics of Utting, D. 2003. "Prevention through Family and Parenting Programs." Injuries in Eldoret, Kenya."East African Medical Journal 72 (11): 706­10. In Early Prevention of Adult Antisocial Behaviour, ed. D. P. Farrington Ozanne-Smith, J., K. Ashby, S. Newstead, V. Z. Stathakis, and A. and J. W. Coid, 243­64. Cambridge, U.K.: Cambridge University Press. Clapperton. 2004. "Firearm Related Deaths: The Impact of Regulatory Villaveces, A., P. Cummings, V. E. Espitia, T. D. Koepsell, B. McKnight, and Reform." Injury Prevention 10: 280­86. A. L. Kellermann. 2000. "Effect of a Ban on Carrying Firearms on Payne, J. E., T. V. Berne, R. L. Kaufman, and R. Dubrowskij. 1993. Homicide Rates in 2 Colombian Cities." Journal of the American "Outcome of Treatment of 686 Gunshot Wounds of the Trunk at Medical Association 283 (9): 1205­9. Los Angeles County­USC Medical Center." Journal of Trauma 34 (2): Viscusi, W. K. 1993. "The Value of Risks to Life and Health." Journal of 276­81. Economic Literature 31 (4): 1912­46. Interpersonal Violence | 769 Waters, H., A. Hyder, Y. Rajkotia, S. Basu, and A. Butchart. 2004. The Childhood." International Journal of Obesity and Related Metabolic Economic Dimensions of Interpersonal Violence. Geneva: World Health Disorders 26: 1075­82. Organization, Department of Injuries and Violence Prevention. Wolf, R., L. Daichman, and G. Bennett. 2002. "Abuse of the Elderly." In WHO (World Health Organization). 1999. Report of the Consultation on World Report on Violence and Health, ed. E. G. Krug, L. L. Dahlberg, Child Abuse Prevention. WHO/HSC/PVI/99.1. Geneva: WHO. J. A. Mercy, A. B. Zwi, and R. Lozano, 125­45. Geneva: World Health Organization. WHO Global Consultation on Violence and Health. 1996. Violence: A Public Health Priority. WHO/EHA/SPI.POA.2. Geneva: WHO. World Bank. 2004. "List of World Economies." Washington, DC: World Bank. http://www.worldbank.org. Williams, K. R., N. G. Guerra, and D. S. Elliott. 1997. Human Development Zwi, A. B., R. Garfield, and A. Loretti. 2002. "Collective Violence." In World and Violence Prevention: A Focus on Youth. Boulder: University of Report on Violence and Health, ed. E. G. Krug, L. L. Dahlberg, J. A. Colorado, Center for the Study and Prevention of Violence, Institute Mercy, A. B. Zwi, and R. Lozano, 213­39. Geneva: World Health of Behavioral Science. Organization. Williamson, D. F., T. J. Thompson, R. F. Anda, W. H. Dietz, and V. J. Felitti. 2002. "Body Weight and Obesity in Adults and Self-Reported Abuse in 770 | Disease Control Priorities in Developing Countries | Mark L. Rosenberg, Alexander Butchart, James Mercy, and others Chapter 41 Water Supply, Sanitation, and Hygiene Promotion Sandy Cairncross and Vivian Valdmanis Water supply in the context of this chapter includes the supply together with data on existing levels of coverage to derive esti- of water for domestic purposes, excluding provision for irriga- mates of the potential effects of water supply and excreta dis- tion or livestock. Sanitation is used here in the narrow sense of posal on the burden of disease, globally and by region, and with excreta disposal, excluding other environmental health inter- cost data to derive cost-effectiveness estimates. ventions such as solid waste management and surface water drainage. The effect of these other measures on disease burden is WATER SUPPLY largely confined to urban areas and is considerably less than that of water supply, sanitation, and hygiene promotion What constitutes a perfectly satisfactory water supply to some (Cairncross and others 2003). More fundamentally, expendi- consumers leaves others, even in developing countries, consid- ture on solid waste disposal and drainage is rarely seen as ering themselves unserved. In much of rural Africa, a hand forming part of a portfolio of investments in public health or pump 500 meters from the household is a luxury, but most res- competing with public health investments. Rather, it is general- idents in urban Latin America would not consider themselves ly perceived by decision makers as comparable with other served by a water supply unless they had a house connection. investments in municipal infrastructure and services, such as In Asia, urban planners would consider a community served if roads or public transportation, which are not considered to be there were sufficient standposts on the street corner; however, public health interventions. if the water only flows for a few hours per week, producing This chapter focuses on water supply, excreta disposal, and lengthy nighttime queues, the residents may regard this situa- hygiene promotion and considers the costs and benefits of each tion as a lack of service and opt to buy water expensively from in turn. Water supply and sanitation can be provided at various itinerant vendors. As these examples illustrate, water supply is levels of service, and those levels have implications for benefits. not a single, well-defined intervention, such as immunization, Water supply and sanitation offer many benefits in addition to but can be provided at various levels of service with varying improved health, and those benefits are considered in detail benefits and differing costs. because they have important implications for the share of the cost that is attributable to the health sector. From the point of view of their effect on burden of disease, the main health ben- Levels of Service and Their Costs efit of water supply, sanitation, and hygiene is a reduction in Many public health workers unfamiliar with the water sector diarrheal disease, although the effects on other diseases are sub- assume that the most important characteristic of a water sup- stantial. In the concluding sections, the percentage reductions ply is its improved quality. However, most of the benefit is arrived at in the discussion throughout the chapter are used attributable to improved convenience of access to water in 771 quantity. Moreover, global statistics are not available on the Cost per capita (US$) coverage and costs of provision of water in terms of its quality. 160 Africa Asia The Global Water Supply and Sanitation Assessment 2000 Report 144 140 Latin America and (WHO and UNICEF 2000), the most recent compilation of the Caribbean global statistics on water supply, changed the way that such 120 data are compiled, from the previous unreliable estimates by 102 100 provider agencies to consumers' responses in population-based 92 surveys. The change required a departure from the old defini- 80 tion of reasonable access to safe water, because most consumers 64 60 55 cannot tell whether their water supply is safe. They can, how- 48 49 41 ever, state the type of technology involved, and that fact was 40 34 36 31 used to define a new indicator of improved water supply. In the 23 21 22 20 17 main, improved water supplies could be expected to provide water of better quality and with greater convenience than tra- 0 ditional not improved sources. The report treated the following House Standpost Borehole Dug well Rainwater connection technologies as improved: household connection, public stand- Type of supply pipe, borehole, protected (lined) dug well, protected spring, Source: WHO/UNICEF 2000. and rainwater collection. Unprotected wells and springs, ven- dors, and tanker-trucks were considered unimproved. Bottled Figure 41.1 Median Construction Cost of Water Supply Facilities for water was also considered unimproved because of concerns Africa, Asia, and Latin America and the Caribbean about the quantity of water supplied, not because of concerns over the water quality. Reasonable access was defined as the availability of at least costs of house connections relate almost exclusively to urban 20 liters per capita per day from a source within 1 kilometer of areas because such connections are only rarely provided in the user's dwelling. Within the broad category of those with smaller communities. The smaller size of rural communities reasonable access to an improved water supply, two significantly means that piped systems in general--and house connections different levels of service can be distinguished: in particular--will tend to be more expensive per capita there than in urban areas. An overall unit cost figure of US$150, · house connections just above the highest of the three continental medians, is · public or community sources. therefore taken for house connections in the cost-effectiveness In most settings, these subcategories correspond to very dif- calculations. ferent levels of water consumption, different amounts of time For public water points corresponding to improved water spent collecting water, and as discussed in later sections, differ- supply, hydrogeological and other constraints mean that the ent health benefits. cheapest technology is not feasible in every community. A cost The Global Water Supply and Sanitation Assessment 2000 figure of US$40 per capita is about the middle of the range Report also gives median construction costs per person served offered by different technologies (standpost, borehole, and for the various technologies in the three main regions of the dug well) providing this level of service for each continent developing world. These costs are shown in figure 41.1. (figure 41.1) and, therefore, seems reasonable for this level of However, local conditions, such as the size of the community to service, although it can be expected to vary between US$15 and be served and the presence of suitable aquifers, can cause US$65 or more, depending on local conditions. The range of tremendous variations in the unit cost of water supply. costs reported by individual countries for the Global Water For a community of given size, there are no significant Supply and Sanitation Assessment 2000 Report varied by more returns to scale in the number of house connections made. than an order of magnitude. Most of the investment in major works must be made before In calculating the cost-effectiveness of investment in water house connections can be offered, so that the marginal cost of supplies, one must amortize these capital costs over an appro- each connection is only a fraction of the total. For those and priate lifetime. Most major components of an urban water other reasons, water supply is a natural monopoly requiring supply system have a potential lifetime of 50 years or more, but "lumpy" investments, which makes the unit costs difficult to a prudent utility would aim to amortize them within about calculate. 20 years. A reasonable basis for calculation, for both urban and The cost of house connections may be representative in rural supplies, is to allow an amount of 5 percent of the capital Latin America and the Caribbean, where they are often pro- cost as an annual straight-line amortization of the construction vided in rural areas. In Asia and Africa, however, the reported cost of the water supply. 772 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis Construction costs do not represent the full cost of water Cairncross and Kinnear (1992) found that vendor prices supplies. The Global Water Supply and Sanitation Assessment increased with the time required to collect the water, showing 2000 Report also gives median reported production costs per that households pay more as the alternative of collecting water cubic meter for urban (house connection) water supplies as themselves becomes more burdensome. If the amount paid to US$0.20 for Asia and US$0.30 for Africa and Latin America the vendor for bringing the water is divided by the time saved and the Caribbean. If we assume a mean daily water consump- from collecting it, the implicit value that people ascribe to their tion of 100 liters per capita by those with household connec- time can be calculated. Whittington, Mu, and Roche (1990), tions, those figures give annual per capita operation and main- working in rural Kenya, showed in this way that the implicit tenance costs of US$7.30 and US$10.95, respectively, or 8 to value of the time saved was roughly US$0.38 per hour, very 10 percent of the capital cost of construction. In this chapter, a close to the average imputed wage rate for such households of generic figure of US$10 is used for the annual per capita oper- US$0.35 per hour. ation and maintenance cost. Because the poorest urban households typically spend more Reliable figures for the annual maintenance costs for rural than 90 percent of their household budget on food, the money water supplies are harder to find, particularly because much of they spend on water is sacrificed from their food budget the maintenance is carried out by the volunteer labor of villagers. (Cairncross and Kinnear 1992). The provision of water more Arlosoroff and others (1987), after reviewing a wide range of cheaply thus offers a substantial nutritional benefit to the rural water supply projects in various countries, concluded that poorest. with a centralized maintenance system, the annual per capita cost of maintenance of a hand pump­based supply can range Assessing the Time Saved. The cost of water collection in from US$0.50 to US$2.00,while well-planned,community-level rural areas is usually in time and effort rather than in money maintenance can bring that figure down as low as US$0.05 per paid to vendors. The saving in time and drudgery underlies capita per year. A nominal annual figure of US$1.00 per capita is many social benefits. Given the relevance of the time-saving therefore used in this chapter. A similar figure can be applied to benefit to water supply policy and the fact that the benefit is urban public standposts, for which volunteer labor is less forth- usually uppermost in the mind of the consumer, it is remark- coming but transport costs are lower. This maintenance cost able how few data have been collected on the amounts of time represents 2.5 percent of the construction cost arrived at above. spent collecting water. Working in 334 study sites in Kenya, Tanzania, and Uganda, Thompson and others (2001) found a mean distance from The Time-Saving Benefit rural unpiped households to their water sources of 622 meters. Benefits to health are not normally foremost in the minds of In urban areas, the distance was only 204 meters, but queuing those provided with new water supplies. An exhaustive study of at the tap meant that a water collection journey took almost the economics of rural water supply by the World Bank con- as long. cluded that "the most obvious benefit is that water is made Feachem and others (1978) found in 10 villages of the available closer to where rural households need it. . . . It is not densely populated lowlands of Lesotho that the installation of clear that rural populations think much about the relationship a water supply had saved the average adult woman 30 minutes between water and health" (Churchill and others 1987, 21­22). per day. In one-third of the villages, the saving per woman was more than an hour a day. Lesotho has many springs, so that The Value of Time. The saving in time and drudgery of carry- time saving is likely to be on the low side compared with Africa ing water home from the source is substantial, and several rea- as a whole. sons exist to attribute a money value to it. The most powerful These time-saving benefits are confirmed by the Multi- argument for the money value of poor women's time is that Indicator Cluster Surveys of the United Nations Children's households often pay others to deliver their water, or pay to col- Fund (UNICEF). A recent analysis of the responses in 23 lect from nearby rather than from more distant sources that are African countries has produced a more representative account free of charge. Thompson and others (2001) found that, of of water collection journey times in that continent (G. Keast, urban East African households lacking a piped supply, the pro- UNICEF, personal communication 2003). Nearly half the portion paying for water had increased from 53 percent to households interviewed (44 percent) required a journey of 80 percent over 30 years. In a survey of 12 sites in 10 countries, more than 30 minutes to collect water, implying that the Zaroff and Okun (1984) found that households were spending women in such households spent an hour or more each day in a median of over 20 percent of their income on the purchase of water collection. At almost any reasonable level of service, most water from vendors. The prices charged by vendors are typically of that time would be saved by an improved water supply. more than 10 times--and can be up to 50 times--the normal In Asia, an Indian national survey for UNICEF found that tariff charged by the formal water supply utility. women spent an average of 2.2 hours per day collecting water Water Supply, Sanitation, and Hygiene Promotion | 773 from rural wells (Mukherjee 1990). A study in Sri Lanka, which the water sector. The function of the health sector is one of reg- is generally considered to be well provided with water sources, ulation, advocacy, and provision of supplementary inputs, as found that 10 percent of women had to travel more than 1 kilo- appropriate, to ensure that potential health benefits of water meter to their nearest source (Mertens and others 1990). supply are realized to the optimal extent. For example, the regulatory role of the health sector in qual- Valuation of the Time-Saving Benefit. Putting a precise figure ity surveillance of drinking water is well known and widely on the money value of the time of poor people is a tricky task, accepted. Substantial and largely unexploited additional poten- even for the most self-confident economist. In 1987, Churchill tial is present in this role if quality is interpreted in the wider and others took US$0.125 per hour as an illustrative but not sense of quality of service rendered by the water supply utility, unrealistic figure. To take the same figure today could hardly be in terms not only of water quality but also of quantity, conti- described as extravagant. Assuming this valuation of an hour of nuity, coverage, control of sanitary hazards, and cost. Those time--and that a water supply bestows a mean saving of only other aspects, as will be argued in the following sections, are no 15 minutes per person per day--yields a conservative estimate less important for health. of the value of the time-saving benefit of US$11.40 per year. Where a regulatory role is not available to the health sector The data presented earlier indicate that, at least in Africa, the or agencies concerned with public health, advocacy can be no true figure is nearer to double that amount, enough to justify less cost-effective. For example, connection charges are a major the full construction cost of a dug well or borehole supply in a barrier to house connections for low-income groups. In many single year. In Latin America and the Caribbean, costs are cities of the developing world, the individual connection higher, and time savings may be less, but rural incomes are also charge is about a month's basic wage. Advocacy of lower con- higher--and so, therefore, is the value of people's time. Little nection charges, with the amount recovered from the monthly doubt exists that, in all three regions of the developing world, water tariffs, can therefore help achieve an increase in the num- the value of time saved is sufficient on its own to justify both ber of people who have house connections and who can bene- the investment costs (at any reasonable rate of amortization) fit from the corresponding health gain at no cost to the public and the operation and maintenance costs of water supplies. purse. Finally, the health sector can provide important comple- Even in settings where water vending is not common, con- mentary services, such as hygiene promotion and promotion of tingent valuation surveys have widely demonstrated a willing- low-cost sanitation to increase coverage; because of the nature ness to pay for water supplies, particularly at the level of serv- of such services, the water sector, with its focus on technology, ice of house connections (World Bank Water Demand is ill-equipped to offer them. Research Team 1993). In general, such measured willingness to The unit costs of such regulation and advocacy are minimal. pay has exceeded the cost of providing the supplies, and pay- One example is the case of UNICEF's participation over the ment to vendors often exceeds it by many times. past 30 years in India's rural water supply program. UNICEF's investment has represented no more than 1 percent of the total, Policy Implications. Whether the consumers actually pay for but its influence has played a central part in the evolution of the the full value of the time-saving benefit, it is what makes water technical and institutional model of the program that supplies supplies popular and largely it is what motivates politicians to water to 1 in 10 members of the human race. invest in them. More than half the total annual investment in An example of the effectiveness of such measures is pro- water supply in the developing countries of Africa, Asia, and vided by the interventions of the Mexican Ministry of Health in Latin America and the Caribbean is from domestic sources June 1991. Fostered by fear of the devastating effects of cholera, (WHO and UNICEF 2000). Most of the investment is from the these measures included the chlorination of water supplied for public sector. In general, investments in water supply-- human consumption and the prohibition of sewage irrigation whether by the governments of developing countries or by of fruit and vegetables. As a result, the incidence of diarrhea in external support agencies--do not come from health sector children under five years of age fell from 4.5 to 2.2 episodes per budgets and are not compared with other health interventions child-year, and the corresponding mortality rate fell from 101.6 when investment decisions are made, even though health ben- to 62.9 per 100,000 children (Gutiérrez and others 1996). efits do arise from water supply improvements. The current rate of annual investment per capita in water Water supply is thus a health-related intervention that supply and sanitation, including both national investment and comes without cost to the budgets of the health sector. external aid funds, is reportedly US$2.25 in Asia, US$7.53 in Although it undoubtedly offers health benefits, it has a suffi- Africa, and US$8.87 in Latin America and the Caribbean cient economic and political rationale in other social benefits (WHO and UNICEF 2000). One percent of the water sector's associated with time saving. The health benefits are a positive investment would, therefore, be US$0.02 to US$0.10 per capita. externality to this rationale. However, this fact does not mean If each ministry of health in the developing world were to that the authorities responsible for public health should ignore invest such a sum in public health advocacy and regulation 774 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis related to water supply, the sector's performance, at least where insect vectors of disease (for example, if a more reliable supply low-income groups are concerned, could be transformed. It is averts the need for the water-storage vessels in which dengue hard to put a figure on the health effects of such investment, vectors breed), though that will depend on the precise life cycle but the Mexican example suggests that they would be substan- of the parasite involved and the preferred breeding sites and tial. For the sake of cost-effectiveness estimation, such spending behavior of the vector. is arbitrarily assumed to have the effect of ensuring improved water supplies for an additional 10 percent of the population to Classification and Burden of Water-Related Diseases. Before which it refers. Bradley's classification can be applied to diseases (rather than transmission routes), it requires a small adjustment (Cairncross and Feachem 1993) to allow for the fact that prac- Direct Health Effects tically all potentially waterborne infections that are transmitted The full list of water-related infections is large and varied, but by the feco-oral route can potentially be transmitted by other most are only marginally affected by water supply improve- means (contamination of fingers, food, fomites, field crops, ments. The first effort to simplify the relationship between other fluids, flies, and so on) all of which are water-washed water supplies and health in developing countries was made by routes. In addition to the feco-oral infections, a number of David Bradley (White, Bradley, and White 1972), who devel- infections of the skin and eyes can be considered water oped a classification of disease transmission routes in terms of washed but not waterborne. The final classification is shown in whether they were table 41.1. The classification can now be used to assess how the disease · waterborne, in the strict sense in which the pathogen is burden prevented by water supply is distributed among disease ingested in drinking water groups. Bradley himself did this, a time long before the · water-washed--that is, favored by inadequate hygiene con- disability-adjusted life year (DALY) had been invented as a unit ditions and practices and susceptible to control by improve- of benefit measurement (White, Bradley, and White 1972, 191). ments in hygiene He used official statistics on the number of cases of each disease · water-based, referring to transmission by means of an diagnosed and treated by health services in East Africa and aquatic invertebrate host combined them with notional percentages by which morbidity · water-related insect vector routes, involving an insect vector and mortality caused by each condition could be expected to that breeds in or near to water. fall if water supply were "excellent." Those notional reductions were based on subjective assess- Whereas the prevention of waterborne disease transmission ments of the literature available at the time and were described requires improvements in water quality, water-washed trans- by their author as "little more than guesses," but it is hard to mission is interrupted by improvements in the availability-- prove many of them seriously at fault, even today. A selection is and hence the quantity--of water used for hygiene and the pur- presented in table 41.2. poses to which it is put. Water supply may affect water-based The result of these calculations was that the feco-oral disease transmission (for example, if it reduces the need for people to group accounted for 91 percent of the deaths preventable by enter schistosomiasis-infected water bodies) or water-related water supply, 50 percent of inpatient bed nights, and 33 percent Table 41.1 The Bradley Classification of Water-Related Infections Transmission route Description Disease group Examples Waterborne The pathogen is in water Feco-oral Diarrheas, dysenteries, that is ingested typhoid fever Water-washed Person-to-person transmission Skin and eye Scabies, trachoma (or water-scarce) because of a lack of water infections for hygiene Water-based Transmission via an aquatic Water-based Schistosomiasis, intermediate host (for example, guinea worm a snail) Water-related Transmission by insects that Water-related Dengue, malaria, insect vector breed in water or bite near insect vector trypanosomiasis water Source: Cairncross and Feachem 1993. Water Supply, Sanitation, and Hygiene Promotion | 775 Table 41.2 Percentage Reductions in Disease Rates because this group can be transmitted by both waterborne and Assumed by Bradley water-washed routes. It is important for the water engineer to know whether scarce funding should be spent on improved Percentage reduction expected water treatment and measures to protect water quality or Diagnosis from excellent water supply instead on providing a limitless supply of water at a high level Most diarrhea and dysentery 50 of access and convenience and encouraging its use for Typhoid fever 80 improved hygiene practices. We need to know, that is, whether Paratyphoid, other Salmonella 40 the feco-oral infections endemic in poor communities are Trachoma 60 mainly waterborne or mainly water washed. Scabies 80 Moreover, the fact that some diarrheal diseases are still Skin and subcutaneous infections 50 prevalent in communities with a high level of water supply serv- Urinary schistosomiasis 80 ice indicates that water supply alone cannot completely prevent Intestinal schistosomiasis 40 these diseases. A further question then, is this: by how much do Malaria 0 water supply improvements reduce diarrheal diseases? Numerous studies have sought to answer these questions, Source: White, Bradley, and White 1972. but they are hard to answer rigorously, for several reasons. First, it is almost impossible, ethically and politically, to randomize the intervention. Where the intervention is an improvement in of outpatient consultations. Rosen and Vincent (2001) have the level of access to water, it cannot be blinded; no placebo made a similar calculation for the whole of Africa in 1990 and exists for a standpost. Where quasi-experimental studies have found that the feco-oral group accounted for 85 percent of the been used--opportunistically exploiting an intervention preventable DALYs. When measured in terms of deaths or allocated by political or technical means--significant con- DALYs, feco-oral infections account for the vast majority of the founding has frequently been found (Briscoe, Feachem, and impact, because of the high mortality caused by diarrheal dis- Rahaman 1985). eases among young children. Most deaths from diarrheal dis- Confounding has been especially intractable in studies in eases are of children younger than age five, and most of those which the allocation of facilities has been on a household basis, are among children younger than two. A child death averted is so that the exposure groups are self-selected--for instance, worth 30 DALYs. Varley, Tarvid, and Chao (1998) have calcu- studies in which individual households that have chosen to lated that for diarrhea morbidity reduction to have the same install a private tap are compared with others that have chosen effect in DALYs as averting one such death, it would have to not to do so. The former households are likely to be wealthier, prevent 115,000 child-days of diarrhea. After the diarrheal dis- better educated, and more conscious of hygiene than their eases, the next most important category in terms of DALYs neighbors, so it would not be surprising if they were also more (12 percent of the total) is the water-based group, primarily likely do many other things that protect their families from schistosomiasis. The purely water-washed diseases, mainly skin feco-oral disease. The more sophisticated studies have used infections, represent a more conspicuous portion only when multivariate models to control for confounding, but where rel- compared in terms of the burden placed on health services by ative risks are low and the exposure groups are self-selected, inpatients or outpatients. even those models do not guarantee that confounding is elim- How representative is this African breakdown of the devel- inated (Cairncross 1990). oping world as a whole? Diarrheal disease among poor com- A further difficulty arises from the fact that cases of feco- munities is cosmopolitan. A global review of studies of the inci- oral disease in a given community cannot be considered inde- dence of diarrhea morbidity could find no clear geographic or pendent events, because such diseases are infectious. The sam- climatic trend (Bern and others 1992), so the burden of disease ple size, it can be argued, is the number of such villages rather is no doubt similar around the developing world. The second than the number of individuals enrolled in the study. Yet a most important disease group is represented by schistosomia- number of important studies in the literature compare a single sis, which is absent from much of Asia and Latin America. The intervention area with only one control area. relative importance of feco-oral disease is, therefore, likely to be Other epidemiological weaknesses exist in the data. Blum still greater in the poor communities of Asia and the Western and Feachem (1983) reviewed 50 studies of the health effect of Hemisphere than it is in Africa. water supply and sanitation projects and noted that every one contained one or more of these basic errors of methodology. A Epidemiological Questions and Problems. The predominant further weakness in the evidence for the effect of water supply contribution of feco-oral diseases to the burden of disease on diarrheal disease burden is that most of it relates to attributable to water supply raises an important question, diarrheal disease morbidity, and significant assumptions are 776 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis Table 41.3 Median Reductions in Diarrhea Morbidity Reported from Different Water Supply and Sanitation Interventions Number of rigorous studies Median reduction in Intervention (object of from which morbidity diarrheal morbidity improvement) reductions could be calculated (percent) Water quality only 4 15 Water quantity only 5 20 Water quantity and quality 2 17 Sanitation only 5 36 Water and sanitation 2 30 Hygiene promotion only 6 33 Source: Esrey and others 1991. needed to extrapolate such evidence to an effect on diarrheal New Guinea that use of a household tap was associated with mortality. 56 percent less diarrhea than use of public standposts provid- ing water of good quality. Effect on Diarrheal Disease. Esrey and Habicht (1985) and Esrey and others (1991) reviewed the same literature from a Conditions for Health Effect. Providing a public water point different perspective. Though conscious of the methodological appears to have little effect on health, even where the water pro- shortcomings of most studies, they sought to assess the overall vided is of good quality and replaces a traditional source that reductions in diarrheal disease that water supply could be was heavily contaminated with fecal material. By contrast, expected to cause. They applied a number of criteria of epi- moving the same tap from the street corner to the yard pro- demiological rigor and took the median reduction in morbidity duces a substantial reduction in diarrheal morbidity. How is reported from each type of intervention. Their conclusions are this pattern to be understood? summarized in table 41.3. The first step to an explanation is an understanding that For more than a decade, this review has remained the most most endemic diarrheal disease is transmitted by water-washed authoritative on the subject. However, the small reductions in routes and is not waterborne. Although waterborne epidemics disease that it reports for water supply conceal an important of diarrheal diseases such as cholera and typhoid have been heterogeneity. Though these overall results are frequently notorious in the history of public health, the endemic pattern quoted, the following remark by Esrey and others (1991, 613) of transmission seems to be different, particularly in poor com- has usually been overlooked: munities. Five types of evidence support this view: In the studies reporting a health benefit, the water supply · Negative health impact studies. As mentioned earlier, Esrey was piped into or near the home, whereas in those studies and Habicht (1985) and Esrey and others (1991) cite a num- reporting no benefit, the improved water supplies were pro- ber of studies of the health impact of water supplies in tected wells, tubewells, and standpipes. which water quality improvements have failed to have a sig- nificant effect on diarrheal disease incidence. In the studies in the two reviews by Esrey and Habicht · Food microbiology. Studies of the microbiology of foods in (1985) and Esrey and others (1991) in which the water supply developing countries--particularly the weaning foods fed was provided in the home, the median reduction in diarrheal to children in the age group most susceptible to diarrheal disease is 49 percent (from 12 studies), and the reduction from disease--have shown such food to be far more heavily con- the two better studies is 63 percent. Those reductions are sev- taminated with fecal bacteria than is drinking water eral times greater than the overall median impacts in table 41.3. (Lanata 2003), even when the water has been stored in The 63 percent figure will be used in the burden of disease cal- open pots. culations that follow. In the two better studies, the members of · Seasonality of diarrhea. In countries with a seasonal varia- the comparison group were using not an unimproved water tion in temperature, bacterial diarrheas peak in the warmer supply, but a protected water source away from the home. The season, whereas viral diarrheas peak in the winter. This pat- reductions they found are, therefore, in addition to those tern suggests that the bacterial pathogens show environ- resulting from a public standpost level of service. mental regrowth at some stage in their transmission route, Some subsequent studies have confirmed this pattern. For which means that they must have a nutritional substrate. example, Bukenya and Nwokolo (1991) showed in Papua Water is, thus, a less likely vehicle than food. Water Supply, Sanitation, and Hygiene Promotion | 777 · Fly-control studies. Trials in rural Asia and Africa have savings offered by this level of service mean that people are shown that fly control can reduce diarrheal disease inci- willing to pay more for them. Moreover, collecting revenue dence by 23 percent (Chavasse and others 1999). from households with private connections is far simpler than · Hand-washing studies. A recent systematic review of the collecting it from public taps because the sanction of discon- effect of hand washing with soap has shown that this simple nection can be used against households that default on pay- measure is associated with a reduction of 43 percent in diar- ment of the tariff. rheal disease and 48 percent in diarrheas with the more life- Calculating the burden of disease associated with inade- threatening etiologies (Curtis and Cairncross 2003). quate water supply requires a figure for the reduction associ- ated with the levels of service for which coverage statistics are Those five types of evidence suggest that domestic available. The following burden of disease calculations are hygiene--particularly food and hand hygiene--is the principal based on a reduction of 17 percent from an improved public determinant of endemic diarrheal disease rates and not drink- water supply (table 41.3) and of a further 63 percent from ing water quality. house connections. The second step is an understanding of how the level of The effect of water supply improvements (and of hygiene service and convenience of a water supply influence such practices such as hand washing) on diarrhea mortality can be hygiene practices in the home. Taking the amount of water expected to be at least as great as--and probably greater than-- used per capita as an indicator of hygiene changes, other things their effect on morbidity for several reasons. A theoretical argu- being equal, one finds that providing a source of water closer to ment for this improvement pattern is given by Esrey, Feachem, the home--and therefore more convenient to use--has very and Hughes (1985) in terms of infectious doses. Esrey and little effect on water consumption unless the old source was others (1991) also reported a median reduction of 65 percent more than 1 kilometer (30 minutes' roundtrip journey) away in diarrhea mortality attributable to water supply, sanitation, or from the user's dwelling (Feachem and others 1978). both in three studies, compared with 22 percent from 49 stud- However, water consumption doubles or triples when house ies of morbidity. The effect of hand washing on life-threatening connections are provided (White, Bradley, and White 1972), diarrheas--shigellosis, typhoid, cholera, and hospitalized and reason exists to believe that much of the additional con- cases--is greater than that on diarrhea morbidity as a whole sumption is used for hygiene purposes. For example, Curtis (Curtis and Cairncross 2003). Finally, the two known direct and others (1995) found that provision of a yard tap nearly studies in the literature of the effect of house connections on doubled the odds of a mother washing her hands after cleaning diarrhea mortality ("Serviço Especial da Saúde Pública," an her child's anus and more than doubled the odds that she unpublished study in Palmares, Pernambuco, Brazil, cited by would wash any fecally soiled linen immediately. Wagner and Lanoix 1959; Victora and others 1988) found In conclusion, water supplies are likely to have an effect on reductions of 65 percent (relative to a public standpost) and diarrheal disease when they lead to hygiene behavior change-- 80 percent (relative to various communal sources, some that is, when the old source of water was more than 30 minutes' polluted), respectively. roundtrip away or when house connections are provided. By a happy coincidence, then, the health benefits of water Effect on Other Disease Categories. Water supplies have a supply are most likely to be realized in exactly those cases in beneficial effect on a number of disease groups other than diar- which the time-saving benefit is greatest--when the old source rhea, although the corresponding burden of disease is far less. of water is farthest away, and when the new one is on the plot The median reductions in morbidity from other water-related of the individual household. Though water supplies offering conditions, reported by Esrey and others (1990), are shown in house connections are more expensive, the additional time table 41.4. Table 41.4 Median Reductions in Morbidity Associated with Improved Water Supply and Sanitation: Conditions Other Than Diarrhea, Related Most Closely to Water Supply All studies Better studies Median reduction Median reduction Range Disease Number of studies (percent) Number of studies (percent) (percent) Dracunculiasis 7 76 2 78 75­81 Schistosomiasis 4 73 3 77 59­87 Trachoma 13 50 7 27 0­79 Source: Esrey and others 1990. 778 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis To be effective in controlling schistosomiasis, the water sup- time, and second, whether they are adequately maintained. ply must be so convenient as to discourage water contact for . . . Pit latrines would, from the viewpoint of health rather laundry and bathing. It is unlikely that this level of convenience than convenience, approximate the same rating as a water- can be achieved without house connections. borne sewerage system. (Feachem and others 1983, 49­50) Evidence suggests that water availability and hygiene can produce substantial reductions in trachoma (Emerson and The group therefore judged it most appropriate not to dis- others 2000). Because the reductions come from hygiene tinguish between sanitation technologies and to consider all of improvements such as hand and face washing, they are also them as providing adequate access to sanitation as long as likely to be greatest with house connections. Dracunculiasis is they were private or shared (but not public) and hygienically affected by water quality, but the simplest improved water sup- separated human excreta from human contact. This definition ply is adequate to prevent transmission. was followed in the Global Water Supply and Sanitation Conflicting evidence exists about whether water supply or Assessment 2000 Report, which accepted only sewerage, septic improved water-washed hygiene affects the transmission of tanks with soakaways, pour-flush latrines, and pit latrines as intestinal helminths. On one hand, Henry (1981) found in an improved technologies. Service or bucket latrines and latrines intervention study in St. Lucia that piped water supplies were with an open pit were not accepted. The effect of technology associated with a 30 percent reduction in ascariasis among chil- type on health benefit is discussed further in the sections dren under age three over a two-year period. On the other hand, that follow. Han and others (1988) showed in Burma that an intervention Public latrines, however, do not provide an adequate solu- to promote hand washing with soap had no effect on prevalence tion to the excreta disposal needs of a community. Quite apart or intensity of infection with Ascaris spp. However, the poten- from the notorious and widespread inadequacies in their tial contribution of water supply to reducing the burden of dis- maintenance, they are not usually accessible at night or by the ease through its effect on these other infections is relatively elderly, by those with disabilities, or--if there is an entry minor when compared with its effect on diarrheal disease. charge--by young children. Thus, some promiscuous defeca- tion continues to be practiced, particularly by children, in communities where public latrines are the only level of service EXCRETA DISPOSAL available. Figure 41.2 shows the regional median construction costs In much the same way as with water supply, care is needed to per capita of the various sanitation technologies found by the ensure that different people who talk about sanitation are refer- Global Water Supply and Sanitation Assessment 2000 Report. ring to the same thing. When the WHO-UNICEF Joint Although the simple, on-site systems tend to be cheaper than Monitoring Program was compiling the Global Water Supply systems such as sewerage and septic tanks, the difference is less and Sanitation Assessment 2000 Report (WHO and UNICEF than might be expected. For example, a World Bank survey in 2000), a major effort was needed to persuade some of the Latin several developing countries found the mean cost of conven- American partners that a pit latrine, considered a status symbol tional sewerage to be 10 times that for on-site systems such as in much of rural Africa, was an acceptable form of excreta dis- improved pit latrines and pour-flush toilets (Kalbermatten, posal. In some countries, even engineered sewerage systems are Julius, and Gunnerson 1982). It is likely that the off-site costs considered unacceptable if not connected to a functioning of sewered systems and the cost of the additional water wastewater treatment plant. needed for them to function have not been fully included in national reports to the Global Water Supply and Sanitation Levels of Service, Technologies, and Their Costs Assessment 2000 Report. For the purposes of calculating cost- effectiveness, a construction cost of US$60 per capita seems A wide range of technologies is used, particularly for settings in adequate for basic sanitation facilities (a household pit latrine, which low-cost solutions are required, and this variation has ventilation-improved latrine, or a pour-flush toilet) in any led some to inquire whether the different types of latrine might region of the developing world. Taking a relatively short life- confer differing health benefits. In the early 1980s, the World time of five years for a latrine and straight-line amortization Bank established a Technology Advisory Group for low-cost gives an annual cost of US$12 per capita per year. In such a sanitation, and this question was among those it was asked to short lifetime, very little maintenance is normally required, investigate. Using field studies and a thorough literature review, other than occasional cleaning; the cost of maintenance is, the group concluded that all types of systems can be operated therefore, considered to be included in the amortized annual hygienically, and that cost. The greatest determinants of the efficacy of alternative facil- That said, it should be borne in mind that substantially ities are, first, whether they are used by everyone all the cheaper solutions are often feasible, such as the "15 taka latrine" Water Supply, Sanitation, and Hygiene Promotion | 779 Cost per capita (US$) 180 Africa Asia 160 160 160 154 Latin America and the Caribbean 140 120 120 112 115 104 100 91 80 60 60 57 60 60 52 50 50 52 39 40 26 20 0 Sewer Small bore Septic tank Pour-flush Ventilation-improved Simple pit connection sewer latrine latrine Sanitation systems Source: WHO/UNICEF 2000. Figure 41.2 Median Construction Cost of Sanitation Technologies in Africa, Asia, and Latin America and the Caribbean (costing only US$0.27 per household) developed in Bangladesh, Table 41.5 Benefits of Latrine Ownership as Perceived by which includes a pour-flush pan made of tin sheet and an odor- 320 Households in Rural Benin and insect-proof seal made of flexible plastic pipe. (Average importance Benefit rating, scale 1­4) Social Benefits Avoid discomforts of the bush 3.98 Like water supply, sanitation offers a number of social benefits Gain prestige from visitors 3.96 in addition to direct health gains, which tend to feature more Avoid dangers at night 3.86 prominently in the minds of the users. This outcome is illus- Avoid snakes 3.85 trated by the responses given by a sample of householders in Reduce flies in compound 3.81 rural Benin when asked to rate the importance they ascribed to Avoid risk of smelling or seeing 3.78 the various benefits of latrines on a scale of 1 to 4 (table 41.5). feces in bush Health-related benefits (shown bold in table 41.5) were rarely Protect my feces from enemies 3.71 mentioned spontaneously and generally rated among the less Have more privacy to defecate 3.67 important benefits. Keep my house or property clean 3.59 With sanitation as with water supply, strong gender differ- Feel safer 3.56 ences exist in the perception of the social benefits of sanitation. Save time 3.53 For male heads of household in Benin as in other countries Make my house more comfortable 3.50 around the world, enhanced social status figures highly among Reduce my family's health 3.32 the benefits of latrine ownership, whereas for women, security, care expenses convenience, and aesthetic factors count for more. Women who Leave a legacy for my children 3.16 lack sanitation often risk sexual harassment on the way to and Have more privacy for household affairs 3.00 from their defecation site. In some cultural settings, women are Make my life more modern 2.97 constrained to go out for defecation and urination only during Feel royal 2.75 the hours of darkness, effectively becoming prisoners of day- Make it easier to defecate 2.62 light. Though no systematic study has been made of the health because of age or sickness implications of such practices, they are likely to include an Be able to increase my tenants' rent 1.17 increased prevalence of urinary tract infections. The emancipa- For health (spontaneous mention) 1.27 tion that a latrine bestows on such women cannot lightly be dismissed. Source: Jenkins 1999. 780 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis Willingness to Pay. The governments of developing countries Policy Implications. There are important externalities to cannot afford to provide heavily subsidized sanitation to all--or households' investment in sanitation. Households are pro- even to the majority--of their populations. The 2.6 billion peo- tected from their own feces by their sanitation facilities, but so, ple in Africa, Asia, and Latin America who do have adequate too, are their neighbors, and this factor is probably more sanitation--53 percent of the population of those regions-- important in epidemiological terms. If households are not fully have paid most of the cost themselves. Even those of the urban aware of the health benefit--or if much of it accrues to oth- poor who do not have sanitation have expressed a willingness to ers--a case exists for public intervention to increase coverage pay for its full cost--or at least the local cost (excluding major because these externalities exist. interceptor sewers and treatment works,if required)--in a num- This public intervention need not be in the form of subsidy. ber of surveys, as long as credit is available on reasonable terms Strong arguments can be marshaled against a subsidy for low- to smooth the cash flow (Altaf 1994). With regard to the rural cost sanitation (Cairncross 2003a). Subsidy limits the number of poor, the success of well-conceived sanitation promotion pro- facilities that are built to the size of the subsidy budget; it encour- grams in achieving coverage close to 100 percent, without a sub- ages the design and marketing of unaffordable sanitation sys- stantial subsidy, in some of the poorest rural communities in the tems; it frequently leads to capture by the better-off, who install world (Allan 2003) shows that people are willing to pay for san- expensive toilets while the poor go without; and it distorts the itation if a suitable product is offered to them on suitable terms. market, diverting the efforts of latrine builders who would Why then do 2.4 billion people still lack sanitation? Several otherwise be seeking to meet the needs of low-income groups. factors constrain the expression of the existing demand. The intervention can be by regulation. National and local The constraint most frequently mentioned by unserved governments have substantial regulatory powers that can be householders is cost, but this factor is usually more a perceived used to increase sanitation coverage without significantly constraint than an objective one, for several reasons. First, many increasing costs or public expenditure. For example, more than households are unaware of the true cost of latrines in their area, 90 percent of households in the town of Bobo Dioulasso, or the lower-cost models are not offered because local suppliers Burkina Faso, have their own latrine (Traoré and others 1994) as and artisans do not know about them or are attracted by the a direct result of the local administration's practice in the past of greater margins to be made on the more expensive technolo- withdrawing rights of land tenure from owners who did not gies. Second, the high cost of capital to the poor rules out their build a latrine on their plot within a specified time.Another reg- borrowing the cost of a latrine, which to them would be a sub- ulatory intervention is to enforce the obligation of landlords to stantial investment. Third, they may be wary of investing in a provide sanitation for their tenants. property that belongs to their landlord, lest it be used as an An alternative strategy is to provide support to the market- excuse for a rent increase or even eviction. They may also feel, ing of sanitation. This strategy can be undertaken in a number with some reason, that it is for the landlord to make the invest- of ways that are not feasible for the existing producers, mainly ment, rather than themselves, and they may be waiting for the artisan builders and small component manufacturing work- landlord to do so. This belief has a similar effect to the common shops. Those interventions would aim principally at overcom- misapprehension of citizens, often encouraged by politicians, ing the constraints to the expression of effective demand for that the local government is responsible for sanitation and sanitation and could include the following: will eventually come to their aid; in either case, the outcome is inaction. · advertising and other forms of promotion Other constraints include lack of ready access to necessary · facilitation of building regulation approval techniques and skills or to specific building materials and · brokerage to put potential purchasers in touch with components. Where the skills exist locally, residents may lack providers confidence in the quality of work and value for money offered · quality assurance and guarantee schemes by the local artisans, or they may not know how to contact the · training in low-cost construction techniques and in right artisans. In many urban areas, local building regulations marketing make low-cost sanitation technologies illegal. · centralized production of essential components Those constraints are compounded by the fragmentation of · provision of pit emptying and desludging services. governmental responsibility for sanitation. Often it is devolved to local governments with little capacity to implement sanita- Promotion of improved hygiene practices, including appro- tion improvements. At the national level, one ministry may be priate use and maintenance of the sanitation facilities, is responsible for sewerage and another for low-cost technolo- another possible intervention by the public sector. All of those gies; one may be responsible for construction, another for pro- measures will help increase sanitation coverage and health motion, and a third for enforcing building codes and planning benefits and are appropriate interventions for the health sector. regulations. The costs of several of them are recoverable (after an initial Water Supply, Sanitation, and Hygiene Promotion | 781 launch period) as fees, so that public intervention need not grams also promoted domestic hygiene practices in addition to require public expenditure. the construction and use of latrines. In Bangladesh, all (and in Zimbabwe, most) of the costs of latrine construction were paid Costs of Promotion. The costs of promotion and administra- by the population themselves. tion found in two government-run rural sanitation programs The programs in Bangladesh and Zimbabwe were particu- documented by the World Bank were US$16.80 (Zimbabwe) larly successful and well managed. The promotion cost is taken and $20.00 (the Philippines) per latrine, respectively as US$2.50 per capita for cost-effectiveness calculations, which (Cairncross 1992). Because these costs are largely fixed, the cost is slightly above the higher of the two, to allow for the imper- per unit falls as the number of units built increases. Unit costs fections of sanitation programs in the real world. will therefore be high in relatively unsuccessful programs. Successful programs, on the other hand, often engender the construction of more latrines than they can account for, which Direct Health Benefits also gives an upward bias to the promotional costs per unit Evidence supports the claim that improved excreta disposal built. For example, for every latrine built by Lesotho's rural helps prevent a number of diseases, including diarrhea, intes- sanitation program in the late 1980s, four others were built tinal worm parasites, and trachoma. Of these, the effect that independently but as a result of its promotional activities. accounts for the largest burden of DALYs is that on diarrheal More recently, successful sanitation programs managed by disease. nongovernmental organizations (NGOs) have documented slightly lower unit costs for promotion. For example, the Diarrheal Disease. The effect of sanitation on diarrhea mor- Zimbabwean NGO AHEAD (Applied Health Education and bidity has already been mentioned. Table 41.3 shows the results Development), working through district-level health staff and of Esrey and others' (1991) review, attributing a median reduc- a network of community health clubs, achieved the construc- tion in incidence of 36 percent to sanitation. Although this fig- tion of 3,400 latrines in Makoni district within two years at a ure is the median of the five "better" studies, it must be inter- total promotional cost of US$45,660, or US$13.43 per unit, preted with great care because almost all the known studies on equivalent to US$2.24 per household member served the health effects of sanitation are observational studies that (Waterkeyn 2003). In Bangladesh, WaterAid and its partner, a use self-selected exposure groups. Confounding by a sense of local NGO named VERC (Village Education Resource Centre), hygiene is likely to be a significant problem in any such study. have developed an approach that has successfully achieved 100 From Brazil to Bangladesh, the owners of latrines have been percent sanitation coverage and the elimination of open defe- observed to behave more hygienically than their neighbors in cation in more than 100 villages in six districts at a cost of US$8 practices such as hand washing that are not affected by the per household, or US$1.50 per capita (Allan 2003). Both pro- presence of a latrine (Hoque and others 1995--see table 41.6; Table 41.6 Factors Associated with Hand-Washing Behavior by 90 Women in Bangladesh Hand-washing behavior Ratio of prevalences of observed after defecation good practice (95 percent Associated factor Good Poor confidence interval) Uses own sanitary latrine Yes 22 11 1.73 No 22 35 (1.15­2.59) Uses tubewell water exclusively Yes 18 10 1.53 No 26 36 (1.03­2.29) Owns agricultural land Yes 36 24 2.25 No 8 22 (1.20­4.22) Believes that washing hands prevents diseases Yes 26 27 1.01 No 21 18 (0.66­1.55) Source: Hoque and others 1995. 782 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis Strina and others 2003). It is thus impossible to prove, except own. However, that conclusion is based on only two studies, by an intervention study, that any health benefit associated with and the percentage reductions found in the individual studies latrine ownership is due to the latrine and not to the hygiene of each type of intervention exhibit a wide range. Reflection on habits of latrine owners. how in practice each of the two interventions interrupts the The overall reduction in diarrhea from sanitation quoted by transmission of fecal-oral pathogens would suggest that their Esrey and others (1991) likely disguises considerable hetero- effects would be largely independent: whereas water supply geneity in terms of the context rather than the type of sanita- helps prevent contamination of drinking water, hands, and tion technology. For example, sanitation is likely to have a food, excreta disposal helps prevent contamination of the greater effect on diarrheal disease in high-density urban areas, household yard and surroundings, including children's play where open defecation leads to gross fecal pollution of the areas. Esrey and others (1990) reported three other studies in neighborhood, and less effect in rural communities, where all which sanitation and water supply had a greater effect together but the youngest children use communal defecation sites some than individually, but the reductions in diarrhea incidence in distance away from their homes. those studies could not be calculated. For example, Moraes and others (2003), working in urban For the purpose of burden of disease calculations, therefore, favelas in northeast Brazil, found that diarrhea incidence the effects of water supply and sanitation improvements on among children in households with a toilet was half that in diarrhea are considered here to be independent and additive, households that did not have one. This comparison is likely to which has the advantage of simplicity. be affected by confounding because the households with toilets were a self-selected group. Comparison between communities Effect on Other Disease Categories. The first evidence for the is less likely to be affected by confounding, but Moraes and oth- health benefits of excreta disposal related not to its effect on ers found a greater reduction. The mean incidence of diarrhea diarrheal disease but on intestinal helminths. in young children in communities with sewers was only one- A prolonged series of in-depth studies from 1920 to 1930 by third of that in the communities that, for administrative and researchers of the Rockefeller Foundation established beyond technical reasons, did not have sanitary drainage. doubt that promiscuous defecation, especially in the household Thus, although the quality of the studies reviewed by Esrey surroundings and particularly by children, played a major role and others (1991) was in general poor and the range of reduc- in the transmission of Ascaris spp., Trichuris spp., and hook- tions wide, little doubt exists that excreta disposal can be asso- worms in a range of settings from Panama to China and the ciated with significant reductions in diarrhea morbidity. southeastern United States. By implication, the use of sanitary Studies showing that proximity to open or overflowing sewers toilets should interrupt transmission by that route. (Moraes and others 2003), failure to dispose hygienically of However, more recent attempts to measure the reductions children's stools (Traoré and others 1994), or the presence of in parasite prevalence or intensity attributable to improved excreta on the ground in the household compound (Bukenya sanitation have often suffered from the same shortcomings as and Nwokolo 1991) is a risk factor for fecal-oral infections pro- the studies of their impact on diarrheal disease; many have vide supporting evidence for the likely effect of sanitation been cross-sectional studies and, therefore, subject to con- infrastructure, particularly in urban settings, on diarrheal dis- founding. ease transmission. Esrey and others (1991), in reviewing this literature, found In conclusion, there are some reasons, such as the likeli- that water supply and sanitation reduced the prevalence of hood of confounding, to believe that Esrey and others' (1991) ascariasis by a median of 28 percent (range 0 to 83 percent) and median reduction is an overestimate, but reasons exist also to of hookworm infection by 4 percent (0 to 100 percent). Those believe that the reductions measured were not as great as they reductions are likely caused by the sanitation rather than by the might have been had the provision of sanitation been water-supply improvements. Indeed, three of the nine positive accompanied by hygiene promotion to ensure that the facili- studies of ascariasis and three of the five positive studies of ties were fully and appropriately used (especially by young chil- hookworm involved sanitation alone. It is also likely that the dren) and maintained. A systematic review of the effect of san- effect of excreta disposal on Trichuris infection is similar to that itation on diarrheal disease is urgently required. Meanwhile, on ascariasis (Henry 1981). and on balance, Esrey and others' median reduction of 36 per- Much emphasis has been placed in recent years on cent in diarrhea incidence is the most authoritative estimate chemotherapy as a control intervention for intestinal available. helminths, particularly the chemotherapy of schoolchildren. However, that option is not always sustainable because the Interaction with Water Supply. The results of Esrey and oth- children are quickly reinfected by the eggs and larvae that ers' (1991) review suggest that the effect of water supply and remain in the environment. Sanitation, particularly school sanitation combined is no greater than that of either on its sanitation, has been adopted by the major international donor Water Supply, Sanitation, and Hygiene Promotion | 783 agencies as an integral component of the FRESH (Focusing · a description of the target population (in terms of their level Resources on Effective School Health) framework to ensure its of education and other factors) adequate to permit a judg- sustainability. ment of the relevance of the study to other contexts. A study in Bangladesh (Mascie-Taylor and others 1999) sug- gested that chemotherapy was more cost-effective (though less Only three studies were found to meet all four criteria. One effective) as a helminth control intervention than a health edu- (Stanton and Clemens 1987) dealt with environmental hygiene cation program that included the promotion of sanitation. promotion and raises some doubts--although the hygiene However, the health education program was excessively labor behavior of the intervention group was better than the control, intensive and, therefore, expensive; it involved the constant both were significantly worse than they had been before the deployment of six health educators and a supervisor in each intervention. study area of only 550 households, resulting in a cost of A subsequent review of 31 studies (Cave and Curtis 1999) Tk 1600 (US$30) per household, compared with Tk 330 found 5 more studies that could be considered methodologi- (US$6) per year for chemotherapy. That cost compares with cally sound, but none showed a clear effect on behavior. Of a the total cost of US$8 per family for WaterAid's successful "100 further 11 studies of "reasonable" rigor, only two showed a percent sanitation" approach in rural Bangladesh (Allan 2003). major effect on behavior. Whereas the promotion of sanitation is a one-time cost, the Shortcomings also exist in the cost data. Many costings are cost of chemotherapy is a recurrent annual expenditure. based on budget forecasts and not on real expenditures. Even Allowing for such a sanitation promotion initiative once every when actual expenditures are used, major difficulties exist in five years--and using the chemotherapy costing of Mascie- apportioning the overhead costs that make up a significant Taylor and others (1999)--sanitation promotion is more proportion of the total. Health educators and the resources cost-effective against helminths in Bangladesh than is they use (such as vehicles) are rarely dedicated exclusively to chemotherapy. If the cost were apportioned between the effect health education. A further problem in the derivation of unit on diarrheal disease and the effect on helminths, sanitation costs is agreeing on the denominator, which can be the number would be far more cost-effective than chemotherapy. of people attending health education sessions, the number of Sanitation can also help prevent trachoma. More than members in their households, or the number of people in the 70 percent of the incidence of this infection has been shown to target catchment area. For those reasons, different analysts are be caused by flies, mainly of the species Musca sorbens, which likely to derive different unit costs from the same data; indeed, breeds preferentially in scattered human feces. Pit latrines have the same authors have on occasion arrived at widely differing been shown to reduce the population of these flies by depriving unit cost figures from the same data. them of their breeding sites (Emerson and others 2004). Time adds a further dimension to this discussion. Do inter- ventions to promote hygiene behavior change have to be implemented continuously, or at least annually, if their effect is HYGIENE PROMOTION to be sustained, or are such changes self-sustaining? To a greater degree than with water supply and sanitation, lam- entably little reliable evidence exists on the cost or the effec- Sustainability tiveness of interventions to change hygiene behavior and still We will take the last question first. Wilson and Chandler (1993) less on the relative cost-effectiveness of different approaches to returned after two years to a population in which a four-month the design of such interventions. intervention to promote hand washing with soap had included provision of free soap. They found that 79 percent of mothers, The Shortage of Evidence the original target group, had continued the practice despite With regard to effectiveness, Loevinsohn (1990) reviewed the fact that they now had to buy the soap. health education interventions in developing countries and Further evidence of the sustainability of new hygiene applied four relatively modest criteria of scientific rigor to the behaviors was found by Cairncross and Shordt (2003) in a col- 67 published studies he found: laborative study with partner organizations in six developing countries in Africa and South Asia. Target populations of pre- · a description of the intervention in sufficient detail to allow vious hygiene promotion projects were visited at 12-month its replication intervals, and various indicators of hygiene behavior were · an objective outcome measure, based either on health status assessed and compared. In four of the six countries, indicators or on behavior change for populations in which the intervention had ended relatively · a control group and a sample size greater than two clusters recently were also compared with those in areas where the last or 60 individuals intervention had ended several years previously. Those two 784 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis types of comparison, with the various indicators assessed in Health Clubs approach was examined, it was successful in each country, allowed a total of 46 comparisons to be made. increasing the prevalence of hand washing with soap among Only in three such comparisons was there any indication of a the club members by 6 percent and 37 percent, respectively, and falling-off of hygiene with time since the intervention ended; in it was successful in reducing the prevalence of open defecation one case, the falling-off was attributable to the deteriorating by 29 percent and 98 percent, respectively. The marginal cost of condition of the latrines from wear and tear rather than to a the intervention, which used existing health staff, was US$4.00 decline in compliance. per club member, or an average of US$0.67 per member of an In some cases, new hygiene practices have become stronger affected household. Including the salaries of staff members or more prevalent after the ending of external intervention to would roughly double the figure to about US$1.40 per capita. promote them, as they become self-propagating and consoli- Those figures can be compared with an estimate of US$5.00 dated in the community's material culture (Allan 2003). per mother (in 1982 dollars) by Phillips and others (1987) It is likely that hygiene promotion activities need to be based on a review of several programs. Assuming that roughly repeated from time to time--say, every five years--but are not 1 in 10 members of the population are mothers of young chil- required on a continuous basis. It follows from this observation dren, this cost is equivalent to about US$0.50 per capita. For that calculations of cost-effectiveness should take into account cost-effectiveness analysis, a nominal cost of US$1.00 per cap- the morbidity and mortality averted not only during the imple- ita is, therefore, taken because it is roughly the midpoint of the mentation of the intervention, but also for a number of years-- range of recent estimates. perhaps five--thereafter. Effect on Diarrhea Costs Esrey and others (1991) found only six studies of the effect of Cases in which the costs as well as the effectiveness of hygiene hygiene promotion interventions on diarrhea morbidity, with a promotion programs have been documented objectively are median reduction of 33 percent. A subsequent review by few indeed. In the absence of suitable data, Varley, Tarvid, and Huttly, Morriss, and Pisani (1997) arrived at a similar result-- Chao (1998) calculated a costing for a typical program from a median reduction of 35 percent. first principles, arriving at a cost of US$3 (range US$2 to US$3) The interventions promoting the single hygiene practice of per household per year, or US$0.60 per capita. washing one's hands with soap tended to achieve greater reduc- One of the few cases in which data exist is a program in urban tions in disease than those that promoted several different Burkina Faso described by Borghi and others (2002). Their data behaviors. That finding was confirmed by a systematic review show that the total cost to the provider of the three-year inter- of the literature on hand washing (Curtis and Cairncross vention was US$0.65 per capita, or US$4.54 per seven-person 2003), which concluded that hand washing with soap--and household, after deducting the cost of the international research interventions to promote it--could reduce diarrhea morbidity component. Of this total, 63 percent is composed of adminis- by 43 percent and life-threatening diarrhea by 48 percent. tration and undifferentiated start-up costs of the project. Most Because the effect of diarrhea prevention in DALYs is mainly of the remaining costs were accounted for in roughly equal attributable to the prevention of diarrhea deaths, the higher of measure by house-to-house visits, discussions in health centers, these two figures is more appropriate for calculating the effect hygiene lessons in schools, and street theater presentations. of hygiene promotion on the burden of disease. Additional costs were incurred by the 18.5 percent of house- It is not surprising that interventions advocating more holds that complied, practicing improved hygiene as a result of behavior changes should have less effect, because numerous the program, amounting to US$8 per household per year. messages dilute one another in the minds of the target audi- More than 90 percent of that sum was the cost of soap for hand ence. Because some of the interventions in the systematic washing. review were planned without an adequate prior program of However, on the basis of the observed increase in prevalence formative research, it is possible that they could have had a still of hand washing with soap, the intervention was estimated to greater effect if they were better conceived. have averted sufficient diarrhea morbidity and mortality to save US$2.80 per household per year (US$15 per compliant household per year) in direct costs of medical care and indirect Effect on Respiratory Infections costs attributable to lost productivity. Of this total, 93 percent Reasons exist to believe that hand washing with soap could be represented the lost future productivity associated with the a cost-effective intervention not only against diarrheal diseases, deaths of young children. but also for the prevention of acute respiratory infections Waterkeyn (2003) provides an example from rural (ARIs). The intervention is plausible, given what is known Zimbabwe. In the two districts in which the Community about the transmission routes of ARIs, and there is also Water Supply, Sanitation, and Hygiene Promotion | 785 epidemiological evidence, in that all six published studies of the Table 41.7 Assumed Reductions in Diarrhea Attributable to effect of hand washing on ARIs show a significant reduction Water Supply, Sanitation, and Hygiene Promotion (Cairncross 2003b). Reduction in Corresponding These two disease groups are the most important causes of Intervention diarrhea (percent) relative risk child mortality worldwide, and respiratory infections also Water supply cause significant adult mortality, for which no alternative pre- ventive intervention is yet available, field-tested, and ready for Public source 17 1.20 implementation. A randomized, controlled trial of the efficacy Additional, for house 63 2.70 connection of hand-washing promotion on an ARI outcome is an urgent priority for future research. Excreta disposal 36 1.56 Hygiene promotion 48 1.92 Source: Authors. Interactions with Water Supply and Sanitation It can be argued that there is little point in encouraging people to wash their hands if they do not have access to water or to use These assumptions can be compared as follows with the a latrine if they do not have one. assumptions underlying a previous calculation of the global The argument has only limited validity where sanitation is burden of disease from water, sanitation, and hygiene (Prüss concerned; an important role for any hygiene promotion is to and others 2002; WHO 2002). For that calculation, the follow- promote sanitation itself. With regard to water, in the studies ing seven scenarios were considered: reviewed by Curtis and Cairncross (2003), the reductions in disease achieved by hand washing in settings with indoor VI. No improved water supply or basic sanitation piped water supply were not significantly different from those Va. Basic sanitation only achieved elsewhere. Given that the rationale is ambivalent and Vb. Improved water supply only the evidence inconclusive, the simplest plausible assumption is IV. Improved water supply and basic sanitation that the effects of water supply, sanitation, and hygiene pro- III. Improved water supply and basic sanitation plus house motion on diarrhea are independent and additive to one connection water supply, or improved hygiene or water another. disinfected at point of use II. "Regulated" water supply (presumably house connec- tion) and full sanitation I. Ideal situation, corresponding to absence of disease EFFECT ON BURDEN OF DISEASE transmission through water, sanitation, and hygiene. The effect of water supply, sanitation, and hygiene on the global Scenario II is essentially the position prevailing in industri- burden of disease can now be estimated, in two stages. First, the al countries. Leaving out scenarios I and III, which apply to evidence presented in this chapter is used to arrive at the reduc- only a small proportion of the population, the following sce- tions in diarrhea that are expected to result from the various narios are broadly equivalent to the categories considered combinations and levels of service and that are assumed for the earlier in this chapter: calculation. Then, these figures are applied to the coverage lev- els for individual countries and the burden of diarrheal disease VI. No improved water or sanitation prevailing in the different regions of the world. Because such a Va. Sanitation only calculation has been done before by Prüss and others (2002) Vb. Improved water supply (public source) from rather different premises, it was desirable to examine the IV. Both improved water supply and sanitation comparability of the results. II. House connection water supply, and sanitation. In the Prüss model, the relative risks associated with transi- Assumptions: Reductions in Diarrheal Disease tion from scenarios Va and Vb to VI are taken as 1.26 and 1.60, In summary of the discussion of health effects in this chapter, respectively, comparable with the figures of 1.20 and 1.56 in water supply, sanitation, and hygiene promotion are considered table 41.7. However, Prüss and others (2002) assume equal risks to be associated, under typical conditions, with the reductions in scenarios IV and Va, whereas a relative risk of 1.20 follows in diarrheal disease morbidity shown in table 41.7. These from the assumption in this chapter that the effects of water reductions are considered to be independent of one another, supply and sanitation are independent. The Prüss model so that the relative risks for several interventions can be assumes a relative risk of 1.54 between scenarios III and IV, multiplied. corresponding to the diarrhea reduction of 35 percent from 786 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis hygiene promotion found by Huttly, Morriss, and Pisani To allow for the uncertainty in their assumptions, Prüss and (1997). Scenario III is essentially a theoretical construct, and others (2002) calculated the burden of disease attributable to between it and scenario II a further relative risk of 1.8 is water supply, sanitation, and hygiene using two approaches. assumed (in what Prüss and others term their realistic The realistic approach used the assumptions described above approach), on the basis of some recent trials of home disinfec- and shown in figure 41.3. The minimal approach assumed no tion of water, giving a total of 2.76 between scenarios IV and II. difference in risk between scenarios II and III. Given the ideal The latter figure is close to the corresponding value of 2.70 and hypothetical nature of scenario I and the low probability of implied by the assumptions made here, for different reasons. intensive hygiene promotion being funded for a population Scenario I, like scenario III, is included not because it is preva- that already benefits from high levels of water supply and san- lent in reality, but to illustrate a point. Its equivalent would itation provision, we consider the model on the right of figure be the generalized and effective implementation of a well- 41.3 as optimistic and prefer to take for our more realistic conceived hygiene promotion intervention. Because such approach the less ambitious baseline of house connections and hygiene promotion has hardly ever been provided to whole full sanitation, which approximates the current position in populations, it is similarly hypothetical. From that perspective, most of Western Europe and North America. This approach the corresponding relative risks of 2.5 (Prüss and others 2002) responds to recent calls for "baselines and counterfactuals and 1.92 (table 41.7) are of a similar order of magnitude. which should include alternative, operationalizable policy/pro- The similarity of the two sets of assumptions, based on gram options (including the status quo)" (Ezzati 2003, 458). It rather different premises, is illustrated in figure 41.3. also has the advantage of providing an estimate of burden of Prüss and others` Present model model VI: No water No water supply or supply or sanitation sanitation 1.60 1.26 1.56 1.20 Va: Basic Vb: Improved Basic Improved sanitation water supply sanitation water supply 1.00 1.20 IV: Improved water Improved and water and basic basic sanitation sanitation 1.54 III: Piped water supply or other 2.76 2.70 improvement (1.80) II: House House connection and connection and sanitation sanitation 2.50 (1.92) I: Ideal--no House connection, disease sanitation, and transmission hygiene promotion Source: Authors and Prüss and others 2002. Note: The numbers show relative risk of diarrhea in upper relative to lower boxes. Relative risks in parentheses are set to 1.0 for the minimal version of the Prüss model and for the realistic version of the present model. Figure 41.3 Comparison of Assumptions Made by Prüss and others (2002) and in this chapter. Water Supply, Sanitation, and Hygiene Promotion | 787 disease to which the industrial countries contribute only a the present model. The figures for the burden of disease attrib- negligible amount. utable to deficient water supply, sanitation, and hygiene in the industrial countries of Europe, North America, and the Pacific are very different, but the global totals are remarkably similar. Calculation of Burden of Disease It should be no surprise to find that the attributable burden Prüss and others (2002) worked with water and sanitation cov- in the industrial (that is, low-mortality) countries of Europe, erage data for individual countries (WHO and UNICEF 2000) North America, and the Pacific is zero or very close to zero. The to derive distributions of the population in each region realistic model was deliberately designed to take as its baseline between five of the seven scenarios, as shown in table 41.8. the conditions prevailing in those countries. This finding does They then combined these figures with the relative risks in fig- not mean that no diarrheal disease in those countries can be ure 41.3 and diarrhea incidence and case fatality rates from attributed to deficient water supply, sanitation, or hygiene; Murray and Lopez (1996) to derive estimates of the number of rather, it means that the baseline there is the current condition, DALYs attributable to water supply, sanitation, and hygiene in because no realistic policy option is available to reduce the bur- each region and mortality subregion. The results are shown, for den of such disease in the immediate future. their realistic and minimal models, in the first two columns of Table 41.10 shows the two realistic assessments of DALYs table 41.9. The realistic estimates are those presented in the attributable to water supply, sanitation, and hygiene in terms of World Health Report 2002 (WHO 2002, 225). percentages of the total DALYs in each region and subregion. Using the same spreadsheets but the relative risks on the Again, the two estimates are close. The proportion of the total right of figure 41.3, we derive the results in the third and fourth disease burden attributable to water, sanitation, and hygiene columns of table 41.9 for the optimistic and realistic versions of is greatest in the high-mortality countries of the Eastern Table 41.8 Distribution of the Population between Scenarios of Water Supply and Sanitation Provision (percent) Scenario Region (mortality in children and adults) II IV Va Vb VI African Child high, adult high 0 54 5 6 35 Child high, adult very high 0 42 10 9 38 American (Western Hemisphere) Child very low, adult very low 99.8 0 0 0 0.2 Child low, adult low 0 76 1 9 14 Child high, adult high 0 68 0 7 25 Eastern Mediterranean Child low, adult low 0 83 5 8 4 Child high, adult high 0 66 0 16 18 European Child very low, adult very low 100 0 0 0 0 Child low, adult low 0 79 8 1 12 Child low, adult high 0 94 5 0 1 Southeast Asian Child low, adult low 0 70 3 7 19 Child high, adult high 0 35 0 53 12 Western Pacific Child very low, adult very low 100 0 0 0 0 Child low, adult low 0 42 1 33 24 Source: Prüss and others 2002. 788 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis Table 41.9 Distribution of DALYs Attributable to Diarrhea Caused by Poor Water Supply, Sanitation, and Hygiene by Subregion, According to Various Assumptions (thousands) Region (mortality in WHO 2002 Prüss 2002 Present model Present model children and adults) (realistic) (minimal) (optimistic) (realistic) African Child high, adult high 6,916 6,198 6,747 5,727 Child high, adult very high 11,720 10,473 11,402 9,678 American Child very low, adult very low 61 61 49 1 Child low, adult low 1,290 1,143 1,232 1,009 Child high, adult high 756 673 725 613 Eastern Mediterranean Child low, adult low 629 548 599 482 Child high, adult high 8,303 7,318 7,983 6,653 European Child very low, adult very low 66 66 52 0 Child low, adult low 550 483 528 426 Child low, adult high 121 105 115 91 Southeast Asian Child low, adult low 1,241 1,096 1,195 982 Child high, adult high 18,487 16,595 17,856 15,545 Western Pacific Child very low, adult very low 27 27 21 0 Child low, adult low 3,991 3,574 3,619 3,303 Total, industrial countries 825 742 765 518 Total, developing countries 53,333 47,618 51,358 43,992 Global total 54,158 48,360 52,123 44,510 Source: See Acknowledgments. Mediterranean region, reaching 6 to 7 percent of the total. They The annual costs used for water supply included both the are followed by the high-mortality countries of Southeast Asia amortized construction cost and operation and maintenance and Africa, where the water and sanitation complex accounts costs. Given that investments in water supply and sanitation for 4 to 5 percent of the total. Globally, improvements in water are made largely by other sectors (and for other motives) supply, sanitation, and hygiene could eliminate 3 to 4 percent than health, an alternative cost-effectiveness estimate is made of the global burden of disease. that is based only on the costs of regulation, advocacy, and promotion. The other assumptions used to calculate the cost- Cost-Effectiveness effectiveness of improved water supply--of house connections, The assumptions regarding effect on diarrheal disease are sum- of sanitation, and of hygiene promotion--other than those set marized in table 41.7. Because the effect on diarrheal disease out above, are as described by Varley, Tarvid, and Chao (1998). accounts for the vast majority of the effect, no effort is made to The key parameters are as follows: apportion the costs between their effectiveness in preventing the other diseases affected by water supply, sanitation, and · proportion of population under age five: 17 percent hygiene. The costs derived in this chapter are summarized in · diarrhea incidence: five cases per child under age five per table 41.11. year Water Supply, Sanitation, and Hygiene Promotion | 789 Table 41.10 DALYs Due to Diarrhea Attributable to Poor · median age at onset of disease: 1 year Water Supply, Sanitation, and Hygiene by Subregion, as a · average duration: 8 days Percentage of Total DALYs · case fatality rate: 0.5 percent · coverage by oral rehydration therapy: 30 percent Region WHO 2002 Present model (mortality in children and adults) (realistic) (realistic) · oral rehydration therapy reduction in case fatality rate: 50 percent African Child high, adult high 4.7 3.9 On this basis, we arrived at the cost-effectiveness values in Child high, adult very high 5.6 4.6 table 41.12. American (Western Hemisphere) All of these figures underestimate the cost-effectiveness of Child very low, adult very low 0.1 0.0 investments in water and sanitation, for several reasons: Child low, adult low 1.6 1.2 Child high, adult high 4.3 3.5 · The effects of these interventions on diseases other than Eastern Mediterranean diarrhea have not been taken into account; they seem to be relatively minor for water supply but may be substantial if Child low, adult low 2.7 2.1 hand washing proves to affect ARI. Child high, adult high 7.3 5.9 · Effects on diarrhea mortality, which account for 98 percent European of the DALYs, are likely to be greater than the reductions in Child very low, adult very low 0.1 0.0 morbidity shown in table 41.7. Child low, adult low 1.4 1.1 · The cost figures have generally been taken so as to be suffi- Child low, adult high 0.2 0.2 cient for all contexts, whereas water supply and sanitation Southeast Asian can be implemented more cheaply in favorable settings-- such as where a convenient aquifer or reliable rainfall exists. Child low, adult low 2.0 1.6 · Potential economies exist in combining the interventions; Child high, adult high 5.2 4.3 for example, sanitation promotion can be combined with Western Pacific hygiene promotion and water pipes laid with sewers. Child very low, adult very low 0.2 0.0 · The current global initiative to promote hand washing, Child low, adult low 1.7 1.4 involving commercial marketing expertise, may identify Total, industrial countries 0.4 0.2 more cost-effective approaches to hygiene promotion. Total, developing countries 4.3 3.5 · If a sustainable low-cost sanitation industry can be devel- oped, it will have an interest in promoting its own product. Global total 3.7 3.0 Source: See Acknowledgments. As they stand, the cost-effectiveness values above, except for house connections and construction of latrines, are well below the US$150/DALY cutoff value proposed by the World Bank Table 41.11 Costs Assumed for Cost-Effectiveness Calculations (US$ per capita) Operation and Construction cost Amortization Amortized annual maintenance cost Intervention (US$ per capita) lifetime (years) cost (US$ per capita) (US$ per capita) Water supply House connections 150.00 20 7.50 10.00 Hand pump or standpost 40.00 20 1.00 1.00 Water regulation and advocacy US$0.02 to US$0.10 per capita per year Sanitation 60.00 5 12.00 n.a. Sanitation promotion 2.50 5 0.50 n.a. Hygiene promotion 1.00 5 0.20 n.a. Source: Authors. n.a. not applicable. 790 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis Table 41.12 Cost-Effectiveness of Water Supply, Sanitation, Bukenya, G. B., and N. Nwokolo. 1991. "Compound Hygiene, Presence of and Hygiene Promotion (US$/DALY) Standpipe, and Risk of Childhood Diarrhea in an Urban Settlement in Papua New Guinea." International Journal of Epidemiology 20 (2): Intervention Cost-effectiveness 534­39. Cairncross, S. 1990. "Health Impacts in Developing Countries: New Water supply Evidence and New Prospects." Journal of the Institution of Water and Hand pump or standpost 94.00 Environmental Management 4 (6): 571­77. House connection 223.00 . 1992. "Sanitation and Water Supply: Practical Lessons from the Decade." Water and Sanitation Discussion Paper 9, World Bank, Water sector regulation and advocacy 47.00 Washington, DC. Basic sanitation . 2003a. "Sanitation in the Developing World: Current Status and Construction and promotion 270.00 Future Solutions." International Journal of Environmental Health Research 13 (Suppl. 1): S123­31. Promotion only 11.15 . 2003b. "Handwashing with Soap: A New Way to Prevent ARIs?" Hygiene promotion 3.35 Tropical Medicine and International Health 8 (8): 677­79. Source: Authors. Cairncross, S., and R. Feachem. 1993. Environmental Health Engineering in the Tropics. 2nd ed. Chichester, U.K.: John Wiley & Sons. (1993) as a criterion of cost-effectiveness. Allowing only for Cairncross, S., and J. Kinnear. 1992. "Elasticity of Demand for Water in Khartoum, Sudan." Social Science and Medicine 34 (2): 183­89. the cost component that should fall to the health sector puts Cairncross, S., D. O'Neil, A. McCoy, and D. Sethi. 2003. Health, them all well within this ceiling. For comparison, the cost- Environment, and the Burden of Disease: A Guidance Note. London: effectiveness of promoting oral rehydration therapy, the Department for International Development. principal other measure available to prevent diarrhea mortali- Cairncross, S., and K. Shordt. 2003. "It Does Last! Some Findings from a ty, has been estimated at US$23/DALY. The cost-effectiveness of Multi-Country Study of Hygiene Sustainability." Waterlines 22 (3): 4­7. promoting sanitation and hygiene as derived above (US$11.15 Cave, B., and V. Curtis. 1999. "Effectiveness of Promotional Techniques in and US$3.35, respectively, per DALY) compares favorably with Environmental Health." WELL Study 165, London School of Hygiene that figure. and Tropical Medicine for Department for International Development. Chavasse, D. C., R. P. Shier, O. A. Murphy, S. R. Huttly, S. N. Cousens, and ACKNOWLEDGMENTS T. Akhtar. 1999. "Impact of Fly Control on Childhood Diarrhea in Pakistan: Community-Randomised Trial." Lancet 353 (9146): 22­25. The calculations of the burden of disease were made by Dr. D. Churchill, A. A., D. de Ferranti, R. Roche, C. Tager, A. A. Walters, and A. Campbell-Lendrum, using spreadsheets derived by Annette Yazer. 1987. "Rural Water Supply and Sanitation: Time for a Change." World Bank Discussion Paper 18, Washington, DC, World Bank. Prüss-Üstün. Their collaboration is gratefully acknowledged. Curtis, V., and S. Cairncross. 2003. "Effect of Washing Hands with Soap on Diarrhea Risk in the Community: A Systematic Review." Lancet Infectious Diseases 3 (5): 275­81. REFERENCES Curtis, V., B. Kanki, T. Mertens, E. Traore, I. Diallo, F. Tall, and S. Cousens. 1995. "Potties, Pits and Pipes: Explaining Hygiene Behaviour in Allan, S. 2003. "The WaterAid Bangladesh/VERC 100% Sanitation Burkina Faso." Social Science and Medicine 41 (3): 383­93. Approach; Cost, Motivation and Subsidy." M.Sc. dissertation, London Emerson, P. M., S. Cairncross, R. L. Bailey, and D. C. Mabey. 2000. "Review School of Hygiene and Tropical Medicine. of the Evidence Base for the `F' and `E' Components of the SAFE Altaf, M. A. 1994. "Household Demand for Improved Water and Strategy for Trachoma Control." Tropical Medicine and International Sanitation in a Large Secondary City: Findings from a Study in Health 5 (8): 515­27. Gujranwala, Pakistan." Habitat International 18 (1): 45­55. Emerson, P. M., S. W. Lindsay, N. Alexander, M. Bah, S.-M. Dibba, H. B. Arlosoroff S., G. Tchannerl, D. Gray, W. Journey, A. Karp, O. Langenegger, Faal, and others. 2004. "Role of Flies and Provision of Latrines in and R. Roche. 1987. Community Water Supply: The Handpump Option. Trachoma Control: Cluster-Randomised Controlled Trial." Lancet 363: Washington, DC: World Bank. 1093­98. Bern, C., J. Martines, I. de Zoysa, and R. I. Glass. 1992. "The Magnitude of Esrey, S. A., R. G. Feachem, and J. M. Hughes. 1985. "Interventions for the the Global Problem of Diarrheal Disease: A Ten-Year Update." Bulletin Control of Diarrheal Diseases among Young Children: Improving of the World Health Organization 70 (6): 705­14. Water Supplies and Excreta Disposal Facilities." Bulletin of the World Blum, D., and R. G. Feachem. 1983. "Measuring the Impact of Water Health Organization 63 (4): 757­72. Supply and Sanitation Investments on Diarrheal Diseases: Problems of Esrey, S. A., and J-P. Habicht. 1985. The Impact of Improved Water Supplies Methodology." International Journal of Epidemiology 12 (3): 357­65. and Excreta Disposal Facilities on Diarrheal Morbidity, Growth, and Borghi, J., L. Guinness, J. Ouedraogo, and V. Curtis. 2002. "Is Hygiene Mortality among Children. Cornell International Nutrition Promotion Cost-Effective? A Case Study in Burkina Faso." Tropical Monograph Series 15. Ithaca, NY: Division of Nutritional Sciences, Medicine and International Health 7 (11): 960­69. Cornell University. Briscoe, J., R. G. Feachem, and M. M. Rahaman. 1985. "Measuring the Esrey, S. A., J. B. Potash, L. Roberts, and C. Shiff. 1990. "Health Benefits Impact of Water Supply and Sanitation Facilities on Diarrhea from Improvements in Water Supply and Sanitation: Survey and Morbidity: Prospects for Case-Control Methods." Offset publication Analysis of the Literature on Selected Diseases." WASH Technical WHO/CWS/85.3, World Health Organization, Geneva. Report 66, Environmental Health Project, Rosslyn, VA, for USAID. Water Supply, Sanitation, and Hygiene Promotion | 791 . 1991. "Effects of Improved Water Supply and Sanitation on for the Prevention of Diarrhea. London: Evaluation and Planning Ascariasis, Diarrhea, Dracunculiasis, Hookworm Infection, Centre for Health Care. Schistosomiasis, and Trachoma." Bulletin of the World Health Prüss, A., D. Kay, L. Fewtrell, and J. Bartram. 2002. "Estimating the Burden Organization 69 (5): 609­21. of Disease from Water, Sanitation, and Hygiene at a Global Level." Ezzati, M. 2003. "Complexity and Rigour in Assessing the Health Environmental Health Perspectives 110 (5): 537­42. Dimensions of Sectoral Policies and Programmes." Bulletin of the Rosen, S., and J. R. Vincent. 2001. "Household Water Resources and Rural World Health Organization 81 (6): 458­59. Productivity in Sub-Saharan Africa: A Review of the Evidence." African Feachem, R. G., D. J. Bradley, H. Garelick, and D. D. Mara. 1983. Sanitation Economic Policy Discussion Paper 69, John F. Kennedy School of and Disease: Health Aspects of Excreta and Wastewater Management. Government, Harvard University, Cambridge, MA. Chichester, U.K.: John Wiley & Sons. Stanton, B. F., and J. D. Clemens. 1987. "An Educational Intervention for Feachem, R. G., E. Burns, S. Cairncross, A. Cronin, P. Cross, D. Curtis, and Altering Water-Sanitation Behaviors to Reduce Childhood Diarrhea in others. 1978. Water, Health, and Development: An Interdisciplinary Urban Bangladesh: II. A Randomized Trial to Assess the Impact of the Evaluation. London: Tri-Med Books. Intervention on Hygienic Behaviors and Rates of Diarrhea." American Gutiérrez, G., R. Tapie-Conyer, H. Guiscafré, H. Reyes, H. Martínez, and J. Journal of Epidemiology 125 (2): 292­301. Kumate. 1996."Impact of Oral Rehydration and Selected Public Health Strina, A., S. Cairncross, M. L. Barreto, C. Larrea, and M. S. Prado. 2003. Interventions on Reduction of Mortality from Childhood Diarrheal "Childhood Diarrhea and Observed Hygiene Behavior in Salvador, Diseases in Mexico." Bulletin of the World Health Organization 74 (2): Brazil." American Journal of Epidemiology 157 (11): 1032­38. 189­97. Thompson, J., I. T. Porras, J. K. Tumwine, M. R. Mujwahuzi, M. Katui- Han, A. M., T. Hlaing, M. L. Kyin, and T. Saw. 1988. "Hand Washing Katua, N. Johnstone, and L. Wood. 2001. Drawers of Water II: 30 Years Intervention to Reduce Ascariasis in Children." Transactions of the of Change in Domestic Water Use and Environmental Health in East Royal Society of Tropical Medicine and Hygiene 82 (1): 153. Africa. London: International Institute for Environment and Henry, F. J. 1981. "Environmental Sanitation Infection and Nutritional Development. Status of Infants in Rural St. Lucia, West Indies." Transactions of the Traoré, E., S. Cousens, V. Curtis, T. Mertens, F. Tall, A. Traoré, and others. Royal Society of Tropical Medicine and Hygiene 75 (4): 507­13. 1994. "Child Defecation Behaviour, Stool Disposal Practices, and Hoque, B. A., D. Mahalanabis, M. J. Alam, and M. S. Islam. 1995. "Post- Childhood Diarrhea in Burkina Faso: Results from a Case-Control Defecation Handwashing in Bangladesh: Practice and Efficiency Study." Journal of Epidemiology and Community Health 48 (3): 270­75. Perspectives." Public Health 109 (1): 15­24. Varley, R. C. G., J. Tarvid, and D. N. W. Chao. 1998. "A Reassessment of the Huttly, S. R. A., S. S. Morriss, and V. Pisani. 1997. "Prevention of Diarrhea Cost-Effectiveness of Water and Sanitation Interventions in in Young Children in Developing Countries." Bulletin of the World Programmes for Controlling Childhood Diarrhea." Bulletin of the Health Organization 75 (2): 165­74. World Health Organization 76 (6): 617­31. Jenkins, M. W. 1999. "Sanitation Promotion in Developing Countries: Victora, C. G., P. G. Smith, J. P. Vaughan, L. C. Nobre, C. Lombardi, A. M. Why the Latrines of Benin Are Few and Far Between." Ph.D. thesis, Teixeira, and others. 1988. "Water Supply, Sanitation, and Housing in University of California­Davis, Department of Civil and Relation to the Risk of Infant Mortality from Diarrhea." International Environmental Engineering. Journal of Epidemiology 17 (3): 651­54. Kalbermatten, J. D., D. S. Julius, and C. G. Gunnerson. 1982. Appropriate Wagner, E. G., and J. N. Lanoix. 1959. Water Supply for Rural Areas and Sanitation Alternatives: A Technical and Economic Appraisal. Baltimore: Small Communities. WHO Monograph Series 42. Geneva: World Johns Hopkins University Press. Health Organization. Lanata, C. F. 2003. "Studies of Food Hygiene and Diarrheal Disease." Waterkeyn, J. 2003."Cost-Effective Health Promotion: Community Health International Journal of Environmental Health Research 13 (Suppl. 1): Clubs." In Proceedings of the 29th WEDC Conference, Abuja, Nigeria. S175­83. Loughborough, U.K.: Water Engineering and Development Centre. Loevinsohn, B. P. 1990. "Health Education Interventions in Developing White, G. F., D. J. Bradley, and A. U. White. 1972. Drawers of Water: Countries: A Methodological Review of Published Articles." Domestic Water Use in East Africa. Chicago: University of Chicago International Journal of Epidemiology 19 (4): 788­94. Press. Mascie-Taylor, C. G. N., M. Alam, R. M. Montanari, R. Karim, T. Ahmed, Whittington, D., X. Mu, and R. Roche. 1990. "Calculating the Value of E. Karim, and S. Akhtar. 1999. "A Study of the Cost-Effectiveness of Time Spent on Collecting Water: Some Estimates for Ukunda, Kenya." Selective Health Interventions for the Control of Intestinal Parasites in World Development 18 (2): 269­80. Rural Bangladesh." Journal of Parasitology 85 (1): 6­11. WHO (World Health Organization). 2002. Reducing Risks, Promoting Mertens, T. E., M. A. Fernando, T. F. Marshall, B. R. Kirkwood, Healthy Life: World Health Report 2002. Geneva: WHO. S. Cairncross, and A. Radalowicz. 1990. "Determinants of Water WHO and UNICEF (World Health Organization and United Nations Quality, Availability, and Use in Kurunegala, Sri Lanka." Tropical Children's Fund). 2000. Global Water Supply and Sanitation Assessment Medicine and Parasitology 41 (1): 89­97. 2000 Report. Geneva: WHO with UNICEF. Moraes, L. R. S., J. A. Cancio, S. Cairncross, and S. Huttly. 2003. "Impact of Wilson, J. M., and G. N. Chandler. 1993. "Sustained Improvements in Drainage and Sewerage on Diarrhea in Poor Urban Areas in Salvador, Hygiene Behaviour amongst Village Women in Lombok, Indonesia." Brazil." Transactions of the Royal Society of Tropical Medicine and Transactions of the Royal Society of Tropical Medicine and Hygiene 87 Hygiene 97 (2): 153­58. (6): 615­16. Mukherjee, N. 1990. People, Water, and Sanitation: What They Know, World Bank. 1993. World Development Report 1993: Investing in Health. Believe, and Do in Rural India. New Delhi: National Drinking Water New York: Oxford University Press. Mission, Government of India. World Bank Water Demand Research Team. 1993."The Demand for Water Murray, C. J. L., and A. D. Lopez. 1996. Global Health Statistics. Cambridge, in Rural Areas: Determinants and Policy Implication." World Bank MA: Harvard School of Public Health for WHO and World Bank. Research Observer 8 (1): 47­70. Phillips, M. A., R. G. A. Feachem, and A. Mills. 1987. Options for Diarrhoel Zaroff, B., and D. A. Okun. 1984. "Water Vending in Developing Disease Control: The Cost and Cost-Effectiveness of Selected Interventions Countries." Aqua 5: 284­95. 792 | Disease Control Priorities in Developing Countries | Sandy Cairncross and Vivian Valdmanis Chapter 42 Indoor Air Pollution Nigel Bruce, Eva Rehfuess, Sumi Mehta, Guy Hutton, and Kirk Smith Access to modern energy sources has been described as a energy ladder, carrying out more activities with fuels and "necessary, although not sufficient, requirement for economic appliances that are increasingly efficient, clean, convenient, and and social development" (IEA 2002). It is, therefore, of great more expensive. The pace of progress, however, is extremely concern that almost half the world's population still relies for slow, and for the poorest people in Sub-Saharan Africa and its everyday household energy needs on inefficient and highly South Asia, there is little prospect of change. polluting solid fuels, mostly biomass (wood, animal dung, and Illustrated in figures 42.1 and 42.2 are findings for Malawi crop wastes) and coal. and Peru, respectively, from Demographic and Health Surveys The majority of households using solid fuels burn them in (ORC Macro 2004). The examples are selected from available open fires or simple stoves that release most of the smoke into national studies with data on main cooking fuel use to repre- the home. The resulting indoor air pollution (IAP) is a major sent the situation in poor African and South American coun- threat to health, particularly for women and young children, tries. The main rural and urban cooking fuels are illustrated in who may spend many hours close to the fire. Furthermore, figures 42.1a and 42.2a; the findings are then broken down the reliance on solid fuels and inefficient stoves has other, nationally by level of education of the principal respondent far-reaching consequences for health, the environment, and (woman of childbearing age) in figures 42.1b and 42.2b, and in economic development. urban areas by her level of education in figures 42.1c and 42.2c. Biomass is predominantly, though not exclusively, a rural fuel: indeed, in many poor African countries, biomass is the NATURE, CAUSES, AND BURDEN OF CONDITION main fuel for close to 100 percent of rural homes. Marked socioeconomic differences (indicated by women's education) About 3 billion people still rely on solid fuels, 2.4 billion on exist in both urban and rural areas. During the 1990s, use of biomass, and the rest on coal, mostly in China (IEA 2002; traditional fuels (biomass) in Sub-Saharan Africa increased as Smith, Mehta, and Feuz 2004). There is marked regional varia- a percentage of total energy use, although in most other parts tion in solid fuel use, from less than 20 percent in Europe and of the world the trend has generally been the reverse (World Central Asia to 80 percent and more in Sub-Saharan Africa and Bank 2002). South Asia. In many poorer countries, the increase in total energy use This issue is inextricably linked to poverty. It is the poor accompanying economic development has occurred mainly who have to make do with solid fuels and inefficient stoves, and through increased consumption of modern fuels by better-off many are trapped in this situation: the health and economic minorities. In Sub-Saharan Africa, however, the relative consequences contribute to keeping them in poverty, and their increase in biomass use probably reflects population growth in poverty stands as a barrier to change. Where socioeconomic rural and poor urban areas against a background of weak (or circumstances improve, households generally move up the negative) national economic growth. Reliable data on trends in 793 a. Primary household fuel use in urban and rural areas a. Primary household fuel use in urban and rural areas Wood, straw Wood, straw, dung Charcoal Charcoal Kerosene Kerosene Electricity Electricity, gas 0 20 40 60 80 100 0 20 40 60 80 100 Percentage Percentage Urban Rural Urban Rural b. Primary household fuel use, by level of education of respondent b. Primary household fuel use, by level of education of respondent Wood, straw Wood, straw, dung Charcoal Charcoal Kerosene Kerosene Electricity Electricity, gas 0 20 40 60 80 100 0 20 40 60 80 100 Percentage Percentage Primary or less Secondary or higher Primary or less Secondary or higher c. Primary household fuel use in urban areas, by level of education of c. Primary household fuel use in urban areas, by level of education of respondent respondent Wood, straw Wood, straw, dung Charcoal Charcoal Kerosene Kerosene Electricity Electricity, gas 0 20 40 60 80 100 0 20 40 60 80 100 Percentage Percentage Primary or less Secondary or higher Primary or less Secondary or higher Source: Unpublished data derived from Demographic and Health Survey. Source: Unpublished data derived from Demographic and Health Survey. Figure 42.1 Patterns of Household Fuel Use in Malawi, 2000 Figure 42.2 Patterns of Household Fuel Use in Peru, 2000 household energy use are not available for most countries. to 5 percent in rural areas, and it is expected to reach 36 percent Information is available from India, where the percentage of nationally and 12 percent for rural homes by 2016. rural homes using firewood fell from 80 percent in 1993­94 to International Energy Agency projections to 2030 show that, 75 percent in 1999­2000 (D'Sa and Narasimha Murthy 2004). although a reduction in residential biomass use is expected in Nationally, liquid petroleum gas (LPG) use increased from 9 to most developing countries, in Africa and South Asia the decline 16 percent over the same period, with a change from 2 percent will be small, and the population relying on biomass will 794 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others increase from 2.4 billion to 2.6 billion, with more than 50 per- biomass and coal smoke, see Saksena, Thompson, and Smith cent of residential energy consumption still derived from this (2004). source(OECD and IEA 2004). The number of people without Fewer studies of personal exposure have been done than of access to electricity is expected to fall from 1.6 billion to 1.4 bil- area pollution, mainly because measurement of personal PM lion. Because electricity is used by poor households for lighting typically requires wearing a pump, a cumbersome procedure. and not as a cleaner substitute for cooking, electrification will CO can be measured more easily and has been used as a proxy: not, at least in the short to medium term, bring about time-weighted (for example, 24-hour average) CO correlates substantial reductions in IAP. well with PM if a single main biomass stove is used (Naeher and others 2001). Time-activity and area pollution information can also be combined to estimate personal exposure (Ezzati Levels of Pollution and Exposure and Kammen 2001). These various methods indicate that per- Biomass and coal smoke emit many health-damaging pollu- sonal 24-hour PM10 exposures for cooks range from several tants, including particulate matter (PM),1 carbon monoxide hundred g/m3 to more than 1,000 g/m3 (Ezzati and Kammen 2001), with even higher exposures during cooking (CO), sulfur oxides, nitrogen oxides, aldehydes, benzene, and (Smith 1989). Few studies have measured personal PM expo- polyaromatic compounds (Smith 1987). These pollutants sures of very young children: one study in Guatemala found mainly affect the lungs by causing inflammation, reduced ciliary levels a little lower than those of their mothers (Naeher, clearance, and impaired immune response (Bruce, Perez- Leaderer, and Smith 2000). Padilla, and Albalak 2000). Systemic effects also result, for example, in reduced oxygen-carrying capacity of the blood because of carbon monoxide, which may be a cause of intrauter- ine growth retardation (Boy, Bruce, and Delgado 2002). Health Impacts of IAP Evidence is emerging, thus far only from developed countries, A systematic review of the evidence for the impact of IAP on a of the effects of particulates on cardiovascular disease (Pope wide range of health outcomes has recently been carried out and others 2002, 2004). (Smith, Mehta, and Feuz 2004; see table 42.1). This review Saksena, Thompson, and Smith (2004) have recently com- identified three main outcomes with sufficient evidence to piled data on several of the main pollutants associated with include in the burden-of-disease calculations and a range of various household fuels from studies of homes in a wide range other outcomes with as yet insufficient evidence. of developing countries. Concentrations of PM10, averaged Studies for the key outcomes used in the burden-of-disease over 24-hour periods, were in the range 300 to 3,000 (or more) calculations--acute lower respiratory infection (ALRI), micrograms per cubic meter ( g/m3). Annual averages have chronic obstructive pulmonary disease (COPD), and lung not been measured, but because these levels are experienced cancer--had to be primary studies (not reviews or reanalyses), almost every day of the year, the 24-hour concentrations can written or abstracted in English (and for lung cancer, Chinese), be taken as a reasonable estimate. By comparison, the U.S. that reported an odds ratio and variance (or sufficient data to Environmental Protection Agency's annual air pollution stan- estimate them) and provided some proxy for exposure to dard for PM10 is 50 g/m3, one to two orders of magnitude indoor smoke from the use of solid fuels for cooking and heat- lower than levels seen in many homes in developing countries. ing purposes. During cooking, when women and very young children spend A limitation of almost all studies has been the lack of meas- most time in the kitchen and near the fire, much higher levels urement of pollution or exposure: instead, proxy measures of PM10 have been recorded--up to 30,000 g/m3 or more. have been used, including the type of fuel or stove used, time With use of biomass, CO levels are generally not as high in spent near the fire, and whether the child is carried on the comparison, typically with 24-hour averages of up to 10 parts mother's back during cooking. The studies do not, therefore, per million (ppm), somewhat below the World Health provide data on the exposure-response relationship, although a Organization (WHO) guideline level of 10 ppm for an eight- recent study from Kenya has gone some way to addressing this hour period of exposure. Much higher levels of CO have been omission (Ezzati and Kammen 2001). recorded, however. For example, a 24-hour average of around In some countries, household fuels carry locally specific 50 ppm was found in Kenyan Masai homes (Bruce and others risks. It has been estimated that more than 2 million people in 2002), and one Indian study reported carboxyhemoglobin lev- China suffer from skeletal fluorosis, in part resulting from use els similar to those for active cigarette smokers (Behera, Dash, of fluoride-rich coal (Ando and others 1998). Arsenic, another and Malik 1988). The health effects of chronic exposure of contaminant of coal, is associated with an increased risk of young children and pregnant women to levels of CO just lung cancer in China (Finkelman, Belkin, and Zheng 1999). below current WHO guidelines have yet to be studied. There has been concern, however, that reducing smoke could For additional information on levels of other pollutants in increase risk of vectorborne disease, including malaria. Some Indoor Air Pollution | 795 Table 42.1 Status of Evidence Linking Biomass Fuels and Coal with Child and Adult Health Outcomes Health outcome Age Status of evidence Sufficient evidence for burden-of-disease calculation Acute lower respiratory infections Children 5 years Strong. Some 15­20 observational studies for each condition, from developing Chronic obstructive pulmonary disease Adult women countries. Evidence is consistent (significantly elevated risk in most though not all studies); the effects are sizable, plausible, and supported by evidence from outdoor Lung cancer (coal exposure) Adult women air pollution and smoking. Chronic obstructive pulmonary disease Adult men Moderate-I. Smaller number of studies, but consistent and plausible. Lung cancer (coal exposure) Adult men Not yet sufficient evidence for burden-of-disease calculation Lung cancer (biomass exposure) Adult women Moderate-II. Small number of studies, not all consistent (especially for asthma, Tuberculosis Adult which may reflect variations in definitions and condition by age), but supported by studies of outdoor air pollution, smoking, and laboratory animals. Asthma Child and adult Cataracts Adult Adverse pregnancy outcomes Perinatal Tentative. Adverse pregnancy outcomes include low birthweight and increased Cancer of upper aerodigestive tract Adult perinatal mortality. One or a few studies at most for each of these conditions, not all consistent, but some support from outdoor air pollution and passive-smoking Interstitial lung disease Adult studies. Ischemic heart disease Adult Several studies from developed countries have shown increased risk for exposure to outdoor air pollution at much lower levels than IAP levels seen in developing countries. As yet, no studies from developing countries. Source: Smith, Mehta, and Feuz 2004. studies have shown that biomass smoke can repel mosquitoes countries with a 1999 per capita gross national product (GNP) and reduce biting rates (Palsson and Jaenson 1999; Paru and greater than US$5,000 had made a complete transition either others 1995; Vernede, van Meer, and Alpers 1994). Few studies to electricity or cleaner liquid and gaseous fuels or to fully have examined the impact of smoke on malaria transmission: ventilated solid fuel devices. To account for differences in one from southern Mexico found no protective effect of smoke exposure caused by variation in the quality of stoves, they (adjusted odds ratio 1.06 [0.72­1.58]; Danis-Lozano and applied a ventilation factor (VF), set from 1 for no ventilation others 1999), and another from The Gambia found that wood to 0 for complete ventilation. In China, a VF of 0.25 was used smoke did not protect children in areas of moderate transmis- for child health outcomes and 0.5 for adult outcomes, reflect- sion (Snow and others 1987). ing a period of higher exposure (to open fires) before the widespread introduction of chimney stoves. Countries with a 1999 GNP per capita greater than US$5,000 were assigned a Method Used for Determining Attributable Disease Burden VF of 0, and all other countries a value of 1, reflecting the very low rates of use of clean fuels or effective ventilation tech- Smith, Mehta, and Feuz (2004) have provided a full explana- nologies. The authors obtained the final point estimate for tion of the calculation of the disease burden associated with exposure by multiplying the percentage of solid fuel use by the IAP. Summarized here are the methods they used to estimate VF. They arbitrarily assigned an uncertainty range of 5 per- the two most critical components of these calculations: the cent to the estimates. number of people exposed and the relative risks. Exposure. The absence of pollution or exposure measure- Risk. Smith, Mehta, and Feuz (2004) carried out meta- ment in health studies required use of a binary classification: analyses for the three health outcomes with sufficient evidence the use or nonuse of solid fuels. The authors obtained esti- (table 42.2). They used fixed-effects models and sensitivity mates of solid fuel use for 52 countries from a range of analysis that took account of potential sources of heterogene- sources, mostly household surveys, and statistical modeling ity, including the way in which exposure was defined and was used for countries with no data (the majority) (Smith, whether adjustment had been made for confounders (Smith, Mehta, and Feuz 2004). They assumed, conservatively, that all Mehta, and Feuz 2004). 796 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others Table 42.2 Summary of Relative Risk Estimates for Health ages count less than years lost in the prime of adult life. Age Outcomes Used in Burden-of-Disease Estimates weighting makes little difference to the DALYs lost per death up to age five; how much it affects the DALY cost of adult deaths 95 percent depends on the age distribution of deaths from COPD. Because Health Age and Number of Relative confidence outcome sex group studies risk interval these are likely to occur at age 45 or beyond, the DALY losses are underestimated compared with estimates without age weight- ALRI Children 5 years 8 2.3 1.9­2.7 ing that follow the usual practice in this volume. COPD Women 30 years 8 3.2 2.3­4.8 Men 30 yearsa 2 1.8 1.0­3.2 Lung cancer Women 30 years 9 1.9 1.1­3.5 Other Effects of Household Energy Use (coal) Men 30 years 3 1.5 1.0­2.5 in Developing Countries Sources: Smith, Mehta, and Feuz 2004. A number of other health impacts--for example, burns from a. Because of the limited quantity and quality of available evidence, the male COPD relative risk and range have been fixed to include 1.0 (no effect) as the lower estimate. open fires--were not assessed because the burden-of-disease assessment process allowed inclusion of only those health effects resulting directly from pollution. Children are at risk of burns and scalds, resulting from falling into open fires and The Burden of Disease from Solid Fuel Use knocking over pots of hot liquid (Courtright, Haile, and Kohls Information on the proportions exposed and risk of key disease 1993; Onuba and Udoidiok 1987). Modern fuels are not always outcomes was combined with total burden-of-disease data to safe either, because children are also at risk of drinking obtain the population attributable fractions associated with kerosene, which is often stored in soft drink bottles (Gupta and IAP (WHO 2002b). Globally, solid fuels were estimated to others 1998; Reed and Conradie 1997; Yach 1994). account for 1.6 million excess deaths annually and 2.7 percent Families--mainly the women and children--can spend of disability-adjusted life years (DALYs) lost, making them the many hours each week collecting biomass fuels, particularly second most important environmental cause of disease, after where environmental damage and overpopulation have made contaminated water, lack of sanitation, and poor hygiene them scarce. This time could be spent more productively on (table 42.3). Approximately 32 percent of this burden (DALYs) child care and household or income-generating tasks. There are occurs in Sub-Saharan Africa, 37 percent in South Asia, and also risks to health from carrying heavy loads and dangers from 18 percent in East Asia and the Pacific. In developing countries mines, snake bites, and violence (Wickramasinghe 2001). with high child and adult mortality, solid fuel use is the fourth Inefficient stoves waste fuel, draining disposable income if fuel most important risk factor behind malnutrition, unsafe sex, is bought. Although women carry out most of the household and lack of water and sanitation, and it is estimated to account activities requiring fuels, they often have limited control over for 3.7 percent of DALYs lost (WHO 2002b). how resources can be spent to change the situation (Clancy, Overall, there are more female deaths but similar numbers Skutsch, and Batchelor 2003). These conditions can combine of male and female DALYs (table 42.3b). The reason can be to restrict income generation from home-based activities that found by looking further at the health outcomes. Deaths and require fuel energy (for example, processing and preparing DALYs from ALRI in children under five years of age are food for sale). slightly greater for males (table 42.3c). Women experience Homes that are heavily polluted and dark can hinder pro- twice the DALYs and three times the deaths from COPD (male ductivity of householders, including children doing homework smoking-attributable COPD deaths excluded). Far fewer cases and others engaged in home-based income-generating activities of lung cancer are attributable to IAP, but women experience such as handicrafts. In many poor homes, lighting is obtained about three times the burden of men. from the open fire and simple kerosene wick lamps, which pro- Table 42.3 also shows how the poorest regions of the world vide poor light and add to pollution. carry by far the greatest burden, particularly for ALRI. More Solid fuel use has important environmental consequences. than half of all the deaths and 83 percent of DALYs lost attrib- Domestic use of solid fuels in high-density rural and urban utable to solid fuel use occur as a result of ALRI in children environments contributes to outdoor air pollution. Many low- under five years of age. In high-mortality areas, such as Sub- income urban populations rely on charcoal, the production of Saharan Africa, these estimates indicate that approximately which can place severe stress on forests. The use of wood as fuel 30 percent of mortality and 40 percent of morbidity caused by can contribute to deforestation, particularly where it is com- ALRI can be attributed to solid fuel use, as can well over half of bined with population pressure, poor forest management, and the deaths from COPD among women. Because they derive clearance of land for agriculture and building timber. Damage from WHO risk assessments, these estimates include age to forest cover can increase the distance traveled to obtain wood weights, such that years of life lost at very young or advanced and can result in the use of freshly cut (green) wood, dung, and Indoor Air Pollution | 797 Table 42.3 Deaths and DALYs Lost Because of Solid Fuel Use a. Overall Total Deaths DALYs burden World Bank region (thousands) (thousands) (percent) East Asia and the Pacific 540 7,087 18.4 Europe and Central Asia 21 544 1.4 Latin America and the Caribbean 26 774 2.0 Middle East and North Africa 118 3,572 9.3 South Asia 522 14,237 36.9 Sub-Saharan Africa 392 12,318 32.0 World 1,619 38,532 100.0 b. All causes, by sex Deaths (thousands) DALYs (thousands) World Bank region Male Female All Male Female All East Asia and the Pacific 152 388 540 3,028 4,060 7,087 Europe and Central Asia 9 13 21 251 293 544 Latin America and the Caribbean 12 14 26 368 405 774 Middle East and North Africa 57 61 118 1,849 1,724 3,572 South Asia 218 304 522 6,641 7,596 14,237 Sub-Saharan Africa 211 181 392 6,901 5,417 12,318 World 658 961 1,619 19,037 19,495 38,532 c. From ALRI (children under age five) Deaths (thousands) DALYs (thousands) World Bank region Male Female All Male Female All East Asia and the Pacific 40 41 81 1,502 1,535 3,036 Europe and Central Asia 7 6 13 235 204 439 Latin America and the Caribbean 8 7 15 324 281 605 Middle East and North Africa 51 44 95 1,794 1,571 3,365 South Asia 177 178 355 6,228 6,278 12,506 Sub-Saharan Africa 198 153 351 6,777 5,191 11,967 World 481 429 910 16,860 15,058 31,918 d. From COPD (men and women 30 years and over) Deaths (thousands) DALYs (thousands) World Bank region Male Female All Male Female All East Asia and the Pacific 105 338 443 1,461 2,430 3,891 Europe and Central Asia 2 7 9 16 89 104 Latin America and the Caribbean 4 7 11 44 125 168 Middle East and North Africa 6 17 23 55 153 208 South Asia 41 126 167 410 1,314 1,724 Sub-Saharan Africa 13 28 41 124 227 351 World 171 522 693 2,110 4,336 6,446 Source: Modified by authors to World Bank regions, from Smith, Mehta, and Feuz 2004. 798 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others twigs, which are more polluting and less efficient. In some urban women in implementing changes, can promote gender equality communities, poverty and supply problems are resulting in the and empower women (Goal 3). Household energy interven- use of plastic and other wastes for household fuel (IEA 2002). tions can also contribute to eradicating extreme poverty Stoves with inefficient combustion produce relatively more (Goal 1) through health improvements, time saving, and better products of incomplete combustion, such as methane, which environments for education and facilitating income generation have a markedly higher global-warming potential than carbon (WHO 2004a). dioxide (Smith, Uma, and others 2000). It has, therefore, been argued that, although the energy use and greenhouse gas emis- sions from homes in developing countries are small relative to Interventions the emissions generated in industrial countries, cleaner and Although the main focus of this chapter is IAP, the many other more efficient energy systems could provide the double benefit ways in which household energy can affect health and develop- of reduced greenhouse gas emissions (with opportunities for ment emphasize why interventions should aim to achieve a carbon trading) and improved health through reduced IAP range of benefits, including the following: (Wang and Smith 1999). The evidence available for assessing these effects, which · reduced levels of IAP and human exposure together could have a substantial influence on health and eco- · increased fuel efficiency nomic development, is patchy at best. This area is important · reduced time spent collecting fuel and using inefficient for research (Larson and Rosen 2002). stoves · reduced stress on the local environment · increased opportunities for income generation INTERVENTIONS AND POLICY · contribution to an overall improvement in the quality of the home environment--in particular, the working environ- The uses of energy in the home--for example, for cooking ment and conditions for women. and keeping warm and as a focus of social activities--have important attributes that are specific to the locality, culture, Interventions for reducing IAP can be grouped under three and individual households and are often associated with headings: those acting on the source of pollution, those improv- established traditions and deeply held beliefs. Encouraging the ing the living environment (aspects of the home), and changes use of cleaner and more efficient energy technologies by pop- to user behaviors (table 42.4). ulations that are among the poorest in the world has not been It should not be assumed that an intervention that reduces easy, but recent years have seen progress being made with IAP will necessarily achieve other aims listed previously. For respect to suitable technology that meets the needs of house- example, in colder areas, an enclosed stove with a flue that holds and with respect to the development of supportive reduces IAP may reduce radiant heat and light, forcing house- policy. holds to use other fuels for those purposes. If not addressed with households, such problems may well result in disappoint- Poverty Reduction and the Millennium Development Goals ing reductions in IAP exposure, poor acceptance of interven- tions, and lack of motivation to maintain them. Given the close relationship between socioeconomic condi- tions and solid fuel use, poverty reduction must be a key ele- ment of policy to alleviate IAP. The United Nations Policy Instruments Millennium Development Goals set targets for poverty eradi- cation, improvements in health and education, and environ- Although a range of interventions is available, poor households mental protection; they represent the accepted framework for face many barriers to their adoption, and enabling policy is the world community to achieve measurable progress (United needed (table 42.5). This area of practice is complex and evolv- Nations Statistics Division 2003). Although reducing IAP ing, often requiring solutions that are highly setting specific. can contribute to achieving a number of these goals, it is par- ticularly relevant to reducing child mortality (Goal 4) from ALRI. INTERVENTION COSTS AND EFFECTIVENESS Goal 7, Target 9, aims at integrating sustainable develop- ment into country policies and programs. The proportion of The cost-effectiveness analysis discussed in this chapter is based population using solid fuels has been adopted as an indicator on recent work by Mehta and Shahpar (2004). The key compo- for Target 9. Alleviating drudgery resulting from collecting fuel nents of this analysis are described here, with particular and using inefficient stoves, together with the involvement of emphasis on the underlying assumptions. Indoor Air Pollution | 799 Table 42.4 Interventions for Reducing Exposure to IAP Source of pollution Living environment User behaviors Improved cooking devices Improved ventilation Reduced exposure through operation of source · Improved biomass stoves without flues · Hoods, fireplaces, and chimneys built into the · Fuel drying · Improved stoves with flues attached structure of the house · Using pot lids to conserve heat · Windows and ventilation holes (such as in Alternative fuel-cooker combinations · Properly maintaining stoves and chimneys and roof), which may have cowls to assist other appliances · Briquettes and pellets extraction · Charcoal Reductions by avoiding smoke Kitchen design and placement of the stove · Kerosene · Keeping children away from smoke--for · Kitchen separate from house to reduce example, in another room (if available and safe · Liquid petroleum gas exposure of family (less so for cook) to do so) · Biogas, producer gas · Stove at waist height to reduce direct exposure · Solar cookers (thermal) of cook leaning over fire · Other low-smoke fuels · Electricity Reduced need for the fire · Insulated fireless cooker (haybox) · Efficient housing design and construction · Solar water heating Source: Modified from Ballard-Tremeer and Mathee 2000. Table 42.5 Policy Instruments for Promoting Implementation of Effective Household Energy Interventions Policy instruments Examples Applications Information, Schools Learning about household energy, health, and development should be integrated in school curricula, education, and particularly in countries where these topics are a priority for health and economic development. This goal communication can be achieved through programs such as the WHO Global School Health Initiative, which promotes environmental health education, including education about IAP. Media Local and national radio, television, and newspapers can be used to raise awareness and disseminate information on technologies and opportunities to support implementation, such as promotions and microcredit. These media can be directed at a range of audiences, including decision makers, professionals, and the public where radio is widely used. Community education Opportunities such as adult literacy programs can be used to raise awareness and share experience of interventions, and innovative methods can be used (for example theater). Taxes and subsidies Tax on fuels and Reduced tax on fuels and appliances may promote development of distribution networks and uptake, and it appliances may be seen as efficient if there is evidence of health, education, and economic benefits. Subsidy on fuels and General (for example, national) subsidies on fuels such as kerosene have been applied to promote use by poor appliances households. Subsidies have been found to be inefficient instruments, however, often benefiting the better off rather than the poor. Time-limited subsidy on specific products (for example, clean fuel appliances, connection to grid) may be a useful method for promoting initial uptake, generating demand, and thereby providing market conditions for lower prices and more consistent quality. Regulation and Air quality standards Although some developing countries have air quality standards for urban air, none have them for indoor air in legislation settings where solid fuels are widely used. Routine monitoring and enforcement is not practical, but it may be useful to set standards and targets linked to specific assessments. For more routine use, information from censuses and surveys, such as fuel type, stove type, and venting for smoke, offers a practical alternative for setting air quality standards for IAP in developing countries. Design standards for Design standards can be applied to safety (prevention of burns, gas leaks, and explosions); venting of appliances emissions; and efficiency. Although such standards may be difficult to enforce in an informal economy, they could become valuable with wider-scale production. Direct expenditures Public program Large-scale public provision of appliances, such as improved stoves or clean-fuel appliances, has generally provision of appliances been found unsuitable. Some form of targeted provision or partial subsidy where households have made informed choices and commit to cost sharing may be useful to stimulate demand and act in favor of equity. 800 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others Table 42.5 Continued Policy instruments Examples Applications Funding of finance Experience has shown that credit is most likely to be made available and adopted for energy applications schemes that contribute directly to productive, income-generating activities (such as food processing for sale). Meeting everyday cooking and space-heating needs is seen as a lower priority. Good opportunities may exist where biomass fuel is purchased and where cost saving combines with other valued benefits, such as increased prestige and cleaner kitchens. Support for such schemes, mainly in the form of raising awareness, skills training in managing funds, and seed funding (the main source of funds being from users) may be cost-effective. Research and Surveys Surveys of fuel and appliance use, knowledge of risks to health, willingness to pay for interventions, development knowledge of and confidence in credit schemes, and the like are important for planning interventions. Development and Evaluation of interventions should be conducted in a range of settings, using harmonized methods, if evaluation of possible, that allow local flexibility but permit comparison with other types of interventions and other interventions locations. Studies of health Stronger and better-quantified evidence of the effects on health of reducing IAP, which includes exposure effects measurement, is required not only for key outcomes such as ALRI, but also for other health outcomes for which evidence is currently tentative. Research capacity Capacity for carrying out a wide range of research--from national and local surveys, to monitoring and development evaluation of interventions, to more complex health studies--requires strengthening in those countries where the problems associated with household energy and IAP are most pressing. Source: Authors. Costs Experience indicates that successful interventions are sustain- Intervention costs have a number of components, the relative able in local markets, implying that the consumer pays the importance of which will vary with the type of fuel and device majority of initial and recurrent costs. The contributions of (box 42.1). the government, utilities, nongovernmental organizations The level of costs incurred by consumers and others, includ- (NGOs), and the commercial sector will depend on many fac- ing government, depends not only on the type of intervention tors, including the type of intervention and fuel, location but also on how it is delivered, supplied, and adopted. (urban or rural), existing level of supply and distribution Box 42.1 Cost Components for Household Energy Interventions · Fuels, which vary from zero (in direct cash terms,though · Additional appliances--for example, an LPG storage not in opportunity cost) for collected biomass to a bottle has a moderately high initial cost but should last U.S. dollar or so per week for kerosene and several for many years. U.S. dollars per week for electricity (where used for · Maintenance costs,which vary from zero for a three-stone cooking). fire up to modest, but not negligible, costs of repairing · Stove appliances, which vary from zero for a simple (and periodically replacing) woodstoves and chimneys. three-stone fire (stones arranged on the floor to sup- Appliances for using kerosene, LPG, and electricity also port cooking pots, with the fire lit between the require maintenance and periodic replacement. stones), to US$50 (and in some cases more than · Program costs, which apply to various aspects of US$100) for a good-quality woodstove with a chim- provision of energy services, particularly LPG and elec- ney and up to several hundred U.S. dollars for a bio- tricity, but may also include costs of, for example, gas installation. establishing more sustainable biomass reserves and administrative costs. Source: Authors. Indoor Air Pollution | 801 Box 42.2 Cost Issues in Switching to Cleaner Fuels for a "Typical" Poor Kenyan Family Ruth1 and her family live 3 kilometers from a small town have been able to find the money to buy the gas bottle and on the main road about one hour by bus from Kisumu. cooker. They are subsistence farmers, with a small income from She talks with her husband about LPG, and although selling vegetables, from irregular laboring work obtained quite supportive, her husband thinks they cannot afford it. by her husband, and from making and selling handicrafts. They could spend a little more on fuel, but income is Ruth, a mother of five, cooks over a three-stone fire using irregular. Why abandon free fuel when they are so poor? mostly wood, which she collects every other day from Ruth thinks she could earn more money from her handi- plots up to two hours walking distance from home. She crafts in the time she saves collecting wood. On balance, spends 8 to 12 hours each week collecting wood. Ruth and they reckon they could probably afford the cost of the gas her family use about 2 liters of kerosene each week for if they could be sure of more regular income, but they do wick lamps and for cooking. They use dry cell batteries for not know where they could find the money to pay for the the radio; grid electricity runs nearby, but connection is cooker and bottle. far too expensive. In all, the family spends an equivalent of Ruth then learns about a revolving fund set up by her US$1 to US$2 per week on fuel and batteries. women's group with the help of an NGO. If she can make Through her women's group, Ruth hears that a few small regular payments, she and her husband could get a families are using LPG, now available at a nearby petrol loan to buy the stove and gas bottle next year. But they station. The women say it is very quick and easy to use, have never saved before, and what if they need money for and it keeps pots, clothes, and walls clean. The women and medicines or for the children at school? Will they be able children seem to feel better, with less cough, runny eyes, to keep saving each week to make sure they have enough and headaches. But those families run small shops and to refill the gas bottle when needed? 1. Not her real name. Source: Authors. networks, and support for credit (for example, seed funds and Box 42.2 illustrates how these various issues can influence the fund capital) and targeted subsidies. decisions of a "typical" poor rural African household consider- Some degree of market support may be required to stimu- ing transition from gathered biomass to predominant use of a late demand and to encourage adoption by poor households, commercial fuel (LPG). particularly those using three-stone fires (and other simple stoves) and collected biomass, because those methods do not incur direct monetary costs. Some countries have applied sub- Effectiveness sidies on fuels such as kerosene to assist poor families, but Most evidence available for assessing intervention effectiveness general subsidies are now considered to be an inefficient instru- deals with the effect on IAP levels and in some cases personal ment for this purpose (von Schirnding and others 2002). exposure. No experimentally derived evidence is available, Targeted subsidy and small-scale credit may be more appropri- however, on the effect of reducing IAP exposure on incidence ate ways of helping poor families acquire new household of ALRI or the course of COPD in adults. A randomized trial energy technologies and can have low default rates. Experience of an improved chimney stove is currently under way in shows, however, that households are more likely to access cred- Guatemala, focusing on ALRI in children up to 18 months of it for directly productive (with regard to income) uses of age (Dooley 2003). A cohort study in Kenya by Ezzati and energy, rather than for everyday cooking and space-heating Kammen (2001) describes significant exposure-response rela- needs. Because the latter are the most important sources of IAP, tionships for all acute respiratory infections--and for ALRI more promotion of other benefits is needed, such as improved specifically--associated with the use of traditional and family health; fuel cost savings; time saved by faster cooking improved woodstoves and charcoal. However, those effect esti- and reduced need for biomass; greater prestige; and cleaner mates require confirmation because the study has small num- homes, clothes, and utensils. A number of these benefits may bers of children (93 children under age five, living in 55 homes). result in reduced expenditure or increased income generation. For the other major health outcome, lung cancer, Lan and 802 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others others (2002) reported adjusted hazard ratios of 0.59 (95 per- Personal exposures were usually found to have been reduced cent confidence interval: 0.49 to 0.71) for men and 0.54 (0.44 proportionately less than area pollution levels. For example, in to 0.65) for women using improved coal stoves compared with Kenya, where hoods with flues achieved a 75 percent reduction traditional open coal fires in a 16-year retrospective cohort in 24-hour mean kitchen PM3.5 and CO, the woman's mean study in rural China. 24-hour CO exposure was reduced by only 35 percent (Bruce Measuring evidence on reductions in pollution and expo- and others 2002). Similar results were found for child expo- sure is nonetheless an important step in assessing effectiveness. sures in a study of improved wood stoves in Guatemala (Bruce Summarized here are the main findings of studies that have and others 2004). We are aware of only one study that has used measured pollution levels in homes using traditional open direct measurement of personal particulate exposure in very fires, various improved stoves, kerosene, and LPG (see also young children (Naeher, Leaderer, and Smith 2000). This study, Saksena, Thompson, and Smith 2004) and one that examined also in Guatemala, reported mean 10- to 12-hour (daytime) the effect of rural electrification in South Africa (Rollin and PM2.5 levels for children under 15 months of age of 279 g/m3 others 2004). ( SD of 19.5) for the open fire and 170 g/m3 ( 154) for the plancha stoves, a 40 percent reduction. Effect of Improved Stoves. In East Africa, cheap improved stoves without flues, burning either wood or charcoal, are pop- Impact of Cleaner Fuels. Good evidence shows that kerosene ular. These wood-burning stoves can reduce kitchen pollution and LPG can deliver much lower levels of pollution, although it by up to 50 percent, but levels still remain high (Ezzati, is important to determine the extent to which the cleaner fuel Mbinda, and Kammen 2000). Charcoal emits much less PM is substituting for biomass. For example, a study in rural (but with a higher CO-to-PM ratio than wood), and stoves Guatemala comparing LPG with open fires and plancha chim- such as the Kenyan jiko yield particulate levels in the region of ney stoves found that LPG-using households typically also used 10 percent of those from wood fires. an open fire for space heating and cooking with large pots. As a In a number of Asian and Latin American countries, result, the plancha stoves achieved the lowest pollution levels in improved stoves with flues have been promoted quite exten- that setting (Albalak and others 2001). Still, a number of stud- sively, although many such stoves are found to be in poor con- ies, mainly from India, show that introducing kerosene and dition after a few years. Some studies from India have shown LPG dramatically reduces kitchen pollution, which perhaps minimal or small reductions in PM (Ramakrishna 1988; Smith, reflects different cooking requirements and less need for space Aggarwal, and Dave 1983). Other studies, from Nepal, have heating. In rural Tamil Nadu, two-hour (mealtime) kitchen res- shown reductions of about two-thirds, although the very high pirable PM levels of 76 g/m3 using kerosene and of 101 g/m3 baseline levels mean that homes with stoves still recorded total using gas contrasted with levels of 1,500 to 2,000 g/m3 using suspended particulate values of 1,000 to 3,000 g/m3 during wood and animal dung (Parikh and others 2001). Personal cooking (Pandey and others 1990; Reid, Smith, and Sherchand (cook) 24-hour exposure to respirable PM was 132 g/m3 with 1986). Results from Latin American countries are similar, the use of kerosene as opposed to 1,300 and 1,500 g/m3, respec- although the IAP levels are generally lower. Studies have shown tively, with the use of wood and dung (Balakrishnan and others that plancha-type stoves (made of cement blocks, with a metal 2002). Other studies confirm those findings, for example, with plate and flue) reduce PM by 60 to 70 percent and by as much the use of gas in Mexico (Saatkamp,Masera,and Kammen 2000). as 90 percent when they are in good condition. Typical 24-hour Delivering electricity to rural homes requires extensive infra- PM levels (PM10, PM3.5 [respirable], and PM2.5 have variously structure, and most poor people with access to electricity can been reported) with open fires of 1,000 to 2,000 g/m3 have afford to use it only for lighting and running low-demand elec- been reduced to 300 to 500 g/m3, and in some cases to less trical appliances. Without marked improvements in socioeco- than 100 g/m3 (Albalak and others 2001; Brauer and others nomic conditions, electrification has little potential to bring 1996; Naeher, Leaderer, and Smith 2000). One study from about substantial reductions in IAP. South Africa is one of the Mexico found little difference between homes with open fires few countries with a large rural population traditionally and with improved stoves (Riojas-Rodriguez and others 2001), dependent on biomass that has the resources for rural electrifi- but the 16-hour levels of PM10 at about 300 g/m3 with open cation. An investigation of three rural villages with similar fires were relatively low. socioeconomic characteristics, two not electrified and one elec- Improved stoves with flues have so far had little success in trified, in the North West province found that 3.6 years (aver- Sub-Saharan Africa, although recent work developing hoods age) after connection to the grid, 44 percent of the electrified with flues for highly polluted Kenyan Masai homes reported homes had never used an electric cooker (Rollin and others reductions in 24-hour mean respirable PM of 75 percent from 2004). Only 27 percent of electrified homes cooked primarily more than 4,300 g/m3 to about 1,000 g/m3 (Bruce and with electricity; the remainder used a mix of electricity, others 2002). kerosene, and solid fuels. Despite the mixed fuel use, households Indoor Air Pollution | 803 cooking with electricity had the lowest pollution levels. Overall, Costs were estimated separately for cleaner fuel and homes in the electrified village had significantly lower 24-hour improved stove programs, using an "ingredients" approach mean respirable PM and CO levels and significantly lower mean (Johns, Baltussen, and Hutubessy 2003) and a costing template 24-hour CO exposure for children under 18 months of age than developed by WHO (2003). In summary, all the ingredients-- homes in the nonelectrified villages. including administrative, training, and operational costs-- necessary to set up and maintain a given program must be added Effect of Other Interventions. Little systematic evaluation has up. For regional estimates, costs of all traded goods were in U.S. been made of other interventions listed in table 42.4. dollars, whereas nontraded (local) costs were estimated in local Investigation of the potential of improving ventilation has, currency and converted to U.S. dollars using relevant exchange overall, shown that although enlarging eaves can be quite effec- rates. All costs were annualized using a 3 percent discount tive (Bruce and others 2002), removing smoke generally rate. Costs for tradable goods are scaled, using region-specific requires a well-functioning flue or chimney. Behavioral standardized price multipliers to reflect the increasing costs of changes are currently the subject of an intervention study in expanding coverage caused by higher transportation costs to South Africa (Barnes and others 2004a, 2004b). more remote areas (Johns, Baltussen, and Hutubessy 2003). Price multipliers were not applied to improved stoves because they tend to be manufactured locally with mainly local materials. Cost-Effectiveness Analysis Program costs were found to make up a small proportion of the Although clean fuels can be expected to have a greater health overall intervention costs. Savings from averted health care costs effect than improved stoves (even those with flues), clean fuels are not included; because many of these cases currently go may be too expensive and inaccessible for many poor commu- untreated, it can be argued that including treatment costs could nities over the short to medium term. Furthermore, even result in inflated cost-effectiveness ratios (CERs). though clean fuels may be the best longer-term goal, an inter- mediate stage of improved biomass stoves may promote change Effectiveness and Health Outcome Assumptions. For this by raising awareness of benefits and thus creating demand by analysis, cleaner fuels are assumed to remove exposure com- improving health, saving time, and mitigating poverty. For pletely, whereas improved stoves are assumed to reduce expo- those reasons, this cost-effectiveness analysis (CEA) examines sure by 75 percent (ventilation factor of 0.25). The effect on both improved biomass stove and clean fuel options in the health of the exposure reduction will vary from region to following scenarios: region, because it depends on current levels of exposure as well as region-specific rates of morbidity and mortality. A number · access to improved stoves (stoves with flues that vent smoke of assumptions have been made about households in carrying to the exterior), with coverage of 95 percent out analyses at the regional level. First, regional estimates of · access to cleaner fuels (LPG or kerosene), with coverage of household composition (numbers of people, by age group and 95 percent sex) and, hence, the effect of interventions on exposure and · part of the population with access to cleaner fuels (50 per- health apply at the level of individual households. Second, the cent) and part with improved stoves (45 percent). age distribution of household members is similar in exposed and nonexposed groups; for example, the number of children In each case, the intervention is compared with the current per household is the same irrespective of household fuel use level of coverage of the respective technology or fuel. and ventilation characteristics. That assumption is likely to be conservative, since poorer, more polluted homes will typically Cost Assumptions. The assumptions for costs include pro- have higher fertility and more children under five; all other fac- gram costs, fixed costs (including stoves), and recurrent fuel tors being equal, such households would therefore experience a costs. Household costs for each region were drawn from the higher burden of disease from IAP exposure. most comprehensive estimates available in the literature (von The health outcomes included are ALRI and COPD, because Schirnding and others 2002; Westoff and Germann 1995). For they were responsible for nearly all of the 1.6 million deaths LPG, costs include the initial price of a cooker and cylinder and attributable to IAP. The risk estimates used are those derived the recurrent refill costs. Assumed household annual costs, dis- from the meta-analyses, as summarized in table 42.2. Smoking counted at 3 percent, range from US$1 to US$10 for improved is an important confounding variable for COPD, particularly stoves and from US$3 to US$4 for kerosene or up to US$30 for with men, because they generally smoke more than women do LPG. Recurrent costs of fuel were found to be the most signif- in developing countries. At present, information is sparse on the icant cost for the cleaner fuel interventions. Wood fuel costs are independent effect of solid fuel use on COPD in the presence of estimated at US$0.25 per week and assumed to be the same for smoking. To avoid possible overestimation of the impact of IAP traditional and improved stoves. on COPD, attributable fractions for COPD from solid fuel use 804 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others were applied to disease burdens remaining after removal of 50 percent). In Latin America and the Caribbean, kerosene has smoking-attributable burdens (Ezzati and Lopez 2004). Current the most favorable CER, followed by kerosene in combination estimates of exposure are used in combination with estimates of with improved stoves. When the 50 percent and 80 percent cov- disease burden to obtain region-specific disease burdens for erage scenarios are compared, large differences in the ratio are exposed and unexposed populations. Regional patterns of dis- seen in regions where coverage for that intervention is already ease for 2000 have been used, including incidence, mortality, substantial, and there is much less health gain at lower levels of remission, duration, and case-fatality rate, obtained from WHO coverage. Where no result is given, the specified coverage of the (2004b). In contrast to the estimates of burden in table 42.3, no intervention has already been reached. age weighting has been used in the cost-effectiveness analysis. Multivariate sensitivity analysis was conducted to assess the Health impacts are discounted at 3 percent. effect of uncertainty in cost and effectiveness estimates. Costs were assumed to vary with a standard deviation of 5 percent, Implementation Period. The implementation period is and effectiveness by the range of the confidence interval 10 years, although effects have been evaluated over 100 years in around the relative risk for each health endpoint. Results for order to approximate the benefits for an entire population Southeast Asia are shown in figure 42.3: the "clouds," or uncer- cohort. Thus, health effects are calculated for a cohort with a tainty regions, illustrate the range of possible point estimates typical age structure for the population concerned that experi- emerging from the sensitivity analysis. This example is repre- ences the intervention for 10 years. It is assumed that after sentative because other regions show essentially similar results. 100 years, all of the cohort (including children born during the Despite the uncertainty, the ranking of the interventions 10-year implementation period) will have died. remains the same (Mehta and Shahpar 2004). The implementation period has critical implications, partic- ularly in situations in which it takes several years to establish an Discussion. Results of this cost-effectiveness analysis indicate intervention (for example, developing local markets for clean- that an improved biomass stove is the most cost-effective inter- er fuel), in which there are high start-up costs, and in which vention for South Asia and Sub-Saharan Africa, the two regions disease prevention is experienced in the distant future. This is with the highest solid fuel­related disease burden. This finding especially true for chronic health effects (for example, COPD) is important given International Energy Agency projections to that result from exposure over many years. If the intervention 2030, which indicate that biomass will remain the principal is implemented and exposure reduced for only 10 years, the household fuel for the poor in South Asia and Sub-Saharan disease burden is effectively deferred by 10 years, whereas Africa and that actual numbers of users will increase over that longer-term implementation would result in many more cases period (IEA 2002). Cleaner fuels (particularly kerosene) are the being averted. For this analysis, using the 10-year intervention most cost-effective options for East Asia and the Pacific, the scenario specified for the Disease Control Priorities Project-2, other region with a high burden of solid fuel­related disease. incident cases are deferred by 10 years. For COPD, it is assumed Cleaner fuels, in particular LPG, appear relatively costly for that reduced exposure results in a milder form of COPD, South Asia and Sub-Saharan Africa, but circumstances in indi- accounted for by using a lower severity weighting. vidual countries may vary considerably and in ways that make Findings for Cost-Effectiveness Analysis. Findings from the this fuel much more cost-effective. Sudan, for example, has CEA are expressed, for the four intervention scenarios with dif- abundant cheap supplies of LPG and favorable excise arrange- fering coverage (50 percent, 80 percent, 95 percent), by region, ments for imported appliances, which would result in a lower as (a) total healthy years gained in each region, (b) CERs in U.S. CER for LPG than in other countries in the region. Furthermore, dollars per healthy year gained, and (c) healthy years gained per as will be discussed later, costs and benefits from the user's per- US$1 million (table 42.6). For all regions, the cleaner fuels yield spective will differ markedly, depending on whether the starting the greatest gain in healthy years, but improved stoves also have point is free fuel collection or purchased biomass fuel. a significant effect. The largest total population gains in healthy In interpreting the results, one should bear in mind the years are in Sub-Saharan Africa and South Asia for all types of assumptions underlying the CEA. Much of the evidence indi- interventions and in East Asia and the Pacific (mainly China) cates that, although improved biomass stoves may reduce for cleaner fuels. kitchen pollution by up to 75 percent, the reduction in exposure In the two regions with the largest burden of disease attrib- of women and children is typically no more than 30 to 40 percent utable to solid fuel use (Sub-Saharan Africa and South Asia), (equivalent to VF of 0.6 to 0.7). Achievement of the 75 percent CERs are lowest (most favorable) for improved stoves, reduction in exposure (VF 0.25) assumed for this analysis is although in both regions kerosene has CERs just over twice consistent only with well-designed and -maintained chimney those of improved stoves. In East Asia and the Pacific, kerosene stoves that meet most of the cooking and heating energy needs is most cost-effective, followed by improved stove and of the household and high population coverage (to avoid expo- clean fuel combinations and then by LPG (for coverage over sure from neighbors and others). Those conditions may be Indoor Air Pollution | 805 Table 42.6 Intervention Scenarios for World Bank Regions a. Healthy years gained Sub- Latin America Middle East East Asia Coverage Saharan and the and North Europe and and the Intervention (percent) Africa Caribbean Africa Central Asia South Asia Pacific LPG 50 22,160,000 160,000 n.a. n.a. 44,810,000 2,560,000 80 60,370,000 4,670,000 15,570,000 1,330,000 149,300,000 228,710,000 95 75,630,000 11,260,000 22,510,000 4,810,000 184,940,000 568,640,000 Kerosene 50 22,160,000 160,000 n.a. n.a. 44,810,000 2,560,000 80 60,370,000 4,670,000 15,570,000 1,330,000 149,300,000 228,710,000 95 75,630,000 11,260,000 22,520,000 4,810,000 184,940,000 568,640,000 Improved stove 50 18,010,000 n.a. n.a. n.a. 48,880,000 1,120,000 80 40,270,000 1,380,000 6,630,000 n.a. 101,670,000 6,980,000 95 51,540,000 2,600,000 11,640,000 n.a. 128,380,000 32,760,000 Combined (with stove) LPG 69,250,000 8,650,000 19,540,000 3,230,000 170,340,000 427,350,000 Kerosene 69,250,000 8,650,000 19,540,000 3,230,000 170,340,000 427,350,000 b. Cost-effectiveness ratios (US$ per healthy year gained) Sub- Latin America Middle East East Asia Coverage Saharan and the and North Europe and and the Intervention (percent) Africa Caribbean Africa Central Asia South Asia Pacific LPG 50 715 1,405 n.a. n.a. 542 1,695 80 518 783 756 1,221 312 115 95 518 814 762 1,321 314 100 Kerosene 50 84 631 n.a. n.a. 63 225 80 60 115 95 183 36 14 95 60 106 95 167 36 12 Improved stove 50 25 n.a. n.a. n.a. 15 297 80 21 947 457 n.a. 13 587 95 20 1,101 368 n.a. 13 327 Combined (with stove) LPG 295 761 606 1,375 177 83 Kerosene 45 296 220 507 26 25 c. Healthy years gained per US$1 million Sub- Latin America Middle East East Asia Coverage Saharan and the and North Europe and and the Intervention (percent) Africa Caribbean Africa Central Asia South Asia Pacific LPG 50 1,400 710 n.a. n.a. 1,840 590 80 1,930 1,280 1,320 820 3,210 8,680 95 1,930 1,230 1,310 760 3,190 10,040 Kerosene 50 11,970 1,580 n.a. n.a. 16,000 4,440 80 16,600 8,690 10,500 5,470 27,850 72,840 95 16,620 9,470 10,560 6,000 27,680 85,840 Improved stove 50 39,640 n.a. n.a. n.a. 67,330 3,360 80 47,940 1,060 2,190 n.a. 74,750 1,700 95 49,510 910 2,720 n.a. 76,300 3,060 Combined (with stove) LPG 3,390 1,310 1,650 5,660 5,660 12,020 Kerosene 22,250 3,380 4,550 38,590 35,590 40,730 Source: Authors. n.a. not applicable because the specified coverage of the intervention has already been reached. 806 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others Average annual cost in millions of international dollars household budgets change with various interventions and 25,000 approaches to implementation. The calculations have been undertaken for whole regions 20,000 and provide no indication of how CERs differ among countries and specific communities. As local data on exposure, risk 15,000 factors, health outcomes, and intervention effectiveness become available, similar analyses should be conducted at 10,000 national and subnational levels. Averted treatment costs have not been included on the 5,000 grounds that most users of solid fuel are poor and have limited access to health services; many do not seek medical care for ALRI 0 and even fewer do so for COPD.Inclusion of averted costs would 0 50 100 150 200 250 300 increase cost-effectiveness. However, efforts to raise awareness Average annual gain in healthy years in millions about health risks and the importance of seeking care for ALRI Propane and LPG Kerosene Improved stoves (and COPD), which should accompany an intervention pro- Combination: LPG and improved stoves gram, may increase care seeking and costs to the consumer. As Combination: kerosene and improved stoves more complete information becomes available, future CEAs Source: Mehta and Shahpar 2004. should include treatment costs, with the option of allowing for an increase in care seeking associated with the intervention. Figure 42.3 Multivariate Sensitivity Analysis for Three Types of Interpretation of the results of this CEA, particularly with Interventions and Combined Intervention Scenarios, Southeast Asia respect to comparisons with other types of intervention, needs Region to acknowledge that, although public organizations and other agencies will (or may) have some involvement in funding achievable and should be the goal, but they are not currently intervention programs, most of the cost of market-based inter- widespread. The relative cost-effectiveness advantage for ventions will be borne by households and those involved in improved stoves over cleaner fuel reported here should there- production and marketing. Furthermore, it is hoped that, in fore be viewed as relating more to what might be achievable addition to reducing IAP, interventions (and the means of with good biomass stoves rather than to what is currently being accessing them) will have other positive effects, including on achieved. The assumption that kerosene and LPG are equally household budgets, in creating opportunities for income gen- clean and achieve zero exposure (VF = 0) presumes, at the very eration and empowering women in decisions about how least, the use of high-quality kerosene fuel and pressurized energy is used. The promotion of market-based solutions burners. In many places, kerosene is of low quality, and the types implies new opportunities for artisans and entrepreneurs, but of kerosene stoves and lamps used result in poor combustion. also the loss of traditional employment. The balance sheet for Cost comparisons for the various fuels also need careful interventions is therefore complex, is specific to the setting, and consideration. For example, the cost of solid fuel has been will evolve as markets and enterprise develop. assumed to be constant for traditional open fires and improved stoves. As a general assumption this is reasonable, because the efficiency of "improved" stoves varies, and some may even be Cost-Benefit Analysis less fuel efficient than are open fires. However, new stove tech- The CER gives cost per unit of health gained (healthy year) nology is markedly improving efficiency, and some designs based on reduced risk of specified disease outcomes (ALRI, reduce daily fuel consumption by 40 percent or more, resulting COPD). As discussed earlier, however, household energy inter- in savings of time (where fuel is collected) and money (where ventions can affect a wide range of social, economic, and envi- fuel is bought) (Boy and others 2000). ronmental issues, with important implications for health and Transition from biomass (collected free) or charcoal (typi- development. In an economic analysis of water and sanitation cally paid for daily in small amounts) to LPG would almost interventions, Hutton and Haller (2004) found that time saved certainly require changes in saving and budgeting habits for a was the most important benefit. Those other effects cannot poor household (see also box 42.2). Those changes may entail easily be expressed in units of health gain. Cost-benefit analysis arranging a loan to purchase the gas bottle and stove and saving (CBA) offers an alternative approach that may be better suited money for the relatively large, periodic outlay to refill the cylin- to environmental health interventions, given that health argu- der. Such changes are very likely to have other consequences ments alone will not motivate the multiple sectors involved in for the family that should not be overlooked. However, those financing and implementing household energy interventions. consequences are complex and difficult to allow for within the All main benefits in CBA are expressed in a common unit of current CEA framework. Empirical data are required on how monetary value and compared with costs in the cost-benefit Indoor Air Pollution | 807 ratio (CBR). The assessment of costs in CBA would have many sustainability and health, which would be experienced after assumptions and methods in common with CEA. The key dif- many years and by subsequent generations. Larson and Rosen ferences lie in the selection of effects for inclusion as benefits (2002) used a mix of valuation of statistical life and contingent and the methods for valuing them. In principle, there is no rea- valuation methods to examine the CBRs for improved stoves son all the full range of effects discussed earlier could not be with respect to mortality (Guatemala, East Africa) and mor- included (table 42.7), although in practice some, such as glob- bidity (Pakistan), concluding that ratios appeared favorable. al climate effects, may be too uncertain. Where disadvantages of Although they discuss other benefits, those benefits were not interventions are identified, they should also be included. included in their valuations. Their observation that the favor- Benefit valuation presents particular challenges: effects are able CBRs are not reflected in the generally low adoption of highly setting specific; evidence for some is limited, and their improved stoves led them to conclude that the information effects poorly quantified; and valuation in monetary units of required for assessing household demand correctly is not cur- benefits, such as lost working time averted for women, is diffi- rently available. cult because women frequently are unpaid or work in informal markets. As a result, methods of valuation based on human IMPLEMENTATION OF CONTROL STRATEGIES: capital may not be suitable, and alternative approaches such as LESSONS FROM EXPERIENCE contingent valuation, in which communities are involved in agreeing on market values for nontradable commodities, may The past 30 to 40 years have seen many diverse programs on be preferable. A related issue is valuing benefits that relate to household energy, from small-scale NGO- and community-led Table 42.7 Possible Data Requirements for Quantifying Benefits Impact category Variables or elements to identify Direct benefits related to specific health outcomes Expenditure and time for health care­seeking Health service use of those with diseases caused by IAP (number of cases, visits or days per case) Health service use of those having accidents or injuries due to reasons related to fuel use: Direct: burns, poisoning Indirect: injuries in collecting fuel Access features to get to health services (distance, mode of transport, time; average visits per case) Other consumption related to health care­seeking Time loss of seeking health care, both of the patient and of those accompanying patient Other direct benefits in and around the home Time gained owing to less illness and death Activities of those with diseases caused by IAP Impact of disease on activities (time input, productivity) Value of time of various occupations Time saving of changed technology Reduced time spent collecting fuel Reduced time spent cooking and on other tasks requiring fire or stove Value of time of various occupations Income-generating activities achieved through increased time Impact on household cleanliness and hygiene and need for cleaning Change in household environment and production Effect of improved lighting on evening activities (education, production) Effect of availability of electricity and other fuels on household production activities Impact on ergonomics related to cooking Consequences of process of acquiring new Increased confidence in capacity of the household to save for immediate or future needs technology and related changes More involvement of women in decision making with respect to changes in household energy use and related issues Indirect benefits related to the environment Local environment Impact of fuel scarcity on local environment, average fuel collection time Increased risk of environmental effects (such as soil fertility) or disasters (such as flooding, landslides) Global environment Contribution of local area to greenhouse gases Source: Authors. 808 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others initiatives to ambitious national programs, the largest of which sumers purchasing stoves at close to full cost (Smith and has been the installation of some 200 million improved stoves others 1993). Among the key features of the Chinese program in rural China. Although few have been subjected to rigorous reported to have contributed to its success are decentralization evaluation, an assessment has been made of the Indian national of administration; a commercialization strategy that provided stove program (box 42.3; ESMAP and World Bank 2001); the subsidies to rural energy enterprise development and quality Chinese national stove program (box 42.4; Sinton and others control through the central production of critical components, 2004; Smith and others 1993); and LPG promotion (box 42.5; such as parts of the combustion chamber; and engagement of UNDP and ESMAP 2002). Experience with a number of smaller local technical institutions in modifying national stove designs initiatives has also been reported--for example, the ceramic to local needs. National-level stove competitions were held and metal stoves in East Africa, which have proved popular and among counties for contracts, ensuring local interest and provided local employment (Njenga 2001), and improved allowing the best-placed counties to proceed first; financial stove interventions in Guatemala (UNDP and ESMAP 2003). payments were provided to counties only after completion of Implementation of the Chinese national program differed an independent review of their achievements. No large flows substantially from that in India. Although the Chinese rural of funds came from the central government (in contrast, for populations concerned are poor, they do have greater effective example, to India); local governments provided the major purchasing power than the poor in many developing countries, financial contributions. As a result, delays and other problems allowing development of a program with the majority of con- associated with transferring large amounts of money have been Box 42.3 Key Features and Lessons from India's National Stove Program The Indian National Programme of Improved Cookstoves much less. Evaluation of the program identified four main was established in 1983 with goals common to many such problems: initiatives: · Most states placed inadequate emphasis on commer- · conserving fuel cialization, now seen as crucial for effective and · reducing smoke emissions in the cooking area and sustainable uptake. improving health conditions · Overall, there was insufficient interaction with users, · reducing deforestation self-employed workers, and NGOs, so the designs did · limiting the drudgery of women and children and not meet needs of households, and there was very poor reducing cooking time acceptance of user training. · improving employment opportunities for the rural · Quality control for installation and maintenance of the poor. stove and its appropriate use was lacking. · High levels of subsidy (about 50 percent of the stove Although the Ministry of Non-Conventional Energy cost) were found to reduce household motivation to Sources was responsible for planning, setting targets, and use and maintain the stove. approving stove designs, state-level agencies relayed this information to local government agencies or NGOs. A Some more successfully managed areas of the program technical backup unit in each state trained rural women or focused resources on technical assistance, research and unemployed youths to become self-employed workers to development, marketing, and information dissemination. construct and install the stoves. Recently, the government of India decentralized the pro- Between 1983 and 2000, the program distributed more gram and transferred all implementation responsibility to than 33 million improved stoves. Despite extensive gov- state level. Since 2000, the program promotes only durable ernment promotion efforts, improved stoves now account cement stoves with chimneys that have a minimum life for less than 7 percent of all stoves. Among those that have span of five years. The introduction of these stoves will been adopted, poor quality and lack of maintenance have make adhesion to technical specifications and quality con- resulted in a life span of two years at most and typically trol much easier. Source: Authors, based on ESMAP and World Bank 2001. Indoor Air Pollution | 809 Box 42.4 Household Effects of China's National Improved Stove Program In 2002, an independent multidisciplinary evaluation was not among households using combinations of fuels undertaken by a team of U.S. and Chinese researchers to that included coal or LPG), improved stoves showed evaluate (a) implementation methods used to promote significantly lower PM4 and CO concentrations than improved stoves; (b) commercial stove production and traditional stoves. marketing organizations that were created; and (c) effects · In both children and adults, coal use was associated with of the program on households, including health, stove per- higher levels of exposure (as measured by CO in exhaled formance, socioeconomic factors, and monitoring of breath) and improved biomass stoves with lower levels. indoor air quality. The first two objectives were assessed Reported childhood asthma and adult respiratory dis- through a facility survey of 108 institutions at all levels. The ease were negatively associated with use of improved third objective was assessed through a household survey of stoves and good stove maintenance. These results nearly 4,000 households in three provinces: Zheijang, should, however, be treated as indicative because of Hubei, and Shaanxi. Key findings were as follows: limited sample size. · The household survey revealed highly diverse fuel Overall, several important conclusions emerge with usage patterns: 28 and 34 different fuel combinations relevance to future improved stove programs: were used in kitchens in winter and summer, respec- tively. Most households owned at least one or more · A wide range of combinations of different fuel and coal and one or more biomass stoves. Of the biomass stove types may limit the effect of an improved stove stoves 77 percent, but only 38 percent of the coal stoves, program. were classified as improved. On average, improved · Given the importance of space heating, making avail- stoves had a mean efficiency of 14 percent, which is able an improved biomass stove for cooking may not be well below the program target of between 20 and 30 per- a sufficient strategy to reduce IAP. Improved coal stoves cent, but above the mean efficiency of 9 percent for need to be promoted among rural Chinese households. traditional stoves. · Even among households using improved stoves, PM4 · With respect to air quality (measured with PM4, the and CO levels were higher than Chinese national "thoracic fraction" of particulate matter, and CO), coal indoor air standards, implying that a large fraction of stoves showed significantly higher concentrations than China's rural population is still chronically exposed to biomass stoves during the summer, but not during the pollution levels substantially above those determined winter. Among households using biomass fuels (but by the Chinese government to harm human health. Source: Authors, based on Sinton and others 2004. avoided. The Chinese program succeeded in shifting norms: awareness of costs and benefits. McDade (2004) has recently most biomass stoves now available on the market have flues identified a number of key lessons emerging from experience and other technical features that classify them as improved. with the promotion of LPG markets (box 42.5). Experience in the promotion of LPG has also been reported, Electrification has an important role in development (IEA for example, from the Indian Deepam Scheme (ESMAP and 2002). Evidence from South Africa suggests that communities World Bank 2004; UNDP and ESMAP 2002) and from the LPG with grid access experience lower IAP exposure (Rollin and Rural Energy Challenge (UNDP 2005). The latter initiative, others 2004). Electricity is not expected to bring about large developed by UNDP and the World LPG Association in 2002, reductions in IAP exposure in most low-income countries, is promoting the development of new, viable markets for LPG however, because most poor households can afford it only for in developing countries. Key elements include developing part- uses such as lighting and running entertainment appliances and nerships in countries; enabling regulatory environments that not for cooking and space heating. The International Energy facilitate LPG business development and product delivery; Agency has recently carried out a detailed review of electrifica- reducing barriers, for example, by introducing smaller (more tion, including the issues involved in supply and cost recovery affordable) gas bottles; and raising government and consumer among poor (and especially rural) communities (IEA 2002). 810 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others Box 42.5 Key Lessons Learned in the Promotion of New Markets for LPG in Developing Countries · LPG can be affordable outside of urban areas, where expansion. These issues need to be addressed by raising wood fuel is currently purchased. On the other hand, awareness among consumers and strengthening regula- for many consumers who do not participate in the tory environments. monetized economy, it will be premature to promote · Appliances for a range of end uses required by con- LPG markets. sumers must be available. · One-time subsidies on appliances could be a good use · Government leadership is essential, backed up by of government (or other) resources. policy that sets the basic parameters for successful mar- · Microcredit initiatives should emphasize the cost- ket expansion and avoids conflict between, for exam- saving and productive potential and should seek to ple, subsidies on competing fuels that undermine package both the gas (and appliances) and the efforts to promote LPG markets. financing. · Specific initiatives, such as integrated energy centers (as · Concerns about safe handling, cylinder refilling, and in Morocco and South Africa) offer an effective means transportation can be serious barriers to market of developing markets in rural areas. Source: Authors, based on McDade 2004. The key lessons from experience with interventions to date Management Assistance Programme (UNDP and ESMAP may be summarized as follows: 2003) found that, despite the almost total reliance of the rural population on biomass, a marked lack of national policy, lead- · Too often, intervention technologies have been developed ership, and coordinated action existed in relation to household without adequate reference to users' needs and, as a result, energy. Countries need to develop mechanisms for action and have been poorly used and maintained or abandoned. coordination in light of local needs, available institutional Consequently, it is important to involve users--particularly capacity, and leadership potential. women--in assessing needs and developing suitable interventions. · Sustainable adoption should also be promoted through greater availability of a choice of appropriately priced inter- THE RESEARCH AND DEVELOPMENT AGENDA ventions through local commercial outlets (artisans, shops, WHO has, through a process involving multistakeholder meet- markets). This situation will come about only if demand is ings and reviews, developed some consensus on research and sufficient and if producers and distributors recognize this development priorities for household energy, IAP, and health demand. (see for example WHO 2002a). Effective coordination is a pre- · All too commonly, communities most at risk exhibit low requisite because of the need for input from, and collaboration awareness, low demand, and poverty (often extreme between, many different organizations and "actors" that have poverty). A combination of user involvement and market generally not previously worked in partnership on this issue. approaches is needed, supported by the promotion and One recent response to this need has been the establishment availability of targeted subsidies or microcredit facilities or of the Partnership for Clean Indoor Air, following the both. The nature and extent of such financial support should Johannesburg World Summit on Sustainable Development in depend on the purchasing power of the community. 2002 (EPA 2004; http://www.pciaonline.org/). · Local initiatives such as those outlined above must be led by The evidence base on health effects requires further national (and subnational) policy that acknowledges the strengthening, particularly to quantify the effect of a measured contributions of a range of actors (government, business, reduction in IAP exposure on the risk of key outcomes (for NGOs, and so on) and sectors (energy, health, environment, example, ALRI). A randomized controlled trial is currently finance, and so on) and that results in coordinated action. under way in Guatemala, focusing primarily on ALRI in chil- The instruments listed in table 42.5 should be considered dren up to 18 months of age (Dooley 2003); however, at least when developing national policy. one other such trial on another continent would be desirable. In a recent review of the situation in Guatemala, the United Also required are observational studies for outcomes for Nations Development Programme and Energy Sector which few studies currently exist, including tuberculosis, low Indoor Air Pollution | 811 birthweight and perinatal mortality, cataracts, asthma, and as well as through work on regional and national indica- cardiovascular disease. A small number of such studies are in tors conducted under the Global Initiative on Children's progress, but further effort is required, with perinatal outcomes Environmental Health Indicators (WHO 2004e). Future being a particular priority. reporting will need to be further refined by taking into account Despite limitations in the evidence on health effects, what is differences in cooking practices (for example, type of stove and known about the health, social, and economic consequences of cooking location), as well as in fuel use for lighting and heating. current patterns of household energy use in poor countries is Advocacy for stronger action, internationally and in coun- of sufficient concern to press ahead with an active program tries, is required. Products and guidance for a range of audiences of research and development regarding interventions. This should be prepared, with clear messages on the extent of the activity should address both the technology (and associated problem, the population groups most affected, what works, and knowledge and behavior) and the approaches taken for imple- what should be avoided. Tools such as the recently published mentation. Although some development and innovation in guidelines on estimating the national burden of disease from technology and fuels (for example, clean fuels derived from solid fuels will help provide local evidence to argue for greater biomass) are likely to be valuable, the single greatest challenge attention and action (Desai, Mehta, and Smith 2004). is to promote wider access to--and adoption of--existing knowledge and interventions. Projects and programs currently CONCLUSIONS in progress or being developed should be carefully evaluated using quantitative and qualitative methods to assess a range of IAP from solid fuel use is responsible for a large burden of dis- effects. Work is currently under way to develop suitable meth- ease among the world's poorest and most vulnerable popula- ods and tools for this purpose (WHO 2005). Experience and tions. Inefficient and polluting household energy systems hold lessons learned need to be disseminated widely to ensure that back development through resulting ill health, constraints on they reach governments, donors, researchers, NGOs, and com- women's time and income generation, environmental impacts, munities. As part of this effort, WHO is developing a resource and other factors. Although there is a trend toward cleaner and for countries that offers information on the effectiveness of more efficient energy with increasing prosperity, little improve- interventions as well as the enabling factors that facilitate long- ment is in prospect for more than 2 billion of the world's poor- term, sustained adoption and use of suitable improved tech- est people, particularly in South Asia and Sub-Saharan Africa. nologies in different settings (WHO 2004c). The number of people relying on traditional biomass is Economic assessment, including cost-effectiveness analysis, actually expected to increase until 2030. has a valuable part to play. Critical issues resulting from limited Although the development of new energy technologies has evidence have been identified about estimations and assump- a part to play in addressing this problem, many effective inter- tions for costs, exposure reductions, health effects, and averted ventions are already available. The single greatest challenge is treatment costs, as well as the current inability to assess national to dramatically increase the access of poor households to and subnational cost-effectiveness. CBA may be more suitable cleaner and more efficient household energy systems. Much for interventions in this and similar areas but will require better valuable experience has been gained from successful--and description of environmental, social, and economic effects and unsuccessful--programs in household energy over the past further development of valuation methods. New health studies three to four decades. Despite this experience, coherent, and broadly based evaluations of interventions should help fill evidence-based policy is lacking in most of the countries con- some of these gaps. cerned, where the lessons from experience now need to be Determination of the macroeconomic costs to countries of implemented. Implementation will require greater awareness current household energy use and the potential gains resulting of the problem at international and national levels, provision from change to more efficient and cleaner options could sub- of support for national collaborative action, and a focus on stantially add to the case for action. supporting appropriate, mainly market-based interventions. Monitoring progress requires the development and testing Better information is crucial to this effort, including of standard indicators for use in such policy documents as the stronger evidence of the health effects of IAP exposure; assess- World Development Report and for routine application at ment of the social, economic, and environmental benefits of national and subnational levels. The Millennium Development interventions; and indicators to monitor progress. Economic Goal Indicator on the proportion of the population using solid analysis can help bring the case for action into policy, but it fuels is a key starting point, and WHO, the reporting agency, is needs to be applied at country level and to include a wider working to broaden the monitoring of this indicator through range of benefits. Results from analysis at the regional level international surveys, such as demographic and health surveys show that interventions can be cost-effective, particularly (ORC Macro 2004), the Multiple Indicator Cluster Survey improved stoves, as long as these interventions can deliver sub- (UNICEF 2004), and the World Health Survey (WHO 2004d), stantial exposure reductions in practice. This conclusion, as 812 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others well as its qualification, is important given the expectation that Boy, E., N. G. Bruce, and H. Delgado. 2002. "Birthweight and Exposure to biomass will remain the principal household fuel in many Kitchen Wood Smoke during Pregnancy." Environmental Health Perspectives 110 (1): 109­14. developing countries for more than 20 years. The balance of Boy, E., N. G. Bruce, K. R. Smith, and R. Hernandez. 2000. "Fuel Efficiency effort and resources put into promoting cleaner biomass inter- of an Improved Wood Burning Stove in Rural Guatemala: Implications ventions rather than cleaner fuels, or vice-versa, will be an for Health, Environment, and Development." Energy for Sustainable important policy issue for many countries and for the interna- Development 4 (2): 21­29. tional community (Smith 2002). Brauer, B., K. Bartlett, J. Regaldo-Pineda, R. Perez-Padilla. 1996. "Assessment of Particulate Concentrations from Domestic Biomass With a range of innovative projects and programs under Combustion in Rural Mexico." Environmental Science and Technology way in a number of countries and regions of the world, now is 30: 104­9. an important time to focus attention and effort on achieving Bruce, N. G., E. Bates, R. Nguti, S. Gitonga, J. Kithinji, and A. Doig. 2002. the health, social, and economic gains that should result from "Reducing Indoor Air Pollution through Participatory Development in improvements in household energy systems in developing Rural Kenya." In Proceedings of 9th International Conference on Indoor Air Quality and Climate, Monterey, CA, 590­95. countries. Bruce, N. G., J. P. McCracken, R. Albalak, M. Schei, K. R. Smith, V. Lopez, and C. West. 2004. "The Impact of Improved Stoves, House Construction, and Child Location on Levels of Indoor Air Pollution and Exposure in Young Guatemalan Children." Journal of Exposure NOTE Analysis and Environmental Epidemiology 14 (Suppl. 1): S110­17. Bruce, N. G., R. Perez-Padilla, and R. Albalak. 2000. "Indoor Air Pollution 1. Particles are typically described according to the aerodynamic diam- in Developing Countries: A Major Environmental and Public Health eter, and although the devices used to separate particles of a given size do Challenge." Bulletin of the World Health Organization 78 (9): 1078­92. not yield a very sharp cutoff, this classification is functionally useful because smaller particles are able to penetrate farther into the lungs. Total Clancy, J. S., M. Skutsch, and S. Batchelor. 2003. The Gender-Energy- suspended particles (TSP) include suspended particles of all sizes. Poverty Nexus: Finding the Energy to Address Gender Concerns in Commonly defined smaller particles include PM10 (up to 10 microns Development. Project report CNTR998521. London: Department for diameter); respirable PM (includes all very small particles, about 50 per- International Development. cent of those 4 microns in diameter, and none above 10 microns in diam- Courtright, P., D. Haile, and E. Kohls. 1993. "The Epidemiology of Burns eter); and PM2.5 (up to 2.5 microns in diameter). in Rural Ethiopia." Journal of Epidemiology and Community Health 47 (1): 19­22. Danis-Lozano, R., M. H. Rodriguez, L. Gonzalez-Ceron, and M. Hernandez-Avila. 1999. "Risk Factors for Plasmodium Vivax Infection REFERENCES in the Lacandon Forest, Southern Mexico." Epidemiology of Infection 122 (3): 461­69. Albalak, R., N. G. Bruce, J. P. McCracken, and K. R. Smith. 2001. "Indoor Desai, M. A., S. Mehta, and K. R. Smith. 2004. Indoor Smoke from Solid Respirable Particulate Matter Concentrations from an Open Fire, Fuels: Assessing the Environmental Burden of Disease at National and Improved Cookstove, and LPG/Open Fire Combination in a Rural Local Levels. Environmental Burden of Disease Series 4. Geneva: World Guatemalan Community." Environmental Science and Technology 35 Health Organization. (13): 2650­55. Dooley, E. E. 2003. "New Stoves for Better Children's Health?" Ando, M., M. Tadano, S. Asanuma, K. Tamura, S. Matsushima, T. Environmental Health Perspectives 111 (1): A33. Watanabe, and others. 1998. "Health Effects of Indoor Fluoride Pollution from Coal Burning in China." Environmental Health D'Sa, A., and K. V. Narasimha Murthy. 2004. "LPG as a Cooking Fuel Perspectives 106 (5): 239­44. Option for India." Energy for Sustainable Development 8 (3): 91­106. Balakrishnan, K., J. Parikh, S. Sankar, R. Padmavathi, K. Srividya, V. EPA (U.S. Environmental Protection Agency). 2004. Partnership for Clean Venugopal, and others. 2002. "Daily Average Exposures to Respirable Indoor Air. Washington, DC: EPA. http://www.epa.gov/iaq/pcia.html. Particulate Matter from Combustion of Biomass Fuels in Rural ESMAP (Energy Sector Management Assistance Programme) and World Households of Southern India." Environmental Health Perspectives Bank. 2001. Indoor Air Pollution: Energy and Health for the Poor, Issue 110 (11): 1069­75. 5 (September). Delhi: World Bank. Ballard-Tremeer, G., and A. Mathee. 2000. "Review of Interventions to ------. 2004. Clean Household Energy for India: Reducing the Risks to Reduce the Exposure of Women and Young Children to Indoor Air Health. Delhi: World Bank. Pollution in Developing Countries." Paper prepared for USAID/WHO Ezzati, M., and D. M. Kammen. 2001."Quantifying the Effects of Exposure International Consultation on Household Energy, Indoor Air to Indoor Air Pollution from Biomass Combustion on Acute Pollution and Health, Washington, DC, May 4­6. Respiratory Infections in Developing Countries." Environmental Barnes, B., A. Mathee, L. Shafritz, L. Krieger, L. Sherburne, and M. Favin. Health Perspectives 109 (5): 481­88. 2004a. "Testing Selected Behaviours to Reduce Indoor Air Pollution Ezzati, M., and A. Lopez. 2004. "Mortality and Morbidity Due to Smoking Exposure in Young Children."Health Education Research 19 (5): 543­50. and Oral Tobacco Use: Global and Regional Estimates for 2000." In Barnes, B., A. Mathee, L. Shafritz, L. Krieger, and S. A. Zimicki. 2004b. "A Comparative Quantification of Health Risks: The Global Burden of Behavioural Intervention to Reduce Child Exposure to Indoor Air Disease Due to Selected Risk Factors, ed. M. Ezzati, A. D. Lopez, A. Pollution: Identifying Possible Target Behaviours." Health Education Rodgers, and C. J. L. Murray. Geneva: World Health Organization. and Behaviour 13 (3): 306­17. Ezzati, M., M. B. Mbinda, and D. M. Kammen. 2000. "Comparison of Behera, D., S. Dash, and S. Malik. 1988. "Blood Carboxyhaemoglobin Emissions and Residential Exposure from Traditional and Improved Levels Following Acute Exposure to Smoke of Biomass Fuel." Indian Cookstoves in Kenya." Environmental Science and Technology 34 (4): Journal of Medical Research 88 (December): 522­24. 578­83. Indoor Air Pollution | 813 Finkelman, R. B., H. E. Belkin, and B. Zheng. 1999. "Health Impacts of Pope, C. A. III, M. L. Hansen, R. W. Long, K. R. Nielsen, N. L. Eatough, W. Domestic Coal Use in China." Proceedings of the National Academy of E. Wilson, and D. J. Eatough. 2004. "Ambient Particulate Air Pollution, Science 96 (7): 3427­31. Heart Rate Variability, and Blood Markers of Inflammation in a Panel Gupta, S.,Y. C. Govil, P. K. Misra, R. Nath, and K. L. Srivastava. 1998."Trends of Elderly Subjects." Environmental Health Perspectives 112 (3): 339­45. in Poisoning in Children: Experience at a Large Referral Teaching Ramakrishna, J. 1988. "Patterns of Domestic Air Pollution in Rural India." Hospital." National Medical Journal of India 11 (4): 166­68. Ph.D. dissertation. University of Hawaii, Honolulu. Hutton, G., and L. Haller. 2004. Evaluation of the Costs and Benefits of Reed, R. P., and F. M. Conradie. 1997. "The Epidemiology and Clinical Water and Sanitation Improvements at the Global Level. Geneva: World Features of Paraffin (Kerosene) Poisoning in Rural African Children." Health Organization. Annals of Tropical Paediatrics 17 (1): 49­55. IEA (International Energy Agency). 2002. "Energy and Poverty." In World Reid, H., K. R. Smith, and B. Sherchand. 1986. "Indoor Smoke Exposures Energy Outlook 2002. Paris: International Energy Agency. from Traditional and Improved Cookstoves: Comparisons among Rural Johns, B., R. Baltussen, and R. Hutubessy. 2003. "Programme Costs in Nepali Women." Mountain Research and Development 6 (4): 293­304. the Economic Evaluation of Health Interventions. Cost Effectiveness Riojas-Rodriguez, H., P. Romano-Riquer, C. Santos-Burgoa, and K. R. and Resource Allocation 1 (1). http://www.resource-allocation.com/ Smith. 2001. "Household Firewood Use and the Health of Children content/1/1/1. and Women of Indian Communities in Chiapas, Mexico." International Lan, Q., R. S. Chapman, D. M. Schreinemachers, L. Tian, and X. He. 2002. Journal of Occupational and Environmental Health 7 (1): 44­53. "Household Stove Improvement and Risk of Lung Cancer in Xuanwei, Rollin, H., A. Mathee, N. G. Bruce, J. Levin, and Y. E. R. von Schirnding. China." Journal of the National Cancer Institute 94 (11): 826­35. 2004. "Comparison of Indoor Air Quality in Electrified and Un- Larson, B. A., and S. Rosen. 2002. "Understanding Household Demand for Electrified Dwellings in Rural South African Villages." Indoor Air 14 (3): Indoor Air Pollution Control in Developing Countries." Social Science 208­16. and Medicine 55 (4): 571­84. Saatkamp, B. D., O. R. Masera, and D. M. Kammen. 2000. "Energy and McDade, S. 2004. "Fueling Development: The Role of LPG in Poverty Health Transitions in Development: Fuel Use, Stove Technology, and Reduction and Growth." Energy for Sustainable Development 8 (3): Morbidity in Jarácuaro, Mexico." Energy for Sustainable Development 74­81. 4 (2): 5­14. Mehta, S., and C. Shahpar. 2004. "The Health Benefits of Interventions to Saksena, S., L. Thompson, and K. R. Smith. 2004. "Indoor Air Pollution Reduce Indoor Air Pollution from Solid Fuel Use: A Cost-Effectiveness and Exposure Database: Household Measurements in Developing Analysis." Energy for Sustainable Development 8 (3): 53­59. Countries." http://ehs.sph.berkeley.edu/hem/page.asp?id=33. Naeher, L., B. Leaderer, and K. R. Smith. 2000. "Particulate Matter and Sinton, J. E., K. R. Smith, J. Peabody, L. Yaping, Z. Ziliang, R. Edwards, and Carbon Monoxide in Highland Guatemala: Indoor and Outdoor G. Quan. 2004. "An Assessment of Programs to Promote Improved Levels from Traditional and Improved Wood Stoves and Gas Stoves." Household Stoves in China." Energy for Sustainable Development 8 (3): Indoor Air 10 (3): 200­205. 33­52. Naeher, L., K. R. Smith, B. Leaderer, L. Neufeld, and D. Mage. 2001. Smith, K. R. 1987. Biofuels, Air Pollution, and Health: A Global Review. New "Carbon Monoxide as a Tracer for Assessing Exposures to Particulate York: Plenum Press. Matter in Wood and Gas Cookstove Households in Highland ------. 1989. "Dialectics of Improved Stoves." Economic and Political Guatemala." Environmental Science and Technology 35 (3): 575­81. Weekly 24 (10): 517­22. Njenga, B. K. 2001. "Upesi Rural Stoves Project." In Generating ------. 2002. "In Praise of Petroleum?" Science 298: 1847. Opportunities: Case Studies on Energy and Women, ed. G. V. Karlsson Smith, K. R., A. L. Aggarwal, and R. M. Dave. 1983. "Air Pollution and and S. Misana, 45­51. Washington, DC: United Nations Development Rural Biomass Fuels in Developing Countries: A Pilot Village Study Programme. in India and Implications for Research and Policy." Atmospheric OECD and IEA (Organisation for Economic Co-operation and Environment 17 (11): 2343­62. Development and International Energy Agency). 2004. "Energy and Smith, K. R., S. Mehta, and M. Feuz. 2004. "Chapter 18: Indoor Smoke Development." In World Energy Outlook 2004. Paris: OECD and IEA. from Household Use of Solid Fuels." In Comparative Quantification of Onuba, O., and E. Udoidiok. 1987. "The Problems of Burns and Health Risks: The Global Burden of Disease Due to Selected Risk Factors, Prevention of Burns in Developing Countries." Burns 13 (5): 382­85. ed. M. Ezzati, A. D. Lopez, A. Rodgers, and C. J. L. Murray, vol 2, ORC Macro. 2004. "Demographic and Health Survey." ORC Macro. 1435­93. Geneva: World Health Organization. http://www.measuredhs.com/. Smith, K. R., G. Shuhua, H. Kun, and Q. Daxiong. 1993. "One Hundred Palsson, K., and T. G. Jaenson. 1999. "Plant Products Used as Mosquito Million Improved Cookstoves in China: How Was It Done?" World Repellants in Guinea Bissau, West Africa." Acta Tropica 72 (1): 39­52. Development 21 (6): 941­61. Pandey, M., R. Neupane, A. Gautam, and I. Shrestha. 1990. "The Smith, K. R., R. Uma, V. V. N. Kishore, J. Zhang, V. Joshi, and M. A. K. Effectiveness of Smokeless Stoves in Reducing Indoor Air Pollution in Khalil. 2000. "Greenhouse Implications of Household Stoves: An a Rural Hill Region of Nepal." Mountain Research and Development Analysis for India." Annual Review Energy Environment 25: 741­63. 10 (4): 313­20. Snow, R. W., A. K. Bradley, R. Hayes, P. Byass, and B. M. Greenwood. 1987. Parikh, J., K. Balakrishnan, V. Laxmi, and B. Haimanti. 2001. "Exposures "Does Woodsmoke Protect against Malaria?" Annals of Tropical from Cooking with Biofuels: Pollution Monitoring and Analysis for Medical Parasitology 81 (4): 449­51. Rural Tamil Nadu, India." Energy 26: 949­62. UNDP (United Nations Development Programme). 2005. "LP Gas Rural Paru, R., J. Hii, D. Lewis, and M. P. Alpers. 1995. "Relative Repellancy Challenge." http://www.undp.org/energy/lpg.htm. of Woodsmoke and Topical Applications of Plant Products against UNDP and ESMAP (United Nations Development Programme and Mosquitoes." Papua New Guinea Medical Journal 38 (3): 215­21. Energy Sector Management Assistance Programme). 2002. India: Pope, C. A. III, R. T. Burnett, M. J. Thun, E. E. Calle, D. Krewski, K. Ito, and Household Energy, Indoor Air Pollution, and Health. Delhi: UNDP and G. D. Thurston. 2002. "Lung Cancer, Cardiopulmonary Mortality, and World Bank. Long-Term Exposure to Fine Particulate Air Pollution." Journal of the ------. 2003. Health Impacts of Traditional Fuel Use in Guatemala. American Medical Association 287 (9): 1132­41. Washington, DC: UNDP and World Bank. 814 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others UNICEF (United Nations Children's Fund). 2004. "Monitoring the ------. 2002b. Reducing Risks, Promoting Healthy Life: World Health Situation of Women and Children: The Multiple Indicator Cluster Report 2002. Geneva: WHO. Survey." http://www.childinfo.org. ------. 2003. Making Choices in Health: WHO Guide to Cost-Effectiveness United Nations Statistics Division. 2003. "Millennium Development Analysis. Geneva: WHO. Goals." http://millenniumindicators.un.org/unsd/mi/mi_goals.asp. ------. 2004a. "Indoor Air Thematic Briefing 1: Indoor Air Pollution, Vernede, R., M. M. van Meer, and M. P. Alpers. 1994. "Smoke as a Form of Household Energy and the Millennium Development Goals." Geneva: Personal Protection against Mosquitoes: A Field Study in Papua New WHO. http://www.who.int/indoorair/info/en/iabriefing1rev.pdf. Guinea." Southeast Asian Journal of Tropical Medicine and Public Health ------. 2004b. "Burden of Disease Statistics. Burden of Disease Unit." 25 (4): 771­75. Geneva: WHO. http://www.who.int/research/en/. von Schirnding, Y. E. R., N. G. Bruce, K. R. Smith, G. Ballard-Tremeer, M. ------. 2004c. "Evidence for Policy Makers: Indoor Air Pollution." Ezzati, and K. Lvovsky. 2002. "Addressing the Impact of Household Geneva: WHO. http://www.who.int/indoorair/policy/en/. Energy and Indoor Air Pollution on the Health of the Poor: Implications for Policy Action and Intervention Measures." Paper pre- ------. 2004d. "The World Health Survey." Geneva: WHO. http://www3. pared for the Commission on Macroeconomics and Health, who.int/whs/. WHO/HDE/HID/02.9. World Health Organization, Geneva. ------. 2004e. "From Theory to Action: Implementing the WSSD Global Wang, X., and K. R. Smith. 1999. "Secondary Benefits of Greenhouse Gas Initiative on Children's Environmental Health Indicators." Geneva: Control: Health Impacts in China." Environmental Science and WHO. http://www.who.int/ceh/publications/924159188_9/en/. Technology 33 (18): 3056­61. ------. 2005. "Development of a Catalogue of Methods: Indoor Air Pol- Westoff, B., and D. Germann. 1995. Stove Images: A Documentation of lution." Geneva: WHO. http://www.who.int/indoorair/interventions/ Improved and Traditional Stoves in Africa, Asia and Latin America. methodology/en/. Brussels: Commission of the European Communities, Directorate- Wickramasinghe, A. 2001. "Gendered Sights and Health Issues in the General for Development. Paradigm of Biofuel in Sri Lanka." Energia News 4 (4): 14­16. WHO (World Health Organization). 2002a. "Addressing the Links World Bank. 2002. "Energy Efficiency and Emissions." In World between Indoor Air Pollution, Household Energy, and Human Health: Development Indicators. Washington, DC: World Bank. Based on the WHO-USAID Global Consultation on the Health Impact Yach, D. 1994. "Paraffin Poisoning: Partnership Is the Key to Prevention." of Indoor Air Pollution and Household Energy in Developing South African Medical Journal 84 (11): 717. Countries." Meeting Report WHO/HDE/HID/02.10. WHO, Geneva. Indoor Air Pollution | 815 Chapter 43 Air and Water Pollution: Burden and Strategies for Control Tord Kjellstrom, Madhumita Lodh, Tony McMichael, Geetha Ranmuthugala, Rupendra Shrestha, and Sally Kingsland Environmental pollution has many facets, and the resultant food hygiene; respiratory diseases related to severe indoor air health risks include diseases in almost all organ systems. Thus, pollution from biomass burning; and vectorborne diseases a chapter on air and water pollution control links with chapters with a major environmental component, such as malaria. on, for instance, diarrheal diseases (chapter 19), respiratory These three types of diseases each contribute approximately diseases in children and adults (chapters 25 and 35), cancers 6 percent to the updated estimate of the global burden of dis- (chapter 29), neurological disorders (chapter 32), and cardio- ease (WHO 2002). vascular disease (chapter 33), as well as with a number of chap- As the World Health Organization (WHO) points out, out- ters dealing with health care issues. door air pollution contributes as much as 0.6 to 1.4 percent of the burden of disease in developing regions, and other pollu- NATURE, CAUSES, AND BURDEN OF AIR tion, such as lead in water, air, and soil, may contribute 0.9 per- AND WATER POLLUTION cent (WHO 2002). These numbers may look small, but the contribution from most risk factors other than the "top 10" is Each pollutant has its own health risk profile, which makes within the 0.5 to 1.0 percent range (WHO 2002). summarizing all relevant information into a short chapter dif- Because of space limitations, this chapter can give only ficult. Nevertheless, public health practitioners and decision selected examples of air and water pollution health concerns. makers in developing countries need to be aware of the poten- Other information sources on environmental health include tial health risks caused by air and water pollution and to know Yassi and others (2001) and the Web sites of or major reference where to find the more detailed information required to handle works by WHO, the United Nations Environment Programme a specific situation. This chapter will not repeat the discussion (UNEP), Division of Technology, Industry, and Economics about indoor air pollution caused by biomass burning (http://www.uneptie.org/); the International Labour Organi- (chapter 42) and water pollution caused by poor sanitation at zation (ILO), the United Nations Industrial Development the household level (chapter 41), but it will focus on the prob- Organization (UNIDO; http://www.unido.org/), and other rel- lems caused by air and water pollution at the community, evant agencies. country, and global levels. Table 43.1 indicates some of the industrial sectors that can Estimates indicate that the proportion of the global burden pose significant environmental and occupational health risks of disease associated with environmental pollution hazards to populations in developing countries. Clearly, disease control ranges from 23 percent (WHO 1997) to 30 percent (Smith, measures for people working in or living around a smelter may Corvalan, and Kjellstrom 1999). These estimates include be quite different from those for people living near a tannery or infectious diseases related to drinking water, sanitation, and a brewery. For detailed information about industry-specific 817 Table 43.1 Selected Industrial Sectors and Their Contribution to Air and Water Pollution and to Workplace Hazards Industrial sector Air Water Workplacea Base metal and iron ore mining PM Toxic metal sludge Silica Cement manufacturing PM Sludge Silica Coalmining and production PM, coal dust Sludge Coal dust, silica Copper smelting Arsenic Arsenic Arsenic, cadmium Electricity generation PM, SO2 Hot water SO2 Foundries PM Solvents Silica, solvents Iron and steel smelting PM Sludge Carbon monoxide, nickel Lead and zinc smelting PM, SO2, lead, cadmium, arsenic Lead, cadmium, arsenic PM, SO2, lead, cadmium, arsenic Meat processing and rendering Odor High biological oxygen demand Infections Oil and gas development SO2, carcinogens Oil Hydrocarbons Pesticide manufacturing Pesticides and toxic intermediates Pesticides and toxic intermediates Pesticides and toxic intermediates Petrochemicals manufacturing SO2 Oil Hydrocarbons Petroleum refining SO2 Sludge, hydrocarbons Hydrocarbons Phosphate fertilizer plants PM Nutrients Pulp and paper mills Odor High biological oxygen demand, mercury Chlorine Tanning and leather finishing Odor Chromium, acids Chromium, acids Textile manufacturing Toxic dyes Source: World Bank 1999. a. In all the cases, the workplaces are subject to risk of injury, noise, dust, and excessively hot or cold temperatures. pollution control methods, see the Web sites of industry sector charcoal, natural gas, and so on); solvents; paints; glues; and organizations, relevant international trade union organiza- other products commonly used at work or at home. Volatile tions, and the organizations listed above. organic compounds include such chemicals as benzene, toluene, methylene chloride, and methyl chloroform. Emis- sions of nitrogen oxides and hydrocarbons react with sunlight Air Pollution to eventually form another secondary pollutant, ozone, at Air pollutants are usually classified into suspended particulate ground level. Ozone at this level creates health concerns, unlike matter (PM) (dusts, fumes, mists, and smokes); gaseous pollu- ozone in the upper atmosphere, which occurs naturally and tants (gases and vapors); and odors. protects life by filtering out ultraviolet radiation from the sun. Suspended PM can be categorized according to total sus- pended particles: the finer fraction, PM10, which can reach the Sources of Outdoor Air Pollution. Outdoor air pollution is alveoli, and the most hazardous, PM2.5 (median aerodynamic caused mainly by the combustion of petroleum products or diameters of less than 10.0 microns and 2.5 microns, respec- coal by motor vehicles, industry, and power stations. In some tively). Much of the PM2.5 consists of secondary pollutants countries, the combustion of wood or agricultural waste is created by the condensation of gaseous pollutants--for exam- another major source. Pollution can also originate from indus- ple, sulfur dioxide (SO2) and nitrogen dioxide (NO2). Types of trial processes that involve dust formation (for example, from suspended PM include diesel exhaust particles; coal fly ash; cement factories and metal smelters) or gas releases (for wood smoke; mineral dusts, such as coal, asbestos, limestone, instance, from chemicals production). Indoor sources also con- and cement; metal dusts and fumes; acid mists (for example,sul- tribute to outdoor air pollution, and in heavily populated areas, furic acid); and pesticide mists. the contribution from indoor sources can create extremely Gaseous pollutants include sulfur compounds such as SO2 high levels of outdoor air pollution. and sulfur trioxide; carbon monoxide; nitrogen compounds Motor vehicles emit PM, nitric oxide and NO2 (together such as nitric oxide, NO2, and ammonia; organic compounds referred to as NOx), carbon monoxide, organic compounds, such as hydrocarbons; volatile organic compounds; polycyclic and lead. Lead is a gasoline additive that has been phased out in aromatic hydrocarbons and halogen derivatives such as alde- industrial countries, but some developing countries still use hydes; and odorous substances. Volatile organic compounds leaded gasoline. Mandating the use of lead-free gasoline is are released from burning fuel (gasoline, oil, coal, wood, an important intervention in relation to health. It eliminates 818 | Disease Control Priorities in Developing Countries | Tord Kjellstrom, Madhumita Lodh, Tony McMichael, and others Box 43.1 The Bhopal Catastrophe The Bhopal plant, owned by the Union Carbide The dominating nonlethal effects of this emission were Corporation, produced methyl isocyanate, an intermedi- severe irritation of the eyes, lungs, and skin. Effects on ate in the production of the insecticide carbaryl. On the nervous system and reproductive organs were also December 2, 1984, a 150,000-gallon storage tank contain- reported. The reaction of methyl isocyanate with water ing methyl isocyanate apparently became contaminated had a corrosive effect on the respiratory tract, which with water, initiating a violent reaction and the release of resulted in extensive necrosis, bleeding, and edema. a cloud of toxic gas to which 200,000 people living near Treatment was impeded by the unknown and disputed the plant were exposed. Low wind speed and the high composition of the gas cloud and a lack of knowledge vapor pressure of methyl isocyanate exacerbated the sever- about its health effects and about antidotes. ity of toxic exposure, resulting in the immediate death of at least 6,000 people. Source: Dhara and Dhara 2002. vehicle-related lead pollution and permits the use of catalytic dust or hazardous fumes at the worksite (table 43.1). Such converters, which reduce emissions of other pollutants. industries include coalmining, mineral mining, quarrying, and Catastrophic emissions of organic chemicals, as occurred in cement production. Developed countries have shifted much of Bhopal, India, in 1984 (box 43.1), can also have major health their hazardous production to developing countries (LaDou consequences (McGranahan and Murray 2003; WHO 1999). 1992). This shift creates jobs in the developing countries, but at Another type of air pollution that can have disastrous con- the price of exposure to air pollution resulting from outdated sequences is radioactive pollution from a malfunctioning technology. In addition, specific hazardous compounds, such nuclear power station, as occurred in Chernobyl in 1986 as asbestos, have been banned in developed countries (Kazan- (WHO 1996). Radioactive isotopes emitted from the burning Allen 2004), but their use may still be common in developing reactor spread over large areas of what are now the countries of countries. Belarus, the Russian Federation, and Ukraine, causing thou- sands of cases of thyroid cancer in children and threatening to Impacts on Health. Epidemiological analysis is needed to cause many cancer cases in later decades. quantify the health impact in an exposed population. The major pollutants emitted by combustion have all been associ- Exposure to Air Pollutants. The extent of the health effects of ated with increased respiratory and cardiovascular morbidity air pollution depends on actual exposure. Total daily exposure and mortality (Brunekreef and Holgate 2002). The most is determined by people's time and activity patterns, and it famous disease outbreak of this type occurred in London in combines indoor and outdoor exposures. Young children and 1952 (U.K. Ministry of Health 1954), when 4,000 people died elderly people may travel less during the day than working prematurely in a single week because of severe air pollution, adults, and their exposure may therefore be closely correlated followed by another 8,000 deaths during the next few months with air pollution levels in their homes. Children are particu- (Bell and Davis 2001). larly vulnerable to environmental toxicants because of their In the 1970s and 1980s, new statistical methods and possibly greater relative exposure and the effects on their improved computer technology allowed investigators to study growth and physiological development. mortality increases at much lower concentrations of pollutants. Meteorological factors, such as wind speed and direction, A key question is the extent to which life has been shortened. are usually the strongest determinants of variations in air pol- Early loss of life in elderly people, who would have died soon lution, along with topography and temperature inversions. regardless of the air pollution, has been labeled mortality dis- Therefore, weather reports can be a guide to likely air pollution placement, because it contributes little to the overall burden of levels on a specific day. disease (McMichael and others 1998). Workplace air is another important source of air pollution Long-term studies have documented the increased cardio- exposure (chapter 60). Resource extraction and processing vascular and respiratory mortality associated with exposure industries, which are common in developing countries, emit to PM (Dockery and others 1993; Pope and others 1995). Air and Water Pollution: Burden and Strategies for Control | 819 A 16-year follow-up of a cohort of 500,000 Americans living in and 1 percent of respiratory infections, adding up to 7.9 mil- different cities found that the associations were strongest with lion disability-adjusted life years based on mortality only. This PM2.5 and also established an association with lung cancer burden of disease occurs primarily in developing countries, mortality (Pope and others 2002). Another approach is ecolog- with China and India contributing the most to the global bur- ical studies of small areas based on census data, air pollution den. Eastern Europe also has major air pollution problems, and information, and health events data (Scoggins and others in some countries, air pollution accounts for 0.6 to 1.4 percent 2004), with adjustments for potential confounding factors, of the total disability-adjusted life years from mortality. including socioeconomic status. Such studies indicate that the The global burden of disease caused by lead exposure mortality increase for every 10 micrograms per cubic meter includes subtle changes in learning ability and behavior and ( g per m3) of PM2.5 ranges from 4 to 8 percent for cities in other signs of central nervous system damage (Fewthrell, developed countries where average annual PM2.5 levels are 10 Kaufmann, and Preuss 2003). WHO (2002) concludes that to 30 g/m3. Many urban areas of developing countries have 0.4 percent of deaths and 0.9 percent (12.9 million) of all similar or greater levels of air pollution. disability-adjusted life years may be due to lead exposure. The major urban air pollutants can also give rise to signifi- cant respiratory morbidity (WHO 2000). For instance, Romieu and others (1996) report an exacerbation of asthma among Water Pollution children in Mexico City, and Xu and Wang (1993) note an Chemical pollution of surface water can create health risks, increased risk of respiratory symptoms in middle-aged non- because such waterways are often used directly as drinking smokers in Beijing. water sources or connected with shallow wells used for drink- In relation to the very young, Wang and others (1997) find ing water. In addition, waterways have important roles for that PM exposure, SO2 exposure, or both increased the risk of washing and cleaning, for fishing and fish farming, and for low birthweight in Beijing, and Pereira and others (1998) find recreation. that air pollution increased intrauterine mortality in São Paulo. Another major source of drinking water is groundwater, Other effects of ambient air pollution are postneonatal which often has low concentrations of pathogens because the mortality and mortality caused by acute respiratory infections, water is filtered during its transit through underground layers as well as effects on children's lung function, cardiovascular and of sand, clay, or rocks. However, toxic chemicals such as arsenic respiratory hospital admissions in the elderly, and markers for and fluoride can be dissolved from the soil or rock layers into functional damage of the heart muscle (WHO 2000). Asthma groundwater. Direct contamination can also occur from badly is another disease that researchers have linked to urban air pol- designed hazardous waste sites or from industrial sites. In the lution (McConnell and others 2002; Rios and others 2004). United States in the 1980s, the government set in motion the Ozone exposure as a trigger of asthma attacks is of particular Superfund Program, a major investigation and cleanup pro- concern. The mechanism behind an air pollution and asthma gram to deal with such sites (U.S. Environmental Protection link is not fully known, but early childhood NO2 exposure may Agency 2000). be important (see, for example, Ponsonby and others 2000). Coastal pollution of seawater may give rise to health hazards Leaded gasoline creates high lead exposure conditions in because of local contamination of fish or shellfish--for urban areas, with a risk for lead poisoning, primarily in young instance, the mercury contamination of fish in the infamous children. The main concern is effects on the brain from low- Minamata disease outbreak in Japan in 1956 (WHO 1976). level exposure leading to behavioral aberrations and reduced or Seawater pollution with persistent chemicals, such as polychlo- delayed development of intellectual or motoric ability (WHO rinated biphenyls (PCBs) and dioxins, can also be a significant 1995). Lead exposure has been implicated in hypertension in health hazard even at extremely low concentrations (Yassi and adults, and this effect may be the most important for the lead others 2001). burden of disease at a population level (WHO 2002). Other pollutants of concern are the carcinogenic volatile organic Sources of Chemical Water Pollution. Chemicals can enter compounds, which may be related to an increase in lung can- waterways from a point source or a nonpoint source. Point- cer, as reported by two recent epidemiological studies (Nyberg source pollution is due to discharges from a single source, such and others 2000; Pope and others 2002). as an industrial site. Nonpoint-source pollution involves many Urban air pollution and lead exposure are two of the envi- small sources that combine to cause significant pollution. For ronmental hazards that WHO (2002) assessed as part of its instance, the movement of rain or irrigation water over land burden-of-disease calculations for the World Health Report picks up pollutants such as fertilizers, herbicides, and insecti- 2002. The report estimates that pollution by urban PM causes cides and carries them into rivers, lakes, reservoirs, coastal as much as 5 percent of the global cases of lung cancer, 2 per- waters, or groundwater. Another nonpoint source is storm- cent of deaths from cardiovascular and respiratory conditions, water that collects on roads and eventually reaches rivers or 820 | Disease Control Priorities in Developing Countries | Tord Kjellstrom, Madhumita Lodh, Tony McMichael, and others lakes. Table 43.1 shows examples of point-source industrial neurological disease or more subtle functional damage to the chemical pollution. nervous system (Murata and others 2004). Paper and pulp mills consume large volumes of water and Runoff from farmland, in addition to carrying soil and sed- discharge liquid and solid waste products into the environ- iments that contribute to increased turbidity, also carries nutri- ment. The liquid waste is usually high in biological oxygen ents such as nitrogen and phosphates, which are often added in demand, suspended solids, and chlorinated organic com- the form of animal manure or fertilizers. These chemicals cause pounds such as dioxins (World Bank 1999). The storage and eutrophication (excessive nutrient levels in water), which in- transport of the resulting solid waste (wastewater treatment creases the growth of algae and plants in waterways, leading to sludge, lime sludge, and ash) may also contaminate surface an increase in cyanobacteria (blue-green algae). The toxics waters. Sugar mills are associated with effluent characterized by released during their decay are harmful to humans. biological oxygen demand and suspended solids, and the efflu- The use of nitrogen fertilizers can be a problem in areas ent is high in ammonium content. In addition, the sugarcane where agriculture is becoming increasingly intensified. These rinse liquid may contain pesticide residues. Leather tanneries fertilizers increase the concentration of nitrates in groundwa- produce a significant amount of solid waste, including hide, ter, leading to high nitrate levels in underground drinking hair, and sludge. The wastewater contains chromium, acids, water sources, which can cause methemoglobinemia, the life- sulfides, and chlorides. Textile and dye industries emit a liquid threatening "blue baby" syndrome, in very young children, effluent that contains toxic residues from the cleaning of which is a significant problem in parts of rural Eastern Europe equipment. Waste from petrochemical manufacturing plants (Yassi and others 2001). contains suspended solids, oils and grease, phenols, and ben- Some pesticides are applied directly on soil to kill pests in zene. Solid waste generated by petrochemical processes con- the soil or on the ground. This practice can create seepage to tains spent caustic and other hazardous chemicals implicated groundwater or runoff to surface waters. Some pesticides are in cancer. applied to plants by spraying from a distance--even from air- Another major source of industrial water pollution is min- planes. This practice can create spray drift when the wind car- ing. The grinding of ores and the subsequent processing with ries the materials to nearby waterways. Efforts to reduce the use water lead to discharges of fine silt with toxic metals into water- of the most toxic and long-lasting pesticides in industrial coun- ways unless proper precautions are taken, such as the use of tries have largely been successful, but the rules for their use in sedimentation ponds. Lead and zinc ores usually contain the developing countries may be more permissive, and the rules of much more toxic cadmium as a minor component. If the cad- application may not be known or enforced. Hence, health risks mium is not retrieved, major water pollution can occur. from pesticide water pollution are higher in such countries Mining was the source of most of the widespread cadmium (WHO 1990). poisoning (Itai-Itai disease) in Japan in 1940­50 (Kjellstrom Naturally occurring toxic chemicals can also contaminate 1986). groundwater, such as the high metal concentrations in under- Other metals, such as copper, nickel, and chromium, are ground water sources in mining areas. The most extensive essential micronutrients, but in high levels these metals can be problem of this type is the arsenic contamination of ground- harmful to health. Wastewater from mines or stainless steel water in Argentina, Bangladesh (box 43.2), Chile, China, India, production can be a source of exposure to these metals. The Mexico, Nepal, Taiwan (China), and parts of Eastern Europe presence of copper in water can also be due to corrosion of and the United States (WHO 2001). Fluoride is another drinking water pipes. Soft water or low pH makes corrosion substance that may occur naturally at high concentrations in more likely. High levels of copper may make water appear parts of China, India, Sri Lanka, Africa, and the eastern bluish green and give it a metallic taste. Flushing the first water Mediterranean. Although fluoride helps prevent dental decay, out of the tap can minimize exposure to copper. The use of lead exposure to levels greater than 1.5 milligrams per liter in drink- pipes and plumbing fixtures may result in high levels of lead in ing water can cause pitting of tooth enamel and deposits in piped water. bones. Exposure to levels greater than 10 milligrams per liter Mercury can enter waterways from mining and industrial can cause crippling skeletal fluorosis (Smith 2003). premises. Incineration of medical waste containing broken Water disinfection using chemicals is another source of medical equipment is a source of environmental contamina- chemical contamination of water. Chlorination is currently the tion with mercury. Metallic mercury is also easily transported most widely practiced and most cost-effective method of disin- through the atmosphere because of its highly volatile nature. fecting large community water supplies. This success in disin- Sulfate-reducing bacteria and certain other micro-organisms in fecting water supplies has contributed significantly to public lake, river, or coastal underwater sediments can methylate health by reducing the transmission of waterborne disease. mercury, increasing its toxicity. Methylmercury accumulates However, chlorine reacts with naturally occurring organic mat- and concentrates in the food chain and can lead to serious ter in water to form potentially toxic chemical compounds, Air and Water Pollution: Burden and Strategies for Control | 821 Box 43.2 Arsenic in Bangladesh The presence of arsenic in tube wells in Bangladesh This number increases to 46 million to 57 million if the because of natural contamination from underground geo- WHO guideline level of 10 micrograms per liter is used. logical layers was first confirmed in 1993. Ironically, the The most common sign of arsenic poisoning in United Nations Children's Fund had introduced the wells Bangladesh is skin lesions characterized by hyperkeratosis in the 1960s and 1970s as a safe alternative to water con- and melanosis. Other effects reported, but not epidemio- taminated with microbes, which contributed to a heavy logically confirmed, include cancer (particularly of the diarrheal disease burden. Estimates indicate that 28 mil- skin, lungs, and bladder); liver damage; diabetes; hyper- lion to 35 million people of Bangladesh's population of tension; and reproductive effects (spontaneous abortions 130 million are exposed to arsenic levels exceeding and stillbirths). Cancer and vascular effects are the domi- 50 micrograms per liter, the prescribed limit for drinking nating effects in other arsenic-polluted areas (WHO water in Bangladesh (Kinniburgh and Smedley 2001). 2001). Source: Authors. known collectively as disinfection by-products (International bone diseases of chronic cadmium poisoning (Itai-Itai disease), Agency for Research on Cancer 2004). and the circulatory system diseases of nitrate exposure (methe- moglobinemia) and lead exposure (anemia and hypertension). Exposure to Chemical Water Pollution. Drinking contami- Acute exposure to contaminants in drinking water can cause nated water is the most direct route of exposure to pollutants irritation or inflammation of the eyes and nose, skin, and gas- in water. The actual exposure via drinking water depends on trointestinal system; however, the most important health the amount of water consumed, usually 2 to 3 liters per day for effects are due to chronic exposure (for example, liver toxicity) an adult, with higher amounts for people living in hot areas or to copper, arsenic, or chromium in drinking water. Excretion of people engaged in heavy physical work. Use of contaminated chemicals through the kidney targets the kidney for toxic water in food preparation can result in contaminated food, effects, as seen with chemicals such as cadmium, copper, mer- because high cooking temperatures do not affect the toxicity of cury, and chlorobenzene (WHO 2003). most chemical contaminants. Pesticides and other chemical contaminants that enter Inhalation exposure to volatile compounds during hot waterways through agricultural runoff, stormwater drains, and showers and skin exposure while bathing or using water for industrial discharges may persist in the environment for long recreation are also potential routes of exposure to water pollu- periods and be transported by water or air over long distances. tants. Toxic chemicals in water can affect unborn or young chil- They may disrupt the function of the endocrine system, result- dren by crossing the placenta or being ingested through breast ing in reproductive, developmental, and behavioral problems. milk. The endocrine disruptors can reduce fertility and increase the Estimating actual exposure via water involves analyzing the occurrence of stillbirths, birth defects, and hormonally level of the contaminant in the water consumed and assessing dependent cancers such as breast, testicular, and prostate can- daily water intake (WHO 2003). Biological monitoring using cers. The effects on the developing nervous system can include blood or urine samples can be a precise tool for measuring total impaired mental and psychomotor development, as well as exposure from water, food, and air (Yassi and others 2001). cognitive impairment and behavior abnormalities (WHO and International Programme on Chemical Safety 2002). Examples Health Effects. No published estimates are available of the of endocrine disruptors include organochlorines, PCBs, global burden of disease resulting from the overall effects of alkylphenols, phytoestrogens (natural estrogens in plants), and chemical pollutants in water. The burden in specific local areas pharmaceuticals such as antibiotics and synthetic sex hor- may be large, as in the example cited in box 43.2 of arsenic in mones from contraceptives. Chemicals in drinking water can drinking water in Bangladesh. Other examples of a high also be carcinogenic. Disinfection by-products and arsenic local burden of disease are the nervous system diseases of have been a particular concern (International Agency for methylmercury poisoning (Minamata disease), the kidney and Research on Cancer 2004). 822 | Disease Control Priorities in Developing Countries | Tord Kjellstrom, Madhumita Lodh, Tony McMichael, and others INTERVENTIONS less hazardous materials. Interventions at the level of the state of the environment would include air quality monitoring The variety of hazardous pollutants that can occur in air or linked to local actions to reduce pollution during especially water also leads to many different interventions. Interventions polluted periods (for example, banning vehicle use when pol- pertaining to environmental hazards are often more sustain- lution levels reach predetermined thresholds). Interventions at able if they address the driving forces behind the pollution at the exposure level include using household water filters to the community level rather than attempt to deal with specific reduce arsenic in drinking water as done in Bangladesh. Finally, exposures at the individual level. In addition, effective meth- interventions at the effect level would include actions by health ods to prevent exposure to chemical hazards in the air or services to protect or restore the health of people already show- water may not exist at the individual level, and the only feasi- ing signs of an adverse effect. ble individual-level intervention may be treating cases of illness. Figure 43.1 shows five levels at which actions can be taken to Interventions to Reduce Air Pollution prevent the health effects of environmental hazards. Some Reducing air pollution exposure is largely a technical issue. would label interventions at the driving force level as policy Technologies to reduce pollution at its source are plentiful, as instruments. These include legal restrictions on the use of a are technologies that reduce pollution by filtering it away from toxic substance, such as banning the use of lead in gasoline, or the emission source (end-of-pipe solutions; see, for example, community-level policies, such as boosting public transporta- Gwilliam, Kojima, and Johnson 2004). Getting these technolo- tion and reducing individual use of motor vehicles. gies applied in practice requires government or corporate Interventions to reduce pressures on environmental quality policies that guide technical decision making in the right include those that limit hazardous waste disposal by recycling direction. Such policies could involve outright bans (such as hazardous substances at their site of use or replacing them with requiring lead-free gasoline or asbestos-free vehicle brake lin- ings or building materials); guidance on desirable technologies (for example, providing best-practice manuals); or economic Driving force Action instruments that make using more polluting technologies more Population growth Economic policy expensive than using less polluting technologies (an example of Economic development Social policy the polluter pays principle). Technology Clean technologies Examples of technologies to reduce air pollution include the use of lead-free gasoline, which allows the use of catalytic con- Pressure verters on vehicles' exhaust systems. Such technologies signifi- Production cantly reduce the emissions of several air pollutants from vehi- Consumption Hazard management cles (box 43.3). For trucks, buses, and an increasing number of Waste release smaller vehicles that use diesel fuel, improving the quality of State the diesel itself by lowering its sulfur content is another way to reduce air pollution at the source. More fuel-efficient Natural hazards Environmental Resource availability vehicles, such as hybrid gas-electric vehicles, are another way improvement Pollution levels forward. These vehicles can reduce gasoline consumption by about 50 percent during city driving. Policies that reduce Exposure "unnecessary" driving, or traffic demand management, can External exposure Education also reduce air pollution in urban areas. A system of congestion Absorbed dose Awareness fees, in which drivers have to pay before entering central urban Target organ dose raising areas, was introduced in Singapore, Oslo, and London and has Effect been effective in this respect. Power plants and industrial plants that burn fossil fuels use Well-being Morbidity Treatment a variety of filtering methods to reduce particles and scrubbing Mortality methods to reduce gases, although no effective method is cur- rently available for the greenhouse gas carbon dioxide. High chimneys dilute pollutants, but the combined input of pollu- tants from a number of smokestacks can still lead to an over- Source: Kjellstrom and Corvalan 1995. load of pollutants. An important example is acid rain, which is Figure 43.1 Framework for Environmental Health Interventions caused by SO2 and NOx emissions that make water vapor in the Air and Water Pollution: Burden and Strategies for Control | 823 Box 43.3 Air Pollution Reduction in Mexico City Mexico City is one of the world's largest megacities, with monoxide, NOx, and hydrocarbon emissions. In 1997, nearly 20 million inhabitants. Local authorities have leaded gasoline was completely phased out. The annual acknowledged its air quality problems since the 1970s. The average concentration of lead in the air in the worst- emissions from several million motor vehicles and thou- polluted area was reduced from 1.2 g/m3 in 1990 to less sands of industries created major concerns about health than 0.1 g/m3 in 2000. Surveys of blood lead levels in effects. Annual average particulate matter (PM10) levels children showed reductions from 200 to 100 g/liter dur- of 50 to 100 g/m3 have been measured in the worst- ing the same period, implying that the intervention had polluted central area and can be associated with annual protected thousands of children from lead poisoning. mortality excess of 15 to 30 percent. Even if only 20 per- Another key concern was SO2 emissions from industry cent of the population were exposed to such high levels, and diesel vehicles. Heavy fuel oil was phased out in the that exposure would account for 6,000 to 12,000 addi- mid 1990s, and the sulfur content of diesel was reduced. In tional deaths per year. To tackle the problem, Mexico City addition, power plants and some industry shifted to natu- started air quality monitoring and health studies in the ral gas in the early 1990s. The result was a 90 percent 1980s. High-risk groups were the 2.2 million children, reduction of SO2 in ambient air in five years. 250,000 street vendors, and 250,000 commercial drivers. Air quality standards, emission standards for vehicles, After 20 years of policies and actions, interventions for and other technical actions to reduce air emissions were better health have borne fruit. tightened during the 1990s, contributing to downward The first intervention was lead-free gasoline in 1990, trends of carbon monoxide, NOx, and ozone levels. Levels which enabled the government to require catalytic con- of emissions were reduced by half at some sites, resulting verters on new cars, thus dramatically reducing carbon in an estimated reduction of 3,000 to 6,000 excess deaths. Sources: Fernandez 2002; McMichael, Kjellstrom, and Smith 2001; WHO 2000. atmosphere acidic (WHO 2000). Large combined emissions Interventions to Reduce Water Pollution from industry and power stations in the eastern United States Water pollution control requires action at all levels of the hier- drift north with the winds and cause damage to Canadian archical framework shown in figure 43.1. The ideal method to ecosystems. In Europe, emissions from the industrial belt abate diffuse chemical pollution of waterways is to minimize across Belgium, Germany, and Poland drift north to Sweden or avoid the use of chemicals for industrial, agricultural, and and have damaged many lakes there. The convergence of air domestic purposes. Adapting practices such as organic farming pollutants from many sources and the associated health effects and integrated pest management could help protect waterways have also been documented in relation to the multiple fires in (Scheierling 1995). Chemical contamination of waterways Indonesia's rain forest in 1997 (Brauer and Hisham-Hashim from industrial emissions could be reduced by cleaner produc- 1998); the brown cloud over large areas of Asia, which is mainly tion processes (UNEP 2002). Box 43.4 describes one project related to coal burning; and a similar brown cloud over central aimed at effectively reducing pollution. Europe in the summer, which is caused primarily by vehicle Other interventions include proper treatment of hazardous emissions. waste and recycling of chemical containers and discarded prod- Managing air pollution interventions involves monitoring ucts containing chemicals to reduce solid waste buildup and air quality, which may focus on exceedances of air quality leaching of toxic chemicals into waterways. A variety of techni- guidelines in specific hotspots or on attempts to establish a spe- cal solutions are available to filter out chemical waste from cific population's average exposure to pollution. Sophisticated industrial processes or otherwise render them harmless. modeling in combination with monitoring has made it possi- Changing the pH of wastewater or adding chemicals that floc- ble to start producing detailed estimates and maps of air pollu- culate the toxic chemicals so that they settle in sedimentation tion levels in key urban areas (World Bank 2004), thus provid- ponds are common methods. The same principle can be used ing a powerful tool for assessing current health impacts and at the individual household level. One example is the use of estimated changes in the health impacts brought about by iron chips to filter out arsenic from contaminated well water in defined air pollution interventions. Bangladeshi households (Kinniburgh and Smedley 2001). 824 | Disease Control Priorities in Developing Countries | Tord Kjellstrom, Madhumita Lodh, Tony McMichael, and others Box 43.4 Water Pollution Control in India In 1993, the Demonstration in Small Industries for environmental regulations in a cost-effective manner. Reducing Wastes Project was started in India with support Pressure from the public to improve environmental from the United Nations Industrial Development performance and the need to conserve water, especially Organization. International and local experts initiated during the summer, added urgency to the project. The waste reduction audits in four pulp and paper plants, four company implemented 24 waste minimization options, textile dyeing and finishing factories, and four pesticide with 13 additional options under consideration, resulting production units. The experts identified priority areas, in net annual savings of about US$160,000. The payback estimated the likely reduction in the pollutant load, and period for the implemented options was less than seven came up with more than 500 pollution prevention months, and the annual savings will continue. options. The 12 companies spent a total of US$300,000 to The project demonstrated that waste minimization can implement pollution prevention options and saved US$3 cut pollution and business costs at the same time, espe- million in raw materials and wastewater treatment costs. cially when the environmental protection effort is directed The most impressive savings were in the pulp and paper toward the production process itself rather than to end- sector. For instance, the Ashoka Pulp and Paper Company of-pipe treatment. The key to success lies in the sustained participated in the project with the dual objectives involvement of local experts and committed factory of reducing production costs and complying with managers. Source: United Nations 1997. INTERVENTION COSTS AND others (1995) does not report the extent to which the various COST-EFFECTIVENESS interventions were implemented in existing pollution control or public health programs, and many of the most cost-effective This chapter cannot follow the detailed format for the eco- interventions are probably already in wide use. The review did nomic analysis of different preventive interventions devised for create a good deal of controversy in the United States, because the disease-specific chapters, because the exposures, health professionals and nongovernmental organizations active in the effects, and interventions are too varied and because of the environmental field accused the authors of overestimating lack of overarching examples of economic assessments. the costs and underestimating the benefits of controls over Nevertheless, it does present a few examples of the types of chemicals (see, for example, U.S. Congress 1999). analyses available. Costs and Savings in Relation to Pollution Control Comparison of Interventions A number of publications review and discuss the evidence A review of more than 1,000 reports on cost per life year saved on the costs and benefits of different pollution control in the United States for 587 interventions in the environment interventions in industrial countries (see, for example, U.S. and other fields (table 43.2) evaluated costs from a societal per- Environmental Protection Agency 1999). For developing coun- spective. The net costs included only direct costs and savings. tries, specific data on this topic are found primarily in the Indirect costs, such as forgone earnings, were excluded. Future so-called gray literature: government reports, consultant costs and life years saved were discounted at 5 percent per year. reports, or reports by the international banks. Interventions with a cost per life year saved of less than or equal to zero cost less to implement than the value of the lives saved. Air Pollution. Examples of cost-effectiveness analysis for Each of three categories of interventions (toxin control, fatal assessing air quality policy include studies carried out in injury reduction, and medicine) presented in table 43.2 Jakarta, Kathmandu, Manila, and Mumbai under the World includes several extremely cost-effective interventions. Bank's Urban Air Quality Management Strategy in Asia The cost-effective interventions in the air pollution area (Grønskei and others 1996a, 1996b; Larssen and others 1996a, could be of value in developing countries as their industrial 1996b; Shah, Nagpal, and Brandon 1997). In each city, an emis- and transportation pollution situations become similar to sions inventory was established, and rudimentary dispersion the United States in the 1960s. The review by Tengs and modeling was carried out. Various mitigation measures for Air and Water Pollution: Burden and Strategies for Control | 825 Table 43.2 Median Cost Per Life Year Saved, Selected Relatively Low-Cost Interventions (1993 U.S. dollars) Intervention Cost per life year saved Toxin control Control coal-fired power plant emissions through high chimneys and other means 0 Reduce lead in gasoline from 1.1 to 0.1 grams per gallon 0 Ban amitraz pesticide on apples 0 Introduce a chloroform emission standard at selected pulp mills 0 Control SO2 by desulfuring residual fuel oil 0 Initiate sedimentation, filtration, and chlorination of drinking water 4,200 Introduce radon remediation in homes with levels greater than 21.6 picocuries per liter 6,100 Ban asbestos in brake linings 29,000 Set arsenic emission standards at selected copper smelters 36,000 Fatal injury reduction Make motorcycle helmet laws mandatory 0 Install automatic seat belts in cars 0 Require bad drivers to attend driving improvement schools 0 Pass a law requiring smoke detectors in homes 0 Improve standards for concrete construction 0 Ban residential growth in tsunami-prone areas 0 Make seat belt use in cars mandatory 69 Install smoke detectors in airplane lavatories 30,000 Medicine Require all common types of early childhood vaccinations 0 Implement annual stool colon cancer screening for people age 55 and older 0 Introduce detoxification or methadone maintenance for heroin addicts 0 Screen newborns for phenylketonuria 0 Recommend cervical cancer screening every three years for women age 65 and older 0 Introduce universal prenatal care for expectant mothers 0 Vaccinate all citizens against influenza 140 Screen men age 45­54 for hypertension 5,200 Institute annual mammography and breast examinations for women age 40­64 17,000 Perform three-vessel coronary artery bypass surgery for severe angina 23,000 Source: Based on Tengs and others 1995. Note: The fatal injury reduction and medicine categories are included for comparison purposes. reducing PM10 and health impacts were examined in terms of tive, but the World Bank has developed a method to take these reductions in tons of PM10 emitted, cost of implementation, considerations into account. The costs of different air quality time frame for implementation, and health benefits and their improvement policies are explored in relation to a baseline associated cost savings. Some of the abatement measures that investment and the estimated health effects of air pollution. A have been implemented include introducing unleaded gaso- comparison will indicate the cost-effectiveness of each policy. line, tightening standards, introducing low-smoke lubricants The World Bank has worked out this "overlay" approach in for two-stroke engine vehicles, implementing inspections of some detail for the energy and forestry sectors in the analogous vehicle exhaust emissions to address gross polluters, and reduc- case of greenhouse gas reduction strategies (World Bank 2004). ing garbage burning. Transportation policies and industrial development do not Water Pollution. The costs and benefits associated with inter- usually have air quality considerations as their primary objec- ventions to remove chemical contaminants from water need to 826 | Disease Control Priorities in Developing Countries | Tord Kjellstrom, Madhumita Lodh, Tony McMichael, and others be assessed on a local or national basis to determine specific pollution damage costs are the actual payments for victims' needs, available resources, environmental conditions (includ- compensation and the cost of environmental remediation. The ing climate), and sustainability. A developing country for which compensation costs are based on court cases or government substantial economic analysis of interventions has been carried decisions and can be seen as a valid representation of the eco- out is China (Dasgupta, Wang, and Wheeler 1997; Zhang and nomic value of the health damage in each case. As table 43.3 others 1996). shows, controlling the relevant pollutants would have cost far Another country with major concerns about chemicals less than paying for damage caused by the pollution. (arsenic) in water is Bangladesh. The arsenic mitigation pro- A few studies have analyzed cost-benefit aspects of air pollu- grams have applied various arsenic removal technologies, but tion control in specific cities. Those analyses are based mainly on the costs and benefits are not well established. Bangladesh has modeling health impacts from exposure and relationships adopted a drinking water standard of 50 g/L (micrograms per between doses and responses. Voorhees and others (2001) find liter) for arsenic in drinking water. The cost of achieving the that most studies that analyzed the situation in specific urban lower WHO guideline value of 10 g/L would be significant. areas used health impact assessment to estimate impacts avoided An evaluation of the cost of lowering arsenic levels in drinking by interventions. Investigators have used different methods for water in the United States predicts that a reduction from 50 to valuing the economic benefits of health improvements, includ- 10 g/L would prevent a limited number of deaths from blad- ing market valuation, stated preference methods, and revealed der and lung cancer at a cost of several million dollars per death preference methods. The choice of assumptions and inputs sub- prevented (Frost and others 2002). stantially affected the resulting cost and benefit valuations. Alternative water supplies need to be considered when the One of the few detailed studies of the costs and benefits of air costs of improving existing water sources outweigh the bene- pollution control in a specific urban area (Voorhees and others fits. Harvesting rainwater may provide communities with safe 2000) used changing nitric oxide and NO2 emissions in Tokyo drinking water, free of chemicals and micro-organisms, but during 1973­94 as a basis for the calculations. The study did not contamination from roofs and storage tanks needs to be con- use actual health improvement data but calculated likely health sidered. Rainwater collection is relatively inexpensive. improvements from estimated reductions in NO2 levels and published dose-response curves. The health effects included ECONOMIC BENEFITS OF INTERVENTIONS respiratory morbidity (as determined by hospital admissions and medical expenses), and working days lost for sick adults, One of the early examples of cost-benefit analysis for chemical and maternal working days lost in the case of a child's illness. pollution control is the Japan Environment Agency's (1991) The results indicated an average cost-benefit ratio of 1 to 6, with study of three Japanese classical pollution diseases: Yokkaichi a large range from a lower limit of 3 to 1 to an upper limit of 1 asthma, Minamata disease, and Itai-Itai disease (table 43.3). to 44. The estimated economic benefits of reductions in nitric This analysis was intended to highlight the economic aspects of oxide and NO2 emissions between 1973 and 1994 were consid- pollution control and to encourage governments in developing erable: US$6.78 billion for avoided medical costs, US$6.33 bil- countries to consider both the costs and the benefits of indus- lion for avoided lost wages of sick adults, and US$0.83 billion trial development. The calculations take into account the 20 or for avoided lost wages of mothers with sick children. 30 years that have elapsed since the disease outbreaks occurred Blackman and others' (2000) cost-benefit analysis of and annualize the costs and benefits over a 30-year period. The four practical strategies for reducing PM10 emissions from Table 43.3 Comparison of Actual Pollution Damage Costs and the Pollution Control Costs That Would Have Prevented the Damage, for Three Pollution-Related Disease Outbreaks, Japan (¥ millions, 1989 equivalents) Pollution damage costs Pollution Main Pollution Health Livelihood Environmental disease pollutant control costs damage damage remediation Total Yokkaichi asthma SO2, air pollution 14,800 21,000 Not estimated Not estimated 21,000 (1,300)a Minamata disease Mercury, water pollution 125 7,670 4,270 690 12,630 Itai-Itai disease Cadmium, water and soil pollution 600 740 880 890 2,510 Source: Japan Environment Agency 1991. Note: US$1 ¥150. a. Based on actual compensation payments to a fraction of the population. The larger figure is what it would have cost to compensate all those who were affected. Air and Water Pollution: Burden and Strategies for Control | 827 traditional brick kilns in Ciudad Juárez in Mexico suggests that, Agency (1991) estimates the national economic impact of pol- given a wide range of modeling assumptions, the benefits of lution control legislation and associated interventions. During three control strategies would be considerably higher than the the 1960s and early 1970s, when the government made many of costs. Reduced mortality was by far the largest component of the major decisions about intensified pollution control inter- benefits, accounting for more than 80 percent of the total. ventions, Japan's gross domestic product (GDP) per capita was Pandey and Nathwani (2003) applied cost-benefit analysis growing at an annual rate of about 10 percent, similar to that to a pollution control program in Canada. Their study pro- of the rapidly industrializing countries in the early 21st century. posed using the life quality index as a tool for quantifying the At that time, Japan's economic policies aimed at eliminating level of public expenditure beyond which the use of resources bottlenecks to high economic growth, and in the mid 1960s, is not justified. The study estimated total pollution control industry was spending less than ¥50 billion per year on pollu- costs at US$2.5 billion per year against a monetary benefit of tion control equipment. By 1976, this spending had increased US$7.5 billion per year, using 1996 as the base year for all to almost ¥1 trillion per year. The ¥5 trillion invested in pollu- cost and benefit estimates. The benefit estimated in terms of tion control between 1965 and 1975 accounted for about avoided mortality was about 1,800 deaths per year. 0.9 percent of the increase in GDP per capita during this El-Fadel and Massoud's (2000) study of urban areas in period. The Japan Environment Agency concluded that the Lebanon shows that the health benefits and economic benefits stricter environmental protection legislation and associated of reducing PM concentration in the air can range from US$4.53 major investment in pollution control had little effect on the million to US$172.50 million per year using a willingness-to-pay overall economy, but that the resulting health benefits are likely approach. In that study, the major monetized benefits resulted cumulative. from reduced mortality costs. Aunan and others (1998) assessed the costs and benefits of implementing an energy saving and air pollution control pro- Air gram in Hungary. They based their monetary evaluation of The broadest analysis of the implementation of control strate- benefits on local monitoring and population data and took gies for air pollution was conducted by the U.S. Environmental exposure-response functions and valuation estimates from Protection Agency in the late 1990s (Krupnick and Canadian, U.S., and European studies. The authors valued the Morgenstern 2002). The analysis developed a hypothetical sce- average total benefits of the interventions at US$1.56 billion nario for 1970 to 1990, assuming that the real costs for pollu- per year (with 1994 as the base year), with high and low bounds tion control during this period could be compared with the at US$7.6, billion and US$0.4 billion, respectively. They esti- benefits of reduced mortality and morbidity and avoided dam- mated the cost-benefit ratio at 1 to 3.4, given a total cost of age to agricultural crops brought about by the reduction of interventions of US$0.46 billion per year. Many of the benefits major air pollutant levels across the country during this period. resulted from reduced mortality in the elderly population and The study estimated reduced mortality from dose-response from reduced asthma morbidity costs. relationships for the major air pollutants, assigning the cost of Misra (2002) examined the costs and benefits of water pol- each death at the value of statistical life and the cost of mor- lution abatement for a cluster of 250 small-scale industries in bidity in relation to estimated health service utilization. The Gujarat, India. Misra's assessment looked at command-and- study used a variety of costing methods to reach the range of control, market-based solutions and at effluent treatment as likely present values presented in table 43.4. It assumed that the alternatives. In a cost-benefit analysis, Misra estimated the net reduction of air pollution resulted from the implementation of present social benefits from water pollution abatement at the federal Clean Air Act of 1970 and associated state-level reg- the Nandesari Industrial Estate at Rs 0.550 billion at 1995­96 ulations and air pollution limits. market prices using a 12 percent social discount rate. After The analysis showed a dramatically high cost-benefit ratio making corrections for the prices of foreign exchange, and inspired debate about the methodologies used and the unskilled labor, and investment, the figure rose to Rs 0.62 bil- results. One major criticism was of the use of the value of lion. It rose still further to about Rs 3.1 billion when distribu- statistical life for each death potentially avoided by the tional effects were taken into account. reduced air pollution. A recalculation using the life-years-lost method reduced the benefits for deaths caused by PM from IMPLEMENTATION OF CONTROL STRATEGIES: US$16,632 billion to US$9,100 billion (Krupnick and LESSONS OF EXPERIENCE Morgenstern 2002). The recalculated figure is still well above the fifth percentile estimate of benefits and does not under- The foregoing examples demonstrate that interventions to mine the positive cost-benefit ratio reported. Thus, if a devel- protect health that use chemical pollution control can have oping country were to implement an appropriate control an attractive cost-benefit ratio. The Japan Environment strategy for urban air pollution, it might derive significant 828 | Disease Control Priorities in Developing Countries | Tord Kjellstrom, Madhumita Lodh, Tony McMichael, and others Table 43.4 Present Value of Monetary Benefits and Costs Associated with Implementation of the U.S. Clean Air Act, 1970­90 (1990 US$ billions) Present value, Present Present value, Category Pollutant 5th percentile value, mean 95th percentile Mortality PM 2,369 16,632 40,957 Mortality Lead 121 1,339 3,910 Chronic bronchitis PM 409 3,313 10,401 IQ reduction Lead 271 399 551 Other morbidity Several 227 337 501 Soil damage PM 6 74 192 Visibility reduction PM 38 54 71 Agricultural damage Ozone 11 23 35 Total benefits All 3,452 22,171 56,618 Total costs All Not estimated 523 Not estimated Net benefits (total benefits total costs) All Not estimated 21,648 Not estimated Source: Krupnick and Morgenstern 2002. economic benefits over the subsequent decades. The country's RESEARCH AND DEVELOPMENT AGENDA level of economic development, local costs, and local benefit valuations will be important for any cost-benefit assessment. Even though a good deal of information is available about the WHO's (2000) air quality guidelines are among the documents health risks of common air and water pollutants, further that provide advice on analytical approaches. research is needed to guide regulations and interventions. The pollutants that were most common in developed countries in the past are still major problems in developing countries; how- Water ever, direct application of the experiences of developed coun- We were unable to find an analysis for water similar to the tries may not be appropriate, because exposed populations in broad analysis presented for air, but the examples of water pol- developing countries may have a different burden of preexist- lution with mercury, cadmium, and arsenic described earlier ing diseases, malnutrition, and other factors related to poverty. indicate the economic benefits that can be reaped from effec- Research on specific vulnerabilities and on relevant dose- tive interventions against chemical water pollution. Since the response relationships for different levels of economic develop- pollution disease outbreaks of mercury and cadmium poison- ment and for various geographic conditions would therefore ing in Japan, serious mercury pollution situations have been be valuable for assessing risks and targeting interventions. In identified in Brazil, China, and the Philippines, and serious addition, global chemical exposure concerns, such as endocrine cadmium pollution has occurred in Cambodia, China, the Lao disruptors in air, water, and food, require urgent research to People's Democratic Republic, and Thailand. Arsenic in establish the need for interventions in both industrial and groundwater is an ongoing, serious problem in Bangladesh, developing countries. India, and Nepal and a less serious problem in a number of An important research topic is to clearly describe and quan- other countries. tify the long-term health effects of exposure to air pollution. WHO has analyzed control strategies for biological water The existing literature indicates that long-term exposure may pollution and water and sanitation improvements in relation to have more adverse health effects than short-term exposure the Millennium Development Goals (Hutton and Haller 2004). and, hence, have higher cost implications. Another topic is to The analysis demonstrated the considerable benefits of water assess the health issue pertaining to greenhouse gases and and sanitation improvements: for every US$1 invested, the eco- climate change, which are related to the same sources as urban nomic return was in the range of US$5 to US$28 for a number air pollution (Intergovernmental Panel on Climate Change of intervention options. Careful analysis of the same type is 2001). Research and policy analysis on how best to develop required for populations particularly vulnerable to chemical interventions to reduce health risks related to climate change water pollution to assess whether control of chemical pollution need to be considered together with the analysis of other air can also yield significant benefits. pollutants. Air and Water Pollution: Burden and Strategies for Control | 829 In addition, to improve analysis of the economic costs of CONCLUSION: PROMISES AND PITFALLS health impacts, better estimates are needed of the burden of disease related to chemical air and water pollution at local, Evidence shows that a number of chemicals that may be national, and global levels. Cost-effectiveness analysis of air released into the air or water can cause adverse health effects. and water pollution control measures in developing countries The associated burden of disease can be substantial, and invest- needs to be supported by further research, as cost levels and ment in research on health effects and interventions in specific benefit valuations will vary from country to country, and populations and exposure situations is important for the devel- solutions that are valid in industrial countries may not work opment of control strategies. Pollution control is therefore an as well in developing countries. Strategies for effective air and important component of disease control, and health profes- water resource management should include research on the sionals and authorities need to develop partnerships with other potential side effects of an intervention, such as in sectors to identify and implement priority interventions. Bangladesh, where tube wells drilled to supply water turned Developing countries face major water quantity and quality out to be contaminated with arsenic (see box 43.2). Research challenges, compounded by the effects of rapid industrializa- is also needed that would link methodologies for assessing tion. Concerted actions are needed to safely manage the use of adverse health effects with exposure and epidemiological stud- toxic chemicals and to develop monitoring and regulatory ies in different settings to permit the development of more guidelines. Recycling and the use of biodegradable products precise forecasting of the health and economic benefits of must be encouraged. Technologies to reduce air pollution at the interventions. source are well established and should be used in all new indus- The variety of health effects of urban air pollution and the trial development. Retrofitting of existing industries and power variety of sources create opportunities for ancillary effects that plants is also worthwhile. The growing number of private need to be taken into account in economic cost-effectiveness motor vehicles in developing countries brings certain benefits, and cost-benefit analysis. These are the beneficial effects of but alternative means of transportation, particularly in rapidly reducing air pollution on other health risks associated with the growing urban areas, need to be considered at an early stage, as sources of air pollution. For example, if the air pollution from the negative health and economic impacts of high concentra- transportation emissions is reduced by actions that reduce the tions of motor vehicles are well established. The principles use of private motor vehicles by, say, providing public trans- and practices of sustainable development, coupled with local portation, not only are carbon dioxide levels reduced; traffic research, will help contain or eliminate health risks resulting crash injuries, noise, and physical inactivity related to the from chemical pollution. International collaboration involving widespread use of motor vehicles also decline (Kjellstrom and both governmental and nongovernmental organizations can others 2003). guide this highly interdisciplinary and intersectoral area of One of the key challenges for policies and actions is to find disease control. ways to avoid a rapid buildup of urban air pollution in coun- tries that do not yet have a major problem. The health sector needs to be involved in assessing urban planning, the location REFERENCES of industries, and the development of transportation systems and needs to encourage those designing public transportation Aunan, K., G. Patzay, H. A. Aaheim, and H. M. Seip. 1998. "Health and Environmental Benefits from Air Pollution Reductions in Hungary." and housing to ensure that new sources of air pollution are not Science of the Total Environment 212: 245­68. being built into cities. Bell, M. L., and D. I. Davis. 2001. "Reassessment of the Lethal London Fog Decades of economic and industrial growth have resulted in of 1952: Novel Indicators of Acute and Chronic Consequences of Acute lifestyles that increase the demands on water resources simul- Exposure to Air Pollution." Environmental Health Perspectives 109 taneous with increases in water pollution levels. Conflicts (Suppl. 3): 389­94. between household, industrial, and agricultural water use are Blackman, A., S. Newbold, J. S. Shih, and J. Cook. 2000. "The Benefits and Costs of Informal Sector Pollution Control: Mexican Brick Kilns." a common public health problem (UNESCO 2003). The Discussion Paper 00­46, Resources for the Future, Washington, DC. developing countries need to avoid the experiences of water Brauer M., and J. Hisham-Hashim. 1998. "Indonesian Fires: Crisis and pollution and associated disease outbreaks in industrial coun- Reaction." Environmental Science and Technology 32: 404A­7A. tries. Strategies to ensure sufficient pollution control must Brunekreef, B., and S. T. Holgate. 2002. "Air Pollution and Health." Lancet be identified at the same time as strategies to reduce water con- 360: 1233­42. sumption. High water use depletes supplies and increases Constantinides, G. 2000. Cost-Benefit Analysis Case Studies in Eastern salinity in groundwater aquifers, particularly in coastal regions. Africa for the GPA Strategic Action Plan on Sewage. Nairobi: United Nations Environmental Programme. http://www.gpa.unep.org/ The impact of climate change must also be taken into consid- documents/other/casestudies/east_africa_case_studies_final_draft. eration (Vorosmarty and others 2000). pdf. 830 | Disease Control Priorities in Developing Countries | Tord Kjellstrom, Madhumita Lodh, Tony McMichael, and others Dasgupta, S., H. Wang, and D. Wheeler. 1997. "Surviving Success: Policy LaDou, J. 1992. "The Export of Hazards to Developing Countries." In Reform and the Future of Industrial Pollution in China." Working Occupational Health in Developing Countries, ed. J. Jeyaratnam, Paper 1856, World Bank, Washington, DC. 340­60. Oxford, U.K.: Oxford University Press. Dhara, V. R., and R. Dhara. 2002. "The Union Carbide Disaster in Bhopal: Larssen, S., F. Gram, L. O. Hagen, H. Jansen, X. Olsthoorn, R. V. Aundhe, A Review of Health Effects." Archives of Environmental Health 57 (5): and U. Joglekar. 1996a. URBAIR Urban Air Quality Management 391­404. Strategy in Asia: Greater Mumbai Report. Washington, DC: World Bank. Dockery, D. W., C. A. Pope, X. Xu, J. D. Spengler, J. H. Ware, M. E. Fay, and Larssen, S., F. Gram, L. O. Hagen, H. Jansen, X. Olsthoorn, R. Lesaca, and others. 1993. "An Association between Air Pollution and Mortality in others. 1996b. URBAIR Urban Air Quality Management Strategy in Six U.S. Cities." New England Journal of Medicine 329 (24): 1753­59. Asia: Metro Manila Report. Washington, DC: World Bank. El-Fadel, M., and M. Massoud. 2000. "Particulate Matter in Urban Areas: McConnell, R., K. Berhane, F. Gilliland, S. J. London, T. Islam, W. J. Health-Based Economic Assessment." Science of the Total Environment Gauderman, and others. 2002. "Asthma in Exercising Children 257: 133­46. Exposed to Ozone: A Cohort Study." Lancet 359 (9304): 386­91. Fernandez, A. 2002. Mexico City: Air Pollution Case Study. Environment McGranahan, G., and F. Murray. 2003. "Air Pollution and Health in course. Cambridge, MA: Harvard University. http://courses.dce. Rapidly Developing Countries." London: Earthscan. harvard.edu/environment/week10em1_new.html. McMichael, A. J., H. R. Anderson, B. Brunekreef, and A. Cohen. 1998. Fewthrell, L., R. B. Kaufmann, and A. Preuss. 2003. Assessing the "Inappropriate Use of Daily Mortality Analyses to Estimate Longer- Environmental Burden of Disease at the National and Local Level: Lead. Term Mortality Effects of Air Pollution." International Journal of Environmental Burden of Disease Series 2. Geneva: World Health Epidemiology 27: 450­53. Organization. McMichael, A. J., T. Kjellstrom, and K. Smith. 2001. "Environmental Frost, F. J., K. Tollestrup, G. F. Craun, R. Raucher, J. Chwirka, and J. Stomp. Health." In International Public Health, ed. M. H. Merson, R. E. Black, 2002. "Evaluation of Costs and Benefits of a Lower Arsenic MCL." and A. J. Mills, 379­438. Gaithersburg, MD: Aspen. Journal AWWA (American Water Works Association) 94 (3): 71­82. Misra, S. 2002. "An Empirical Investigation of Collective Action Grønskei, K. E., F. Gram, L. O. Hagen, S. Larssen, H. Jansen, X. Olsthoorn, Possibilities for Industrial Water Pollution Abatement: Case Study of a and others. 1996a. URBAIR Urban Air Quality Management Strategy in Cluster of Small-Scale Industries in India." World Bank Economists' Asia: Jakarta Report. Washington, DC: World Bank. Forum 2: 89­113. ------. 1996b. URBAIR Urban Air Quality Management Strategy in Asia: Murata, K., P. Weihe, E. Budtz-Jorgensen, P. J. Jorgensen, and P. Grandjean. Kathmandu Valley Report. Washington, DC: World Bank. 2004. "Delayed Brainstem Auditory Evoked Potential Latencies in Gwilliam, K., M. Kojima, and T. Johnson. 2004. Reducing Air Pollution from 14-Year-Old Children Exposed to Methylmercury." Journal of Transport. Washington, DC: World Bank. Pediatrics 144: 177­83. Hutton, G., and L. Haller. 2004. Evaluation of the Costs and Benefits of Nyberg, F., P. Gustavsson, L. Jarup, T. Bellander, N. Berglind, R. Jacobsson, Water and Sanitation Improvements at the Global Level. WHO/ and others. 2000. "Urban Air Pollution and Lung Cancer in SDE/WSH/04.04. Geneva: World Health Organization. http://www. Stockholm." Epidemiology 11: 487­95. who.int/water_sanitation_health/wsh0404/en/. Pandey, M. D., and J. S. Nathwani. 2003. "Canada Wide Standard for Intergovernmental Panel on Climate Change. 2001. Climate Change 2001. Particulate Matter and Ozone: Cost-Benefit Analysis Using a Life Geneva: World Meteorological Organization, Intergovernmental Panel Quality Index." Risk Analysis 23 (1): 55­67. on Climate Change. http://www.ipcc.ch. Pereira, L. A., D. Loomis, G. M. Conceição, A. L. Braga, R. M. Arcas, K. S. International Agency for Research on Cancer. 2004. Some Drinking Water Kishi, and others. 1998. "Association between Air Pollution and Disinfectants and Contaminants, Including Arsenic. Monograph 84. Intrauterine Mortality in São Paulo, Brazil." Environmental Health Lyon, France: International Agency for Research on Cancer. Perspectives 106: 325­29. Japan Environment Agency. 1991. Pollution in Japan--Our Tragic Ponsonby, A. L., D. Couper, T. Dwyer, A. Carmichael, A. Kemp, and Experience (in Japanese, with English translation available). Tokyo: J. Cochrane. 2000. "The Relation between Infant Indoor Environment Japan Environment Agency, Study Group for Global Environment and and Subsequent Asthma." Epidemiology 11: 128­35. Economics, Office of Planning and Research. Pope, C. III, R. Burnett, M. Thun, E. Calle, D. Krewski, K. Ito, and G. D. Kazan-Allen, L. 2004. "The Asbestos War." International Journal of Thurston. 2002."Lung Cancer, Cardiopulmonary Mortality, and Long- Occupational and Environmental Health 9: 173­93. Term Exposure to Fine Particulate Air Pollution." Journal of the Kinniburgh, D. G., and P. A. Smedley, eds. 2001. Arsenic Contamination of American Medical Association 287 (9): 1132­41. Groundwater in Bangladesh. BGS Technical Report WC/00/19. Pope, C. III, M. J. Thun, M. M. Namboodiri, D. W. Dockery, J. S. Evans, F. E. Keyworth, U.K.: British Geological Survey; Dhaka: Department of Speizer, and others. 1995. "Particulate Air Pollution as a Predictor Public Health Engineering. of Mortality in a Prospective Study of U.S. Adults." American Journal Kjellstrom, T. 1986. "Itai-Itai Disease." In Cadmium and Health, ed. L. of Respiratory Critical Care Medicine 151 (3, part 1): 669­74. Friberg, G. F. Nordberg, T. Kjellstrom, and C. G. Elinder, vol. 2, 257­90. Rios, J. L. M., J. L. Boechat, C. C. Sant'Anna, and A. T. Franca. 2004. Boca Raton, FL: CRC Press. "Atmospheric Pollution and the Prevalence of Asthma: Study among Kjellstrom, T., and C. Corvalan. 1995."Framework for the Development of Schoolchildren in Two Areas of Rio de Janeiro, Brazil." Annals of Environmental Health Indicators." World Health Statistics Quarterly Allergy, Asthma, and Immunology 92 (6): 629­34. 48: 144­54. Romieu, I., F. Meneses, S. Ruiz, J. J. Sienra, J. Huerta, M. C. White, and R. A. Kjellstrom, T., L. van Kerkhoff, G. Bammer, and T. McMichael. 2003. Etzel. 1996. "Effects of Air Pollution on the Respiratory Health "Comparative Assessment of Transport Risks: How It Can Contribute of Asthmatic Children Living in Mexico City." American Journal of to Health Impact Assessment of Transport Policies." Bulletin of the Respiratory Critical Care Medicine 154: 300­7. World Health Organization 81: 451­57. Scheierling, S. 1995. "Overcoming Agricultural Pollution of Water: The Krupnick, A., and R. Morgenstern. 2002. "The Future of Benefit-Cost Challenge of Integrating Agricultural and Environmental Policies in the Analysis of the CleanAirAct."Annual Review of Public Health 23: 427­48. European Union." Technical Paper 269, World Bank, Washington, DC. Air and Water Pollution: Burden and Strategies for Control | 831 Scoggins, A., T. Kjellstrom, G. Fisher, J. Connor, and N. Gimson. 2004. Wang, X., H. Ding, L. Ryan, and X. Xu. 1997. "Association between Air "Spatial Analysis of Annual Air Pollution and Mortality." Science of the Pollution and Low Birth Weight: A Community-Based Study." Total Environment 321: 71­85. Environmental Health Perspectives 105: 514­20. Shah, J., T. Nagpal, and C. Brandon, eds. 1997. Urban Air Quality WHO (World Health Organization). 1976. Mercury. Environmental Management Strategy in Asia: Guidebook. Washington, DC: World Health Criteria 1. Geneva: WHO. Bank. ------. 1990. Public Health Impact of Pesticides Used in Agriculture. Smith, K. R., C. Corvalan, and T. Kjellstrom. 1999. "How Much Global Ill Geneva: WHO. Health Is Attributable to Environmental Factors?" Epidemiology 10: ------. 1995. Lead, Inorganic. Environmental Health Criteria 165. 573­84. Geneva: WHO. Smith, R. S. 2003. "Naturally Occurring Hazards." Article prepared for ------. 1996. Health Consequences of the Chernobyl Accident: Scientific World Water Day, March 22, World Health Organization, Geneva. Report. Geneva: WHO. http://www.worldwaterday.org/2001/thematic/natural.html. ------. 1997. Health and Environment in Sustainable Development. Tengs, T. O., M. E. Adams, J. S. Pliskin, D. G. Safran, J. E. Siegel, M. C. Document WHO/EHG/97.8. Geneva: WHO. Weinstein, and J. D. Graham. 1995. "Five-Hundred Life-Saving ------. 1999. Public Health and Chemical Incidents: Guidance for National Interventions and Their Cost-Effectiveness." Risk Analysis 15: 369­90. and Regional Policy Makers. Cardiff, U.K.: University of Wales Institute, U.K. Ministry of Health. 1954. Mortality and Morbidity during the London WHO Collaborating Centre for Chemical Incidents. http://www. Fog in December 1952. London: U.K. Ministry of Health. who.int/ipcs/publications/en/Public_Health_Management.pdf. UNEP (United Nations Environment Programme). 2002. "Cleaner ------. 2000. Air Quality Guidelines for Europe. 2nd ed. Copenhagen: Production. Seventh International High-Level Seminar, Prague." WHO. Industry and Environment 25 (3­4): 1­109. ------. 2001. Arsenic and Arsenic Compounds. Environmental Health UNESCO (United Nations Educational, Scientific and Cultural Criteria 224. Geneva: WHO. Organization). 2003. Water for People, Water for Life. Paris: UNESCO. ------. 2002. World Health Report 2002. Geneva: WHO. United Nations. 1997. Success Stories from India: Minimizing Waste by ------. 2003. Guidelines for Drinking Water Quality. 3rd ed. Geneva: DESIRE. Report for the special session of the General Assembly, WHO. http://www.who.int/water_sanitation_health/dwq/guidelines Earth Summit+5, New York, June 23­27. http://www.un.org/esa/ 3rd/en/. earthsummit/unido3.htm. WHO and International Programme on Chemical Safety. 2002. Global U.S. Congress. Senate. Governmental Affairs Committee. 1999. Testimony Assessment of the State of Science of Endocrine Disruptors. Document of Professor Lisa Heinzerling Concerning the Nomination of John D. WHO/PCS/EDC/02.2. Geneva: WHO and the International Graham to Be Administrator of the Office of Budget and Regulatory Programme on Chemical Safety. Affairs, Office of Management and Budget. http://www.citizen.org/ congress/regulations/graham/heinzerling_testimony.html. World Bank. 1999. Pollution Prevention and Abatement Handbook 1998. Washington, DC: World Bank. http://wbln0018.worldbank.org/ U.S. Environmental Protection Agency. 1999. The Benefits and Costs of the essd/essd.nsf/GlobalView/PPAH. Clean Air Act 1990 to 2010. Report to Congress. Washington, DC: U.S. Environmental Protection Agency. http://www.epa.gov/oar/sect812/. ------. 2004. Air Pollution Calculation Toolkit. Washington, DC: World Bank. http://lnweb18.worldbank.org/essd/essdext.nsf/46ByDocName/ ------. 2000. Superfund: 20 Years of Protecting Human Health and ToolkitsGlobalOverlay. the Environment. EPA 540-R-00-007. Washington, DC: U.S. Environmental Protection Agency. http://www.epa.gov/superfund. Xu, X., and L. Wang. 1993."Association of Indoor and Outdoor Particulate Level with Chronic Respiratory Disease." American Review of Voorhees, A. S., S. Araki, R. Sakai, and H. Sato. 2000. "An Ex Post Cost- Respiratory Diseases 148: 1516­22. Benefit Analysis of the Nitrogen Dioxide Air Pollution Control Program in Tokyo." Journal of the Air and Waste Management Yassi, A-L., T. Kjellstrom, T. deKok, and T. Guidotti. 2001. Basic Association 50: 391­410. Environmental Health. New York: Oxford University Press. Voorhees, A. S., R. Sakai, S. Araki, H. Sato, and A. Otsu. 2001."Cost-Benefit Zhang, C., M. Huq, S. Dasgupta, and D. Wheeler. 1996. "Water Pollution Analysis Methods for Assessing Air Pollution Control Programs in Abatement by Chinese Industry: Cost Estimates and Policy Urban Environments: A Review." Environmental Health and Preventive Implications." Working Paper 1630, World Bank, Washington, DC. Medicine 6: 63­73. Vorosmarty, C. J., P. Green, J. Salisbury, and R. B. Lammers. 2000. "Global Water Resources: Vulnerability from Climate Change and Population Growth." Science 289: 283­88. 832 | Disease Control Priorities in Developing Countries | Tord Kjellstrom, Madhumita Lodh, Tony McMichael, and others Chapter 44 Prevention of Chronic Disease by Means of Diet and Lifestyle Changes Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, Courtenay Dusenbury, Pekka Puska, and Thomas A. Gaziano Coronary artery disease (CAD), ischemic stroke, diabetes, and facilities, an outcome that is not surprising, because their rates some specific cancers, which until recently were common only have historically been extremely low in developing countries in high-income countries, are now becoming the dominant with few medical facilities. However, preventing these diseases sources of morbidity and mortality worldwide (WHO 2002). will require changes in behaviors related to smoking, physical In addition, rates of cancers and cardiovascular disease (CVD) activity, and diet; investments in education, food policies, and among migrants from low-risk to high-risk countries almost urban physical infrastructure are needed to support and always increase dramatically. In traditional African societies, encourage these changes (see box 44.1). for example, CAD is virtually nonexistent, but rates among African Americans are similar to those among Caucasian Americans. These striking changes in rates within countries CHRONIC DISEASE PREVENTION over time and among migrating populations indicate that the primary determinants of these diseases are not genetic but In this section, we briefly review dietary and lifestyle changes environmental factors, including diet and lifestyle. Thus, con- that reduce the incidence of chronic disease. The potential siderable research has been aimed at identifying modifiable magnitude of benefit is also discussed. determinants of chronic diseases. Prospective epidemiological studies, some randomized pre- vention trials, and many short-term studies of intermediate Recommended Lifestyle Changes endpoints such as blood pressure and lipids have revealed a Specific changes in diet and lifestyle and likely benefits are good deal about the specific dietary and lifestyle determinants summarized in table 44.1. These relationships and supporting of major chronic diseases. Most of these studies have been con- evidence are summarized here. ducted in Western countries, in part because of the historical importance of these diseases in the West, but also because they Avoid Tobacco Use. Avoidance of smoking by preventing have the most developed research infrastructure. A general initiation or by cessation for those who already smoke is the conclusion is that reducing identified, modifiable dietary and single most important way to prevent CVD and cancer (chap- lifestyle risk factors could prevent most cases of CAD, stroke, ter 46). Avoiding the use of smokeless tobacco will also prevent diabetes, and many cancers among high-income populations a good deal of oral cancer. (Willett 2002). These findings are profoundly important, because they indicate that these diseases are not inevitable con- Maintain a Healthy Weight. Obesity is increasing rapidly sequences of a modern society. Furthermore, low rates of these worldwide (chapter 45). Even though obesity--a body mass diseases can be attained without drugs or expensive medical index (BMI) of 30 or greater--has received more attention 833 Box 44.1 The Insulin Resistance Syndrome In recent years, researchers have recognized the insulin most direct causes are overweight and inactivity, but resistance syndrome (also known as the metabolic syn- dietary factors contribute. Genetic factors, which are drome) as a common contributing factor to the develop- probably beneficial during periods of food shortages, also ment of diabetes, CAD, and some cancers. The syndrome play a role. Recent evidence indicates that the populations is characterized by increased waist circumference, low of Asia, Latin America, and probably Africa are particu- HDL (high-density lipoprotein) cholesterol, high levels of larly susceptible (Dickinson and others 2002; Harris and triglycerides, hypertension, and glucose intolerance. The others 1998). Table 44.1 Convincing and Probable Relationships between Dietary and Lifestyle Factors and Chronic Diseases Dietary and Type 2 Dental Birth Metabolic Sexual lifestyle factors CVD diabetes Cancer disease Fracture Cataract defects Obesity syndrome Depression dysfunction Avoid smoking Pursue physical activity Avoid overweight Diet Consume healthy types of fatsa Eat plenty of fruits and vegetables Replace refined grains with whole grains Limit sugar intakeb Limit excessive calories Limit sodium intake Source: Authors' summary of a review by the WHO and FAO 2003; Bacon and others 2003; Fox 1999; IARC 2002. Note: Bold convincing; Standard probable relation; increase in risk; decrease in risk. a. Replace trans and saturated fats with mono- and polyunsaturated fats, including a regular source of N-3 fatty acids. b. Includes limiting sugar-based beverages. than overweight, overweight (BMI of 25 to 30) is typically even experience elevated mortality from cancers of the colon, breast more prevalent and also confers elevated risks of many dis- (postmenopausal), kidney, endometrium, and other sites eases. For example, overweight people experience a two- to (Calle and others 2003). threefold elevation in the risks of CAD and hypertension and a Many people with a BMI of less than 25 have gained sub- more than tenfold increase in the risk of type 2 diabetes com- stantial weight since they were young adults and are also at pared with lean individuals (BMI less than 23) (Willett, Dietz, increased risk of these diseases, even though they are not and Colditz 1999). Both overweight and obese people also technically overweight (Willett, Dietz, and Colditz 1999). For 834 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others example, in rural China, where the average BMI was less than Aspects of the food supply unrelated to its macronutrient 21 for both men and women, F. B. Hu and others (2000) found composition are also likely to be contributing to the global rise that the prevalence of hypertension was nearly five times in obesity. Inexpensive food energy from refined grains, sugar, greater for those with a BMI of approximately 25 than for the and vegetable oils has become extremely plentiful in most leanest people. Because many Asians are experiencing adverse countries. Food manufacturers and suppliers use carefully consequences of excess body fat with a BMI of less than 25, the researched methods to make products based on these cheap definition of overweight for Asia has recently been expanded to ingredients maximally convenient and attractive. include a BMI of 23 to 25 (WHO 2000). For most people, unless obviously malnourished as an adolescent or young Maintain Daily Physical Activity and Limit Television adult, bodyweight should ideally not increase by more than 2 Watching. Contemporary life in developed nations has or 3 kilograms after age 20 to maintain optimal health (Willett, markedly reduced people's opportunities to expend energy, Dietz, and Colditz 1999). Thus, a desirable weight for most whether in moving from place to place, in the work environ- people should be within the BMI range of 18.5 to 25.0, and ment, or at home (Koplan and Dietz 1999). Dramatic preferably less than 23. reductions in physical activity are also occurring in developing Additional valuable information can be obtained by meas- countries because of urbanization, increased availability of uring waist circumference, which reflects abdominal fat accu- motorized transportation to replace walking and bicycle riding, mulation. In many studies, waist circumference is a strong and mechanization of labor. However, regular physical activity predictor of CAD, stroke, and type 2 diabetes, even after con- is a key element in weight control and prevention of obesity trolling for BMI (Willett, Dietz, and Colditz 1999). A waist (IARC 2002; Swinburn and others 2004). For example, among circumference of approximately 100 centimeters for men and middle-aged West African women, more walking was asso- 88 centimeters for women has been used as the criterion for ciated with a three-unit lower BMI (Sobngwi, Gautier, and the upper limit of the healthy range in the United States, but Mbanya 2003), and in China, car owners are 80 percent more for many people this extent of abdominal fat would be far likely to be obese (Hu 2002). above optimal. Because abdominal circumference is easily In addition to its key role in maintaining a healthy weight, assessed, even where scales may not be available, further work regular physical activity reduces the risk of CAD, stroke, type 2 to develop locally appropriate criteria could be worthwhile. In diabetes, colon and breast cancer, osteoporotic fractures, the meantime, increases of more than 5 centimeters can be osteoarthritis, depression, and erectile dysfunction (table 44.1). used as a basis for recommending changes in activity patterns Important health benefits have even been associated with walk- and diet. ing for half an hour per day, but greater reductions in risk are Views about the causes of obesity and ways to prevent or seen with longer durations of physical activity and more reduce it have been controversial. Diets low in fat and high in intense activity. carbohydrates were believed to limit caloric intake sponta- The number of hours of television watched per day is asso- neously and thus to control adiposity, but such diets have not ciated with increased obesity rates among both children and reduced bodyweight in trials that have lasted for a year or more adults (Hernandez and others 1999; Ruangdaraganon and (Willett and Leibel 2002). Some researchers have suggested that others 2002) and with a higher risk of type 2 diabetes and gall- diets with a high energy density, referring to the amount of stones (F. B. Hu and others 2001; Leitzmann and others 1999). energy per volume, offer an alternative explanation for the This association is likely attributable both to reduced physical observed increases in obesity (Swinburn and others 2004), but activity and to increased consumption of foods and beverages long-term studies have not examined this theory. Sugar- high in calories, which are typically those promoted on televi- sweetened beverages contribute significantly to the overcon- sion. Decreases in television watching reduce weight sumption of calories, in part because calories in fluid form (Robinson 1999), and the American Academy of Pediatrics appear to be poorly regulated by the body (E. A. Bell, Roe, and recommends a maximum of two hours of television watching Rolls 2003). In children, an increase in soda consumption of per day. one serving per day was associated with an odds ratio of 1.6 for incidence of obesity (Ludwig, Peterson, and Gortmaker 2001), Eat a Healthy Diet. Medical experts have long recognized the and in a randomized trial, replacement of a standard soda with effects of diet on the risk of CVD, but the relationship between a zero-calorie diet soda was associated with significant weight diet and many other conditions, including specific cancers, dia- loss (Raben and others 2002). Reductions in dietary fiber and betes, cataracts, macular degeneration, cholelithiasis, renal increases in the dietary glycemic load (large amounts of rapidly stones, dental disease, and birth defects, have been documented absorbed carbohydrates from refined starches and sugar) may more recently. The following list discusses six aspects of diet also contribute to obesity (Ebbeling and others 2003; Swinburn for which strong evidence indicates important health implica- and others 2004). tions (table 44.1). These goals are consistent with a detailed Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 835 2003 World Health Organization (WHO) report (WHO and of fruits and vegetables is low. For example, in northern FAO 2003). China, approximately half the adult population is deficient in folic acid (Hao and others 2003). · Replace saturated and trans fats with unsaturated fats, · Consume cereal products in their whole-grain, high-fiber including sources of omega-3 fatty acids. Replacing saturated form. Consuming grains in a whole-grain, high-fiber form fats with unsaturated fats will reduce the risk of CAD (F. B. has double benefits. First, consumption of fiber from cereal Hu and Willett 2002; Institute of Medicine 2002; WHO and products has consistently been associated with lower risks of FAO 2003) by reducing serum low-density lipoprotein CAD and type 2 diabetes (F. B. Hu, van Dam, and Liu 2001; (LDL) cholesterol. Also, polyunsaturated fats (including the F. B. Hu and Willett 2002), which may be because of both long-chain omega-3 fish oils and probably alpha-linoleic the fiber itself and the vitamins and minerals naturally pres- acid, the primary plant omega-3 fatty acid) can prevent ent in whole grains. High consumption of refined starches ventricular arrhythmias and thereby reduce fatal CAD. In a exacerbates the metabolic syndrome and is associated with case-control study in Costa Rica, where fish intake was higher risks of CAD (F. B. Hu and Willett 2002) and type 2 extremely low, the risk of myocardial infarction was 80 per- diabetes (F. B. Hu, van Dam, and Liu 2001). Second, higher cent lower in those with the highest alpha-linoleic acid consumption of dietary fiber also appears to facilitate intake (Baylin and others 2003). Intakes of omega-3 fatty weight control (Swinburn and others 2004) and helps pre- acids are suboptimal in many populations, particularly if vent constipation. fish intake is low and the primary oils consumed are low in · Limit consumption of sugar and sugar-based beverages. Sugar omega-3 fatty acids (for example, partially hydrogenated (free sugars refined from sugarcane or sugar beets and high- soybean, corn, sunflower, or palm oil). These findings fructose corn sweeteners) has no nutritional value except for have major implications, because changes in the type of oil calories and, thus, has negative health implications for those used for food preparation are often quite feasible and not at risk of overweight. Furthermore, sugar contributes to the expensive. dietary glycemic load, which exacerbates the metabolic syn- Trans fatty acids produced by the partial hydrogenation drome and is related to the risk of diabetes and CAD (F. B. of vegetable oils have uniquely adverse effects on blood Hu, van Dam, and Liu 2001; F. B. Hu and Willett 2002; lipids (F. B. Hu and Willett 2002; Institute of Medicine Schulze and others 2004). WHO has suggested an upper 2002) and increase risks of CAD (F. B. Hu and Willett limit of 10 percent of energy from sugar, but lower intakes 2002); on a gram-for-gram basis, both the effects on blood are usually desirable because of the adverse metabolic effects lipids and the relationship with CAD risk are considerably and empty calories. more adverse than for saturated fat. In many developing · Limit excessive caloric intake from any source. Given the countries, trans fat consumption is high because partially importance of obesity and overweight in the causation of hydrogenated soybean oil is among the cheapest fats avail- many chronic diseases, avoiding excessive consumption able. In South Asia, vegetable ghee, which has largely of energy from any source is fundamentally important. replaced traditional ghee, contains approximately 50 per- Because calories consumed as beverages are less well- cent trans fatty acids (Ascherio and others 1996). regulated than calories from solid food, limiting the con- Independent of other risk factors, higher intakes of trans fat sumption of sugar-sweetened beverages is particularly and lower intakes of polyunsaturated fat increase risk of important. type 2 diabetes (F. B. Hu, van Dam, and Liu 2001). · Limit sodium intake. The principle justification for limiting · Ensure generous consumption of fruits and vegetables and ade- sodium is its effect on blood pressure, a major risk factor quate folic acid intake. Strong evidence indicates that high for stroke and coronary disease (chapter 33). WHO has intakes of fruits and vegetables will reduce the risk of CAD suggested an upper limit of 1.7 grams of sodium per day and stroke (Conlin 1999). Some of this benefit is mediated (5 grams of salt per day) (WHO and FAO 2003). by higher intakes of potassium, but folic acid probably also plays a role (F. B. Hu and Willett 2002). Supplementation with folic acid reduces the risk of neural tube defect preg- nancies. Substantial evidence also suggests that low folic Potential of Dietary and Lifestyle Factors to Prevent acid intake is associated with greater risk of colon--and Chronic Diseases possibly breast--cancer and that use of multiple vitamins Several lines of evidence indicate that realistic modifications of containing folic acid reduces the risk of these cancers diet and lifestyle can prevent most CAD, stroke, diabetes, colon (Giovannucci 2002). Findings relating folic acid intake to cancer, and smoking-related cancers. Less progress has been CVD and some cancers have major implications for many made in identifying practically modifiable causes of breast and parts of the developing world. In many areas, consumption prostate cancers. 836 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others Box 44.2 Success in Finland Finland provides one of the best-documented examples cant improvements were documented in smoking, choles- of a community intervention. In 1972, Finland had the terol, and blood pressure. By 1992, CVD mortality rates world's highest CVD mortality rate. Planners examined for men age 35 to 64 had dropped by 57 percent. The pro- the policy and environmental factors contributing to CVD gram was so successful that it was expanded to include and sought appropriate changes, such as increased avail- other lifestyle-related diseases. Twenty years later, major ability of low-fat dairy products, antismoking legislation, reductions in CVD risk-factor levels, morbidity, and mor- and improved school meals. They used the media; schools; tality were attributed to the project. Recent data show a worksites; and spokespersons from sports, education, and 75 percent decrease in CVD mortality (Puska and others agriculture to educate residents. After five years, signifi- 1998). One line of evidence is based on declines in CAD in coun- INTERVENTIONS tries that have implemented preventive programs. Rates of CAD mortality have been cut in half in several high-income Interventions aimed at changing diet and lifestyle factors countries, including Australia, the United Kingdom, and the include educating individuals, changing the environment, United States. The most dramatic example is that of Finland modifying the food supply, undertaking community interven- (box 44.2). tions, and implementing economic policies. In most cases, Other evidence derives from randomized intervention quantifying the effects of the intervention is difficult, because studies. These often have serious limitations for estimating the behavioral changes may take many years and synergies are potential magnitude of benefits, because typically only one or potentially important but hard to estimate in formal studies. a few factors are modified, durations are usually only a few Substantial nihilism often exists regarding the ability to years, and noncompliance with lifestyle change is often sub- change populations' diets or behaviors, but major changes are stantial. Nevertheless, some examples are illustrative of the possible over extended periods of time. For example, per potential benefit. In two randomized studies among adults capita egg consumption in the United States decreased from at high risk of type 2 diabetes, those assigned to a program approximately 420 to 270 per year between 1940 and 1990 emphasizing dietary changes, weight loss, and physical activity following recommendations for preventing CAD (though in experienced only half the risk of incident diabetes (Knowler reality, the evidence for benefits was meager). Similarly, the and others 2002; Tuomilehto and others 2001). The Lyon prevalence of smoking, despite its being a physically addictive Heart Study, conducted among those with existing heart dis- behavior, halved among men in the United States between ease, found a Mediterranean-type diet high in omega-3 fatty 1965 and 2000. Because changing behaviors related to diet and acids reduced recurrent infarction by 70 percent compared lifestyle require sustained efforts, long-term persistence is with an American Heart Association diet (de Lorgeril and needed. However, opportunities exist that do not require indi- others 1994). vidual behavior changes, and these can lead to more rapid A third approach is to estimate the percentage of disease benefits. that is potentially preventable by reducing multiple behavioral risk factors using prospective cohort studies. Among U.S. Educational Interventions adults, more than 90 percent of type 2 diabetes, 80 percent Efforts to change diets, physical activity patterns, and other of CAD, 70 percent of stroke, and 70 percent of colon cancer aspects of lifestyle have traditionally attempted to educate indi- are potentially preventable by a combination of nonsmoking, viduals through schools, health care providers, worksites, and avoidance of overweight, moderate physical activity, healthy general media. These efforts will continue to play an important diet, and moderate alcohol consumption (Willett 2002). role, but they can be strongly reinforced by policy and environ- Collectively, these findings indicate that the low rates of mental changes. these diseases suggested by international comparisons and time trends are attainable by realistic, moderate changes that are School-Based Programs. School-based programs include the compatible with 21st-century lifestyles. roles of nutrition and physical activity in maintaining physical Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 837 Box 44.3 The Planet Health Program Planet Health, developed for middle school students, in education complement the classroom lessons. Teacher the United States, has an immediate goal of reducing tele- training, student self-assessment using graphs, and stu- vision viewing time with the long-range goal of prevent- dent reflection about enjoyable activities that could ing unhealthy weight gain (Gortmaker and others 1999). replace at least a portion of the time they spend watching Teachers incorporate messages about reducing television television are key elements. This program has reduced tel- watching, nutrition, and increasing fitness into mathe- evision watching and weight in girls (Gortmaker and oth- matics, social studies, science, and language arts lessons. ers 1999). Because the program is integrated into existing Fitness units and periodic "FitChecks" during physical classes, its cost is minimal. Box 44.4 Live for Life® Johnson & Johnson introduced Live for Life in 1979 with were offered US$500 in benefit credits for participation. the goal of making its employees the healthiest in the The program included routine health risk assessment, world (Bly, Jones, and Richardson 1986). In 1993, the health promotion after recovery from a medical event, and company integrated its health and wellness program with support when returning to work after a major illness. Even its disability management, employee assistance, and occu- though the intervention program had little effect on body- pational medicine programs. Instead of using physicians weight, physical fitness did increase. By the end of the and nurses to treat symptoms, the combined program third year, savings to the company were more than sought to use a variety of health professionals to change US$400 per year per employee. individual behavior and improve health status. Employees and mental health (box 44.3). School food services should pro- programs during breaks or after work; improving the physical vide healthy meals, both because they directly affect health and environment to promote activity; and providing healthier because they provide a special opportunity to teach by exam- foods in cafeterias (box 44.4). Worksite health promotion can ple. In many countries, school-based physical education result in a positive return on investment through lower health remains a significant source of physical activity for young peo- costs and fewer sick days. ple. In China, 72 percent of children age 6 to 18 engage in mod- erate to vigorous physical activity for a median of 90 to 100 min- Interventions by Health Care Providers. Controlled inter- utes per week (Tudor-Locke and others 2003). Maintaining vention trials for smoking cessation and physical activity have these programs should be a high priority because they have shown that physician counseling, especially when accompanied likely contributed to the historically low rates of obesity in such by supporting written material, can be efficacious in modifying countries. behavior. Studies of dietary counseling by physicians indicate that even brief messages about nutrition can influence behav- Worksite Interventions. Worksite interventions can effi- ior and that the magnitude of the effect is related to the inten- ciently include a wide variety of health promotion activities sity of the intervention (Pignone and others 2003). Identifying because workers spend a large portion of their waking hours patients who are overweight or obese, or who are gaining and eat a large percentage of their food there. Interventions can weight but are not yet overweight, is an initial step in prevent- include educating employees; screening them for behavioral ing and treating overweight. However, many physicians are not risk factors; offering incentive programs to walk, ride a bicycle, well trained to measure and calculate BMI and identify weight or take public transportation to work; offering exercise problems. 838 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others Box 44.5 Reducing Automobile Use in Brazil Curitiba, Brazil, provides an example of the benefits of a are used to quickly and efficiently transport individuals strategy that reduces automobile use and increases use of from residential neighborhoods to express bus lines. These public transportation. In 1965, city planners adopted a bus lines run almost every 90 seconds and can carry up to master plan that promoted development along designated 270 passengers each. Compared with other Brazilian cities corridors along with a bus system so efficient that it has of its size, Curitiba uses 30 percent less gasoline per capita, virtually eliminated the need for automobiles. Minibuses and its air pollution is among the lowest in the nation. Transportation Policy and Environmental Design developed to discourage private automobile use and to Transportation policies and the design of urban environments promote public transportation, walking, and bicycling (see are fundamental determinants of physical activity and there- box 44.5). Singapore has long been in the lead in relation to fore influence the risks of obesity and other chronic diseases. such efforts: a combination of limiting the number of licenses Countries can take a number of steps to make positive changes. issued, implementing a vehicle quota system, and introducing a road pricing system has limited personal car ownership and Limit the Role of Automobiles. In wealthy countries, the congestion throughout the country. Other nations and regions automobile has strongly influenced the trend toward low- are now enacting similar road pricing systems or congestion density, automobile-based suburban developments, many built taxes. For example, London's congestion charging system levies without sidewalks. These sprawling settlements tend to have a fee of approximately US$8 per day for cars entering central few services within walking distance and are usually not linked London. Since its inception in 2003, the charge has reduced to public transporationt. Dependence on automobiles affects congestion in the city and is expected to channel funds back physical activity, because those who use public transportation into the city's transportation facilities. tend to walk more. In a prospective study in eight provinces in Unfortunately some countries, particularly China, have China, 14 percent of households acquired a car between 1889 taken a different approach to their future transportation needs. and 1997, and the likelihood of men becoming obese during Government initiatives that encourage families to buy automo- the same period was twice as great in households that acquired biles include lowering taxes, simplifying registration proce- a car than in those that did not (A. C. Bell, Ge, and Popkin dures, and allowing foreign financing. In Beijing alone, resi- 2002). dents purchased 400,000 cars in 2003. National policies strongly influence automobile use and dependency. In the United States, low taxes on gasoline, free Promote Walking and Bicycle Riding. Walking or cycling for parking, and wide streets encourage car ownership: almost transportation and leisure are effective and practical means of 92 percent of U.S. households own at least one car, and 59 per- engaging in physical activity and are still the most common cent own two or more cars (Pucher and Dijkstra 2003). In ways to travel in many developing countries. In Bangkok and contrast, in most of Western Europe, narrow streets, limited Manila, only 25 percent of travel is by car, motorcycle, or taxi, parking, and high gasoline prices make the costs of automobile compared with 75 percent by public transportation or walking use almost double those in the United States (Pucher and (Pendakur 2000). In Madras, India, only 8 percent of the pop- Dijkstra 2003). As a result, Europeans walk or bike more and ulation travels by private, motorized transportation; 22 per- use their cars approximately 50 percent less than their cent of people walk; 20 percent bike; and the rest use public American counterparts. Investment in roads rather than in transportation (Pendakur 2000). In China, approximately public transportation creates a vicious cycle: poor public trans- 90 percent of the urban population walks or rides a bicycle to portation systems lead to more dependency on the automobile. work, shopping, or school each day (G. Hu and others 2002). As car use grows, injuries and deaths associated with auto- Walking or biking is more likely to be prevalent in smaller mobile accidents also grow. In China, the number of four- cities--that is, those with 1 million to 5 million people--than wheeled vehicles increased from about 60,000 to more than in larger ones. 50 million between 1951 and 1999, and traffic fatalities Bicycle riding and walking are also important for children's increased from about 6,000 to more than 413,000 (S. Y. Wang health. Most American children do not walk or bike to school, and others 2003). Many innovative strategies have been even when distances are short (box 44.6). In contrast, almost Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 839 Box 44.6 Walking and Cycling to School One of the most effective ways to promote walking and walking routes and proposed solutions for problem areas. cycling is through local schools. The Safe Routes to School The program also sponsors walk- and bike-to-school days, program (http://www.saferoutestoschools.org/), estab- frequent-rider contests, and other promotional events lished in Marin County, California, is a private-public (Staunton, Hubsmith, and Kallins 2003). partnership that created a citywide map of safe biking and 90 percent of Chinese children under 12 walk or ride a bicycle linked to decreases in mental health and social capital to school (Hu 2002). (Frumkin 2002) as well as anger and frustration over long com- In many areas, the shift toward private car use has not yet mutes (Surface Transportation Policy Project 1999). Sprawl begun and can perhaps be forestalled by policies that benefit adversely affects the elderly in particular because they are walkers and cyclists rather than drivers. Such policies include unable to walk to places of interest and many cannot drive. implementing road designs that promote a safe and well-lit Such isolation does not promote good physical or mental environment for walking and cycling, including traffic-calming health. measures to reduce automobile speeds. The so-called smart growth movement has resulted from Many Western European countries have taken steps to concerns about urban sprawl and unsustainable development increase safety for cyclists and walkers. In Germany and the and is encouraging governments worldwide to rethink how Netherlands, bike paths serve as travel routes, not just weekend they develop new areas and redevelop older suburbs and cities. recreational destinations as they do in the United States. The Smart growth principles include mixing land uses, using com- former countries have invested heavily in bike paths and have pact building designs, including a range of transportation and also created extensive car-free areas in cities, with well-lit side- housing choices, building walker-friendly neighborhoods in walks, clearly marked crosswalks, and pedestrian islands that attractive communities with a distinctive sense of place, and have improved safety. Both countries have increased the num- implementing a philosophy of directing development toward ber of bicycle-friendly streets (on which cars are permitted but existing communities and the preservation of open space bicycles have the right of way) and have created systems to sep- (Office of the Administrator 2001) (box 44.7). arate streams of traffic, including cars, pedestrians, and bicy- The involvement of public health practitioners in trans- cles. A meta-analysis of selected traffic-calming studies in many portation planning and building design is becoming more countries reported reductions in traffic speed, accidents, common. In Edinburgh, a health impact assessment conducted injuries, and fatalities and an increase in bicycle use and walk- on proposed options for transportation policy showed the ing (Bunn and others 2003). effects of specific choices on both affluent members of the community and the poor. Its recommendations, now adopted, Design Cities and Towns to Promote Health. Handy and included new spending on pedestrian safety, a citywide bicycle others' (2002) comprehensive assessment of recent research on network, more greenways and park-and-ride programs, and urban planning concludes that a combination of urban design, more rail transportation or bus services. Priorities are to bene- land-use patterns, and transportation systems that promotes fit pedestrians first, cyclists second, public transportation users walking and bicycling will help create active, healthier, and third, freight and delivery people fourth, and car users last. more livable communities. In densely developed cities that Establishing criteria for building design can also lead to have been built around public transportation rather than away increases in physical activity. For example, increasing signage from it, individuals are much more likely to take public transit, promoting stair use, as well as the attractiveness of the facilities walk, or bicycle than in other areas and to weigh less and be less themselves, encourages people to use the stairs (Boutelle and likely to suffer from hypertension (Ewing, Schieber, and Zegeer others 2001) (box 44.8). 2003; Lopez 2004; Saelens, Sallis, and Frank 2003). Those living in walker-friendly neighborhoods also appear to be more mentally healthy and are more likely to know their Improved Food Supply neighbors, to be socially active, and to participate in the politi- People's diets can be enhanced by improving the food supply. cal process (Leyden 2003). In contrast, urban sprawl has been The usual position of the food industry is that it simply 840 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others Box 44.7 Enhancing Urban Life in the Republic of Korea In Seoul, the government is managing growth by creating the balance of employment away from one centralized six satellite communities with high-rise residential build- location to provide a more regional balance. Major ings outside the city center. These communities are expressways are being removed to create parks, sidewalks, intended to become new job-creation centers and to shift and bikeways (http://www.itdp.org/STe/ste6/#seoul). Box 44.8 Promoting Physical Activity in Brazil One successful example of increasing activity is Agita São message is displayed on electricity bills and stickers, and it Paulo, a multilevel physical activity initiative designed for is touted by radio stations and other media outlets. the 34 million citizens of Brazil's São Paulo state (Matsudo After four years, 55.7 percent of those surveyed had and others 2002). The program was launched in 1996 to heard about Agita, 37 percent knew its purpose, and those increase the public's knowledge of the benefits of exercise who knew of the program's purpose were more likely to be and expand participation in physical fitness activities by active. Agita appears to have played a role in increasing encouraging people to do 30 minutes of moderate activity activity in the region (Matsudo and others 2002). It is at least five times a week. As elsewhere, program designers closely linked to a national program to promote healthy perceived a lack of time as the major factor preventing diets and active lifestyles by nutritional content labeling, daily exercise. They chose three settings as places to pro- promotion of healthy diets in schools, communication of mote activity: home (gardening, chores, avoidance of tele- guidelines for healthy eating, and encouragement of inno- vision watching); transportation (walking, taking the vative community-based initiatives (Coitinho, Monteiro, stairs); and leisure time (dancing). Agitol, a prescription and Popkin 2002). for exercise, was developed for physicians to dispense. Its provides whatever consumers demand, but this argument is Drug Administration announced that food manufacturers had misleading, because the industry spends more than US$12 bil- to include trans fatty acid content on the standard food label. lion annually to influence consumer choices just within the Following imposition of this requirement, several large food United States and many times this amount globally. Much of companies said that they would reduce or eliminate trans fats, this sum goes to promote foods with adverse health effects, and and many more are planning to do so (U.S. Food and Drug children are primary targets. Administration 2003). In Mauritius, the government required a change in the commonly used cooking oil from mostly palm Improving Processing and Manufacturing. Altering the oil to soybean oil, which changed people's fatty acid intake and manufacturing process can rapidly and effectively improve reduced their serum cholesterol levels (Uusitalo and others diets because such action does not require the slow process of 1996). Changes in types of fat can often be almost invisible and behavioral change. One example is eliminating the partial inexpensive. Omega-3 fatty acid intakes can be increased by hydrogenation of vegetable oils, which destroys essential incorporating oils from rapeseed, mustard, or soybean into omega-3 fatty acids and creates trans fatty acids. European manufactured foods, cooking oils sold for use at home, or both. manufacturers have largely eliminated trans fatty acids from Selective breeding and genetic engineering provide alternative their food supply by altering production methods. ways to improve the healthfulness of oils by modifying their Regulations can facilitate changes in manufacturing directly fatty acid composition. or indirectly by providing an incentive for manufacturers to When the consumption of processed food is high, a reduc- change their processes. For example, in 2003, the U.S. Food and tion in salt consumption will usually require changes at the Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 841 manufacturing level, because processed food is a major salt Another strategy is to protect consumers from aggressive source. If the salt content of foods is reduced gradually, the marketing of unhealthy foods. Producers spend billions of dol- change is imperceptible to consumers. Coordination among lars a year encouraging children to consume foods that are manufacturers or government regulation is needed; otherwise detrimental to their health. Manufacturers and fast-food chains producers whose foods are lower in salt may be placed at a dis- personify food products with cartoon characters; display food advantage. Unfortunately, good examples are not available. brands on toys; and issue "educational" card games that subvert Another example of improved processing would be to reduce children's natural gift for play, story telling, and make believe. the refining of grain products, which can be done in small, The willingness to limit advertising depends on a country's almost invisible decrements. political culture, but the public clearly distinguishes between advertising aimed at adults and that targeted at children. For example, in the United States, 46 percent of adults surveyed Fortifying Food. Food fortification has eliminated iodine supported restrictions on advertising to children (Blendon deficiency, pellagra, and beriberi in much of the world. In 2002). Restrictions can range from banning advertising to regions where iodine deficiency remains a serious problem, children to limiting the types of products that advertisers may fortification should be a high priority. Folic acid intake is sub- promote to this audience. optimal in many regions of both developing and developed countries. Fortifying foods with folic acid is extremely inex- pensive and could substantially reduce the rates of several Initiatives at the Community Level chronic diseases. Grain products--such as flour, rice, and Nations and regions can promote a variety of initiatives to pasta--are usually the best foods to fortify, and in many coun- encourage greater physical activity and better nutrition. These tries, they are already being fortified with other B vitamins. initiatives are likely to be most effective when they are multi- Since 1998, grain products in the United States have been faceted and coordinated and when they are developed with the fortified with folic acid, which has almost eliminated folate active involvement of individuals and organizations within deficiency, and rates of neural tube defect pregnancies have communities (Puska and others 1998). declined by about 19 percent (Honein and others 2001). Where Many countries are undertaking efforts to educate their intakes of vitamins B12 and B6 are also low and contribute to populations about healthy lifestyles. In the Islamic Republic of elevations of homocysteine, as among vegetarian populations Iran, the Isfahan Healthy Heart Program, a WHO collaborating in India, simultaneous fortification of food with these vitamins center for research and training for CVD control, prevention, should be considered. The effects of fortification on reducing and rehabilitation for cardiac patients, has developed a com- CVD are not considered proven, but the potential benefits are prehensive, integrated community intervention that involves huge; therefore, intervention trials to evaluate the effects of schools, worksites, health care facilities, food services, urban fortification should be a high priority. planners, and the media. Physical activity is promoted by creating safe routes for walking and bicycle riding and Increasing the Availability and Reducing the Cost of Healthy by organizing recreational walking that involves entire families Foods. Policies regarding the production, importation, distri- (http://ihhp.mui.ac.ir). bution, and sale of specific foods can influence their cost and South Africa's Community Health Intervention Pro- availability. Policies may be directed at the focus of agricultur- gramme, a partnership between an insurance company and an al research and the types of production promoted by extension academic institution, has created programs targeted to specific services. Policies often promote grains, dairy products, sugar, age groups, including children and older adults. The program's and beef, whereas those that encourage the production and twice-weekly classes have reduced blood pressure and increased consumption of fruits, vegetables, nuts, legumes, whole grains, strength and balance (Lambert, Bohlmann, and Kolbe- and healthy oils would tend to enhance rather than reduce Alexander 2001) (box 44.9). health. Singapore's Fit and Trim Program uses a multidisciplinary approach to increase physical activity and healthy diets among schoolchildren. Between 1992 and 2000, the rate of obesity Promoting Healthy Food Choices and Limiting Aggressive declined by 13.1 to 16.6 percent for children age 11 to 12 and Marketing to Children. Almost every national effort to 15 to 16 (Toh, Cutter, and Chew 2002) (box 44.10 outlines the improve nutrition incorporates the promotion of healthy food national program for adults). choices, such as fruits, vegetables, and legumes. Ideally, such efforts are coordinated among government groups, retailers, professional groups, and nonprofit organizations, and invest- Economic Policies ment in such efforts should include the careful testing and Economic policies can have important effects on behavior and refining of social-marketing strategies. choices, and these policies have been particularly useful in 842 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others Box 44.9 A Comprehensive Intervention Approach in South Africa The Coronary Risk Factor Study in South Africa (Rossouw results showed an improvement in the community risk and others 1993) tested community interventions at dif- factor profile for CAD in the intervention communities, ferent levels of intensity in two communities with a third especially in relation to blood pressure, smoking, and over- control community. The target population was Caucasian all risk. The results indicate no additional benefit of the South Africans. Interventions included direct media cam- personal intervention for high-risk individuals beyond that paigns,public health messages delivered in a variety of ways, already offered by the mass media program. Estimated per and home mailings. Also included were community acti- capita costs of the heavy intervention program were roughly vities, such as fun walks, public meetings, involvement of four times as much as for the mild intervention program community-based organizations, free screening for blood (US$22 per capita compared with US$5 per capita), and the pressure, small-group personal interventions, and encour- low-intervention community received almost the same agement of food substitution in stores and restaurants. The level of benefits as the high-intervention community. Box 44.10 The Singapore National Healthy Lifestyle Program Because CVD and cancer had become the major causes of In a follow-up survey after six years, cigarette smoking death in Singapore, the government adopted the National had decreased from 34 to 27 percent among men, the pro- Healthy Lifestyle Program in 1992 (Cutter, Tan, and Chew portion of adults who exercised regularly had increased 2001). This coordinated, multisectoral approach involved from 14 to 17 percent, and the prevalence of obesity was government ministries, health professionals, employers, stable. However, hypertension and high LDL cholesterol unions, and community organizations. The program levels had increased modestly. From 1991 to 1999, the age- aimed at improving the social and physical environment so standardized incidence of myocardial infarction declined as to promote healthy living. Healthy diets, regular physi- from 98.2 to 83.0 per 100,000 residents (Mak and others cal exercise, and nonsmoking were emphasized. The pro- 2003) and age-standardized mortality from CAD gram used the mass media; legislative measures to discour- decreased from 60.8 to 47.2 per 100,000 residents. age smoking; and widespread school, workplace, and com- munity health promotion packages. reducing the prevalence of smoking (see chapter 46). Policies and those high in trans fats. Legislation can make this that could influence diet and physical activity deserve careful distinction, providing a modest economic incentive for consideration because they are rarely neutral and often support healthier choices and at the same time conveying important unhealthy behaviors. Consider the following examples: nutritional messages (see chapter 11). · Use of individual automobiles is often subsidized by build- · Subsidies can favor the consumption of less healthy foods, ing and maintaining highways, providing inexpensive such as sugar, refined grains, beef, and high-fat dairy prod- parking, and imposing low taxes on petroleum products ucts as opposed to fruits, vegetables, whole grains, nuts, that do not fully reflect their societal and environmental legumes, and fish. Poland provides a striking example of costs. Increasing taxes on petroleum products and subsidiz- how changes in subsidies can affect health (box 44.11). ing public transportation could have an important effect on Governments often subsidize foods indirectly by sheltering choice of transportation modality, which as noted earlier, them from sales taxes in the recognition that they are essen- has major effects on health. tial; however, this logic should not extend to foods with · Walking, riding bicycles, and using public transportation adverse health effects, such as sugar-sweetened beverages can be promoted by economic policies that, in addition Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 843 Box 44.11 Poland: A Dramatic Decline in Heart Disease After Poland's transition to a democratic government Risk of Coronary Heart Disease in the early 1990s, the government removed large sub- According to Polyunsaturated to Saturated Fat Ratio sidies for butter and lard, and consumption of nonhy- Rate ratio for coronary heart Rate ratio for coronary heart drogenated vegetable fat increased rapidly (Zatonski, disease in Nurses' Health Study disease mortality in Poland 1.2 McMichael, and Powles 1998). The ratio of dietary Poland polyunsaturated to saturated fat increased from 0.33 1990 1992 1.0 in 1990 to 0.56 in 1999, and during this period mor- 1.0 tality rates from CAD dropped by 28 percent (data 1994 0.8 1996 provided by W. Zatonski). Changes in smoking and in 1999 the consumption of fruits and vegetables probably 0.7 0.6 played a minor role in this decrease (see figure). 0.4 0.2 0.33 0.56 0 0.2 0.3 0.4 0.5 0.6 Dietary polyunsaturated to saturated fat ratio Notes: Squares represent data for Poland from 1990 to 1999. Circles are for deciles of polyunsaturated fat to saturated fat and for risk of coronary heart disease in the Nurses' Health Study (Hu 1999), which closely predict the observed changes in Poland. to providing better infrastructure, include discounts on Reducing Saturated Fat Content. In the base case, assuming a transportation fares, provide secure bicycle parking, and 3 percent drop in cholesterol and a US$6 per person cost of the reduce health insurance premiums. intervention, averting one disability-adjusted life year (DALY) would cost as little as US$1,865 in South Asia and as much as COST-EFFECTIVENESS OF INTERVENTIONS US$4,012 in the Middle East and North Africa. The interven- tion's effectiveness could be increased by replacing part of the Only a few studies have described interventions for lifestyle saturated fat with polyunsaturated fat, which has additional diseases in developing countries. beneficial effects mediated by mechanisms other than LDL cholesterol (see tables 44.2 and 44.3). Modeling Likely Interventions Primary targets for reducing lifestyle diseases include changing Replacing Dietary Trans Fat from Partial Hydrogenation the fat composition of the diet, limiting sodium intake, and with Polyunsaturated Fat. We could not use the model for engaging in regular physical activity. saturated fat to estimate the effects of replacing trans fat with Using available data, we calculated a range of estimates polyunsaturated fat because only a small part of the benefit is under given assumptions for the cost-effectiveness of replacing attributable to reducing LDL cholesterol (F. B. Hu and Willett dietary saturated fat with monounsaturated fat, replacing 2002). Trans fats also adversely affect high-density lipoprotein trans fat with polyunsaturated fat, and reducing salt intake. An (HDL) cholesterol, triglycerides, endothelial function, and increase in moderate physical activity by three to five hours per inflammatory markers. In addition, increases in polyunsatu- week is considered likely to lower the risk of many diseases, but rated fat (assuming a mix of N-6 and omega-3 fatty acids) will data to model the cost-effectiveness of this intervention are not reduce LDL cholesterol, insulin resistance, and probably fatal currently available. For further details of methods and assump- cardiac arrhythmias. tions underlying the analyses presented here, see the Web site In calculations that are based only on the adverse effects on version of this book. LDL and HDL, replacing 2 percent of the energy from trans fat 844 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others Table 44.2 Incremental Cost-Effectiveness Ratios, Selected Interventions, by Region (US$/DALY averted) Substituting 2 percent of energy from trans fat with polyunsaturated fat Media campaign 7 percent CAD reduction 40 percent CAD reduction Reducing salt to reduce Intervention Intervention Intervention Intervention content by means saturated fat cost of cost of cost of cost of of legislation plus Region content US$0.50/adulta US$6.00/adult US$0.50/adulta US$6.00/adult public education East Asia and the Pacific 2,769 73 1,583 Cost saving 227 2,056 Europe and Central Asia 2,929 65 1,670 Cost saving 228 2,170 Latin America and the Caribbean 3,297 40 1,865 Cost saving 225 2,476 Middle East and North Africa 4,012 25 2,259 Cost saving 252 3,056 South Asia 1,865 38 1,014 Cost saving 138 1,325 Sub-Saharan Africa 2,356 53 1,344 Cost saving 184 1,766 Source: Authors' calculations. a. Based on the U.S. Food and Drug Administration's analysis of the costs of the intervention in the United States. Table 44.3 Two-Way Sensitivity Analysis of the Costs of the was estimated to have only a modest effect on consumer Intervention to Reduce Saturated Fat Content and of the behavior, as noted earlier, it is having a major effect on manu- Relative Risk Reduction in CAD Events, South Asia facturers' behavior. (US$/DALY averted) The potential for reducing CVD rates by replacing trans fats Relative risk with polyunsaturated fats will depend on the diets of specific Cost per individual reduction in CAD populations. Whereas the intake of trans fat is low in China, it events (percent) US$0.25 US$3.00 US$6.00a is likely to be high in parts of India, Pakistan, and other Asian 10 Cost saving 318 680 countries because of the extraordinarily high content in com- 5 Cost saving 680 1,403 monly used cooking fats. 4b Cost saving 911 1,865 Table 44.2 presents the results of a cost-effectiveness analysis assuming the two different estimates for CAD reduction: 7 per- 1 258 3,572 7,188 cent and 40 percent.We used costs of US$0.50 per adult per year, Source: Authors' calculations. which was the maximal cost in the U.S. Food and Drug a. Threshold analysis reveals that at the base assumption of US$6 for the intervention, no level in the range of assumed CAD reduction is cost saving. Administration analysis,and of US$6.00 per adult per year using b. Threshold analysis reveals that at a cost below US$0.36 per individual and a 4 percent traditional health education approaches. The lower estimate-- reduction in CAD (base assumption), the intervention is cost saving. or one even lower--is possible because trans fat can be elimi- nated at the source rather than depending entirely on changes in with polyunsaturated fat was estimated to reduce CAD by 7 individual behavior. With the lower cost, the smaller effect esti- to 8 percent (Grundy 1992; Willett and Ascherio 1994). mate leads to a cost-effectiveness ratio of between US$25 and Epidemiological studies, which include the contributions of US$73 per DALY averted, depending on the region, and with the the additional causal pathways, suggest a much greater reduc- higher-effect estimate, the intervention can be cost saving. tion, from about 25 to 40 percent (F. B. Hu and others 1997; Oomen and others 2001). Another likely benefit is a reduction Reducing the Salt Content of Manufactured Foods through in the incidence of type 2 diabetes: estimates indicate that the Legislation and an Accompanying Education Campaign. same 2 percent reduction would reduce incidence by 40 percent Table 44.2 shows the base-case cost-effectiveness of a legislated (Salmeron and others 2001). reduction in salt content. The intervention appears to be rela- Because voluntary action by industry (as has nearly been tively cost-effective, with a cost per DALY averted of US$1,325 achieved in the Netherlands) or by regulation (as occurred in in South Asia to US$3,056 in the Middle East and North Africa. Denmark) can eliminate partially hydrogenated fat from the Those regional variations are attributable to differing risk pro- diet, this initiative does not require consumer education, and files across regions as well as to price differentials for the costs the costs can be extremely low. In an analysis required before of treating disease sequelae. implementing food labeling, the U.S. Food and Drug The actual blood pressure reduction from lower salt con- Administration (2003) estimated that trans fat labeling would sumption could vary from the base-case assumption, as could be highly cost-effective. Even though the effect of labeling itself the costs of the education campaign. Table 44.4 shows the Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 845 Table 44.4 Two-Way Sensitivity Analysis of the Costs of the and without risk factors for noncommunicable diseases (phys- Intervention to Reduce Salt Content and Its Effectiveness, ical activity, BMI, and smoking status) and find that a healthi- South Asia er lifestyle of physical activity three times per week, a moderate (US$/DALY averted) BMI, and nonsmoking status reduce health care costs by Blood pressure 49 percent compared with an unhealthy lifestyle. Cost per individual reduction (millimeters of mercury) US$1a US$3 US$6 Cost-Effectiveness of Community-Based Interventions 4 9 308 608 Populationwide and community-based interventions appear to 3 49 448 847 be cost-effective if they reach large populations, address high- 2b 129 727 1,326 mortality and high-morbidity diseases, and are multipronged 1 368 1,565 2,761 and integrated efforts. The full costs of achieving changes in behavior and policy are often complex and difficult to estimate. Source: Authors' calculations. a. Threshold analysis reveals that at a cost of US$1 per individual, a blood pressure reduction Interventions may yield additional spinoff benefits. For would have to be greater than 5 millimeters of mercury for the intervention to be cost saving. At instance, decisions to reduce children's television viewing could the base-case assumption of a cost of US$6 for the intervention, there is no cost saving threshold level of reduction. easily improve school outcomes as well as reduce childhood b. Threshold analysis reveals that at a cost of less than US$0.47 per individual the intervention is obesity. Similarly, increasing walking and bicycle riding for cost-saving. transportation could reduce air pollution. results of lower costs of the education campaign and higher or RESEARCH AND DEVELOPMENT PRIORITIES lower effects of the intervention on blood pressure. These results may argue for initial efforts to focus on reductions in the A number of research and development priorities have been use of salt during the manufacturing process with no public identified: education campaign. The cost-effectiveness of such a change is high and could be augmented with a public education cam- · Conduct randomized trials of the use of folic acid and paign only if needed to support the legislated change. At lower alpha-linoleic acid to prevent CAD in developing countries. implementation costs, the intervention is highly cost-effective, These interventions cost little, and the potential benefits are even with half the assumed effect on blood pressure. large and rapid. · Develop prospective cohort studies of dietary and lifestyle Adopting Physical Activity Interventions. Even though factors in developing and transition countries to refine the health experts believe that physical activity interventions are understanding of risk factors in those contexts. To date, effective in reducing the risk of lifestyle diseases, no studies of almost all such studies have taken place in Europe and their cost-effectiveness are available from developing countries. North America. If people walk voluntarily (the model assumes no opportunity · Develop surveillance systems for chronic diseases and for cost), a net economic benefit would accrue to all segments of major risk factors, such as obesity, in developing countries. the U.S. population. If we project the economic benefits to the · Develop additional multifaceted, community-based entire U.S. population and assume 25 percent compliance by demonstration programs in developing countries to docu- the sedentary population, the voluntary program would gener- ment the feasibility of lifestyle changes and to learn more ate US$6.8 billion in savings (in 2001 U.S. dollars). about effective strategies. · Conduct detailed cost-effectiveness analyses of various pre- vention strategies to modify dietary and lifestyle factors. Aggregate Costs of Obesity and Unhealthy Lifestyles A series of U.S. studies appears to confirm that the avoidable costs of chronic diseases are substantial, although many devel- RECOMMENDED PRIORITY INTERVENTIONS oping countries have not yet experienced the full demands on their health sectors resulting from these conditions. Colditz An overall objective is to develop comprehensive national and (1999) estimates that obesity is responsible for 7 percent of all local plans that take advantage of every opportunity to encour- U.S. direct health care costs and that inactivity is responsible for age and promote healthy eating and active living. These plans an additional 2.4 percent of all health care costs. Indirect costs would involve health care providers; worksites; schools; media; associated with obesity and inactivity account for another urban planners; all levels of food production, processing, and 5 percent of health care costs. Pronk and others (1999) assess preparation; and governments. The goal is cultural change the difference in health care costs between adult patients with in the direction of healthy living. An important element in 846 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others cultural change is national leadership by individuals and by º Implement folic acid fortification if folic acid intake is professional organizations. Specific interventions will depend low. on local physical and cultural conditions and should be based º Ensure that health providers regularly weigh both chil- on careful analysis of existing dietary and activity patterns and dren and adult patients, track their weights over time, their determinants; however, the following interventions can be and provide counseling regarding diet and activity if considered (specific interventions for control of smoking are they are already overweight or if unhealthy weight gain is discussed elsewhere): occurring during adulthood. Those activities should be integrated with programs that address undernutrition. · Physical activity: Health care providers should be encouraged to set a º Develop transportation policies and a physical environ- good example by not smoking, by exercising regularly, ment to promote walking and riding bicycles. This inter- and by eating healthy diets. vention includes constructing sidewalks and protected º Promote healthy foods at worksite food services. bicycle paths and lanes that are attractive, safe, well- Worksites can also promote physical activity by provid- lighted, and functional with regard to destinations. ing financial incentives for using public transportation º Adopt policies that promote livable, walker-friendly or riding bicycles (and by not subsidizing automobiles communities that include parks and are centered around by providing free parking). Providing areas for exercise access to public transportation. during work breaks and showers may be useful. º Encourage the use of public transportation and discour- º Set standards that restrict the promotion of foods high age overdependence on private automobiles. in sugar, refined starch, and saturated and trans fats to º Promote the use of stairs. Building codes can require the children on television and elsewhere. inclusion of accessible and attractive stairways. º Set national standards for the amount of sodium in · Healthy diets: processed foods. º Develop comprehensive school programs that integrate · National campaigns: nutrition into core curricula and healthy nutrition into º Invest in developing locally appropriate health messages school food services. Regional or national standards to related to diet, physical activity, and weight control. This promote healthy eating should be developed for school effort is best done in cooperation with government agen- food services. Programs should also aim at limiting cies, nongovernmental organizations, and professional television watching, in part by promoting attractive organizations so that consistent messages can be used on alternatives. television and radio; at health care settings, schools, and º Work with the agriculture sector and food industries to worksites; and elsewhere. This effort should use the best replace unhealthy fats with healthy fats, including ade- social-marketing techniques available, with messages quate amounts of omega-3 fatty acids. This goal can be continuously evaluated for effectiveness. achieved through a combination of education, regula- º Develop a sustainable surveillance system that monitors tion, and incentives. Specific actions will depend on local weight and height, physical activity, and key dietary sources of fat and on regional production and distribu- variables. tion. For example, in areas where palm oil is dominant, research could focus on developing strains that are lower Implementation of the recommended policies to promote in saturated fat and higher in unsaturated fat through health and well-being is often not straightforward because of selective breeding or genetic alteration. Labeling require- opposition by powerful and well-funded political and economic ments or regulation can be used to discourage or elimi- forces, such as those involved in the tobacco, automobile, food, nate the use of partially hydrogenated vegetable oils and and oil industries (Nestle 2002). The solutions will depend on a to promote the use of nonhydrogenated unsaturated oils country's specific political landscape. However, experiences in instead. many countries indicate that alliances of public interest groups, º Require clear labeling of energy content for all packaged professional organizations, and motivated individuals can over- foods, including fast food. come such powerful interests. Strategies should start with sound º Use tax policies to encourage the consumption of science and can use a mix of mass media, lobbying efforts, and healthier foods. For example, high-sugar sodas could be lawsuits. Also, the food industry is far from monolithic, and ele- fully taxed and not subsidized in the same way as health- ments can often be identified whose interests coincide with ier foods. health promotion, which can create valuable partnerships.As an º Emphasize the production and consumption of healthy example, the willingness of some margarine manufacturers to food products in agriculture support and extension invest in developing products free of trans fatty acids greatly programs. helped the effort to reduce these fats, because these producers Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 847 then became proponents for labeling the trans fat content of Studied Cohort of U.S. Adults." New England Journal of Medicine 348: foods. Protection of children can be a powerful lever because of 1625­38. almost universal concern about their welfare and the recogni- Coitinho, D., C. A. Monteiro, and B. M. Popkin. 2002. "What Brazil Is Doing to Promote Healthy Diets and Active Lifestyles." Public Health tion that they cannot be responsible for the long-term conse- Nutrition 5: 263­67. quences of their diet and lifestyle choices. Colditz, G. A. 1999. "Economic Costs of Obesity and Inactivity." Medicine Science and Sports Exercise 31: S663­67. Conlin, P. R. 1999. "The Dietary Approaches to Stop Hypertension (Dash) CONCLUSIONS Clinical Trial: Implications for Lifestyle Modifications in the Treatment of Hypertensive Patients." Cardiology Review 7: 284­88. Many of the ongoing diet and lifestyle interventions in low- Cutter, J., B. Y. Tan, and S. K. Chew. 2001."Levels of Cardiovascular Disease and middle-income countries are relatively recent, and few Risk Factors in Singapore Following a National Intervention have documented reductions in the rates of major chronic dis- Programme." Bulletin of the World Health Organization 79: 908­15. eases. However, the successes of Finland, Singapore, and many de Lorgeril, M., S. Renaud, N. Mamelle, P. Salen, J. L. Martin, I. Monjaud, other high-income countries in reducing rates of CAD, stroke, and others. 1994. "Mediterranean Alpha-Linolenic Acid­Rich Diet in Secondary Prevention of Coronary artery disease." Lancet 343: and smoking-related cancers strongly suggest that similar ben- 1454­59. (Erratum in Lancet 1995, 345: 738.) efits will emerge in the developing countries. Dickinson, S., S. Colagiuri, E. Faramus, P. Petocz, and J. C. Brand-Miller. 2002. "Postprandial Hyperglycemia and Insulin Sensitivity Differ among Lean Young Adults of Different Ethnicities." Journal of ACKNOWLEDGMENTS Nutrition 132: 2574­79. Ebbeling, C. B., M. M. Leidig, K. B. Sinclair, J. P. Hangen, and D. S. The authors appreciate Hilary Farmer's assistance in preparing Ludwig. 2003. "A Reduced-Glycemic Load Diet in the Treatment of this manuscript. Adolescent Obesity." Archives of Pediatric and Adolescent Medicine 157: 773­79. Ewing, R, R. Schieber, and C. Zegeer. 2003. "Urban Sprawl as a Risk Factor REFERENCES in Motor Vehicle Occupant and Pedestrian Facilities." American Journal of Public Health 93: 1541­45. Ascherio, A., E. Cho, K. Walsh, F. M. Sacks, W. C. Willett, and A. Faruqui. Fox, K. R. 1999."The Influence of Physical Activity on Mental Well-Being." 1996. "Premature Coronary Deaths in Asians" (letter). British Medical Public Health Nutrition 2: 411­18. Journal 312: 508. Frumkin, H. 2002. "Urban Sprawl and Public Health." Public Health Bacon, C. G., M. A. Mittleman, I. Kawachi, E. Giovannucci, D. B. Glasser, Reports 117: 201­17. and E. B. Rimm. 2003. "Sexual Function in Men Older Than 50 Years of Age: Results from the Health Professionals Follow-up Study." Annals Giovannucci, E. 2002. "Epidemiologic Studies of Folate and Colorectal of Internal Medicine 139: 161­68. Neoplasia: A Review." Journal of Nutrition 132: 2350­55S. Ball, D., S. Ellison, J. Adamy, and G. Fowler. 2004. "Recipes without Gortmaker, S. L., K. Peterson, J. Wiecha, A. M. Sobol, S. Dixit, M. K. Fox, Borders?" Wall Street Journal, August 18, 2004, p. 1. and N. Laird. 1999. "Reducing Obesity via a School-Based Interdisciplinary Intervention among Youth: Planet Health." Archives of Baylin, A., E. K. Kabagambe, A. Ascherio, D. Spiegelman, and H. Campos. Pediatric and Adolescent Medicine 153: 409­18. 2003. "Adipose Tissue Alpha-Linolenic Acid and Nonfatal Acute Myocardial Infarction in Costa Rica." Circulation 107: 1586­91. Grundy, S. M. 1992. "How Much Does Diet Contribute to Premature Coronary Heart Disease?" In Atherosclerosis IX: Proceedings of the Ninth Bell, A. C., K. Ge, and B. M. Popkin. 2002. "The Road to Obesity or the International Symposium on Atherosclerosis, ed. O. Stein, S. Eisenberg, Path to Prevention: Motorized Transportation and Obesity in China." and Y. Stein, 471­78. Tel Aviv: Creative Communications. Obesity Research 10: 277­83. Handy, S. L., M. G. Boarnet, R. Ewing, and R. E. Killingsworth. 2002. "How Bell, E. A., L. S. Roe, and B. J. Rolls. 2003. "Sensory-Specific Satiety Is the Built Environment Affects Physical Activity: Views from Urban Affected More by Volume Than by Energy Content of a Liquid Food." Planning." American Journal of Preventive Medicine 23: 64­73. Physiology and Behavior 78: 593­600. Hao, L., J. Ma, M. J. Stampfer, A. Ren, Y. Tian, Y. Tang, and others. 2003. Blendon, R. J. 2002. Welfare of Children in America. Cambridge, MA: "Geographical, Seasonal, and Gender Differences in Folate Status Cogent Research. among Chinese Adults." Journal of Nutrition 133: 3630­35. Bly, J. L., R. C. Jones, and J. E. Richardson. 1986. "Impact of Worksite Harris, M. I., K. M. Flegal, C. C. Cowie, M. S. Eberhardt, D. E. Goldstein, Health Promotion on Health Care Costs and Utilization. Evaluation of R. R. Little, and others. 1998."Prevalence of Diabetes, Impaired Fasting Johnson & Johnson's Live for Life Program." Journal of the American Glucose, and Impaired Glucose Tolerance in U.S. Adults: The Third Medical Association 256: 3235­40. National Health and Nutrition Examination Survey, 1988­1994." Boutelle, K, R. Jeffery, D. McMurray, and K. Schmitz. 2001. "Using Signs, Diabetes Care 21: 518­24. Artwork, and Music to Promote Stair Use in a Public Building." Hernandez, B., S. L. Gortmaker, G. A. Colditz, K. E. Peterson, N. M. Laird, American Journal of Public Health 91: 2004­6. and S. Parra-Cabrera. 1999. "Association of Obesity with Physical Bunn, F., T. Collier, C. Frost, K. Ker, I. Roberts, and R. Wentz. 2003. "Traffic Activity, Television Programs, and Other Forms of Video Viewing Calming for the Prevention of Road Traffic Injuries: Systematic Review among Children in Mexico City." International Journal of Obesity and and Meta-Analysis." Injury Prevention 9: 200­4. Relational Metabolism Disorders 23: 845­54. Buss, D. 2004. "Is the Food Industry the Problem or the Solution?" New Honein, M. A., L. J. Paulozzi, T. J. Mathews, J. D. Erickson, and L. Y. Wong. York Times, August 29, 2004, p. 5. 2001. "Impact of Folic Acid Fortification of the U.S. Food Supply on Calle, E. E., C. Rodriguez, K. Walker-Thurmond, and M. J. Thun. 2003. the Occurrence of Neural Tube Defects." Journal of the American "Overweight, Obesity, and Mortality from Cancer in a Prospectively Medical Association 285: 2981­86. 848 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others Hu. 2002. Ludwig, D. S., K. E. Peterson, and S. L. Gortmaker. 2001."Relation between Hu, F. B., M. F. Leitzmann, M. J. Stampfer, G. A. Colditz, W. C. Willett, and Consumption of Sugar-Sweetened Drinks and Childhood Obesity: A E. B. Rimm. 2001. "Physical Activity and Television Watching in Prospective, Observational Analysis." Lancet 357: 505­8. Relation to Risk for Type 2 Diabetes Mellitus in Men." Archives of Mak, K. H., K. S. Chia, J. D. Kark, T. Chua, C. Tan, B. H. Foong, and others. Internal Medicine 161: 1542­48. 2003. "Ethnic Differences in Acute Myocardial Infarction in Hu, F. B., M. J. Stampfer, J. E. Manson, E. B. Rimm, A. Wolk, G. A. Colditz, Singapore." European Heart Journal 24: 151­60. and others. 1999. "Dietary Intake of Alpha-Linolenic Acid and Risk of Matsudo, V., S. Matsudo, D. Andrade, T. Araujo, E. Andrade, L. Carlos de Fatal Ischemic Heart Disease among Women." American Journal of Oliveira, and G. Braggion. 2002. "Promotion of Physical Activity in a Clinical Nutrition 69: 890­97. Developing Country: The Agita São Paulo Experience." Public Health Nutrition 5: 253­61. Hu, F. B., M. J. Stampfer, J. E. Manson, E. Rimm, G. A. Colditz, B. A. Rosner, and others. 1997. "Dietary Fat Intake and the Risk of Coronary Heart Nestle, M. 2002. Food Politics: How the Food Industry Influences Nutrition Disease in Women." New England Journal of Medicine 337: 1491­99. and Health. Berkeley. CA: University of California Press. Hu, F. B., R. M. van Dam, and S. Liu. 2001. "Diet and Risk of Type 2 Nissinen, A., X. Berrios, and P. Puska. 2001. "Community-Based Diabetes: The Role of Types of Fat and Carbohydrate." Diabetologia 44: Noncommunicable Disease Interventions: Lessons from Developed 805­17. Countries for Developing Ones." Bulletin of the World Health Organization 79: 963­70. Hu, F. B., B. Wang, C. Chen, Y. Jin, J. Yang, M. J. Stampfer, and X. Xu. 2000. "Body Mass Index and Cardiovascular Risk Factors in a Rural Chinese Office of the Administrator. 2001. What Is Smart Growth? Washington, Population." American Journal of Epidemiology 151: 88­97. DC: Environmental Protection Agency. http://www.epa.gov/ smartgrowth/pdf/whtissg4v2.pdf. Hu, F. B., and W. C. Willett. 2002. "Optimal Diets for Prevention of Coronary Heart Disease." Journal of the American Medical Association Oomen, C. M., M. C. Ocke, E. J. Feskens, M. A. van Erp-Baart, F. J. Kok, 288: 2569­78. and D. Kromhout. 2001. "Association between Trans Fatty Acid Intake and 10-Year Risk of Coronary Heart Disease in the Zutphen Elderly Hu, G., H. Pekkarinen, O. Hanninen, Z. J. Yu, H. G. Tian, Z. Y. Guo, Study: A Prospective Population-Based Study." Lancet 357: 746­51. and A. Nissinen. 2002. "Physical Activity during Leisure and Commuting in Tianjin, China." Bulletin of the World Health Pendakur, V. S. 2000. World Bank Urban Transport Strategy Review. Organization 80: 933­38. Yokohama, Japan: Pacific Policy and Planning Associates. IARC (International Agency for Research on Cancer). 2002. Weight Petrella, R., J. Koval, D. Cunningham, and D. Paterson. 2003."Can Primary Control and Physical Activity. Lyon, France: IARC Press. Care Doctors Prescribe Exercise to Improve Fitness: The Step Test Exercise Prescription (STEP) Project." American Journal of Preventive Institute of Medicine. 2002. Dietary Reference Intakes for Energy, Medicine 24: 316­22. Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients): A Report of the Panel on Macronutrients, Pignone, M. P., A. Ammerman, L. Fernandez, C. T. Orleans, N. Pender, Subcommittees on Upper Reference Levels of Nutrients and Interpretation S. Woolf, and others. 2003. "Counseling to Promote a Healthy Diet in and Uses of Dietary Reference Intakes, and the Standing Committee on Adults: A Summary of the Evidence for the U.S. Preventive Services the Scientific Evaluation of Dietary Reference Intakes. Washington, DC: Task Force." American Journal of Preventive Medicine 24: 75­92. National Academy of Sciences. http://www.nap.edu/catalog/ Pratt, M., C. A. Macera, and G. Wang. 2000. "Higher Direct Medical Costs 10490.html. Associated with Physical Inactivity." Physician and Sports Medicine Jones, T. F., and C. B. Eaton. 1994. "Cost-Benefit Analysis of Walking 28:63­70. http://www.physsportsmed.com/issues/2000/10_00/pratt. to Prevent Coronary Heart Disease." Archives of Family Medicine 3: htm. 703­10. Pronk, N. P., M. J. Goodman, P. J. O'Connor, and B. C. Martinson. 1999. Keeler, E. B., W. G. Manning, J. P. Newhouse, E. M. Sloss, and J. Wasserman. "Relationship between Modifiable Health Risks and Short-Term 1989. "The External Costs of a Sedentary Life-Style." American Journal Health Care Charges." Journal of the American Medical Association 282: of Public Health 79: 975­81. 2235­39. Knowler, W. C., E. Barrett-Connor, S. E. Fowler, R. F. Hamman, J. M. Pucher, J., and L. Dijkstra. 2003. "Promoting Safe Walking and Cycling to Lachin, E. A. Walker, and D. M. Nathan. 2002. "Reduction in the Improve Public Health: Lessons from the Netherlands and Germany." Incidence of Type 2 Diabetes with Lifestyle Intervention or American Journal of Public Health 93: 1509­16. Metformin." New England Journal of Medicine 346 (6): 393­403. Puska, P., E. Vartiainen, J. Tuomilehto, V. Salomaa, and A. Nissinen. 1998. "Changes in Premature Deaths in Finland: Successful Long-Term Koplan, J. P., and W. H. Dietz. 1999. "Caloric Imbalance and Public Health Prevention of Cardiovascular Diseases." Bulletin of the World Health Policy." Journal of the American Medical Association 282: 1579­81. Organization 76: 419­25. Lambert, E. V., I. Bohlmann, and T. Kolbe-Alexander. 2001. "`Be Active': Raben, A., T. H. Vasilaras, A. C. Moller, and A. Astrup. 2002. "Sucrose Physical Activity for Health in South Africa." South African Journal of Compared with Artificial Sweeteners: Different Effects on Ad Libitum Clinical Nutrition 14: S12­16. Food Intake and Body Weight after 10 Weeks of Supplementation in Law, M. R., C. D. Frost, and N. J. Wald. 1991. "By How Much Does Dietary Overweight Subjects." American Journal of Clinical Nutrition 76: Salt Reduction Lower Blood Pressure? III--Analysis of Data from 721­29. Trials of Salt Reduction." British Medical Journal 302 (6780): 819­24. Robinson, T. N. 1999. "Reducing Children's Television Viewing to Prevent Leitzmann, M. F., E. B. Rimm, W. C. Willett, D. Spiegelman, F. Grodstein, Obesity: A Randomized Controlled Trial." Journal of the American M. J. Stampfer, and others. 1999. "Recreational Physical Activity and Medical Association 282: 1561­67. the Risk of Cholecystectomy in Women." New England Journal of Rossouw, J. E., P. L. Jooste, D. O. Chalton, E. R. Jordaan, M. L. Medicine 341: 777­84. Langenhoven, P. C. Jordaan, and others. 1993. "Community-Based Leyden, K. 2003. "Social Capital and the Built Environment: The Intervention: The Coronary Risk Factor Study (Coris)." International Importance of Walkable Neighborhoods." American Journal of Public Journal of Epidemiology 22: 428­38. Health 93: 1546­51. Ruangdaraganon, N., N. Kotchabhakdi, U. Udomsubpayakul, C. Lopez, R. 2004. "Urban Sprawl and Risk for Being Overweight or Obese." Kunanusont, and P. Suriyawongpaisal. 2002. "The Association between American Journal of Public Health 94: 1574­79. Television Viewing and Childhood Obesity: A National Survey in Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 849 Thailand." Journal of the Medical Association of Thailand 85 (Suppl. 4): Where Do They Feature in the Health Research Agenda?" Bulletin of S1075­80. the World Health Organization 79: 947­53. Saelens, B. E., J. F. Sallis, and L. D. Frank. 2003. "Environmental Correlates U.S. Food and Drug Administration, Center for Food and Safety and of Walking and Cycling: Findings from the Transportation, Urban Applied Nutrition. 2003. "Food Labeling: Trans Fatty Acids in Design, and Planning Literatures." Annals of Behavioral Medicine 25: Nutrition." Federal Register 68, no. 133, 41433­506 (July 11, 2003). 80­91. http://www.cfsan.fda.gov/~lrd/fr03711a.html (see also http://vm. Salmeron, J., F. B. Hu, J. E. Manson, M. J. Stampfer, G. A. Colditz, E. B. cfsan.fda.gov/~lrd/fr991117.html). Rimm, and W. C. Willett. 2001. "Dietary Fat Intake and Risk of Type 2 Uusitalo, U., E. J. Feskens, J. Tuomilehto, G. Dowse, U. Haw, D. Fareed, and Diabetes in Women." American Journal of Clinical Nutrition 73: others. 1996. "Fall in Total Cholesterol Concentration over Five Years 1019­26. in Association with Changes in Fatty Acid Composition of Cooking Schulze, M. B., J. E. Manson, D. S. Ludwig, G. A. Colditz, M. J. Stampfer, Oil in Mauritius: Cross-Sectional Survey." British Medical Journal 313: W. C. Willett, and F. B. Hu. 2004. "Sugar-Sweetened Beverages, Weight 1044­46. Gain, and Incidence of Type 2 Diabetes in Young and Middle-Aged Wang, G., C. Macera, B. Scudder-Soucie, T. Schmid, M. Pratt, and Women." Journal of the American Medical Association 292: 927­34. D. Buchner. 2004. "Cost Effectiveness of a Bicycle/Pedestrian Trail Selmer, R., I. Kristiansen, A. Haglerod, S. Graff-Iverson, H. Larsen, H. Development in Health Promotion." Preventive Medicine 38: 237­42. Meyer, and others. 2000. "Cost and Health Consequences of Reducing Wang, S. Y., G. B. Chi, C. X. Jing, X. M. Dong, C. P. Wu, and L. P. Li. 2003. the Population Intake of Salt." Journal of Epidemiology and Community "Trends in Road Traffic Crashes and Associated Injury and Fatality in Health 54: 697­702. the People's Republic of China, 1951­1999." Injury Control and Safety Sobngwi, E., J. F. Gautier, and J. C. Mbanya. 2003. "Exercise and the Promotion 10: 83­87. Prevention of Cardiovascular Events in Women" (author reply). New WHO (World Health Organization). 2000. Obesity: Preventing and England Journal of Medicine 348: 77­79. Managing the Global Epidemic. WHO Technical Report 894. Geneva: Staunton, C., D. Hubsmith, and W. Kallins. 2003."Promoting Safe Walking WHO. and Biking to School: The Marin County Success Story." American ------. 2002. The World Health Organization Report 2002: Reducing Risks, Journal of Public Health 93: 1431­34. Promoting Healthy Life. Geneva: WHO. Surface Transportation Policy Project. 1999. "Aggressive Driving: Where WHO and FAO (World Health Organization and Food and Agriculture You Live Matters." Washington, DC. http://www.transact.org/report. Organization of the United Nations). 2003. Diet, Nutrition, and the asp?id=56. Prevention of Chronic Diseases: Report of a Joint WHO/FAO Expert Swinburn, B. A., I. Caterson, J. C. Seidell, and W. P. James. 2004. "Diet, Consultation. Report 916. Geneva: WHO. Nutrition, and the Prevention of Excess Weight Gain and Obesity." Willett, W. C. 2002. "Balancing Lifestyle and Genomics Research for Public Health Nutrition 7: 123­46. Disease Prevention." Science 296: 695­98. Toh, C. M., J. Cutter, and S. K. Chew. 2002. "School-Based Intervention Willett, W. C., and A. Ascherio. 1994. "Trans-Fatty Acids: Are the Effects Has Reduced Obesity in Singapore." British Medical Journal 324: 427. Only Marginal?" American Journal of Public Health 84: 722­24. Tosteson, A., M. Weinstein, M. Hunink, M. A. Mittleman, L. Williams, Willett, W. C., W. H. Dietz, and G. A. Colditz. 1999."Guidelines for Healthy P. Goldman, and L. Goldman. 1997. "Cost-Effectiveness of Population- Weight." New England Journal of Medicine 341: 427­34. Wide Educational Approaches to Reduce Serum Cholesterol Levels." Willett, W. C., and R. L. Leibel. 2002. "Dietary Fat Is Not a Major Circulation 95: 24­30. Determinant of Body Fat." American Journal of Medicine 113 Tudor-Locke, C., B. E. Ainsworth, L. S. Adair, S. Du, and B. M. Popkin. (Suppl. 9B): 47­59S. 2003. "Physical Activity and Inactivity in Chinese School-Aged Youth: World Bank. 2003. Noncommunicable Diseases in Pacific Island Countries: The China Health and Nutrition Survey." International Journal of Disease Burden, Economic Cost, and Policy Options. Nouméa, New Obesity 27: 1093­99. Caledonia: World Bank. Tuomilehto, J., J. Lindstrom, J. G. Eriksson, T. T. Valle, H. Hamalainen, Zatonski, W. A., A. J. McMichael, and J. W. Powles. 1998. "Ecological Study P. Ilanne-Parikka, and others. 2001. "Prevention of Type 2 Diabetes of Reasons for Sharp Decline in Mortality from Ischaemic Heart Mellitus by Changes in Lifestyle among Subjects with Impaired Disease in Poland since 1991." British Medical Journal 316: 1047­51. Glucose Tolerance." New England Journal of Medicine 344: 1343­50. Unwin, N., P. Setel, S. Rashid, F. Mugusi, J. C. Mbanya, H. Kitange, and others. 2001. "Noncommunicable Diseases in Sub-Saharan Africa: 850 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others Chapter 45 The Growing Burden of Risk from High Blood Pressure, Cholesterol, and Bodyweight Anthony Rodgers, Carlene M. M. Lawes, Thomas Gaziano, and Theo Vos High blood pressure, cholesterol, and bodyweight are responsi- investments.Another approach is an evolution of the individual- ble for a large and increasing proportion of the global burden of based strategy in which treatments are targeted to those at high disease. Although historically these risks have been regarded as absolute risk of cardiovascular disease (CVD) rather than those "Western," their impact is now recognized as global: they are with single risk-factor levels above traditional thresholds, such already leading causes of disease in middle-income countries as hypertension or obesity (Jackson and others 1993). Such an and of emerging importance in low-income countries approach appears to be highly cost-effective, with the potential (Ezzati and others 2004; WHO 2002). This chapter presents an to substantially reduce CVD rates when combined with popu- evidenced-based review of the impact of high blood pressure, lationwide interventions (Murray and others 2003). cholesterol, and bodyweight; the cost-effectiveness of relevant interventions; and the economic benefits of interventions. The chapter focuses on personal interventions--that is, those that EPIDEMIOLOGY are mediated largely by interpersonal actions and take place at the individual level. As such, the chapter should be considered Elevated blood pressure, cholesterol, and bodyweight are all as complementary to chapter 44 on lifestyles, which addresses established risk factors for CVD and, in the case of body- populationwide interventions. weight, for other diseases, such as diabetes, certain cancers, Prevention strategies have been broadly classified as indi- and osteoarthritis. The associations between blood pressure vidual based (also known as high risk) or population based (Asia Pacific Cohort Studies Collaboration 1999, 2003a; (Rose 1985). The former typically involve screening to detect Prospective Studies Collaboration 2002); cholesterol (Asia individuals above a certain threshold level of an individual risk Pacific Cohort Studies Collaboration 2003b; Law, Wald, and factor--for example, people with hypertension--followed by Thompson 1994; Prospective Studies Collaboration 1995); personal interventions for those individuals. In contrast, the and body mass index (BMI) (Asia Pacific Cohort Studies population-based approach aims at lowering mean risk-factor Collaboration 2004; Willett and others 1995) and CVD are levels and shifting the population distribution of exposure in direct and continuous from relatively low levels, indicating that a favorable direction (Rose 1985). One example would be by optimal levels are about 115/75 millimeters of mercury reducing salt content in manufactured foods, thereby lowering (mmHg), 3.8 millimoles per liter (mmol/l), and 21 kilograms blood pressure levels on a populationwide basis. Such an per square meter (kg/m2), respectively (figure 45.1). approach has the potential to produce large and lasting changes Although some studies suggest J- or U-shaped associations in disease incidence but requires substantial sociopolitical (Calle and others 1999; Cruickshank 1994; D'Agostino and 851 a. Blood pressure b. Cholesterol c. Body mass index Relative risk of coronary heart disease Relative risk of coronary heart disease Relative risk of coronary heart disease 4.0 4.0 4.0 2.0 2.0 2.0 1.0 1.0 1.0 Hypertension Hyper- Obesity cholesterolemia 0.5 0.5 0.5 110 120 130 140 150 160 170 4.0 5.0 6.0 7.0 8.0 16 20 24 28 32 36 Systolic blood pressure (mmHg) Total cholesterol (mmol/l) Body mass index (kg/m2) Figure 45.1 Continuous Risks of Blood Pressure, Cholesterol, and Body Mass and Coronary Heart Disease Risk others 1991; Farnett and others 1991; Field and others 2001; Iso 12 percent lower stroke and IHD risk and an approximately and others 1989; Kannel, D'Agostino, and Silbershatz 1997; 20 to 30 percent lower diabetes risk. Stewart 1979; Troiano and others 1996), low levels of these risk factors are unlikely to cause CVD. Rather, such associations BURDEN OF THE DISEASE, CONDITION, more likely reflect incipient disease, which itself produces both a fall in risk-factor levels and an increase in CVD risk OR RISK FACTOR (Alderman 1996; Flack and others 1995; MacMahon and others Epidemiological data on blood pressure, cholesterol, and body- 1997; Manson, Willett, and Stampfer 1995; Neaton and weight levels are predominantly available from developed Wentworth 1992; Sleight 1997a, 1997b; Stevens and others countries; however, evidence indicates that these risk factors 1998). No trial evidence points to a J-curve association for are important and increasing in many other countries. Surveys blood pressure, despite including patients with below average in developing countries suggest increases in these risks occur blood pressure (Hansson and others 1999; McMurray and early in the path to industrialization (Bobak and others 1997; McInnes 1992; Pfeffer 1993; Staessen and others 1997). Evans and others 2001; Suh 2001; Wu and others 1996). Good The continuous associations between blood pressure, cho- evidence also documents risk-factor levels rising after people lesterol, and bodyweight and CVD demonstrate the lack of a migrate to more urbanized settings (Poulter and Sever 1994) in biological justification for current threshold levels, such as Africa (Poulter 1999; Poulter, Khaw, and Sever 1988), China those that define hypertension. Indeed, most of the disease (He, Klag, and others 1991; He, Tell, and others 1991), and the burden resulting from these three risk factors occurs in the Pacific islands (Joseph and others 1983; Salmond and others large majority of the population with nonoptimal levels but 1985; Salmond, Prior, and Wessen 1989). The World Health without hypertension, hypercholesterolemia, or obesity. Hence, Organization's Global Burden of Disease study demonstrated this chapter avoids those terms and instead uses high blood that CVD was a leading cause of death in many regions and pressure, high cholesterol, and high bodyweight, defined as that most adults in developed and developing countries have nonoptimal levels of these risk factors (that is, over 115/75 nonoptimal blood pressure, cholesterol, and bodyweight levels mmHg, 3.8 mmol/l, or 21 kg/m2, respectively). (Ezzati and others 2004; WHO 2002). Indeed, even using The strength of the proportional associations of these risk traditional cutoff points, these risk factors are prevalent: of 140 factors with CVD is similar for most population subgroups. subgroups defined by age, sex, and region, 45 percent had a Although they attenuate with age, they remain strong and mean SBP equal to or greater than 140 mmHg, 25 percent positive in the oldest age groups. Overall, in middle-aged had mean cholesterol levels over 5.5mmol/l, and 45 percent had populations, a 10 mmHg lower systolic blood pressure (SBP) is mean BMI levels of at least 25 kg/m2. associated with a roughly 30 to 40 percent lower stroke risk and 20 to 25 percent lower ischemic heart disease (IHD) risk, a 1 mmol/l lower cholesterol level is associated with about a 15 Health Burden to 20 percent lower stroke risk and 20 to 25 percent lower The Global Burden of Disease study assessed the burden attrib- IHD risk, and a 2 kg/m2 lower BMI is associated with an 8 to utable to nonoptimal levels of these risks (table 45.1) (Ezzati 852 | Disease Control Priorities in Developing Countries | Anthony Rodgers, Carlene M. M. Lawes, Thomas Gaziano, and others Table 45.1 Global Burden of Disease Attributable to Nonoptimal Blood Pressure, Cholesterol, and BMI by Region, 2000 High-mortality Low-mortality Developed Condition developing countriesa developing countriesb countriesc World total Attributable deaths (thousands) Blood pressure 1,969 2,205 2,966 7,140 (12.8%) Cholesterol 1,405 849 2,161 4,415 (7.9%) BMI 399 775 1,417 2,591 (4.6%) Attributable DALYs (thousands) Blood pressure 20,630 20,277 23,363 64,270 (4.4%) Cholesterol 15,602 8,609 16,227 40,438 (2.8%) BMI 6,408 11,115 15,892 33,415 (2.3%) Sources: Ezzati and others 2004; WHO 2002b. Note: The burden of disease estimated to be attributable to nonoptimal blood pressure (mean SBP 115 mmHg), cholesterol (mean 3.8 mmol/l), and body mass index (mean 21 kg/m2) in 2000. A, B, C and D designations in specific notes below are as follows: A very low child mortality and very low adult mortality; B low child mortality and low adult mortality; C low child mortality and high adult mortality; D high child mortality and high adult mortality; E high child mortality and very high adult mortality. a. The high-mortality developing countries include those in Africa, America D, the Eastern Mediterranean D, and Southeast Asia D. b. The low-mortality developing countries include those in America B, Eastern Mediterranean B, Southeast Asia B, and the Western Pacific B. c. The developed countries include those in America A, Europe, and the Western Pacific A. and others 2004; WHO 2002). The burden for blood pressure burden can be attributed to those risks. In addition, more than was related to deaths and disability-adjusted life years (DALYs) three-quarters of type 2 diabetes is caused by high bodyweight from IHD, stroke, hypertensive disease, and other CVD; end- (Ezzati and others 2004; WHO 2002). Hence the economic points for cholesterol included IHD and stroke; and endpoints impact of nonoptimal levels of those risks will be at least two- for BMI were IHD, stroke, hypertensive disease, diabetes, thirds that due to CVD and diabetes. A recent report high- certain cancers, and osteoarthritis. Globally, 7.1 million deaths lighted the economic impact of CVD in developing economies, were attributed to high blood pressure in 2000, 4.4 million to noting that a high proportion of the CVD burden occurs high cholesterol, and 2.6 million to high BMI. This burden was among adults of working age (Leeder and others 2004). In shared approximately equally among the sexes. A large fraction Brazil, China, India, Mexico, and South Africa, conservative occurred in middle age, especially in developing countries, and this factor, together with the frequently debilitating nature of nonfatal CVD, accounted for a large number of DALYs. Cardiovascular More of the DALY burden was experienced in developing disease 100% countries than in developed countries, reflecting the large pop- ulations in developing countries and their already high risk- factor levels. In all regions, most CVD is attributable to the Systolic blood Cholesterol over combined effects of high blood pressure, cholesterol, and body- pressure over 3.8 mmol/l 115 mmHg weight levels (figure 45.2). 28% 45% Table 45.2 shows the burden resulting from the overlapping or multicausal etiology of diseases. Analyses of the combined impact of these and other major cardiovascular risks indicate Body mass index 21 kg/m2 that the joint contribution of established risks is responsible for 15% 83 to 89 percent of the IHD burden and 70 to 76 percent of the stroke burden worldwide (Ezzati and others 2003; Ezzati and others 2004). Source: Ezzati and others 2004; WHO 2002. Note: Individual and joint contributions of high blood pressure, cholesterol, and body weight to global cardiovascular burden are shown, with the size of each circle Financial Burden proportional to the size of burden (as measured in DALYs) (WHO 2002). The percentages indicate the attributable burden for each risk factor, and the overlap shows disease caused by joint or mediated effects. The economic impact of high blood pressure, cholesterol, and bodyweight levels can be estimated indirectly using the forego- Figure 45.2 Global CVD Burden Caused by High Blood Pressure, ing data--namely, that more than two-thirds of the CVD Cholesterol, and Bodyweight The Growing Burden of Risk from High Blood Pressure, Cholesterol, and Bodyweight | 853 Table 45.2 Individual and Joint Contributions of Seven Selected Risk Factors to the Burden of CVD by Region Percentage of the Population attributable fractions for individual Overall population Disease regional disease burden risk factors (percentages) attributable fraction (percent) High-mortality developing countries Stroke 1.6 High blood pressure (56), high cholesterol (18), 65­71 high BMI (7), low fruit and vegetable intake (12), physical inactivity (6), tobacco (7), alcohol (2) IHD 3.0 High blood pressure (44), high cholesterol (54), 80­87 high BMI (11), low fruit and vegetable intake (33), physical inactivity (21), tobacco (8), alcohol (4) Low-mortality developing countries Stroke 4.7 High blood pressure (58), high cholesterol (13), 67­74 high BMI (11), low fruit and vegetable intake (10), physical inactivity (5), tobacco (8), alcohol (7) IHD 3.2 High blood pressure (45), high cholesterol (48), 79­87 high BMI (22), low fruit and vegetable intake (31), physical inactivity (22), tobacco (8), alcohol (3) Developed countries Stroke 6.0 High blood pressure (72), high cholesterol (27), 81­86 high BMI (23), low fruit and vegetable intake (12), physical inactivity (9), tobacco (22), alcohol (0) IHD 9.4 High blood pressure (58), high cholesterol (63), 89­93 high BMI (33), low fruit and vegetable intake (28), physical inactivity (22), tobacco (22), alcohol ( 0.2) World Stroke 3.1 High blood pressure (62), high cholesterol (18), 70­76 high BMI (13), low fruit and vegetable intake (11), physical inactivity (7), tobacco (12), alcohol (4) IHD 4.0 High blood pressure (49), high cholesterol (56), 83­89 high BMI (21), low fruit and vegetable intake (31), physical inactivity (22), tobacco (12), alcohol (2) Source: Ezzati and others 2003. Note: See notes to table 45.1 for a breakdown of the regional groupings. estimates indicated that at least 21 million years of future pro- for 2 to 8 percent of all health care expenditures in developed ductive life are lost because of CVD each year. Although no countries. For example, in 1991, 2.5 percent of health care costs detailed data exist on the direct economic burden of the indi- in New Zealand were attributable to obesity (Swinburn and vidual risk factors, the costs of CVD treatment in developing others 1997), and in 1996, US$22 billion was attributed to obe- countries are significant. In South Africa, for example, 2 to sity-related CVD in the United States, equivalent to 17 percent 3 percent of gross domestic product was devoted to the direct of CVD-related health expenditures (G.Wang and others 2002). treatment of CVD, or roughly 25 percent of all health care expenditures (Pestana and others 1996). For many middle- income countries, high body mass is already an important INTERVENTIONS cause of health inequities (Monteiro and others 2004). Current expenditure in developed countries provides an Data on the choice of interventions for blood pressure, choles- indication of possible future expenditure in developing coun- terol, and bodyweight and their effectiveness are now presented. tries. For example, estimated direct and indirect costs of CVD in the United States were US$350 billion in 2003. In 1998, US$109 billion was spent on hypertension, or about 13 percent Choice and Classification of Interventions of the health care budget (Hodgson and Cai 2001). Studies are A variety of population-based and personal interventions limited but suggest that obesity-related diseases are responsible could potentially be used to address the risks associated with 854 | Disease Control Priorities in Developing Countries | Anthony Rodgers, Carlene M. M. Lawes, Thomas Gaziano, and others high blood pressure, cholesterol, and bodyweight. Of the per- chapter discusses the cost-effectiveness of efforts to manage sonal interventions discussed in this section--lifestyle and those without previous CVD, and chapter 33 reviews the man- dietary, pharmacological, and surgical interventions--two agement of those with known vascular disease. A unifying sys- main strategies exist for choosing whom to treat: those above tem targeting treatments at those at highest risk, either with certain threshold values of single risk-factor levels and those CVD or multiple risk factors, is likely to be highly cost-effective above certain values of absolute cardiovascular (or global) risk, because more than 75 percent of events occur in the 5 to 10 which is determined by the levels of multiple factors. percent of people with CVD or specific clusters of risk factors Targeting treatments by levels of a single risk factor (such as (Haq and others 1999; Tosteson and others 1997). hypertension) does not effectively focus on overall risk of The limitations of the individual-risk-factor approach, developing CVD, which is mainly determined by the net effects together with increasing evidence that the thresholds do not of other risk factors. For example, the predicted 10-year CVD have any biological justification, have motivated the adoption risk for someone with an SBP of 140/90 mmHg can vary from of strategies that take other risk factors into account. Although 5 to 50 percent depending on the number of concomitant risk the most complete way of doing so is using the absolute-risk factors. The number of people who would need to be treated to strategy outlined earlier, one intermediate strategy involves prevent an event can therefore vary by an order of magnitude, lowering the thresholds of blood pressure or lipid levels at even if they have the same blood pressure levels. Thus, a treat- which treatment is initiated if one or more additional CVD ment strategy based only on individual risk-factor levels is risk factors, such as diabetes, are present (Chobanian and likely to result in high-risk patients being undertreated and others 2003). many patients at relatively low risk being treated with little absolute benefit, which is unlikely to be the best allocation of scarce health care resources. Intervention Effectiveness The absolute-risk strategy was developed in New Zealand This section summarizes data on the effectiveness of popula- (Jackson and others 1993) and has been adopted extensively tion-based interventions and personal interventions (lifestyle elsewhere, for example, by the British Hypertension Society and dietary interventions and pharmacological and surgical (Ramsay and others 1999) and the Joint Task Force of European interventions). The studies concerned have mainly been con- and other Societies on Coronary Prevention (Wood and others ducted in developed countries. 1998). The absolute CVD risk is estimated using risk assess- ments such as the Framingham risk function (Anderson and Population-Based Interventions. Investigators have under- others 1991) or the Prospective Cardiovascular Munster Study taken a variety of population-based community intervention score (Assmann, Cullen, and Schulte 2002) on the basis of the studies, mostly in developed countries in the 1970s and 1980s number and severity of CVD risk factors. Targeting treatments (for further details see chapter 44). These studies have tended at those at high absolute risk rather than those above arbitrary to be multifactorial projects testing whether comprehensive thresholds ensures a favorable ratio of benefits to risks. It can be community programs could produce favorable changes in such expected to reduce events in the large proportion of people who risk factors as bodyweight, cholesterol, and blood pressure and are, for example, nonhypertensive but who still have nonopti- in CVD morbidity and mortality (Schooler and others 1997). mal blood pressure (Rose 1981). Combinations of personal In general, they included a combination of populationwide interventions targeted at those at high absolute risk also have and individual interventions, including messages disseminated the potential of being highly cost-effective. through local associations, sports clubs, the media, and food The simplest indicator of high absolute risk is established associations; healthy food options at restaurants and worksite CVD, principally myocardial infarction, angina, stroke, or tran- cafeterias; food labeling at supermarkets; face-to-face commu- sient ischemic attack. For example, without preventive treat- nication at meetings and distribution of educational materials; ment, people who have had a myocardial infarction face an smoking restrictions; and competitions to develop healthy annual risk of death from coronary heart disease of about food. Except in Finland, the projects had mixed results, 5 percent (Law, Watt, and Wald 2002). That risk persists although many demonstrated significant effects with respect to indefinitely--probably for the rest of a person's life--and individual components of the interventions. The limitations of varies little with age or sex. many of the projects include inability to detect small but However, many individuals with no history of CVD are at potentially important changes in risk factors, short duration of similar elevated risk for future CVD as a result of constellations intervention and follow-up, and issues with outcome measures. of elevated risks. Thus, the distinction between primary Some have also suggested that those trials with less favorable and secondary prevention is somewhat artificial and could results may have lacked adequate community support and lead to undertreatment of many high-risk individuals. While public policy initiatives (Feinleib 1996; Mittelmark and others recognizing that the distinction is somewhat arbitrary, this 1993; Schooler and others 1997; Susser 1995). The Growing Burden of Risk from High Blood Pressure, Cholesterol, and Bodyweight | 855 A number of population-based interventions have also salt intake lowers blood pressure, with larger blood pressure taken place in developing countries, including the following: reductions in the elderly and in those with higher initial blood pressure levels (Law, Frost, and Wald 1991; Whelton and others · In China, the Tianjin Project showed a significant reduction 1998). An increase in daily fruit and vegetable intake may also in sodium intake in men after three years of intervention, lower blood pressure, and when combined with an increase in and after five years, the prevalence rates of both hyperten- low-fat dairy products and a reduction in saturated and total sion and obesity decreased among 45- to 65-year-olds fat, may lower blood pressure even more (Appel and others (Schooler and others 1997). 1997). Weight reduction lowers blood pressure in proportion · In Chile, the Mirame Project was a three-year intervention to the amount of weight lost (Whelton and others 1998), and program designed to provide and evaluate strategies to pro- physical activity appears to lower blood pressure in a way that mote healthy lifestyles among schoolchildren and their fam- may be independent of weight loss. High levels of alcohol ilies. Nissinen, Berrios, and Puska (2001) report a significant intake are associated with blood pressure elevation, which is positive effect on some risk factors for the intervention reversible by reducing intake (Kaplan 1995). schools. Dietary approaches to lowering total cholesterol and low- · In Mauritius, government-led initiatives resulted in a density lipoprotein (LDL) cholesterol typically involve reduced change in the composition of cooking oil from mostly palm intake of dietary fats, particularly saturated fats. Evidence sug- oil, which is high in saturated fatty acids, to wholly soybean gests a dose-response relationship between saturated fatty acid oil, which is high in unsaturated fatty acids. From 1987 to intake and LDL cholesterol levels (NCEP Expert Panel 2002). 1992, total cholesterol concentrations fell significantly, and Plant sterols and stanols have recently been incorporated into the estimated intake of saturated fatty acids decreased, with foods such as margarine and can reduce LDL cholesterol much of this finding reportedly resulting from the change in by about 10 percent; however, this approach is currently rela- cooking oil (Uusitalo and others 1996). tively expensive (Law 2000). Dietary advice may also suggest increasing the intake of viscous fiber--for instance, in the An effective populationwide intervention draws together form of cereal grains, fruits, and vegetables--because these different kinds of feasible activities that combined produce a dietary sources may enhance the lowering of LDL cholesterol. synergistic effect (Nissinen, Berrios, and Puska 2001; Puska Maintaining bodyweight in the desirable range and engaging in 1999). Even though the projects and trials were undertaken in a moderate physical activity complement these dietary strategies range of different communities and used a variety of methods (NCEP Expert Panel 2002). and interventions, several common themes emerge. Some of Increases in obesity have been related to declines in energy the important elements of a successful program that enables expenditure (for example, reductions in physical activity and individuals to adopt healthier lifestyles include the following: adoption of a more sedentary lifestyle) and a higher intake of energy-dense but micronutrient-poor foods, such as most · clear responsibility for coordinating prevention efforts, with processed foods (WHO 2003b). A variety of trials have recorded credible agencies with good communication methods carry- beneficial health effects, with weight reduction achieved by a ing out long-term education programs combination of interventions (NHLBI Obesity Education · intersectoral collaboration, with multiple messages sourced Initiative Expert Panel 1998). These interventions include from different organizations, including health sector enti- dietary counseling and therapy that involves a decrease in daily ties, nonhealth government agencies, schools, workplaces, caloric intake and a reduction in saturated fats and total fats. religious organizations, and voluntary agencies An increase in physical activity is an important component of · collaboration with the food industry to ensure the availabil- weight-loss therapy. Behavioral strategies revolving around ity of reasonably priced healthier food options, with food self-monitoring of eating habits, stress management, problem labeling that presents relevant information in a clear, reli- solving, and social support may also complement these able, and standardized format approaches. Overall, however, the effects of lifestyle modifica- · realistic multiyear time frames. tions to reduce weight and maintain the weight loss are rela- tively poor, with many reports finding that weight returns to Lifestyle and Dietary Personal Interventions. Many guide- baseline levels after several years. lines have concluded that lifestyle modifications, such as weight loss, healthy diet (such as one rich in potassium and low in Pharmacological and Surgical Personal Interventions. sodium), physical activity, and moderate alcohol consumption Randomized trials have shown that medications to lower blood are effective in reducing blood pressure (see, for example, pressure effectively reduce the risk of stroke, IHD, and heart fail- Chobanian and others 2003). Trials indicate that a reduction of ure. Results from meta-analyses of more than 40 different trials 856 | Disease Control Priorities in Developing Countries | Anthony Rodgers, Carlene M. M. Lawes, Thomas Gaziano, and others published in 2003 included about 210,000 participants and In clinical trials of statins, the relative risk reduction in car- more than 8,000 stroke and 11,000 IHD events (Blood Pressure diovascular events is similar at all levels of baseline cholesterol, Lowering Treatment Trialists' Collaboration 2003; Fox and extending to levels below 5 mmol/l total cholesterol, and is also EUROPA Investigators 2003; Law, Wald, and Rudnicka 2003; consistent among patients who are and are not taking concur- Lawes and others 2004; Pepine and others 2003). The trials may rent blood pressure lowering and other medications (Heart be broadly classified into three groups: (a) drug versus placebo Protection Study Collaborative Group 2002). Similar findings trials, (b) more intensive regimens to lower blood pressure ver- are observed with treatments to lower blood pressure (Progress sus less intensive regimens, and (c) drug versus drug trials. Collaborative Group 2001), indicating that these treatment The drug versus placebo trials achieved the greatest reduc- effects are independent. This finding is plausible, because they tions in blood pressure, and a dose-response relationship was act through different mechanisms and because observational apparent between blood pressure reduction and reduced risk of studies do not suggest a large interaction (Neaton and stroke. Overall, the trials indicated that a 10 mmHg reduction Wentworth 1992). in SBP would result in a 32 percent reduction in stroke risk and The benefits of lowering blood pressure and cholesterol are a 14 percent relative reduction in IHD risk. This finding is con- achieved surprisingly rapidly: for most outcomes,risk appears to sistent with the size of associations observed in cohort studies. be fully reversed within 6 to 18 months of beginning treatment. Clear evidence indicates that all the major drug classes have For example,individuals with cholesterol lowered in the past two similar effects on the risk of stroke and coronary heart disease or more years are at approximately the same coronary heart dis- per mmHg reduction in blood pressure (Blood Pressure ease risk as otherwise identical individuals whose cholesterol has Lowering Treatment Trialists' Collaboration 2003; Lawes and been at that level for decades (Law, Wald, and Thompson 1994). others 2004). The only clear evidence of clinically important, Pharmacological agents for weight loss that have been sub- class-specific effects are with agents that block the renin- ject to randomized controlled trials include dexfenfluramine, angiotensin system, which reduce diabetes incidence by about sibutramine, orlistat, and phentermine/fenfluramine (although one-quarter, and with calcium channel blockers, which reduce the last has been withdrawn because of a reported association heart failure less than other agents (although this result may be between the drugs and valvular heart disease). Overall, trials partly caused by misclassification, because a known side effect suggest only modest weight-loss effects, with an average net of calcium channel blockers is ankle edema, which is a diag- weight loss of 1.5 kg after eight weeks and 2 to 3 kg after one nostic component of heart failure). Because all agents lower year (NHLBI Obesity Education Initiative Expert Panel 1998). blood pressure by about the same modest amount and because A systematic review of orlistat trials indicated a pooled net their effects on blood pressure are additive (Law and others weight loss of 1.2 kg at 12 weeks, 2.9 to 3.4 kg at one year, and 2003), the key issue seems to be which combinations of two or 2.5 to 2.4 kg at two years (O'Meara and others 2001). Results of more drugs should be provided and how long-term adherence a systematic review of trials assessing sibutramine were similar can be maximized. (O'Meara and others 2002), with fewer data available on long- Over the past three decades, numerous trials have assessed term sustained weight loss. the effect of different cholesterol-lowering interventions (Law, Investigators have also undertaken several randomized con- Wald, and Rudnicka 2003; Law, Wald, and Thompson 1994). trolled trials to assess the effects of different surgical interven- The placebo-controlled trials can be broadly classified into tions, generally in individuals with a BMI equal to or greater those testing fibrates, statins, and other interventions (mostly than 35 or 40 kg/m2. Weight loss resulting from gastric bypass dietary interventions, but also some other interventions such as varied from 50 to 100 kg six months to a year following surgery resins and niacin). The statins are the most effective in lower- (NHLBI Obesity Education Initiative Expert Panel 1998). ing total and LDL cholesterol, with reductions of more than Overall, several trials suggest that surgery resulted in about 1 mmol/l in most trials. A good correlation has been found 23 to 37 kg more weight loss than conventional treatment and between reduction in total cholesterol and relative risk reduc- that this loss was maintained for eight years (Clegg and others tion. This finding suggests, as for trials investigating blood 2002). Furthermore, gastric bypass surgery appears to be more pressure lowering, that even though some drugs are more beneficial than gastroplasty or jejunoileal bypass. effective in achieving greater reductions in risk factors, their In relation to compliance and adherence with pharmacolog- effect on disease outcomes is similar per unit reduction of ical therapy, population surveys have demonstrated that, even cholesterol. Overall, a 1 mmol/l reduction in total cholesterol is in industrial countries, high blood pressure is either untreated associated with a 21 percent relative risk reduction in IHD and or inadequately controlled in about 70 to 75 percent of patients a 17 percent reduction in risk of stroke. Again, this finding is and that adherence to medications among patients suffering consistent with the epidemiology, with the proviso that the vast from chronic disease is only about 50 percent (WHO 2003a). majority of strokes in clinical trials were ischemic. The extent of poor adherence is likely to be even greater in The Growing Burden of Risk from High Blood Pressure, Cholesterol, and Bodyweight | 857 developing countries given the relative lack of health services COST-EFFECTIVENESS OF INTERVENTIONS and inequities in access. Pharmacotherapy faces a variety of potential barriers, including the symptomless nature of the Costs include expenditures required to identify and treat risk conditions, a lack of knowledge or denial of risk, the compli- factors as well as expenditures for treating CVD when it is not cated nature of drug regimens, the risk of side effects (real and prevented. Where possible, this chapter deals with the separate perceived), and the costs of treatment. sources of costs for several reasons. First, the costs for identify- Health providers may use multiple strategies to increase ing those requiring treatment vary significantly by level of eco- compliance and adherence. Patient-centered interventions nomic development and by urban versus rural location. In include involving individuals in the decision-making process; many situations in developing countries, such costs will make providing individualized patient education and disease coun- most or all forms of screening beyond a determination of CVD seling and adapting treatment to patients' lifestyles; simplifying history unaffordable. Second, in some countries, such as India, dosing schedules; providing drug information leaflets, medica- that are large producers of generic drugs, prices are reported to tion charts, and special reminder packaging for medications; be lower than in most other drug-producing or -importing holding group sessions for education and family-oriented dis- countries. Third, this approach allows researchers and policy ease management therapies; and implementing automated makers to understand the constituent costs so that they can telephone assessment and self-care education calls with nurse examine where cost reductions may be most beneficial. Fourth, follow-up (Haynes and others 2003). it clarifies what expenditures may be required as a result of Strategies may also aim to increase physician adherence, and changes in decisions about the treatment of risk factors. Finally, interventions may include the use of guidelines, peer review many people in developing countries do not have access to and audit, and prompts to remind physicians to review risks hospitals for acute management of CVD events. Nonetheless, and medications (Ebrahim 1998; NCEP Expert Panel 2002). increased expenditure on treating risk factors may lead to These strategies obviously do not address issues pertaining to significant reductions in the costs of treating subsequent CVD resources and access in poor countries. events for many countries. Ultimately, the net effect is reflected Several trials and overviews have attempted to assess the in cost-effectiveness analyses. Unless otherwise stated, costs are value of different interventions to improve compliance and in 2001 U.S. dollars. adherence; however, issues have arisen in connection with the The costs of personal interventions include the costs of generalizability of the interventions, the low statistical power patient screening (identifying high-risk patients), drugs and in many trials, the lack of description of all parts of interven- their acquisition, clinic visits, health care workers' time, labora- tions, and the assessment of complex interventions without tory tests, and travel. Annual drug costs for medications to assessment of the separate effects of the intervention compo- lower blood pressure and cholesterol vary widely by country nents. Haynes and others' (2003) systematic review concludes and depend on whether generics are available and used. For that, overall, no single approach to improving adherence can be example, according to the International Drug Price Indicator recommended. Simpler treatment regimens can sometimes Guide (Management Sciences for Health 2004), annual costs in improve adherence and treatment outcomes for both short- 2002 of generic 40 mg lovastatin ranged from US$14 in and long-term treatments. Several complex strategies, Barbados to US$217 in Costa Rica, and on-patent statins can including combinations of more thorough patient instructions cost almost a US$1,000 a year in the United States. Because and counseling, easier access to care, reminders, close follow- drug costs vary by up to two orders of magnitude across coun- up, supervised self-monitoring, family therapy, and rewards for tries, results of cost-effectiveness analyses are particularly success can improve adherence and treatment outcomes in sensitive to their input prices. Table 45.3 presents some some patients. However, even the most effective interventions sample prices. The costs of these medications have dropped did not lead to large improvements in adherence or treatment considerably in recent years, and now the annual costs for outcomes and were relatively resource intensive. By contrast, hydrochlorothiazide (25 mg), atenolol (50 mg), and captopril Connor, Rafter, and Rodgers's (2004) systematic review indi- (50 mg), are US$2, US$4, and US$9, respectively (Management cates improved adherence and clinical outcomes with fixed- Sciences for Health 2004). Statins will become increasingly dose combination treatment or unit-of-use packaging. affordable as simvastatin joins lovastatin in coming off patent Few good, evidence-based strategies to improve obesity (2006 in the United States and already off patent in Germany management are currently available, although reminder and the United Kingdom). systems, brief training interventions, shared care, inpatient The estimated number of visits to manage high blood pres- care, and dietitian-led treatments may all be worth further sure and cholesterol, under traditional paradigms, ranges from investigation (Harvey and others 2003). Thus, a clear need for two to six per year at costs ranging from US$3 to US$20 per innovations still exists to help people follow medication visit across the six regions assessed, but note that generally prescriptions as well as dietary and lifestyle advice. many fewer tests and less follow-up is required with a strategy 858 | Disease Control Priorities in Developing Countries | Anthony Rodgers, Carlene M. M. Lawes, Thomas Gaziano, and others Table 45.3 Annual Costs of Selected Cardiovascular Medications Medication United States (2002 US$)a Average international price (2002 US$) Projected polypillb Beta-blocker 32­365 3­15 n.a. Diuretic 6­37 1­3 n.a. Statin 180­864 11­147 n.a. Aspirin 2 1­6 n.a. Angiotensin-converting 65­365 1­19 n.a. enzyme inhibitor Total 285­1,633 17­190 20­40 Sources: U.S. prices: Murray 2004; international prices: Management Sciences for Health 2004. n.a. not applicable. a. Based on average wholesale prices. b. Based on a moderate increase from the sum of the lowest-cost generic components. based on absolute risk. Diagnostic testing for cholesterol in the nity intervention that expects a 4 percent reduction in total United States using a general laboratory is reimbursed at US$6 serum cholesterol and costs US$5 per person annually targeted for total cholesterol, US$16 for a complete lipoprotein choles- would save more than US$2 billion over 25 years of the pro- terol fractionation analysis, and US$6 for triglycerides (Xact gram. When the North Karelia (Puska 1999) estimates were Medicare Services 2003). Point-of-care one-step enzymatic used in a cost-effectiveness analysis in the United States strips that require only a few drops of blood from a finger (Tosteson and others 1997), the cost-effectiveness ratios ranged stick and that can process total cholesterol in minutes cost less from being cost saving to US$88,000 (in 1985 U.S. dollars) per than US$3 per test (Greenland and others 1987). A basic life year saved, depending on the percentage reduction in cho- metabolic panel for those on diuretics or for measuring renal lesterol (1 to 4 percent). function is US$12. The costs attributed to patient time and travel for visits have not been estimated for many countries, but they were recently estimated at US$12 to US$26 per visit Personal Interventions to Lower Blood Pressure or in the United States, depending on age and sex (Prosser and Cholesterol in Developed Countries others 2000). A common finding of cost-effectiveness analyses of primary A review of studies to date highlights several issues regard- prevention of CVD by means of lowering blood pressure and ing cost-effectiveness analyses, including the significant varia- cholesterol is the wide variability in cost-effectiveness tions in terms of calculations of cost per life year saved. The two ratios, depending on underlying risk, age, and costs of medica- most important aspects of the cost-effectiveness of any primary tions. For personal interventions using drug treatment for intervention are the future risk for CVD of the population lowering blood pressure and cholesterol levels, no single cost- treated and the costs of the medications. effectiveness analysis adequately summarizes experience in the developed countries. Collectively, the studies evaluating hypertension treatment in Australia, New Zealand, the United Population-Based Interventions States, and the Scandinavian countries suggest a range of cost- Given the strong association between CVD and high blood effectiveness ratios from US$4,600 to more than US$100,000 pressure, cholesterol, and body mass, most guidelines for those per life year gained when applied to the entire adult population risk factors begin by recommending lifestyle modifications. without further risk stratification (Kupersmith and others Although these benefits can lead to changes in risk factors, their 1995). Compared with the entire population, for those at high effect on CVD events is not well documented. However, on the risk with diastolic blood pressures over 105 mmHg and older basis of assumptions about cholesterol and blood pressure than 45, hypertension treatment can cost as little as a few hun- reduction from population-based lifestyle education programs dred dollars per life year gained or can even be cost saving and given the relatively low cost of the interventions--US$5 to (Johannesson and others 1991). US$17 per person per year (Tosteson and others 1997)--the Investigators have reported that primary prevention with cost-effectiveness of such programs may be reasonable. cholesterol-reducing medications is less attractive overall than However, the cost-effectiveness ratios of these interventions other interventions, such as hypertension treatment, from a were sensitive to the cost of the intervention as well as to the cost-effectiveness perspective, although once again this finding expected reduction in the risk factor. For example, a commu- is likely to differ considerably now that statins are off patent. The Growing Burden of Risk from High Blood Pressure, Cholesterol, and Bodyweight | 859 Reported cost-effectiveness ratios have ranged from US$10,000 both hypertension and increased cholesterol with all three med- to US$2 million per life year gained (Hay, Yu, and Ashraf 1999), ications. Finally, the effects of combination treatment with a whereas dietary interventions for cholesterol reduction are more beta-blocker, diuretic, statin, and aspirin were modeled for four favorable, with ratios of around US$2, 000 per quality-adjusted groups defined on the basis of absolute risk (10-year probability life year (QALY) (Prosser and others 2000).For cholesterol treat- of a cardiovascular event of 5, 15, 25, or 35 percent). ment, Prosser and others (2000) find cost-effectiveness ratios of Intervention effects were based on systematic reviews of US$50,000 per QALY for on-patent statins among those at high- randomized trials or meta-analyses. Population health effects est risk (high cholesterol levels and multiple risk factors) and up caused by the interventions were based on stochastically simu- to US$1.4 million per QALY among low-risk females when com- lating populations on the basis of age, sex, and risk factor dis- pared with dietary strategy alone. The cost per life year gained in tribution of smoking, hypertension, cholesterol, BMI, and the primary prevention trial of the West of Scotland Coronary smoking in the 14 subregions, both with and without the vari- Prevention Study among high-risk individuals treated with ous treatments to determine the effect. The effects of the inter- pravastatin was about US$30,000 (Caro and others 1997). Using vention were then translated into DALYs using a standard mul- the same criteria, Downs and others (1998) find that the cost per tistate modeling tool. Costs include both program-level costs life year saved in the Air Force/Texas Coronary Atherosclerosis (media, training, and administration) and patient-level costs Prevention Study cohort with average cholesterol levels was (medicines, health care visits, diagnostic tests). All costs were more than US$100,000.In general,younger and older age groups based on a standard ingredients approach and on regional esti- tend to have the least favorable cost-effectiveness ratios. For mations. The costs of CVD events were not included. younger groups, this finding probably reflects their overall lower The results are summarized in table 45.4. The incremental risk and the many years of treatment required before realizing a cost-effectiveness ratios for the strategy assessing absolute risk benefit. For the elderly, high cost-effectiveness ratios may reflect and using the triple combination of beta-blocker, statin, and other competing causes of death and the delay of up two aspirin with or without the addition of health education and salt years between treatment and benefit seen in most primary pre- legislation ranged from US$138 per DALY saved (absolute risk vention trials. greater than 35 percent) in the Africa E region to US$4,319 per DALY saved (absolute risk greater than 5 percent) in the Latin America and the Caribbean B region. These estimates are in Personal Interventions to Lower Blood Pressure or international or purchasing-power parity dollars (see chapter 15 Cholesterol in Developing Countries for an explanation). Table 45.4 also shows the approximate No trials of blood pressure, cholesterol, or body mass lowering equivalent costs in U.S. dollars and explains the conversion have been conducted solely in developing countries. As a result, from the WHO-CHOICE estimates. The nonpersonal interven- we have derived assessments of cost-effectiveness by extrapo- tions, including efforts to reduce salt intake in processed foods, lating from results in developed countries presented earlier. were less costly than the personal interventions. Personal Goldman and others (1991) report that a decline in the cost of interventions based on treatment guidelines were cost- lovastatin by 40 percent, once generic, would result in a roughly effective; however, when the strategies for treating high choles- 30 percent reduction in the cost-effectiveness ratio. However, terol or hypertension were compared with the absolute-risk this finding does not take into account that both the underly- approach, they were not favorable and were dominated by the ing epidemiology and the costs can be quite different across latter, meaning that the absolute-risk approach of treating those and within countries and regions. with a greater than 35 percent risk averted more DALYs and cost Murray and others (2003) compare 17 nonpersonal and per- less than either the blood pressure or cholesterol strategies. For sonal health service interventions or combinations of interven- an example of a country-specific analysis, see box 45.1. tions in the 14 epidemiological subregions defined by the Several recent publications have suggested that combination World Health Organization (WHO) as part of its Choosing treatments of medications to lower blood pressure, statin, Interventions That Are Cost-Effective (WHO-CHOICE) initia- aspirin, and perhaps other agents such as folate could more tive. The nonpersonal interventions included health education than halve cardiovascular risk (Wald and Law 2003; WHO through the mass media and legislative efforts to reduce salt 2002; Yusuf 2002). This suggestion is especially relevant for intake, improve blood pressure generally, and reduce cholesterol developing countries, given that suitable components are all and obesity levels. The personal interventions included treat- now off patent. Good evidence indicates that single-pill combi- ment with statins of those above two different cholesterol-level nations increase adherence to drug regimens (Connor, Rafter, thresholds (greater than 6.2 mmol/l or greater than 5.7 mmol/l), and Rodgers 2004) and reduce supply and transport costs. We treatment with beta-blockers and diuretics of those above two used a Markov model to evaluate the cost-effectiveness of such different hypertension thresholds (greater than 160 mmHg or a hypothetical pill or combination packaging of the individual greater than 140 mmHg), and treatment of individuals with medications. We modeled the effect of a pill that included half 860 | Disease Control Priorities in Developing Countries | Anthony Rodgers, Carlene M. M. Lawes, Thomas Gaziano, and others Table 45.4 Comparison of the Cost-Effectiveness of Absolute Risk with Treatment According to Either Blood Pressure or Lipid Targets Alone in Addition to Population-Based Strategies, Selected WHO Regions Incremental cost-effectiveness ratio (cost/DALY saved)a Region Strategy Risk (percent) International $ US$ Africa E Prevention (SL and/or HE) Dominatedb Targeted risk factorsc Dominatedb Absolute riskc (TRI) 35 138 42 25 778 295 15 1,445 639 Latin America and the Prevention (SL) 127 65 Caribbean B Prevention (SL HE) 145 74 Targeted risk factorsc Dominatedb Absolute riskd (TRI SL HE) 35 286 178 25 1,598 1,058 15 2,391 1,664 5 4,319 3,075 Southeast Asia B Prevention (SL) 70 18 Prevention (SL HE) 127 32 Targeted risk factorsc Dominatedb Absolute riskd (TRI SL HE) 35 301 133 25 1,197 578 15 2,094 1,120 5 3,952 2,233 Western Pacific B Prevention (SL) 97 18 Targeted risk factorsc Dominatedb Absolute riskd (TRI SL HE) 35 1,124 423 25 1,278 564 15 2,092 1,042 5 4,028 2,135 Source: Murray and others 2003. B low child mortality and low adult mortality; E high child mortality and very high adult mortality; HE health education through the mass media to reduce cholesterol; SL legislation to decrease the salt content of processed foods, including appropriate labeling and enforcement; TRI treatment with aspirin, beta-blockers, and a statin. a. Costs of prevention and nondrug costs for treatment according to absolute risk are converted at an estimated regional average ratio of exchange rate to purchasing-power parity rate; drug costs are not converted, assuming drugs to be imported at world prices. The share of drug costs in total treatment cost, as a function of risk, is taken from the estimates for India in table 45.6 and assumed to be the same for all regions. b. Dominated strategies were both less effective and more costly than comparator strategies. c. Treating SBP greater than 140 mmHg or 160 mmHg or total cholesterol greater than 5.7 mmol/l or 6.2 mmol/l (220 or 240 mg/dlL). d. Risk refers to 10-year risk of CVD greater than or equal to the number listed. of the standard doses of hydrochlorothiazide, atenolol, lisino- fits seen from a societal perspective and with the intervention pril, lovastatin, and aspirin on overall morbidity and mortality run for 10 years. We calculated estimates for one representative in treating those without prior CVD. We did not include folate country from each region where demographic and risk factor because no randomized trials had shown that it reduced CVD data existed. Unlike the WHO-CHOICE analysis, this analysis events at the time of the analysis. The assumptions of the rela- separated use of the intervention according to those with and tive risk reductions were based on those of Wald and Law without established CVD. Table 45.5 presents the results. (2003). The strategies compared were for treating various high- Table 45.6 shows the breakdown of events averted and costs risk populations (absolute risk for CVD greater than 15, 25, and for India. Even though the absolute numbers differ for other 35 percent over 10 years. We applied the model to a population countries, the relative differences between the different groups of 1 million adults over the age of 35, with the costs and bene- receiving the "polypill" compared with the groups not receiving The Growing Burden of Risk from High Blood Pressure, Cholesterol, and Bodyweight | 861 Box 45.1 Example of Country-Specific Analysis: South Africa In another analysis, researchers (Gaziano 2001) compared than 40 percent had an incremental cost-effectiveness ratio the approach based on absolute risk with blood pressure of US$700 per QALY gained compared with no treatment. guidelines in South Africa. The analysis used country- The absolute risk of CVD greater than 30, 20, and 15 per- specific epidemiology and, where available, applied local cent had larger and increasing cost-effectiveness ratios. cost data. The study compared six strategies for initiating Treatments based on the 1995 South African guidelines and drug treatment--two different blood pressure levels the Joint National Commission VI guidelines were both (160/95 mmHg and 140/90 mmHg) and four different more costly and resulted in fewer QALY gains than the 15 levels of absolute CVD risk over 10 years (40, 30, 20, and percent absolute-risk strategy and were therefore domi- 15 percent)--to a strategy of no treatment. The method- nated by the less costly absolute-risk treatment strategies. ology differed from the WHO-CHOICE study because of Furthermore, the results showed that the cost- the availability of local data. Data on diabetes prevalence effectiveness ratios were quite sensitive to the costs of were included to further refine risk estimates. Also the treatment for hypertension, especially medication costs. actual mix of medications was used to assess costs with Further analysis revealed a threshold point for an annual actual current drug-use patterns, which included the use treatment cost of US$53. Below this threshold, the 40 per- of some nongeneric medications. cent absolute-risk strategy cost less and increased the The table displays the results. The four absolute-risk number of life years gained compared with the no pri- strategies had the four lowest incremental cost-effectiveness mary prevention strategy and is therefore cost saving. In ratios.The strategy of initiating antihypertensive therapy for South Africa, annual treatment with diuretics and beta- those individuals with a predicted 10-year CVD risk greater blockers could be provided for less than US$40. Incremental Cost-Effectiveness Ratios for Selected Hypertension Management Strategies over 10 Years, South Africa Incremental cost-effectiveness ratioa Treatment US$/QALY US$/life year savedb No treatment n.a. n.a. Absolute risk of CVD 40 percent 700 900 Absolute risk of CVD 30 percent 1,600 2,100 Absolute risk of CVD 20 percent 4,900 6,700 Absolute risk of CVD 15 percent 11,000 18,000 Target level 160/95 mmHg (1995 South African guidelines)c Dominatedd Dominatedd Target level 140/90 mmHg (Joint National Commission VI guidelines) Dominatedd Dominatedd Source: Gaziano and others 2005. n.a. not applicable. a. Each strategy's costs and effects are compared with those of the preceding less costly strategy. b. Total and incremental life years not shown. c. Compared with an absolute risk of CVD greater than 15 percent because the 1995 South African guidelines are dominated by the former. d. A dominated strategy is one that is both more expensive and less effective than the preceding strategy to which it is compared. Table 45.5 Incremental Cost-Effectiveness Ratios of a Multidrug Regimen by World Bank Region Compared with a Baseline of No Drug Treatment (2001 US$/DALY) Region 35 percent risk 25 percent risk 15 percent risk 5 percent risk East Asia and the Pacific 830 1,440 2,320 3,820 Europe and Central Asia 940 1,450 1,960 3,620 Latin America and the Caribbean 920 1,470 2,420 3,740 Middle East and North Africa 720 1,290 2,190 4,030 South Asia 670 1,250 1,932 3,020 Sub-Saharan Africa 610 1,170 1,920 2,960 Source: Authors' calculations. Note: The regimen includes aspirin, a beta-blocker, a thiazide diuretic, an angiotensin-converting enzyme inhibitor, and a statin. The risk refers to a 10-year risk of CVD. 862 | Disease Control Priorities in Developing Countries | Anthony Rodgers, Carlene M. M. Lawes, Thomas Gaziano, and others Table 45.6 Polypill Cost-Effectiveness Estimates for a Population of 1 Million Adults at Varying Levels of Risk for CVD Treated for 10 Years in India Absolute risk of a CVD event over 10 years Comparison with 35 25 15 5 Costs and effects no polypill percent percent percent percent Total cost (2001 US$ millions) 23.5 34.5 51.4 92.2 205.2 Profile of total costs Percentage attributable to inpatient stay 12.0 6.0 3.0 1.0 0.3 Percentage attributable to ambulatory care 0 29.0 40.0 49.0 54.0 Percentage attributable to labor 75.0 36.0 21.0 9.0 2.0 Percentage attributable to pharmaceuticals 0 23.0 31.0 38.0 42.0 Percentage attributable to laboratory expenses 12.0 6.0 3.0 1.0 0.0 Effectsa Number of myocardial infarction cases averted n.a. 10,200 14,400 21,300 31,800 Number of stroke cases averted n.a. 5,200 7,000 12,400 19,600 Number of coronary heart disease deaths averted n.a. 10,500 13,500 19,600 25,900 Number of stroke deaths averted n.a. 5,900 7,500 10,500 14,200 Number of life years saved n.a. 39,000 51,000 67,000 98,000 Number of DALYs gained n.a. 41,000 57,000 86,000 134,000 Incremental cost-effectiveness (US$/DALY) n.a. 300 990 1,500 2,430 Source: Authors' calculations. n.a. not applicable. a. Each strategy compared with no polypill. it are similar for all countries. Although the total costs for treat- Distributional and Equity Consequences ing lower-risk patients increase, so do the benefits, and the Failing to translate available evidence from industrial countries overall incremental cost-effectiveness ratio remains relatively about CVD prevention strategies into practicable solutions for favorable. The proportion of costs shifts away from those developing countries would have clear equity implications, attributable to hospital care when no primary prevention is ini- especially when CVD is a large and growing problem in devel- tiated to costs attributable to ambulatory care and pharmaceu- oping countries and when safe and effective interventions that ticals when more lower-risk patients are treated. were once extremely expensive are now available for a few cents a day. Because access to cardiovascular health care in develop- Interventions to Reduce Bodyweight ing countries often depends on patients' ability to pay, the poor No large-scale randomized trials of weight reduction as an iso- would stand to benefit the most from a low-cost intervention lated intervention are available on which to base estimates of such as a polypill. the benefits of weight loss in lowering the risk of coronary Some see CVD as exclusively a disease of the affluent in heart disease. Thus, costs per life year saved would have to be developing countries; however, in many developing coun- modeled to project benefits. In one such analysis, a school- tries, the transition of CVD to becoming a disease of the based educational program to reduce obesity among middle poor has already begun--a transition already seen in devel- school students reported a cost of US$4,300 per QALY (L. oped countries around the world. A recent analysis of the Wang and others 2003). However, this analysis assumed that distribution of major cardiovascular risks by poverty levels the weight loss would be maintained throughout adulthood, has shown that many cardiovascular risks already affect the but the high relapse rates found in weight reduction studies poor in the world's poorest countries (Ezzati and others 2004; do not bear out this assumption (Serdula and others 1999). WHO 2002). Combating the trend requires highly effective, Further research is needed to evaluate the benefits of low-cost solutions relevant for most or all of those at risk in weight reduction in relation to reducing CVD events and the developing countries, in contrast to the investments in high- long-term sustainability of weight loss before reliable cost- tech treatment interventions that have commonly occurred to effectiveness estimates can be made. date. The Growing Burden of Risk from High Blood Pressure, Cholesterol, and Bodyweight | 863 ECONOMIC BENEFITS OF INTERVENTION new paradigm and its advantages over traditional para- digms, such as hypertension treatment. One barrier to In the cost-effectiveness analyses, most of the gains are reported adopting preventive therapy based on absolute risk has in cost savings either from particular interventions, such as been its relative complexity compared with dichotomous decreased hospitalizations resulting from the improved combi- diagnosis-based strategies, such as hypertension­no nation therapy of the polypill, or from a more efficient means hypertension. of screening those at highest risk through an absolute-risk º Develop simple methods for predicting absolute risk approach. Those who do not die from the sequelae of poorly using straightforward, inexpensive, direct measures, such controlled risk factors for CVD suffer from serious chronic ill- as physical examination, clinical history, and on-site tests. ness, such as stroke and congestive heart failure. Those chronic These methods would likely involve low-cost algorithms diseases can result in significant impairments, thereby prevent- completed by a multipurpose health care worker involv- ing those affected from continuing to work and sometimes also ing, for example, the collection of data on age, sex, tobacco requiring the services of other family members, who them- use, blood pressure, waist circumference, and urine selves end up having to leave the workforce. Further losses dipstick results. The development of different levels of resulting from disability include the loss of wages for major screening protocol may also be needed in certain settings. wage earners and their families and the state's losses in terms of º Calibrate existing algorithms for different disease rates disability compensation. Leeder and others (2004) estimate and cardiovascular profiles in developing countries. that in 2000 the cost of CVD disability payments in South º Develop treatment algorithms that can easily be adopted Africa equaled US$70 million. in resource-poor settings by, for example, multipurpose However, many other indirect economic gains or losses are health care workers. not included in the economic analysis, such as gains or losses in º Develop methods for predicting absolute risk on the productivity. Leeder and others (2004) report that, at current basis of the probability of lost healthy life years as well as CVD mortality rates, the potential productive years of life lost the probability of a clinical event. This strategy could (defined as those years between the ages of 35 and 64) will mean developing an index of healthy life years at risk nearly double by 2030. Those later adult working years are par- from a cardiovascular event in the next five years, which ticularly important, given the many years of investment in skills would require taking case fatalities into consideration through formal education and experience that would be lost. and discounting. A major barrier to adopting a strategy Preventing CVD would therefore improve the size and skills of based on absolute risk has been the absence of a time- the workforce and would therefore aid economic development. based measure and, hence, the equal value placed on pre- For those reasons, the Commission for Macroeconomics and venting an event at a young and at an old age. Health has recommended that any intervention that costs less · Develop and evaluate combination treatments: than triple a country's per capita gross domestic product be º Carry out new research on the ideal combinations for regarded as cost-effective (WHO 2001). Many of the combina- different patient groups and populations at different tion cardiovascular preventive approaches outlined in this stages of the health transition. Local initiatives would be chapter comfortably satisfy that criterion. needed to determine the ideal combination of medica- tions based principally on cost, tolerability, and ability to RESEARCH AND DEVELOPMENT AGENDA lower risk-factor levels. One default set of interventions could be an angiotensin-converting enzyme inhibitor The cost-effectiveness data reviewed in this chapter indicate that (for example, enalapril or lisinopril); a diuretic (such as the best use of resources for personal-level interventions for hydrochlorothiazide or chlothalidone); a statin (for preventing CVD mediated by high blood pressure, cholesterol, instance, simvastatin or lovastatin); and low-dose and bodyweight would be combination medications targeted to aspirin. those at high absolute risk. This strategy represents a consider- º Measure the potential costs and benefits of adding other able departure from existing paradigms, such as hypertension active agents, such as vitamins or diabetic medications. treatment. Research and development is therefore required in º Quantify the extent of improved access, acceptability, several areas to develop, implement, and evaluate this strategy. and tolerability for people with symptomatic vascular This research could include several themes as follows: disease who have established indications for those medications. · Refine absolute risk-based treatment in developing country º Evaluate the benefits and costs in developing countries settings: with large-scale clinical trials and demonstration º Evaluate optimal communications to the public and to projects, both among patients who have established health professionals that explain the rationale for this indications (compared with usual care) and among those 864 | Disease Control Priorities in Developing Countries | Anthony Rodgers, Carlene M. M. Lawes, Thomas Gaziano, and others who do not have clear indications but are still at high risk REFERENCES (compared with a placebo). Alderman, M. H. 1996. "Blood Pressure J-Curve: Is It Cause or Effect?" º Evaluate the advantages and disadvantages of a polypill Current Opinion in Nephrology and Hypertension 5 (3): 209­13. versus unit-of-use packs and other novel delivery Anderson, K. M., P. M. Odell, P. W. Wilson, and W. B. Kannel. 1991. strategies. "Cardiovascular Disease Risk Profiles." American Heart Journal 121 · Investigate weight-loss initiatives: (1, part 2): 293­98. Appel, L. J., T. J. Moore, E. Obarzanek, W. M. Vollmer, L. P. Svetkey, F. M. º Develop strategies to improve the effectiveness of per- Sacks, and others. 1997. "A Clinical Trial of the Effects of Dietary sonal interventions to reduce bodyweight in developing Patterns on Blood Pressure: DASH Collaborative Research Group." countries. New England Journal of Medicine 336 (16): 1117­24. º Evaluate the use of gastric surgery for weight loss in the Asia Pacific Cohort Studies Collaboration. 1999. "Determinants of extremely obese in selected settings. Cardiovascular Disease in the Asia Pacific Region: Protocol for a Collaborative Overview of Cohort Studies." CVD Prevention 2: 281­89. · Assess technology: ------. 2003a. "Blood Pressure and Cardiovascular Disease in the Asia º Screen which technologies should be transferred to devel- Pacific Region." Journal of Hypertension 21: 707­16. oping countries on the basis of cost-effectiveness criteria. ------. 2003b. "Cholesterol, Coronary Heart Disease, and Stroke in the º Design new technologies specifically for use by commu- Asia Pacific Region." International Journal of Epidemiology 32: 563­72. nity health workers (for example, point-of-care devices). ------. 2004. "Body Mass Index and Cardiovascular Disease in the Asia- Pacific Region: An Overview of 33 Cohorts Involving 305,000 · Review public and personal health services: Participants." International Journal of Epidemiology 33: 1­8. º Carry out a critical evaluation of community health Assmann, G., P. Cullen, and H. Schulte. 2002. "Simple Scoring Scheme for workers versus trained health professionals in delivering Calculating the Risk of Acute Coronary Events Based on the 10-Year simplified screening and treatment regimens. Follow-up of the Prospective Cardiovascular Munster (PROCAM) Study." Circulation 105: 310­15. º Provide guideline assistance for CVD prevention and Blood Pressure Lowering Treatment Trialists' Collaboration. 2003. "Effects management to regional and country-specific ministers of Different Blood-Pressure-Lowering Regimens on Major of health and policy makers. Cardiovascular Events: Results of Prospectively Designed Overviews of Randomised Trials: Blood Pressure Lowering Treatment Trialists' º Support demonstration projects to determine the limita- Collaboration." Lancet 362 (9395): 1527­35. tions for managing chronic conditions in resource-poor Bobak, M., Z. Skodova, Z. Pisa, R. Poledne, and M. Marmot. 1997. settings. "Political Changes and Trends in Cardiovascular Risk Factors in the Czech Republic, 1985­92." Journal of Epidemiology and Community Health 51 (3): 272­77. Calle, E. E., M. J. Thun, J. M. Pettrelli, C. Rodriguez, and C. Heath. 1999. "Body Mass Index and Mortality in a Prospective Cohort of U.S. CONCLUSIONS Adults." New England Journal of Medicine 341 (15): 1097­1105. Caro, J., W. Klittich, A. McGuire, I. Ford, J. Norrie, and D. Pettitt. 1997. The analyses presented in this chapter indicate that providing "The West of Scotland Coronary Prevention Study: Economic Benefit off-patent blood pressure and cholesterol-lowering medica- Analysis of Primary Prevention with Pravastatin." British Medical Journal 315: 1577­82. tions targeted at those at high absolute risk seems to be a cost- Chobanian, A. V., G. L. Bakris, H. R. Black, W. C. Cushman, L. A. Green, effective strategy. Currently available personal interventions to I. L. Izzo Jr., and others. 2003. "The Seventh Report of the Joint prevent or reduce high BMI are likely to be much less cost- National Committee on Prevention, Detection, Evaluation, and effective. Treatment of High Blood Pressure: The JNC 7 Report." Journal of the American College of Cardiology 289 (19): 2560­72. An approach based on absolute risk will still involve choos- Clegg, A. J., J. Colquitt, M. K. Sidhu, P. Royle, E. Loveman, and A. Walker. ing some level below which people are not recommended for 2002. "The Clinical Effectiveness and Cost-Effectiveness of Surgery for personal treatments, which will leave some people at risk of People with Morbid Obesity: A Systematic Review and Economic progression of vascular disease. This issue exists with current Evaluation." Health Technology Assessment (Winchester, U.K.) 6 (12): 1­153. paradigms and underscores the need for parallel improvements Connor, J., N. Rafter, and A. Rodgers. 2004. "Do Fixed-Dose Combination in population-based prevention. The strategy based on Pills or Unit-of-Use Packaging Improve Adherence? A Systematic absolute risk must be regarded as complementary to popula- Review." Bulletin of World Health Organization 82: 935­39. tionwide initiatives that address the root causes of CVD--in Cruickshank, J. M. 1994. "J-Curve in Antihypertensive Therapy: Does It Exist? A Personal Point of View." Cardiovascular Drugs and Therapy particular, the societal determinants that lead to high salt and 8 (5): 757­60. saturated fat in the diet in relation to high blood pressure and D'Agostino, R. B., A. J. Belanger, W. B. Kannel, and J. M. Cruickshank. cholesterol and high-energy diets coupled with decreasing 1991. "Relation of Low Diastolic Blood Pressure to Coronary Heart physical activity in relation to high bodyweight. Preventing and Disease Death in Presence of Myocardial Infarction: The Framingham Study." British Medical Journal 303 (6799): 385­89. reducing those risks in developing countries will reduce the Downs, J. R., M. Clearfield, S. Weis, E. Whitney, D. R. Shapiro, P. A. Beere, need for medication-based prevention strategies in the coming and others. 1998. "Primary Prevention of Acute Coronary Events decades. with Lovastatin in Men and Women with Average Cholesterol The Growing Burden of Risk from High Blood Pressure, Cholesterol, and Bodyweight | 865 Levels: Results of AFCAPS/TexCAPS--Air Force/Texas Coronary Haynes, R. B., H. McDonald, A. X. Garg, and P. Montague. 2003. Atherosclerosis Prevention Study." Journal of the American Medical "Interventions for Helping Patients to Follow Prescriptions for Association 279 (20): 1615­22. Medications." Cochrane Database of Systematic Reviews (1). Ebrahim, S. 1998. "Detection, Adherence, and Control of Hypertension He, J., M. J. Klag, P. K. Whelton, J. Y. Chen, J. P. Mo, M. C. Qian, and others. for the Prevention of Stroke: A Systematic Review." Health Technology 1991. "Migration, Blood Pressure Pattern, and Hypertension: The Assessment (Winchester, U.K.) 2 (11): i­iv, 1­78. Yi Migrant Study." American Journal of Epidemiology 134 (10): Evans, A., H. Tolonen, H. W. Hense, M. Ferrario, S. Sans, K. Kuulasmaa, and 1085­1101. others. 2001. "Trends in Coronary Risk Factors in the WHO MONICA He, J., G. S. Tell, Y. C. Tang, P. S. Mo, and G. Q. He. 1991. "Effect of Project." International Journal of Epidemiology 30 (Suppl. 1): S35­40. Migration on Blood Pressure: The Yi People Study." Epidemiology 2 (2): Ezzati, M., A. Lopez, A. Rodgers, S. Vander Hoorn, and C. J. L. Murray, eds. 88­97. 2004. Comparative Quantification of Health Risks: Global and Regional Heart Protection Study Collaborative Group. 2002. "MRC/BHF Heart Burden of Disease Attributable to Selected Major Risk Factors. Geneva: Protection Study of Cholesterol Lowering with Simvastatin in 20,536 World Health Organization. High-Risk Individuals: A Randomised Placebo-Controlled Trial." Ezzati, M., S. Vander Hoorn, A. Rodgers, A. D. Lopez, C. D. Mathers, C. J. L. Lancet 360 (9326): 7­22. Murray, and others. 2003. "Estimates of Global and Regional Potential Hodgson, T. A., and L. Cai. 2001. "Medical Care Expenditures for Health Gains from Reducing Multiple Major Risk Factors." Lancet 362: Hypertension, Its Complications, and Its Comorbidities." Medical Care 271­80. 39: 599­615. Farnett, L., C. D. Mulrow, W. D. Linn, C. R. Lucey, and M. R. Tuley. 1991. Iso, H., D. R. Jacobs Jr., D. Wentworth, J. D. Neaton, and J. D. Cohen. 1989. "The J-Curve Phenomenon and the Treatment of Hypertension. Is "Serum Cholesterol Levels and Six-Year Mortality from Stroke in There a Point Beyond Which Pressure Reduction Is Dangerous?" 350,977 Men Screened for the Multiple Risk Factor Intervention Trial." Journal of the American Medical Association 265 (4): 489­95. New England Journal of Medicine 320 (14): 904­10. Feinleib, M. 1996. "New Directions for Community Intervention Studies." Jackson, R., P. Barham, J. Bills, T. Birch, L. McLennan, S. MacMahon, and American Journal of Public Health 86 (12): 696­98. others. 1993. "Management of Raised Blood Pressure in New Zealand: Field, A. E., E. H. Coakley, A. Must, J. L. Spadano, N. Laird, W. H. Dietz, A Discussion Document." British Medical Journal 307: 107­10. and others. 2001. "Impact of Overweight on the Risk of Developing Johannesson, M., L. Borgquist, B. Jonsson, and L. Rastam. 1991. "The Common Chronic Diseases during a 10-Year Period." Archives of Costs of Treating Hypertension: An Analysis of Different Cutoff Internal Medicine 161: 1581­86. Points." Health Policy 18 (2): 141­50. Flack, J. M., J. Neaton, R. Grimm Jr., J. Shih, J. Cutler, K. Ensrud, and Joseph, J. G., I. A. Prior, C. E. Salmond, and D. Stanley. 1983. "Elevation others. 1995. "Blood Pressure and Mortality among Men with Prior of Systolic and Diastolic Blood Pressure Associated with Migration: Myocardial Infarction: Multiple Risk Factor Intervention Trial The Tokelau Island Migrant Study." Journal of Chronic Diseases 36 (7): Research Group." Circulation 92 (9): 2437­45. 507­16. Fox, K. M. and EUROPA (European Trial on Reduction of Cardiac Events Kannel, W. B., R. B. D'Agostino, and H. Silbershatz. 1997. "Blood Pressure with Perindopril in Stable Coronary Artery Disease) Investigators. and Cardiovascular Morbidity and Mortality Rates in the Elderly." 2003. "Efficacy of Perindopril in Reduction of Cardiovascular Events American Heart Journal 134 (4): 758­63. among Patients with Stable Coronary Artery Disease: Randomised, Double-Blind, Placebo-Controlled, Multicentre Trial (the EUROPA Kaplan, N. M. 1995. "Alcohol and Hypertension." Lancet 345 (8965): Study)." Lancet 362 (9386): 782­88. 1588­89. Gaziano, T. A., K. Steyn, D. J. Cohen, M. C. Weinstein, and L. H. Opie. 2005. Kupersmith, J., M. Holmes-Rovner, A. Hogan, D. Rovner, and J. Gardiner. "Cost-Effectiveness Analysis of Hypertension Guidelines in South Africa: 1995. "Cost-Effectiveness Analysis in Heart Disease, Part II: Preventive Absolute Risk versus Blood Pressure Level."Circulation 112 (23): 3569­76. Therapies." Progress in Cardiovascular Diseases 37: 243­71. Goldman, L., M. C. Weinstein, P. A. Goldman, and L. W. Williams. 1991. Law, M. 2000. "Plant Sterol and Stanol Margarines and Health." British "Cost-Effectiveness of HMG-CoA Reductase Inhibition for Primary Medical Journal 320: 861­64. and Secondary Prevention of Coronary Heart Disease." Journal of the Law, M., C. Frost, and N. Wald. 1991. "By How Much Does Dietary Salt American Medical Association 265: 1145­51. Reduction Lower Blood Pressure? III: Analysis of Data from Trials of Greenland, P., J. C. Levenkron, M. G. Radley, J. G. Baggs, R. A. Manchester, Salt Reduction." British Medical Journal 302: 819­24. and N. L. Bowley. 1987. "Feasibility of Large-Scale Cholesterol Law, M. R., N. J. Wald, J. K. Morris, and R. E. Jordan. 2003. "Value of Low Screening: Experience with a Portable Capillary-Blood Testing Dose Combination Treatment with Blood Pressure Lowering Drugs: Device." American Journal of Public Health 77: 73­75. Analysis of 354 Randomised Trials." British Medical Journal 326 (7404): Hansson, L., L. H. Lindholm, T. Ekbom, B. Dahlof, J. Lanke, B. Schersten, 1427. and others. 1999. "Randomised Trial of Old and New Antihypertensive Drugs in Elderly Patients: Cardiovascular Mortality and Morbidity-- Law, M. R., N. J. Wald, and A. R. Rudnicka. 2003. "Quantifying Effect of The Swedish Trial in Old Patients with Hypertension-2 Study." Lancet Statins on Low Density Lipoprotein Cholesterol, Ischaemic Heart 354 (9192): 1751­56. Disease, and Stroke: Systematic Review and Meta-Analysis." British Medical Journal 326 (7404): 1423. Haq, I. U., L. E. Ramsay, W. W. Yeo, P. R. Jackson, and E. J. Wallis. 1999. "Is the Framingham Risk Function Valid for Northern European Law, M. R., N. J. Wald, and S. G. Thompson. 1994. "By How Much and Populations? A Comparison of Methods for Estimating Absolute How Quickly Does Reduction in Serum Cholesterol Concentration Coronary Risk in High Risk Men." Heart 81 (1): 40­46. Lower Risk of Ischaemic Heart Disease?" British Medical Journal 308 (6925): 367­72. Harvey, E. L., A. M. Glenny, S. F. L. Kirk, and C. D. Summerbell. 2003. "Improving Health Professionals' Management and the Organisation Law, M. R., H. C. Watt, and N. J. Wald. 2002. "The Underlying Risk of of Care for Overweight and Obese People." Cochrane Database of Death after Myocardial Infarction in the Absence of Treatment." Systematic Reviews (1). Archives of Internal Medicine 162 (21): 2405­10. Hay, J. W., W. M. Yu, and T. Ashraf. 1999. "Pharmacoeconomics of Lipid- Lawes, C. M. M., D. A. Bennett, V. L. Feigin, and A. Rodgers. 2004. "Blood Lowering Agents for Primary and Secondary Prevention of Coronary Pressure and Stroke: An Overview of Published Reviews." Stroke 35: Artery Disease." Pharmacoeconomics 15: 47­74. 776­85. 866 | Disease Control Priorities in Developing Countries | Anthony Rodgers, Carlene M. M. Lawes, Thomas Gaziano, and others Leeder, S., S. Raymond, H. Greenburg, H. Liu, and K. Esson. 2004. A Race Pfeffer, M. A. 1993. "Angiotensin-Converting Enzyme Inhibition in against Time: The Challenge of Cardiovascular Disease in Developing Congestive Heart Failure: Benefit and Perspective." American Heart Economies. New York: Columbia University. Journal 126 (3, part 2): 789­93. MacMahon, S., A. Rodgers, B. Neal, and J. Chalmers. 1997."Blood Pressure Poulter, N. R. 1999. "Coronary Heart Disease Is a Multifactorial Disease." Lowering for the Secondary Prevention of Myocardial Infarction and American Journal of Hypertension 12 (10, part 2): 92­95S. Stroke." Hypertension 29: 537­38. Poulter, N. R., K. T. Khaw, and P. S. Sever. 1988. "Higher Blood Pressures Management Sciences for Health. 2004. International Drug Price Indicator of Urban Migrants from an African Low-Blood Pressure Population Guide. Washington, DC: Management Sciences for Health. Are Not Due to Selective Migration." American Journal of Hypertension 1 (3 Pt. 3): 143S­45S. Manson, J. E., W. C. Willett, and M. J. Stampfer. 1995. "Body Weight and Mortality among Women." New England Journal of Medicine 333 (11): Poulter, N. R., and P. Sever. 1994. "Blood Pressure in Other Populations: A. 677­85. Low Blood Pressure Populations and the Impact of Rural-Urban Migration." In Textbook of Hypertension, ed. J. Swales, 22­36. Oxford, McMurray, J., and G. T. McInnes. 1992. "The J-Curve Hypothesis." Lancet U.K.: Blackwell Scientific Publications. 339 (8792): 561­62. Progress Collaborative Group. 2001. "Randomised Trial of a Perindopril- Mittelmark, M. B., M. K. Hunt, G. W. Heath, and T. L. Schmid. 1993. Based Blood-Pressure-Lowering Regimen among 6,105 Individuals "Realistic Outcomes: Lessons from Community-Based Research and with Previous Stroke or Transient Ischaemic Attack." Lancet 358 Demonstration Programs for the Prevention of Cardiovascular (9287): 1033­41. Diseases." Journal of Public Health Policy 14 (4): 437­62. Prospective Studies Collaboration. 1995. "Cholesterol, Diastolic Blood Monteiro, C. A., W. L. Conde, B. Lu, and B. M. Popkin. 2004. "Obesity and Pressure, and Stroke: 13,000 Strokes in 45,000 People in 45 Prospective Inequities in Health in the Developing World." International Journal of Cohorts." Lancet 346: 1647­53. Obesity 28: 1181­86. ------. 2002. "Age-Specific Relevance of Usual Blood Pressure to Vascular Murray, C. J. L., J. A. Lauer, R. C. W. Hutubessy, L. Niessen, N. Tomijima, Mortality: A Meta-Analysis of Individual Data for One Million Adults A. Rodgers, and others. 2003. "Reducing the Risk of Cardiovascular in 61 Prospective Studies." Lancet 360: 1903­13. Disease: Effectiveness and Costs of Interventions to Reduce Systolic Prosser, L. A., A. A. Stinnett, P. A. Goldman, L. W. Williams, M. G. Hunink, Blood Pressure and Cholesterol: A Global and Regional Analysis." and L. Goldman. 2000. "Cost-Effectiveness of Cholesterol-Lowering Lancet 361: 717­25. Therapies According to Selected Patient Characteristics." Annals of Murray, L., ed. 2004. Red Book. Montvale, NJ: Thomson Physicians Desk Internal Medicine 132: 769­79. Reference. Puska, P. 1999. "The North Karelia Project: From Community NCEP (National Cholesterol Education Program) Expert Panel. 2002. Intervention to National Activity in Lowering Cholesterol Levels and Third Report of the Expert Panel on Detection, Evaluation, and CHD Risk." European Heart Journal 1 (Suppl.): S9­13. Treatment of High Blood Cholesterol in Adults (Adult Treatment Ramsay, L. E., B. Williams, G. D. Johnston, G. A. MacGregor, L. Poston, J. Panel III). Bethesda, MD: National Institutes of Health, National F. Potter, and others. 1999. "British Hypertension Society Guidelines Heart, Lung, and Blood Institute. for Hypertension Management 1999: Summary." British Medical Neaton, J. D., and D. Wentworth 1992. "Serum Cholesterol, Blood Journal 319 (7210): 630­35. Pressure, Cigarette Smoking, and Death from Coronary Heart Disease: Rose, G. 1981. "Strategy of Prevention: Lessons from Cardiovascular Overall Findings and Differences by Age for 316,099 White Men-- Disease." British Medical Journal 282: 1847­51. Multiple Risk Factor Intervention Trial Research Group." Archives of ------. 1985. "Sick Individuals and Sick Populations." International Internal Medicine 152 (1): 56­64. Journal of Epidemiology 14: 32­38. NHLBI (National Heart, Lung, and Blood Institute) Obesity Education Salmond, C. E., J. G. Joseph, I. A. Prior, D. G. Stanley, and A. F. Wessen. Initiative Expert Panel 1998. Clinical Guidelines on the Identification, 1985. "Longitudinal Analysis of the Relationship between Blood Evaluation, and Treatment of Overweight and Obesity in Adults. Pressure and Migration: The Tokelau Island Migrant Study." American Bethesda, MD: National Institutes of Health, NHLBI. Journal of Epidemiology 122 (2): 291­301. Nissinen, A., X. Berrios, and P. Puska. 2001. "Community-Based Salmond, C. E., I. A. Prior, and A. F. Wessen. 1989."Blood Pressure Patterns Noncommunicable Disease Interventions: Lessons from Developed and Migration: A 14-Year Cohort Study of Adult Tokelauans." Countries for Developing Ones." Bulletin of the World Health American Journal of Epidemiology 130 (1): 37­52. Organization 79 (10): 963­70. Schooler, C., J. W. Farquhar, S. P. Fortmann, and J. A. Flora. 1997. O'Meara, S., R. Riemsma, L. Shirran, L. Mather, and G. ter Riet. 2001. "A "Synthesis of Findings and Issues from Community Prevention Trials." Rapid and Systematic Review of the Clinical Effectiveness and Cost- Annals of Epidemiology 7 (Suppl.): S54­68. Effectiveness of Orlistat in the Management of Obesity." Health Serdula, M., A. Mokad, D. Williamson, D. Galuska, J. Mendlein, and G. Technology Assessment (Winchester, U.K.) 5 (18): 1­81. Heath. 1999. "Prevalence of Attempting Weight Loss and Strategies for ------. 2002. "The Clinical Effectiveness and Cost-Effectiveness of Controlling Weight." Journal of the American Medical Association 282 Sibutramine in the Management of Obesity: A Technology (14): 1353­58. Assessment." Health Technology Assessment (Winchester, U.K.) 6 (6): Sleight, P. 1997a. "Lowering of Blood Pressure and Artery Stiffness." Lancet 1­97. 349 (9048): 362. Pepine, C. J., E. M. Handberg, R. M. Cooper-DeHoff, R. G. Marks, P. ------. 1997b. "Lowering of Blood Pressure and Artery Stiffness." Lancet Kowey, F. H. Messerli, and others. 2003. "A Calcium Antagonist vs. a 349 (9056): 955­56. Non-Calcium Antagonist Hypertension Treatment Strategy for Staessen, J., R. Fagard, L. Thijs, H. Celis, G. Arabidze, W. Birkenhager, and Patients with Coronary Artery Disease--The International Verapamil- others. 1997. "Randomised Double-Blind Comparison of Placebo and Trandolapril Study (INVEST): A Randomized Controlled Trial." Active Treatment for Older Patients with Isolated Systolic Journal of the American Medical Association 290 (21): 2805­16. Hypertension." Lancet 350: 757­64. Pestana, J. A., K. Steyn, A. Leiman, and G. M. Hartzenberg. 1996. "The Stevens, J., J. Cai, E. R. Pamuk, D. F. Williamson, M. J. Thun, and J. L. Direct and Indirect Costs of Cardiovascular Disease in South Africa in Wood. 1998."The Effect of Age on the Association between Body-Mass 1991." South African Medical Journal 86 (6): 679­84. Index and Mortality." New England Journal of Medicine 338 (1): 1­7. The Growing Burden of Risk from High Blood Pressure, Cholesterol, and Bodyweight | 867 Stewart, I. M. 1979. "Relation of Reduction in Pressure to First Myocardial in the Treatment of Hypertension in Older Persons: A Randomized Infarction in Patients Receiving Treatment for Severe Hypertension." Controlled Trial of Nonpharmacologic Interventions in the Elderly Lancet 1 (8121): 861­65. (TONE): TONE Collaborative Research Group." Journal of the Suh, I. 2001. "Cardiovascular Mortality in Korea: A Country Experiencing American Medical Association 279 (11): 839­46. Epidemiologic Transition." Acta Cardiologica 56 (2): 75­81. WHO (World Health Organization). 2001. Macroeconomics and Health: Susser, M. 1995. "The Tribulations of Trials--Intervention in Investing in Health for Economic Development--Report of the Communities." American Journal of Public Health 85 (2): 156­58. Commission on Macroeconomics and Health. Geneva: WHO. http://www.cmhealth.org/. Swinburn, B., T. Ashton, J. Gillespie, B. Cox, A. Menon, D. Simmons, and others. 1997. "Health Care Costs of Obesity in New Zealand." ------. 2002. World Health Report 2002: Reducing Risks, Promoting International Journal of Obesity and Related Metabolic Disorders: Healthy Life. Geneva: WHO. Journal of the International Association for the Study of Obesity 21: ------. 2003a. Adherence to Long-Term Therapies: Evidence for Action. 891­96. Geneva: WHO. Tosteson, A. N. A., M. C. Weinstein, M. G. M. Hunink, M. A. Mittleman, ------. 2003b. Diet, Nutrition, and the Prevention of Chronic Diseases. L. W. Williams, P. A. Goldman, and others. 1997. "Cost-Effectiveness of Geneva: WHO. Populationwide Educational Approaches to Reduce Serum Cholesterol Willett, W. C., J. E. Manson, M. J. Stampfer, G. A. Colditz, B. Rosner, F. E. Levels." Circulation 95: 24­30. Speizer, and others. 1995. "Weight, Weight Change, and Coronary Troiano, R. P., E. A. Frongillo, J. Sobal, and D. A. Levitsky. 1996. "The Heart Disease in Women: Risk within the `Normal' Weight Range." Relationship between Body Weight and Mortality: A Quantitative Journal of the American Medical Association 273 (6): 461­65. Analysis of Combined Information from Existing Studies." Wood, D., G. De Backer, O. Faergeman, I. Graham, G. Mancia, K. Pyorala, International Journal of Obesity and Related Metabolic Disorders 20: and others. 1998. "Prevention of Coronary Heart Disease in Clinical 63­75. Practice: Summary of Recommendations of the Second Joint Task Uusitalo, U., E. J. Feskens, J. Tuomilehto, G. Dowse, U. Haw, D. Fareed, and Force of European and other Societies on Coronary Prevention." others. 1996. "Fall in Total Cholesterol Concentration over Five Years Journal of Hypertension 16: 1404­14. in Association with Changes in Fatty Acid Composition of Cooking Wu, X., Z. Huang, J. Stamler, Y. Wu, Y. Li, A. R. Folsom, and others. 1996. Oil in Mauritius: Cross-Sectional Survey." British Medical Journal 313 "Changes in Average Blood Pressure and Incidence of High Blood (7064): 1044­46. Pressure 1983­1984 to 1987­1988 in Four Population Cohorts in the Wald, N. J., and M. R. Law 2003. "A Strategy to Reduce Cardiovascular People's Republic of China: The PRC-USA Cardiovascular and Disease by More Than 80 Percent." British Medical Journal 326 (7404): Cardiopulmonary Epidemiology Research Group." Journal of 1419. Hypertension 14 (11): 1267­74. Wang, G., Z. J. Zheng, G. Heath, C. Macera, M. Pratt, and D. Buchner. Xact Medicare Services. 2003. Medicare Clinical Laboratory Fee Schedule. 2002. "Economic Burden of Cardiovascular Disease Associated with Camp Hill, PA: Xact Medicare Services. Excess Body Weight in U.S. Adults." American Journal of Preventive Yusuf, S. 2002. "Two Decades of Progress in Preventing Vascular Disease." Medicine 23: 1­6. Lancet 360 (9326): 2­3. Wang, L. Y., Q. Yang, R. Lowry, and H. Wechsler. 2003. "Economic Analysis of a School-Based Obesity Prevention Program." Obesity Research 11: 1313­24. Whelton, P. K., L. J. Appel, M. A. Espeland, W. B. Applegate, W. H. Ettinger Jr., J. B. Kostis, and others. 1998. "Sodium Reduction and Weight Loss 868 | Disease Control Priorities in Developing Countries | Anthony Rodgers, Carlene M. M. Lawes, Thomas Gaziano, and others Chapter 46 Tobacco Addiction Prabhat Jha, Frank J. Chaloupka, James Moore, Vendhan Gajalakshmi, Prakash C. Gupta, Richard Peck, Samira Asma, and Witold Zatonski Cigarette smoking and other forms of tobacco use impose a on the uniquely addictive properties of nicotine. A review of the large and growing global public health burden. Worldwide, effectiveness of tobacco-control policies in reducing tobacco ini- tobacco use is estimated to kill about 5 million people annually, tiation and in increasing cessation follows. A cost-effectiveness accounting for 1 in every 5 male deaths and 1 in 20 female analysis of these interventions is provided. Finally, the con- deaths of those over age 30. On current smoking patterns, straints to implementing tobacco-control policies are discussed. annual tobacco deaths will rise to 10 million by 2030. The 21st century is likely to see 1 billion tobacco deaths, most of them SMOKING TRENDS in low-income countries. In contrast, the 20th century saw 100 million tobacco deaths, most of them in Western countries Tobacco use, in both smoked and nonsmoked forms, is com- and the former socialist economies. mon worldwide. This chapter focuses on smoked tobacco, Hundreds of millions of premature tobacco deaths could chiefly cigarettes and bidis (tobacco hand rolled in the leaf of be avoided if effective interventions were widely applied in another plant, temburi, which is popular in India and parts of low- and middle-income countries. Numerous studies from Southeast Asia), because smoked tobacco is more common-- high-income countries and a growing number from low- and accounting for about 65 to 85 percent of all tobacco produced middle-income countries provide robust evidence that tobacco worldwide (WHO 1997)--and causes more disease and more tax increases, timely dissemination of information about the diverse types of disease than does oral tobacco use. health risks of smoking, restrictions on smoking in public and workplaces, comprehensive bans on advertising and promo- Prevalence tion, and increased access to cessation therapies are effective in A systematic review of 139 studies on adult smoking prevalence reducing tobacco use and its consequences. Cessation by the (Jha and others 2002) found that more than 1.1 billion people 1.1 billion current smokers is central to meaningful reductions worldwide smoke, with about 82 percent of smokers residing in tobacco deaths over the next five decades. New analyses pre- in low- and middle-income countries. Table 46.1 provides an sented here find that higher tobacco taxes could prevent 3 mil- update of these estimates for the population in 2000. Globally, lion tobacco deaths by 2030 among smokers alive today. male smoking far exceeds female smoking, with a smaller gen- Reduced uptake of smoking by children would yield benefits der difference in high-income countries. Smoking prevalence is chiefly after 2050. Price and non-price interventions are, for the highest in Europe and Central Asia, where 35 percent of all most part, highly cost-effective. adults are smokers. This chapter begins with an overview of smoking trends and While overall smoking prevalence continues to increase in tobacco's health consequences, followed by a discussion of the many low- and middle-income countries, many high-income economic rationale for government intervention, with a focus countries have witnessed decreases, most clearly in men. A 869 Table 46.1 Estimated Smoking Prevalence (by Gender) and Number of Smokers, 15 Years of Age and Older, 2000 Smoking prevalence (percent) Total smokers World Bank region Males Females Overall Millions Percentage of all smokers East Asia and the Pacific 63 5 34 429 38 Europe and Central Asia 56 17 35 122 11 Latin America and the Caribbean 40 24 32 98 9 Middle East and North Africa 36 5 21 37 3 South Asia 32 6 20 178 15 Sub-Saharan Africa 29 8 18 56 6 Low- and middle-income economies 49 8 29 920 82 High-income economies 37 21 29 202 18 Source: Authors. study in 36 mostly Western countries, from early 1980 to the health and population policy. Thus, the salient aspects of mid 1990s, suggested that the decrease in smoking prevalence tobacco epidemiology are outlined in this section. observed among men was caused by the higher prevalence in younger age groups of those who have never smoked. Among Key Messages for the Individual Smoker women, there was little overall change in smoking prevalence More than 50 years of epidemiology on smoking-related dis- because the increasing prevalence of smokers in younger eases have led to three key messages for individual smokers cohorts counterbalanced increasing cessation in older age worldwide (Doll and others 2004; Peto and others 2003). groups (Molarius and others 2001). · The eventual risk of death from smoking is high, with about Cessation one-half to two-thirds of long-term smokers eventually Ex-smoking rates are a good measure of cessation at a popula- being killed by their addiction. tion level. In some high-income countries, the prevalence rates · These deaths involve a substantial number of life years for- of ex-smokers have increased over the past two to three gone. About half of all tobacco deaths occur at ages 35 to 69, decades. For example, in the United Kingdom, smoking preva- resulting in the loss of about 20 to 25 years of life, compared lence among males over age 30 fell from 70 percent in the 1950s with the life expectancy of nonsmokers. to 30 percent in 2000; female smoking prevalence fell from 40 · Cessation works: those adults who quit before middle age to 20 percent over the same period. Much of the decrease arose avoid almost all the excess hazards of continued smoking. from cessation. Today, two times as many ex-smokers as smok- ers exist among those age 50 or over. Currently, 30 percent of Worldwide, about 80 percent of deaths among the 2.7 bil- the U.K. male population is made up of former smokers (Peto lion adults over age 30 involve vascular, respiratory, or neoplas- and others 2000). Polish male cessation rates have also tic disease. Smoking is associated with an increase in the increased, partly because of control programs. One of every frequency of many of these diseases, although important dif- four adult Polish males described himself as an ex-smoker ferences exist between and across populations. The following (Zatonski and Jha 2000). In contrast, the prevalence of male ex- discussion focuses on the consequences of smoking on adult smokers in most developing countries is low: 10 percent in mortality. Detailed epidemiological reviews of worldwide mor- Vietnam, 5 percent in India, and 2 percent in China (Jha and tality from smoking are found elsewhere (C. Gajalakshmi and others 2002). Even those low figures may be falsely elevated others 2000; V. Gajalakshmi and others 2003; Gupta and Mehta because they include people who quit because either they were 2000; Liu and others 1998; Niu and others 1998; Peto and too ill to continue or they had early symptoms of tobacco- others 1994). related illness (Martinson and others 2003). Current Mortality and Disability from Smoking HEALTH CONSEQUENCES OF SMOKING Recent updates of indirect estimates of global tobacco mortality (Ezzati and Lopez 2003; M. Ezzati, personal communication, The health consequences of smoking are often assumed to be November 2004) indicate that in 2000, 5.0 million premature widely understood. In fact, ignorance of the magnitude of deaths were caused by tobacco. About half (2.6 million) of tobacco hazards is widespread in terms of both individual those deaths were in low-income countries. Males accounted 870 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others Table 46.2 Tobacco Mortality and Total DALYs by Gender, 2000 (thousands) Tobacco deaths Total DALYs World Bank region Males Females Males Females East Asia and the Pacific 829 274 13,116 4,128 Europe and Central Asia 754 161 12,407 2,686 Latin America and the Caribbean 177 97 2,789 1,613 Middle East and North Africa 97 28 1,676 554 South Asia 768 187 12,397 3,285 Sub-Saharan Africa 105 66 1,659 1,091 Low- and middle-income economies 2,730 813 44,044 13,357 High-income economies 929 548 12,304 6,866 World 3,659 1,361 56,347 20,222 Source: Ezzati and Lopez 2003; Mathers and others 2006. Note: The terms high-income and former socialist economies as used in the text correspond roughly to high-income and Europe and Central Asia regions using the World Bank classification. Low-income countries corresponds roughly to East Asia and the Pacific, Latin America and the Caribbean, Middle East and North Africa, South Asia, and Sub-Saharan Africa. for 3.7 million deaths, or 72 percent of all tobacco deaths. 1990. Similarly, a large increase in female lung cancer at young About 60 percent of male and 40 percent of female tobacco ages was avoided in the United Kingdom, but female lung cancer deaths were of middle-aged persons (ages 35 to 69). at young ages continues to rise in France. In high-income countries and former socialist economies, Future increases in tobacco deaths worldwide are expected the 1 million middle-aged male tobacco deaths were largely to arise from increased smoking by males in developing coun- composed of cardiovascular disease (0.45 million) and lung tries and by women worldwide. Such increases are a product of cancer (0.21 million). In contrast, in low-income countries, the population growth and increased age-specific tobacco mortali- leading causes of death among the 1.3 million male tobacco ty rates, the latter relating to both smoking duration and the deaths were cardiovascular disease (0.4 million), chronic amount of tobacco smoked. Peto and others (1994) have made obstructive pulmonary disease (0.2 million), other respiratory the following calculation: if the proportion of young people disease (chiefly tuberculosis, 0.2 million), and lung cancer taking up smoking continues to be about half of men and one- (0.18 million). The specific numbers of deaths from tobacco tenth of young women, there will be about 30 million new and of total disability-adjusted life years (DALYs) by gender long-term smokers each year. As previously noted, epidemio- and World Bank region are shown in table 46.2. Disability esti- logical studies in developed and developing countries suggest mates are not discussed here; however, disability is highly cor- that half of these smokers will eventually die from smoking. related with mortality in most settings. However, if we conservatively assume that "only" about one- third of smokers die as a result of smoking, then smoking will Past and Future Trends in Mortality eventually kill about 10 million people a year. Thus, for the In high-income and former socialist economies with more 25-year period from 2000 to 2025, there would be about 150 complete and reliable mortality statistics, one can measure million tobacco deaths, or about 6 million deaths per year on the effects of increased smoking prevalence and subsequent average; from 2025 to 2050, there would be about 300 million decreases that have been observed among large numbers of tobacco deaths, or about 12 million deaths per year. adults. These changes are best documented by examining lung Further estimations are more uncertain, but current smok- cancer mortality rates among young adults because lung cancer ing trends and projected population growth indicate that from is not often misclassified with other causes of death at young 2050 to 2100 there will be an additional 500 million tobacco ages and it is almost entirely attributable to smoking. deaths. These projections for the next three to four decades are comparable to retrospective and early prospective epidemiolog- Age-Standardized Lung Cancer Mortality Rates ical studies in China (Liu and others 1998; Niu and others 1998), Age-standardized male lung cancer rates at ages 35 to 44 per which suggest that annual tobacco deaths will rise to 1 million 100,000menintheUnitedKingdomhadfallenfrom18in1950to before 2010 and to 2 million by 2025, when the young adult 4 by 2000. In contrast, comparable French male lung cancer rates smokers of today reach old age. Similarly, results from a large show the reverse pattern (Peto and others 2003; figure 46.1). In retrospective study in India suggest that 1 million annual deaths France,the increase in smoking occurred some decades later than can be expected from male smokers by 2025 (V. Gajalakshmi in the United Kingdom,and declines in smoking began only after and others 2003). With other populations in Asia, Eastern Tobacco Addiction | 871 a. United Kingdom b. France Death rate/100,000 men, age standardizeda Death rate/100,000 men, age standardizeda 18 18 16 16 14 Males 14 Males 12 12 10 10 8 8 6 6 4 Females 4 Females 2 2 0 0 1950 1960 1970 1980 1990 2000 1950 1960 1970 1980 1990 2000 Source: Peto and others 2003. a. Mean of annual rates in component five-year age groups (35­39, 40­44). Figure 46.1 Changes in Lung Cancer Mortality at Age 35 to 44 in the United Kingdom and France, 1950­99 Europe, Latin America, the Middle East, and (less certainly) Tobacco deaths (million) Sub-Saharan Africa showing similar growth in population and 520 500 500 age-specific tobacco death rates, the estimate of some 450 mil- lion tobacco deaths over the next five decades appears plausible. Almost all of these deaths will be among current smokers. 400 340 300 Benefits of Cessation 220 Current tobacco mortality statistics reflect past smoking behav- 200 ior, given the long delay between the onset of smoking and 190 the development of disease. The prevention of a substantial proportion of these tobacco deaths before 2050 requires adult 100 cessation. For example, halving the per capita adult consump- 70 tion of tobacco by 2020 (akin to the declines in adult smoking 0 in the United Kingdom) would avert about 180 million tobacco 1950 2000 2025 2050 deaths. Continuing to reduce the percentage of children who Baseline If proportion of young adults start to smoke will prevent many deaths, but its main effect will If adult consumption taking up smoking halves by 2020 be on mortality rates in 2050 and beyond (figure 46.2; Jha and halves by 2020 Chaloupka 2000a; Peto and Lopez 2001). Source: Jha and Chaloupka 2000a; Peto and Lopez 2001. Substantial evidence indicates that smoking cessation Note: Peto and others (1994) estimate 60 million tobacco deaths between 1950 and 2000 reduces the risk of death from tobacco-related diseases. Among in industrial countries. This figure estimates an additional 10 million tobacco deaths between 1990 and 2000 in developing countries. The figure also assumes no tobacco doctors in the United Kingdom, those who quit smoking before deaths before 1990 in developing countries and minimal tobacco deaths worldwide before 1950. Projections for deaths from 2000 to 2050 are based on Peto and Lopez (2001). the onset of major disease avoided most of the excess hazards of smoking (Doll and others 2004). The benefits of quitting Figure 46.2 Tobacco Deaths in the Next 50 Years under Current were largest in those who quit before middle age (between ages Smoking Patterns 25 and 34 years) but were still significant in those who quit later (between ages 45 and 54 years). have never smoked. In the United Kingdom, among those Cessation before middle age avoids more than 90 percent who stopped smoking, the risk of lung cancer fell steeply with of the lung cancer risk attributable to tobacco, with quitters time since cessation. For men who stopped at ages 60, 50, 40, possessing a pattern of survival similar to that of persons who and 30, the cumulative risks of lung cancer by age 75 were 872 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others Lung cancer mortality (percent) · The addictive nature of tobacco is underappreciated and poorly understood. Although general awareness of risks is Continued 15 smoking better in high-income countries, many people still underes- timate tobacco's danger relative to other health risks, and many smokers fail to fully internalize these risks (Weinstein 1998). 10 · Smokers may impose costs on others from passive tobacco smoke or, more controversially, from higher health care costs (Lightwood and others 2000; Warner 2003). Stopped age 50 5 The reader is referred to more detailed discussions on the welfare economics of tobacco (Barnum 1994; Jha and others Stopped age 30 2000; Peck and others 2000; Warner and others 1995; and Never smoked several background papers in the Disease Control Priorities 0 Project Working Paper Series). We discuss nicotine addiction 45 55 65 75 because this newer evidence has profound implications for Age explaining smoking behavior and for devising control policies. Source: Peto and others 2000. Figure 46.3 Stopping Works: Cumulative Risk of Lung Cancer Nicotine Addiction Mortality in U.K. Males, 1990 rates Before the landmark 1988 U.S. Surgeon General's report, which suggested that cigarettes and other forms of tobacco are addic- tive and that nicotine is the major agent in tobacco responsible 10 percent, 6 percent, 3 percent, and 2 percent, respectively for addiction, the prevailing view was that tobacco use was (Peto and others 2000; figure 46.3). These results have been largely a voluntary behavior or personal choice (Koop 2003). supported by a recent multicenter study of men in four Since that time, clinicians, behavioral scientists, researchers, European countries; for men who quit smoking at age 40, the and public health experts have increasingly recognized manu- study found that the excess lung cancer risk avoided was factured tobacco products as some of the most addictive and 85 percent, 91 percent, and 80 percent in the United Kingdom, deadly dependence-producing substances available. Although Germany, and Italy, respectively (Crispo and others 2004). numerous factors have been identified that can contribute to Smoking cessation is uncommon in most developing countries, the reinforcement of the smoking habit--for example, the syn- but some evidence exists that, among Chinese men, quitting ergistic and independent effects of other compounds in tobacco also reduces the risks of total and vascular mortality (Lam and smoke (such as tar and acetaldehyde) or the sensory and envi- others 2002). ronmental stimuli associated with smoking (such as tobacco advertising)--little debate exists that nicotine is a significant contributor to the development and maintenance of the smok- RATIONALE FOR GOVERNMENT INTERVENTION ing habit (Markou and Henningfield 2003). In most aspects of dependence, nicotine is on par with other powerfully addictive In addition to the public health burden caused by tobacco, drugs, such as heroin and cocaine. Newer evidence has con- an economic rationale exists for government to intervene to verged on the following key points. reduce tobacco use: · Consumers have inadequate information about the health Biological Aspects. Nicotine is a psychoactive drug that trig- consequences of tobacco use (Jha and others 2000; Warner gers a cascade of neurobiological events in the reward areas of and others 1995). Specifically, the decision to initiate smok- the brain and throughout the body that can, in turn, act in con- ing is made primarily by youths, whose ability to make fully cert to reinforce tobacco use (Markou and Henningfield 2003). informed, appropriately forward-looking decisions is Even a short-term exposure to nicotine has been shown to questioned by society in many different contexts (minimum induce long-lasting changes of the excitatory input into the ages for drinking, driving, and voting, for instance). In brain's reward system, which may be an important early step in industrial countries, about 80 percent of adult smokers the path to addiction (Laviolette and van der Kooy 2004). begin smoking before age 20. Even if children and young Notably, in some experimental models, if nicotine's neurobio- adults have information on future risks, they tend to dis- logical effects are blocked pharmacologically, or if nicotine is count that future risk greatly. removed from cigarette smoke, smoking eventually ceases Tobacco Addiction | 873 (Jarvis 2004). The overwhelming property of nicotine that smoking choices, the net economic costs of tobacco are pro- leads to its frequent use is the occurrence of nicotine with- foundly negative (Barnum 1994; Peck and others 2000). While drawal, for which cigarette smoke provides rapid relief. Though some of the methods for such costing have been disputed, each individual differs greatly in his or her sensitivity to nico- newer economic evidence supports the idea that widespread tine dependence, evidence suggests that most adults are sus- hazards of tobacco use lead to major economic costs. Jamison, ceptible to the biological effects of nicotine and tobacco Lau, and Wang (2005) have outlined that male survival explains (Picciotto 2003). income growth independent of changes in physical capital, education, fertility, economic openness, and technical progress. Psychological Aspects. In addition to the unique neurobiolo- Thus, if adult male survival in the former socialist economies gy of nicotine, the ready availability of tobacco influences the of Europe had been that of high-income countries, annual uptake of smoking as well as the development and mainte- growth rates over the past three decades would have been about nance of dependence. With illicit drugs, legal and social barri- 1.4 percent rather than 1 percent, making 1990 per capita ers constantly test a user's drive to consume the drug. In income about 12 percent higher, or an absolute value of contrast, a smoker is presented with nearly ubiquitous US$140 billion. The chief determinant of the mortality gap opportunities and frequent cues to both purchase and use between the former socialist economies and high-income tobacco because of mass marketing and promotion of tobacco countries from 1960 to 1990 is smoking (Peto and others 1994; (Shiffman and West 2003). Young people, who are attracted Zatonski and Jha 2000). More recent economic studies that to many risk behaviors, such as fast driving or binge drinking, have put a value on "statistical life" suggest that smoking cessa- do not "learn" from early smoking in the way that most young tion generates huge benefits. For example, Murphy and Topel people become safer drivers and moderate drinkers as adults (2003) find that in the United States, the value of reduced mor- (Jha and others 2000; O'Malley, Bachman, and Johnston 1988). tality from all causes between 1970 and 1998 amounted to US$2.6 trillion per year, or half of gross domestic product Economics. The traditional economic formulation of costs (GDP) growth during the period. Fully US$1.1 trillion per year and benefits tends not to take into account the unique proper- arose from reduced heart disease, of which at least one-third ties of addiction (see Chaloupka, Tauras, and Grossman 2000 was from reduced smoking and saturated fat in diets (Cutler for a review). Newer models have begun to incorporate factors and Kadiyala 2003; see chapter 15 for a fuller discussion on the such as lack of information, regret, and addiction itself. One economic benefits of disease control). key innovation by Gruber and Koszegi (2001, 2002) permits smokers to be time inconsistent, meaning that, given prefer- INTERVENTIONS TO REDUCE DEMAND ences, smokers would make different decisions at different FOR TOBACCO points in time. This approach, now widely used within the new field of behavioral economics, admits conflict between what Numerous studies, mostly from high-income countries, have smokers would like for themselves today and what they would examined the effect of interventions aimed at reducing the like for themselves in the future. demand for tobacco products on smoking and other kinds of tobacco use. The small but growing number of studies from Implications for Control Programs. These newer economic low- and middle-income countries provide useful lessons models have several implications for control programs. First is about differences in the effect of these interventions between the need for much more aggressive tobacco taxation to deter these countries and high-income countries. The following is a the development of tobacco smoking. Estimates suggest that, review of the effect of price and non-price interventions in in the United States, optimal taxation taking into account reducing demand for smoking, including a discussion of each smoking initiation and nicotine addiction would be at least intervention's effect on initiation and cessation. A more com- US$1 higher per pack (Gruber 2003; Gruber and Koszegi 2002; plete study of the effectiveness of various interventions is avail- Gruber and Mullainathan 2002). The second implication is able elsewhere (Jha and Chaloupka 2000b). that the usual assumption that higher taxes reduce the welfare or satisfaction of continuing smokers may not be true. Higher Tobacco Taxation taxes enhance welfare by acting like an external control device Nearly all governments tax tobacco products. However, signifi- over the time-inconsistent preferences of smokers, which cant differences exist across countries in levels of tobacco taxa- would reduce the likelihood of smoking initiation. tion. Some of these taxes are specific or per unit taxes; others The third implication is that the overall economic benefits are expressed as a percentage of wholesale or retail prices (ad of tobacco control, taking into account addiction, are likely to valorem taxes). As illustrated in figure 46.4, taxes tend to be be substantially positive. Earlier cost-benefit analyses have absolutely higher and account for a greater share of the retail shown that if even modest costs are assigned to uninformed price (two-thirds or more) in high-income countries. In 874 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others Average price or tax per pack (US$) Percentage of tax share to 13 percent shorter smoking duration or a 3.4 percent higher 3.5 80 probability of cessation. Many recent studies from the United States have used 3.0 70 individual-level data to explore differences in the price elastic- 60 ity of cigarette demand by age, with a particular emphasis on 2.5 youth and young adults (Chaloupka, Hu, and others 2000; U.S. 50 2.0 DHHS 2000). Given that most smoking behavior becomes 40 firmly established during teenage years and young adulthood, 1.5 30 interventions that are effective in preventing smoking initiation and the transition to regular, addicted smoking will have sig- 1.0 20 nificant long-term public health benefits. Estimates from these 0.5 10 recent studies conclude that an inverse relationship exists between price elasticity and age, with estimates for youth price 0 0 High-income Upper-middle- Lower-middle- Low-income elasticity of demand up to three times those obtained for adults countries income income countries (Gruber 2003; Ross, Chaloupka, and Wakefield 2001). Several countries countries recent studies have begun to explore the differential effect of Average price Percentage cigarette prices on youth smoking uptake, concluding that Average tax of tax share higher cigarette prices are particularly effective in preventing young smokers from moving beyond experimentation into Source: Authors. regular, addicted smoking (Emery, White, and Pierce 2001; Figure 46.4 Average Cigarette Price, Tax, and Percentage of Tax Ross, Chaloupka, and Wakefield 2001). Share per Pack, by Income Group, 1996 In the United Kingdom and the United States, increases in the price of cigarettes have had the greatest effect on smoking low- and middle-income countries, taxes are generally much among the lowest-income and least educated populations lower and account for less than half of the final price of ciga- (CDC 1994; Townsend, Roderick, and Cooper 1998). rettes. In the United States, federal and state excise taxes on cig- Furthermore, it was estimated that smokers in U.S. households arettes were one-third lower, in real terms, in 1995 than their below median income level are four times more responsive to peak level of the mid 1960s. However, taxes rose sharply over the price increases than smokers in households above median next eight years and stood at US$1.12 per pack as of 2002. income level. In general, estimates of price elasticity for low- Well over 100 studies from high-income countries clearly and middle-income countries are about double those esti- demonstrate that increases in taxes on cigarettes and other mated for high-income countries, implying that significant tobacco products lead to significant reductions in cigarette increases in tobacco taxes in these countries would be effective smoking and other tobacco use (Chaloupka, Hu, and others in reducing tobacco use. 2000). These reductions reflect the combination of increased smoking cessation, reduced relapse, lower smoking initiation, and decreased consumption among continuing tobacco users. Restrictions on Smoking Studies from Canada, the United Kingdom, the United Over the past three decades, as the quantity and quality of States, and many other high-income countries generally esti- information about the health consequences of exposure to mate that the price elasticity of cigarette demand ranges from passive smoking have increased, many governments, especially 0.25 to 0.50, indicating that a 10 percent increase in in high-income countries, have enacted legislation restricting cigarette prices will reduce overall cigarette smoking by 2.5 to smoking in a variety of public places and private worksites. In 5.0 percent (Chaloupka, Hu, and others 2000; U.S. DHHS addition, increased awareness of the consequences of passive 2000). Estimates from a limited number of studies from low- smoke exposure, particularly to children, has led many work- and middle-income countries suggest a greater price elasticity places and households to adopt voluntary restrictions on of 0.8 in such countries. Recent studies using survey data smoking. Although the intent of those restrictions is to reduce have concluded that half or more of the effect of price on over- nonsmokers' exposure to passive tobacco smoke, the policies all cigarette demand results from reducing the number of cur- also reduce smokers' opportunities to smoke. Additional rent smokers (CDC 1994; Wasserman and others 1991). Higher reductions in smoking, especially among youths, will result taxes increase both the number of attempts at quitting smok- from the changes in social norms that are introduced by adopt- ing and the success of those attempts (Tauras 1999; Tauras and ing these policies (U.S. DHHS 1994). Chaloupka 2003). A study in the United States (Taurus 1999) In Western populations, comprehensive restrictions on suggested that a 10 percent increase in price would result in 11 cigarette smoking have been estimated to reduce population Tobacco Addiction | 875 smoking rates by 5 to 15 percent (see review by Woolery, Asma, spent US$12.5 billion on advertising and promotion in the and Sharp 2000) and can also lead to changes in social norms United States alone, the highest spending level reported to date regarding smoking behavior, especially among youths. As with (U.S. Federal Trade Commission 2004). Tobacco advertising higher taxes, these restrictions reduce both the prevalence of efforts worldwide include traditional forms of advertising on smoking and cigarette consumption among current smokers. television, radio, and billboards and in magazines and newspa- Smoking bans in workplaces generally reduce the quantity of pers as well as favorable product placement; price-related cigarettes smoked by 5 to 25 percent and reduce prevalence promotions, such as coupons and multipack discounts; and rates by up to 20 percent (Levy, Friend, and Polishchuk 2001). sponsorship of highly visible sporting and cultural events. No-smoking policies were most effective when strong social Numerous econometric studies, largely from the United norms against smoking helped make smoking restrictions self- Kingdom and the United States, have explored the relationship enforcing. between cigarette advertising and promotional expenditure and cigarette demand. In general, these studies have resulted in Health Information and Counteradvertising mixed findings, with most studies concluding that advertising The 1962 report by the British Royal College of Physicians and has a small positive effect on demand (Chaloupka, Hu, and the 1964 U.S. Surgeon General's Report were landmark tobacco- others 2000; Townsend 1993). However, critics of these studies control events in high-income countries. These publications note that econometric methods, which estimate the effect of a resulted in the first widespread press coverage of the scientific marginal change in advertising expenditures on smoking, are links between smoking and lung cancer. The reports were fol- ill suited for studying the effect of advertising (Chaloupka, lowed, in many countries, by policies requiring health warning Hu, and others 2000; U.S. Federal Trade Commission 2004; labels on tobacco products, which were later extended to Townsend 1993). Approaches used by other disciplines, includ- tobacco advertising. ing survey research and experiments that assess reactions to Research from high-income countries indicates that these and recall of cigarette advertising, do support the hypothesis initial reports and the publicity that followed about the health that increases in cigarette advertising and promotion directly consequences of smoking led to significant reductions in con- and indirectly increase cigarette demand and smoking initia- sumption, with initial declines of between 4 and 9 percent and tion (U.S. DHHS 1994; U.K. Department of Health 1992). longer-term cumulative declines of 15 to 30 percent (Kenkel These studies conclude that cigarette advertising is effective in and Chen 2000; Townsend 1993). Efforts to disseminate infor- getting and retaining children's attention, with the strength mation about the risks of smoking and of other tobacco use in of the association strongly correlated with current smoking low- and middle-income countries have led to similar declines behavior, smoking initiation, and smoking intentions. in tobacco use in those countries (Kenkel and Chen 2000). In Comprehensive advertising and promotion bans on ciga- addition, mass media antismoking campaigns, in many cases rettes provide more direct evidence on the effect of advertising funded by earmarked tobacco taxes, have generated reductions these products (Saffer 2000). One study using data from in cigarette smoking and other tobacco use (Kenkel and Chen 22 high-income countries for the period 1970 through 1992 2000; Saffer 2000). Decreases in smoking prevalence were provides strong evidence that comprehensive bans on cigarette largest in Western countries, where the public is constantly and advertising and promotion led to significant reductions in cig- consistently reminded of the dangers of smoking by extensive arette smoking. The study predicts that a comprehensive set coverage of issues related to tobacco in the news media of tobacco advertising bans in high-income countries could (Molarius and others 2001). reduce tobacco consumption by more than 6 percent, taking In many low- and middle-income countries, a lack of aware- into account price and non-price control interventions (Saffer ness continues to exist about the risks of mortality and disease and Chaloupka 2000). However, the study concludes that par- posed by smoking. For example, a national survey in China in tial bans have little effect on smoking behavior, given that the 1996 found that 61 percent of smokers thought that tobacco tobacco industry can shift its resources from banned media to did them "little or no harm" (Chinese Academy of Preventive other media that are not banned. Medicine 1997). In high-income countries, smokers are aware of the risks, but a recent review of psychological studies found that few smokers judge the size of these risks to be higher and Smoking Cessation Treatments more established than do nonsmokers, and that smokers min- Near-term reductions in smoking-related mortality depend imize the personal relevance of these risks (Weinstein 1998). heavily on smoking cessation. Numerous behavioral smoking cessation treatments are available, including self-help manuals, Bans on Advertising and Promotion community-based programs, and minimal or intensive clinical Cigarettes are among the most heavily advertised and pro- interventions (U.S. DHHS 2000). In clinical settings, moted products in the world. In 2002, cigarette companies pharmacological treatments, including nicotine replacement 876 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others therapies (NRT) and bupropion, have become much more ment of antismuggling laws, and stronger penalties for those widely available in recent years in high-income countries caught violating these laws (Joossens and others 2000). Recent through deregulation of some NRT from prescription to over- analysis suggests that, even in the presence of smuggling, the-counter status (Novotny and others 2000; U.S. DHHS tax increases will reduce consumption and increase revenue 2000). The evidence is strong and consistent that pharmaco- (Merriman, Yurekli, and Chaloupka 2000). logical treatments significantly improve the likelihood of quit- In contrast to the effectiveness of demand-side interven- ting, with success rates two to three times those when pharma- tions, there is much less evidence that interventions aimed at cological treatments are not used (Novotny and others 2000; reducing the supply of tobacco products are as effective in Raw, McNeill, and West 1999; U.S. DHHS 2000). The effective- reducing cigarette smoking (Jha and Chaloupka 1999, 2000a). ness of all commercially available NRT seems to be largely The U.S. experience provides mixed evidence about the effec- independent of the duration of therapy, the setting in which tiveness of limiting youth access to tobacco products in reduc- the therapy is provided, regulatory status (over-the-counter ing youth tobacco use (U.S. DHHS 2000; Woolery, Asma, and versus prescribed therapy), and the type of provider (Novotny Sharp 2000). In addition, the effective implementation and others 2000). Over-the-counter NRT without physician and enforcement of these policies may require infrastructure oversight have been used in many countries for a number of and resources that do not exist in many low- and middle- years with good success. income countries. A preliminary discussion is occurring in Although NRT are successful in treating nicotine addic- Canada about reducing its number of retail outlets for tobacco tion, the markets for NRT and other pharmacological thera- from the current 65,000. Neither the effect of such a move nor pies are more highly regulated and less affordable than are its enforcement costs are well known. Crop substitution and nicotine-containing tobacco products. Recent evidence indi- diversification programs are often proposed as a means of cates that the demand for NRT is related to economic factors, reducing the supply of tobacco. However, little evidence exists including price (Tauras and Chaloupka 2003). Policies that that such programs would significantly reduce the supply of decrease the cost of NRT and increase availability--such as tobacco, given that the incentives for growing tobacco tend to mandating private health insurance coverage of NRT, includ- attract new farmers who would replace those who abandon ing such coverage in public health insurance programs, and tobacco farming (Jacobs and others 2000). Similarly, direct subsidizing NRT for uninsured or underinsured individuals-- prohibition of tobacco production is not likely to be politically would likely lead to substantial increases in the use of these feasible, effective, or economically optimal. Finally, although products. Given the demonstrated efficacy of NRT in treating trade liberalization has contributed to increases in tobacco use smoking, these policies could generate significant increases in (particularly in low- and middle-income countries), restric- smoking cessation. tions on trade in tobacco and tobacco products that violate international trade agreements or draw retaliatory measures (or both) may be more harmful (Taylor and others 2000). INTERVENTIONS TO REDUCE THE SUPPLY OF TOBACCO The key intervention on the supply side is the control of smug- EFFECTIVENESS AND COST-EFFECTIVENESS gling. Recent estimates suggest that 6 to 8 percent of cigarettes OF TOBACCO-CONTROL INTERVENTIONS consumed globally are smuggled (Merriman, Yurekli, and Using a static model of the cohort of smokers alive in 2000, we Chaloupka 2000). Of note, the tobacco industry itself has an estimate the number of deaths attributable to smoking over the economic incentive to smuggle, in part to increase market share next few decades that could be averted by (a) price increases, and decrease tax rates (Joossens and others 2000; Merriman, (b) NRT, and (c) a package of non-price interventions other Yurekli, and Chaloupka 2000). Although differences in taxes than NRT. Cost-effectiveness of these policy interventions was and prices across countries create a motive for smuggling, a calculated by weighing the approximate public sector costs recent analysis comparing the degree of corruption in individ- against the years of healthy life saved, measured in DALYs. The ual countries with price and tax levels found that corruption details of an earlier version of this model have been published within countries is a stronger predictor of smuggling than is previously (Ranson and others 2002). price (Merriman, Yurekli, and Chaloupka 2000). Several gov- ernments are adopting policies aimed at controlling smuggling. In addition to harmonizing price differentials between coun- tries, effective measures include prominent tax stamps and Results of Model Projections warning labels in local languages, better methods for tracking The following is an updated analysis, using higher price cigarettes through the distribution chain, aggressive enforce- increases and a greater effectiveness for NRT than did the Tobacco Addiction | 877 Table 46.3 Reductions in Future Tobacco Deaths among Smokers Alive in 2000 from Price Increases of 10 Percent, 33 Percent, 50 Percent, and 70 Percent by World Bank Region Reduction in number of deaths (millions) 10 percent 33 percent 50 percent 70 percent Baseline price increase price increase price increase price increase smoking-attributable World Bank region deaths (millions) Low High Low High Low High Low High East Asia and the Pacific 173 2.9 8.7 9.6 27.5 14.5 37.5 20.3 46.2 (percent) (1.7) (5.0) (5.5) (15.9) (8.4) (21.7) (11.7) (26.8) Europe and Central Asia 51 0.9 2.6 2.8 8.1 4.3 11.2 6.0 13.8 (percent) (1.7) (5.1) (5.6) (16.0) (8.5) (22.0) (11.8) (27.2) Latin America and the Caribbean 40 0.7 2.1 2.3 6.7 3.5 9.5 4.9 11.6 (percent) (1.8) (5.3) (5.8) (16.8) (8.8) (23.7) (12.3) (29.1) Middle East and North Africa 13 0.2 0.7 0.8 2.2 1.2 3.1 1.6 3.8 (percent) (1.7) (5.2) (5.8) (16.6) (8.7) (23.2) (12.2) (28.5) South Asia 62 0.9 2.6 2.9 8.5 4.4 12.5 6.2 16.0 (percent) (2.4) (8.6) (9.5) (27.7) (14.3) (40.6) (20.1) (52) Sub-Saharan Africa 23 0.4 1.1 1.3 3.7 1.9 5.5 2.7 6.6 (percent) (1.6) (4.9) (5.4) (15.9) (8.2) (23.6) (11.5) (28.5) Low- and middle-income countries 362 6.0 17.9 19.7 56.8 29.8 79.2 41.7 98.2 (percent) (1.6) (4.9) (5.4) (15.7) (8.2) (21.9) (11.5) (27.1) High-income countries 81 0.6 2.6 2.1 8.5 3.2 12.2 4.5 16.2 (percent) (0.8) (3.2) (2.6) (10.6) (4.0) (15.1) (5.6) (20.0) World 443 6.6 20.5 21.8 65.3 33.0 91.5 46.2 114.3 (percent) (1.5) (4.6) (4.9) (14.7) (7.5) (20.7) (10.4) (25.8) Source: Authors' calculations. original (Ranson and others 2002). This analysis is conservative Africa, recent tax increases have doubled the real price of ciga- in its assumptions about effectiveness and generous in its rettes (Guindon, Tobin, and Yach 2002). assumptions about the costs of tobacco control. Potential Effect of Nicotine Replacement Therapies. Potential Effect of Price Increases. With a price increase of Provision of NRT with an effectiveness of 1 percent is predicted 33 percent, the model predicts that 22 million to 65 million to result in the avoidance of about 3.5 million smoking- smoking-attributable deaths will be averted worldwide, which attributable deaths (table 46.4). NRT of 5 percent effectiveness is approximately equivalent to 5 to 15 percent of all smoking- would have about five times the effect. Again, low- and middle- attributable deaths expected among those who smoke in 2000 income countries would account for roughly 80 percent of (see table 46.3). Low- and middle-income countries account the averted deaths. The relative effect of NRT (of 2.5 percent for about 90 percent of averted deaths. East Asia and the Pacific effectiveness) on deaths averted is 2 to 3 percent among indi- alone will account for roughly 40 percent of averted deaths. viduals age 15 to 59 and lower among those age 60 and older Total smoking-attributable deaths averted worldwide range (results not shown). from 33 million to 92 million for a 50 percent price increase and 46 million to 114 million for a 70 percent price increase. A Potential Effect of Non-price Interventions Other Than NRT. 70 percent price increase would avert 10 to 26 percent of all A package of non-price interventions, other than NRT, that smoking-attributable deaths worldwide. decreases the prevalence of smoking by 2 percent is predicted Of the tobacco-related deaths that would be averted by a to prevent about 7 million smoking-attributable deaths (more price increase, 80 percent would be male, reflecting the higher than 1.6 percent of all smoking-attributable deaths among overall prevalence of smoking in men. The greatest relative those who smoked in 2000; see table 46.4). A package of inter- effect of a price increase on deaths averted is among younger ventions that decreases the prevalence of smoking by 10 per- cohorts. Note that these projections use conservative price cent would have an effect five times greater. Low- and middle- increases. In certain countries, such as Poland and South income countries would account for approximately four-fifths 878 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others Table 46.4 Reductions in Future Tobacco Deaths among Smokers Alive in 2000 from Price Increases of 33 Percent, Increased NRT Use, and a Package of Non-price Measures by World Bank Region Reduction in number of deaths (millions) Non-price intervention Baseline 33 percent price increase NRT effectiveness effectiveness smoking-attributable World Bank region deaths (millions) Low elasticity High elasticity 1 percent 5 percent 2 percent 10 percent East Asia and the Pacific 173 9.6 27.5 1.4 6.9 2.8 13.8 (percent) (5.5) (15.9) (0.8) (4.0) (1.6) (8.0) Europe and Central Asia 51 2.8 8.1 0.4 2.1 0.8 4.1 (percent) (5.6) (16.0) (0.8) (4.0) (1.6) (8.1) Latin America and the Caribbean 40 2.3 6.7 0.3 1.7 0.7 3.4 (percent) (5.8) (16.8) (0.8) (4.2) (1.7) (8.5) Middle East and North Africa 13 0.8 2.2 0.11 0.6 0.2 1.1 (percent) (5.8) (16.6) (0.8) (4.2) (1.7) (8.4) South Asia 62 2.9 8.5 0.4 2.2 0.9 4.3 (percent) (9.5) (27.7) (1.4) (7.2) (2.8) (13.9) Sub-Saharan Africa 23 1.3 3.7 0.2 0.9 0.4 1.8 (percent) (5.4) (15.9) (0.8) (4.0) (1.6) (7.9) Low- and middle-income countries 362 19.7 56.8 2.9 14.3 5.7 28.6 (percent) (5.4) (15.7) (0.8) (4.0) (1.6) (7.9) High-income countries 81 2.1 8.5 0.6 3.1 1.2 6.1 (percent) (2.6) (10.6) (0.8) (3.8) (1.5) (7.6) World 443 21.8 65.3 3.5 17.4 6.9 34.7 (percent) (4.9) (14.8) (0.8) (3.9) (1.6) (7.8) Source: Authors. of quitters and averted deaths. The greatest relative effect of Tobacco deaths (millions) non-price interventions on deaths averted would be among 443 425 younger cohorts. 400 408 Figure 46.5 summarizes the potential effect of a set of Death year 2030: 10 million 377 deaths per year independent tobacco-control interventions, using 33 and 328 versus 70 percent price increases (using a high elasticity of 1.2 for 300 low- and middle-income regions and 0.8 for high-income 7 million deaths per year regions), a 5 percent effectiveness of NRT, and a 10 percent reduction from non-price interventions other than NRT. In 200 this cohort of smokers alive in 2000, approximately 443 mil- lion are expected to die in the next 50 years in the absence 100 of interventions. A substantial fraction of these tobacco deaths are avoidable with interventions. Price increases have the greatest effect on tobacco mortality, with the most 0 aggressive price increase of 70 percent having the potential 2000 2010 2020 2030 2040 2050 to avert almost one-quarter of all tobacco deaths. Even a Baseline 33 percent price increase modest price increase of 33 percent could potentially NRT with 5 percent 70 percent price increase prevent 66 million tobacco deaths over the course of the effectiveness Non-price interventions next 50 years. Although NRT and other non-price interven- with 10 percent reduction tions are less effective than price increases, they can still avert a substantial number of tobacco deaths (18 million Source: Authors. Note: Price increases assume a high price elasticity (­1.2 for low- and and 35 million deaths, respectively). The greatest effect of middle-income countries and ­0.8 for high-income countries). these tobacco-control interventions would occur after 2010, but a substantial number of deaths could be avoided even Figure 46.5 Potential Effect of Tax Increases, NRT, and Non-price before then. Interventions on Tobacco Mortality, 2000­50 Tobacco Addiction | 879 Note that no attempt has been made in this analysis to highly sensitive to the actual price of the NRT. NRT with a price examine the effect of combining the various packages of inter- of US$25 have a cost-effectiveness of US$75 per DALY com- ventions (for example, price increases with NRT, or NRT and pared with US$329 for an NRT price of US$150 (data not other non-price interventions). A number of studies have com- shown). pared the effect of price and non-price interventions; few For a given set of assumptions, the variation in the cost- empirical attempts have been made to assess how these inter- effectiveness of each intervention between low- and middle- ventions might interact. Although price increases have been income regions is relatively small and sensitive to the discount found in this analysis to be the most cost-effective antismoking rate (data not shown). All three interventions are most cost- intervention, policy makers should use both price and non- effective in South Asia and Sub-Saharan Africa. The difference price interventions to counter smoking. Non-price measures between low- and middle-income countries and high-income may be required to affect the most heavily dependent smokers, countries is more pronounced. For NRT, the cost per year of for whom medical and social support in stopping will be nec- healthy life gained is 3 to 10 times higher in high-income coun- essary. Furthermore, these non-price measures may be effective tries than elsewhere. For non-price interventions other than in increasing social acceptance and support of tobacco price NRT, the cost in high-income countries is 22 times higher, and increases. for price increases, almost 42 times higher. Of note, the esti- mates of cost-effectiveness are given as wide ranges,which reflect Cost-Effectiveness of Antismoking Interventions. In general, the range of assumptions used. price increases are found to be the most cost-effective anti- For price increases, the high-end estimates are roughly 25 smoking intervention. A 33 percent price increase (our base times the low-end estimates, and this difference is consistent case scenario) could be achieved for a cost of US$13 to US$195 among the regions. For NRT, the high-end estimates are 2.5 to per DALY saved globally, or US$3 to US$42 in low-income 10 times the low-end estimates, varying among the regions. For countries and US$85 to US$1,773 in high-income countries. non-price interventions other than NRT, the high-end esti- Wider access to NRT could be achieved for between US$75 and mates are 20 times the low-end estimates, and this difference is US$1,250 per DALY saved, depending on which assumptions consistent among the regions. are used. Non-price interventions other than NRT could be The cost-effectiveness results can be compared against exist- implemented for between US$233 and US$2,916 per DALY ing studies only for high-income countries because of a lack saved (table 46.5). Thus, NRT and other non-price measures of studies situated elsewhere. Our estimates of deaths avoided are slightly less cost-effective than price increases but remain for a 10 percent price increase are conservative compared with cost-effective in many settings. The cost-effectiveness of NRT is those of Moore (1996) and Warner (1986). Table 46.5 Range of Cost-Effectiveness Values for Price Increase, NRT, and Non-price Interventions, 2000 (2002 U.S. dollars per DALY saved) NRT with Non-price interventions 33 percent price effectiveness of with effectiveness of increase 1 to 5 percent 2 to 10 percent Baseline smoking-attributable Low-end High-end Low-end High-end Low-end High-end World Bank region deaths (millions) estimate estimate estimate estimate estimate estimate East Asia and the Pacific 173 2 30 65 864 40 498 Europe and Central Asia 51 3 42 45 633 55 685 Latin America and the Caribbean 40 6 85 53 812 109 1,361 Middle East and North Africa 13 6 89 47 750 115 1,432 South Asia 62 2 27 54 716 34 431 Sub-Saharan Africa 23 2 26 42 570 33 417 Low- and middle-income countries 362 3 42 55 761 54 674 High-income countries 81 85 1,773 175 3,781 1,166 14,572 World 443 13 195 75 1,250 233 2,916 Source: Authors. 880 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others COMPREHENSIVE TOBACCO-CONTROL use and in improving public health (Farrelly, Pechacek, and PROGRAMS Chaloupka 2003; Townsend, Roderick, and Cooper 1998; U.S. DHHS 1994). In California, for example, the state's compre- In recent years, several governments, mostly in high-income hensive tobacco-control program has produced a rate of countries, have adopted comprehensive programs to reduce decline in tobacco use double that seen in the rest of the United tobacco use, often funded by earmarked tobacco tax revenues. States. The programs generally have similar goals for reducing tobacco The cost of implementing control programs is low. use: Table 46.6 provides the estimated total costs of implementing price and NRT interventions by World Bank region. Current · preventing initiation among youths and young adults estimates of the costs of implementing a comprehensive · promoting cessation among all smokers tobacco-control program range from US$2.50 to US$10 per · reducing exposure to passive tobacco smoke capita in the United States. The U.S. Centers for Disease · identifying and eliminating disparities among population Control and Prevention recommends spending US$6 to US$16 subgroups (U.S. DHHS 1994). per capita for a comprehensive tobacco-control program in the United States (CDC 1999). Canadian spending on tobacco- Furthermore, the programs have one or more of four key control programs was approximately US$1.70 per capita in components: community interventions engaging a diverse set 1996 (Pechmann, Dixon, and Layne 1998). At the highest rec- of local organizations; countermarketing and health informa- ommended spending level (US$16 per capita) in the United tion campaigns; program policies and regulations (such as States, annual funding for a comprehensive tobacco program taxes, restrictions on smoking, bans on tobacco advertising, would equal only 0.9 percent of U.S. public spending, per and access to better cessation treatments); and surveillance and capita, on health. evaluation of potential issues, such as smuggling (U.S. DHHS Evidence from the United States demonstrates that states 1994). Programs have placed differing emphasis on these four with the greatest prevalence of smoking have a greater marginal components, with substantial diversity among the types of effect with their tobacco-control spending, suggesting that the activities supported within each component. Disaggregating potential gains from modest investments in comprehensive current tobacco-control program spending reveals that the tobacco-control measures are large. Each US$10 spent per greatest effect can be achieved through a focus on macro-level capita on tobacco control annually has resulted in a 55 percent changes, such as policy change. Recent analyses from the reduction (variation of 20 to 70 percent) in per capita cigarette United Kingdom and United States clearly indicate that these consumption (Tauras and others 2005). In the United States, comprehensive efforts have been successful in reducing tobacco US$10 translates into 0.03 percent of per capita GDP in 2003. Table 46.6 Estimated Cost of Price Intervention and NRT Programs (2002 U.S. dollars) Cost of NRT (US$25 to US$150) (millions) To treat 1 percent of To treat 5 percent of Cost for price increase (millions) current smokers current smokers Low-end estimate High-end estimate World Bank region GDP (billions) (0.02 percent GDP) (0.05 percent GDP) US$25 US$50 US$150 US$25 US$50 US$150 East Asia and the Pacific 1,802 360 901 1,079 2,158 6,474 5,395 10,791 32,372 Europe and Central Asia 1,136 227 568 318 635 1,906 1,588 3,176 9,529 Latin America and the Caribbean 1,673 335 836 250 500 1,500 1,250 2,500 7,499 Middle East and North Africa 694 139 347 84 169 506 422 843 2,530 South Asia 655 131 327 2,312 1,926 3,853 11,558 2,312 1,926 Sub-Saharan Africa 319 64 159 868 723 1,447 4,340 868 723 Low- and middle-income countries 6,279 1,256 3,138 4,911 6,111 15,686 24,553 20,490 54,579 High-income countries 25,992 5,198 12,996 3,034 2,529 10,114 15,172 3,034 2,529 World 32,271 6,454 16,134 7,945 8,640 25,800 39,725 23,524 57,108 Source: Authors. Tobacco Addiction | 881 CONSTRAINTS TO EFFECTIVE or "dead-weight losses" from earmarking tobacco taxes are TOBACCO-CONTROL POLICIES minimal (Hu, Xu, and Keeler 1998). Furthermore, earmarking tobacco taxes can be justified if governments use the funds to Although substantial evidence exists concerning the effective- benefit those who pay (the benefits principle), provide assured ness of numerous policy interventions to reduce tobacco use, funding for tobacco-control policies and programs, and secure the use of these interventions globally is uneven and limited public support for new or higher tobacco taxes. Earmarked (see a more formal analysis in Chaloupka and others 2001). taxes also have a political function in that they help concentrate World Bank data reveal that ample room exists to increase political winners of tobacco control and thus influence policy. tobacco taxes. In 1995, the average percentage of all govern- Earmarked funds that support broad health and social services ment revenue derived from tobacco tax was 0.63 percent. (such as other disease programs) broaden the political and civil Middle-income countries averaged 0.51 percent of govern- society support base for tobacco control. In Australia, broad ment revenue from tobacco taxes, while lower-income coun- political support from the Ministries of Sports and Education tries averaged only 0.42 percent. An increase in cigarette taxes helped convince the Ministry of Finance that raising tobacco of 10 percent globally would raise cigarette tax revenues by taxes was possible. Indeed, after an earmarked tax was passed, nearly 7 percent, with relatively larger increases in revenues the Ministry of Finance went on to raise tobacco taxes further in high-income countries and smaller increases in revenues in without earmarking (Galbally 1997). Additionally, targeting low- and middle-income countries (Sunley, Yurekli, and revenue from tobacco taxes to other health programs for the Chaloupka 2000). Despite this evidence, price increases have poorest socioeconomic groups could produce double health been underused. Guindon, Tobin, and Yach (2002) studied 80 gains--reduced tobacco consumption combined with countries and found that the real price of tobacco, adjusted for increased access to and use of health services. In China, a purchasing power, fell in most developing countries from 1990 10 percent increase in cigarette taxes would decrease consump- to 2000. tion by 5 percent and would increase government revenue by Why does so much variation exist in tobacco-control 5 percent. The increased earnings could finance a package of policies? The political economy of tobacco control has been essential health services for one-third of China's poorest inadequately studied. A few plausible areas of interest are out- 100 million citizens in 1990 (Saxenian and McGreevey 1996). lined here. First, the recognition of tobacco as a major health Finally, a key pillar in tobacco control that can help over- hazard appears to be the impetus for most of the tobacco- come some of these constraints is the Framework Convention control policies in many high-income countries. Some evi- on Tobacco Control (FCTC). The World Health Assembly dence shows that improved national capacity and local needs of the World Health Organization adopted the FCTC in assessment could increase the likelihood that tobacco-control May 2003. It consists of a series of negotiated protocols within measures will be adopted. For example, econometric analyses a general framework. The first three protocols are negotiations in South Africa geared to local policy requirements substantially covering smuggling, advertising, and treatment of tobacco increased the willingness of the government to implement addiction. Countries agreeing to the negotiated protocols are to tobacco-control policies (Abedian and others 1998). Second, adopt appropriate legislation and, if necessary, implement the tobacco-control budgets are only a fraction of what is required. appropriate measures. As of February 2005, 168 countries had Funding is needed not so much to implement programs as to signed the FCTC, 57 had ratified it, and it had come into force fight off tobacco industry tactics and to build popular support on February 27, 2005. for control. Third, the most obvious constraint to tobacco control is political opposition, which is difficult to quantify. Opposition from the tobacco industry is well organized and CONCLUSION well funded (Pollock 1996). A key tool for addressing political opposition is earmarking Worldwide, only two large and growing causes of death exist. tobacco taxes. Earmarking has been successfully used in several One is HIV-1 infection, and the other is tobacco. On current countries, including Australia, Finland, Nepal, and Thailand. consumption patterns, about 1 billion people in the 21st Of the 48 countries currently in the World Health century will be killed by their addiction to tobacco. Strong Organization's European region, 12 earmark taxes for tobacco evidence shows that tobacco tax increases, the dissemination control and other public health measures. The average level of of information about health risks from smoking, restrictions allocation is less than 1 percent of total tax revenue (WHO on smoking in public places and workplaces, comprehensive 2002). Earmarking does introduce clear restrictions and ineffi- bans on advertising and promotion, and increased access to ciencies on public finance, and for this reason alone most cessation therapies are effective both in reducing tobacco use macroeconomists do not favor earmarking, no matter how and in improving the health of populations. Despite this worthy the cause. However, analysis suggests that the efficiency evidence, these policies, especially higher taxes, have been 882 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others applied aggressively only in a few high-income countries, cov- Ezzati, M., and A. D. Lopez. 2003. "Estimates of Global Mortality ering a small proportion of the world's smokers. Limited Attributable to Smoking in 2000." Lancet 362 (9387): 847­52. implementation of effective tobacco control in developing Farrelly, M. C., T. F. Pechacek, and F. J. Chaloupka. 2003. "The Impact of Tobacco Control Program Expenditures on Aggregate Cigarette Sales: countries is due to political constraints as well as the lack of 1981­2000." Journal of Health Economics 22 (5): 843­59. awareness of the unique effectiveness and cost-effectiveness Gajalakshmi, C. K., P. Jha, K. Ranson, and S. Nguyen. 2000. "Global of these interventions. Patterns of Smoking and Smoking-Attributable Mortality." In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. ACKNOWLEDGMENTS Gajalakshmi, V., R. Peto, T. S. Kanaka, and P. Jha. 2003. "Smoking and Mortality from Tuberculosis and Other Diseases in India: Retrospective Study of 43,000 Adult Male Deaths and 35,000 We thank Allison Gilbert for help with the cost-effectiveness Controls." Lancet 362 (9883): 507­15. analyses and Hellen Gelband, Andra Ghent, and Dhirendra Galbally, R. L. 1997. "Health-Promoting Environments: Who Will Miss Sinha for comments. Out?" Australia and New Zealand Journal of Public Health 21 (4 Spec. No.): 429­30. Gruber, J. 2003. "Government Policy toward Smoking: A New View from REFERENCES Economics." Disease Control Priorities Project Working Paper Series. Paper presented at the Disease Control Priorities Project Nicotine Abedian, I., R. van der Merwe, N. Wilkins, and P. Jha, eds. 1998. The Addiction Workshop, Mumbai, India, September 2003. Economics of Tobacco Control: Towards an Optimal Policy Mix. Cape Gruber, J., and B. Koszegi. 2001. "Is Addiction Rational? Theory and Town, South Africa: Applied Fiscal Research Centre, University of Evidence." Quarterly Journal of Economics 116 (4): 1261­303. Cape Town. ------. 2002. "A Theory of Government Regulation of Addictive Bads: Barnum, H. 1994."The Economic Burden of the Global Trade in Tobacco." Optimal Tax Levels and Tax Incidence for Cigarette Taxation." NBER Tobacco Control 3 (4): 358­61. Working Paper 8777. Cambridge, MA: National Bureau of Economic CDC (U.S. Centers for Disease Control and Prevention). 1994. "Response Research. to Increases in Cigarette Prices by Race/Ethnicity, Income, and Age Gruber, J., and S. Mullainathan. 2002. "Do Cigarette Taxes Make Smokers Groups--United States, 1976­1993." Morbidity and Mortality Weekly Happier?" NBER Working Paper 8872. Cambridge, MA: National Report 43 (26): 469­72. Bureau of Economic Research. ------. 1999. "Best Practices for Comprehensive Tobacco Control Guindon, G. E., S. Tobin, and D. Yach. 2002. "Trends and Affordability of Programs." Atlanta: U.S. Department of Health and Human Services, Cigarette Prices: Ample Room for Tax Increases and Related Health Centers for Disease Control and Prevention, National Center for Gains." Tobacco Control 11 (1): 35­43. Chronic Disease Prevention and Health Promotion, Office on Gupta, P. C., and H. C. Mehta. 2000. "Cohort Study of All-Cause Mortality Smoking and Health. amongst Tobacco Users in Mumbai, India." Bulletin of the World Health Chaloupka, F. J., T. W. Hu, K. E. Warner, R. Jacobs, and A. Yurekli. 2000. Organization 78 (7): 877­83. "The Taxation of Tobacco Products." In Tobacco Control in Developing Hu, T. W., X. Xu, and T. Keeler. 1998. "Earmarked Tobacco Taxes: Lessons Countries, ed. P. Jha and F. Chaloupka. Oxford, U.K.: Oxford University Learned." In The Economics of Tobacco Control, ed. I. Abedian, R. van Press. der Merwe, N. Wilkins, and P. Jha. Cape Town, South Africa: Applied Chaloupka, F., J. P. Jha, M.A. Corrao, V. Costa e Silva, H. Ross, C. Czart, and Fiscal Research Centre, University of Cape Town. D. Yach. 2001. "The Evidence Base for Reducing Mortality from Jacobs, R., H. F. Gale, T. C. Capehart, P. Zhang, and P. Jha. 2000. "The Smoking in Low and Middle Income Countries." Commission on Supply-Side Effects of Tobacco-Control Policies." In Tobacco Control Macroeconomics and Health Working Paper Series. http://www. in Developing Countries, ed. P. Jha, and F. J. Chaloupka. Oxford, U.K.: cmhealth.org/docs/wg5_paper7.pdf. Oxford University Press. Chaloupka, F. J., J. A. Tauras, and M. Grossman. 2000. "The Economics of Jamison, D. T., L. J. Lau, and J. Wang. 2005. "Health's Contribution to Addiction." In Tobacco Control in Developing Countries, ed. P. Jha and Economic Growth in an Environment of Partially Endogenous F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Technical Progress." In Health and Economic Growth: Findings and Chinese Academy of Preventive Medicine. 1997. Smoking in China: 1996 Policy Implications, eds. G. Lopez-Casasnovas, B. Rivera, and L. National Prevalence Survey of Smoking Pattern. Beijing: China Science Currais. Cambridge: MIT Press, 67­91. and Technology Press. Jarvis, M. J. 2004."ABC of Smoking Cessation: Why People Smoke." British Crispo, A., P. Brennan, K. H. Jockel, A. Schaffrath-Rosario, H. E. Medical Journal 328 (7434): 277­79. Wichmann, F. Nyberg, and others. 2004."The Cumulative Risk of Lung Jha, P., and F. J. Chaloupka. 1999. Curbing the Epidemic: Governments and Cancer among Current, Ex- and Never-Smokers in European Men." the Economics of Tobacco Control. Washington, DC: World Bank. British Journal of Cancer 91 (7): 1280­86. ------. 2000a. "The Economics of Global Tobacco Control." British Cutler, D. M., and S. Kadiyala. 2003. "The Return to Biomedical Research: Medical Journal 321 (7257): 358­61. Treatment and Behavioral Effects." In Measuring the Gains of Medical ------, eds. 2000b. Tobacco Control in Developing Countries. Oxford, U.K.: Research: An Economic Approach, ed. K. M. Murphy and R. H. Topel. Oxford University Press. Chicago: University of Chicago. Jha, P., P. Musgrove, F. J. Chaloupka, and A. Yurekli. 2000. "The Economic Doll, R., R. Peto, J. Boreham, and I. Sutherland. 2004. "Mortality in Rationale for Intervention in the Tobacco Market." In Tobacco Control Relation to Smoking: 50 Years' Observation on Male British Doctors." in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: British Medical Journal 328 (7455): 1519­28. Oxford University Press. Emery, S., M. M. White, and J. P. Pierce. 2001. "Does Cigarette Price Jha, P., M. K. Ranson, S. N. Nguyen, and D. Yach. 2002. "Estimates of Influence Adolescent Experimentation?" Journal of Health Economics Global and Regional Smoking Prevalence in 1995 by Age and Sex." 20 (2): 261­70. American Journal of Public Health 92 (6): 1002­6. Tobacco Addiction | 883 Joossens, L., F. J. Chaloupka, D. Merriman, and A.Yurekli. 2000. "Issues in Pechmann, C., P. Dixon, and N. Layne. 1998. "An Assessment of U.S. and the Smuggling of Tobacco Products." In Tobacco Control in Developing Canadian Smoking Reduction Objectives for the Year 2000." American Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford Journal of Public Health 88 (9): 1362­67. University Press. Peck, R., F. J. Chaloupka, P. Jha, J. Lightwood. 2000. "Welfare Analyses of Kenkel, D., and L. Chen. 2000. "Consumer Information and Tobacco Use." Tobacco." In Tobacco Control in Developing Countries, ed. P. Jha and In Tobacco Control in Developing Countries, ed. P. Jha and F. J. F. J. Chaloupka, 131­52. Oxford, U.K.: Oxford University Press. Chaloupka. Oxford, U.K.: Oxford University Press. Peto, R., S. Darby, H. Deo, P. Silcocks, E. Whitley, and R. Doll. 2000. Koop, C. E. 2003. "Tobacco Addiction: Accomplishments and Challenges "Smoking, Smoking Cessation, and Lung Cancer in the U.K. since in Science, Health, and Policy." Nicotine and Tobacco Research 5 (5): 1950: Combination of National Statistics with Two Case-Control 613­19. Studies." British Medical Journal 321 (7257): 323­29. Lam, T. H., Y. He, Q. L. Shi, J. Y. Huang, F. Zhang, Z. H. Wan, and others. Peto, R., and A. D. Lopez. 2001. "The Future Worldwide Health Effects of 2002. "Smoking, Quitting, and Mortality in a Chinese Cohort of Current Smoking Patterns." In Critical Issues in Global Health, ed. C. E. Retired Men." Annals of Epidemiology 12 (5): 316­20. Koop, C. E. Pearson, and M. R. Schwarz. New York: Jossey-Bass. Laviolette, S. R., and D. van der Kooy. 2004. "The Neurobiology of Peto, R., A. D. Lopez, J. Boreham, and M. Thun. 2003. Mortality Nicotine Addiction: Bridging the Gap from Molecules to Behavior." from Smoking in Developed Countries. 2nd ed. Oxford, U.K.: Oxford Nature Reviews 5 (1): 55­65. University Press. Levy, D. T., K. Friend, and E. Polishchuk. 2001. "Effect of Clean Indoor Air Peto, R., A. D. Lopez, J. Boreham, M. Thun, and C. Heath, Jr. 1994. Laws on Smokers: The Clean Air Module of the SimSmoke Computer Mortality from Smoking in Developed Countries, 1950­2000. Oxford, Simulation Model." Tobacco Control 10 (4): 345­51. U.K.: Oxford University Press. Lightwood, J., D. Collins, H. Lapsley, and T. E. Novotny. 2000. "Estimating Picciotto, M. R. 2003. "Nicotine as a Modulator of Behavior: Beyond the the Costs of Tobacco Use." In Tobacco Control in Developing Countries, Inverted U." Trends in Pharmacological Sciences 24 (9): 493­99. ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Pollock, D. 1996. "Forty Years On: A War to Recognise and Win--How the Tobacco Industry Has Survived the Revelations on Smoking and Liu, B. Q., R. Peto, Z. M. Chen, J. Boreham, Y. P. Wu, J. Y. Li, and others. Health." British Medical Bulletin 52 (1): 174­82. 1998. "Emerging Tobacco Hazards in China: 1. Retrospective Proportional Mortality Study of One Million Deaths." British Medical Ranson, M. K., P. Jha, F. J. Chaloupka, and S. N. Nguyen. 2002. "Global and Journal 317 (7170): 1411­22. Regional Estimates of the Effectiveness and Cost-Effectiveness of Price Increases and Other Tobacco Control Policies." Nicotine and Tobacco Markou, A., and J. E. Henningfield. 2003. "Background Paper on the Research 4 (3): 311­19. Neurobiology of Nicotine Addiction." Paper presented at the Disease Control Priorities Project Nicotine Addiction Workshop, Mumbai, Raw, M., A. McNeill, and R. West. 1999. "Smoking Cessation: Evidence- India, September 2003. Based Recommendations for the Healthcare System." British Medical Journal 318 (7177): 182­85. Martinson, B. C., P. J. O'Connor, N. P. Pronk, and S. J. Rolnick. 2003. "Smoking Cessation Attempts in Relation to Prior Health Care Ross, H., F. J. Chaloupka, and M. Wakefield. 2001. "Youth Smoking Uptake Changes: The Effect of Antecedent Smoking-Related Symptoms?" Progress: Price and Public Policy Effects." Research Paper 11. American Journal of Health Promotion 18 (2): 125­32. ImpacTeen, Health Research and Policy Centers, University of Illinois at Chicago. Merriman, D., A. Yurekli, and F. J. Chaloupka. 2000. "How Big Is the Worldwide Cigarette Smuggling Problem?" In Tobacco Control in Saffer, H. 2000. "Tobacco Advertising and Promotion." In Tobacco Control Developing Countries, ed. P. Jha, and F. J. Chaloupka. Oxford, U.K.: in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Oxford University Press. Molarius, A., R. W. Parsons, A. J. Dobson, A. Evans, S. P. Fortmann, K. Saffer, H., and F. Chaloupka. 2000. "Tobacco Advertising: Economic Jamrozik, and others. 2001. "Trends in Cigarette Smoking in 36 Theory and International Evidence." Journal of Health Economics 19 Populations from the Early 1980s to the Mid-1990s: Findings from the (6): 1117­37. WHO MONICA Project." American Journal of Public Health 91 (2): Saxenian, H., and B. McGreevey. 1996. "China: Issues and Options in 206­12. Health Financing." Report 15278-CHA, World Bank, Washington, DC. Moore, M. J. 1996."Death and Tobacco Taxes." RAND Journal of Economics Shiffman, S., and R. West. 2003. "Background Paper on the Psychology 27 (2): 415­28. of Nicotine Addiction." Paper presented at the Disease Control Priorities Project Nicotine Addiction Workshop, Mumbai, India, Murphy, K. M., and R. H. Topel. 2003. "The Economic Value of Medical September 2003. Research." In Measuring the Gains of Medical Research: An Economic Approach, ed. K. M. Murphy and R. H. Topel, 41­73. Chicago: Sunley, E. M., A. Yurekli, and F. J. Chaloupka. 2000. "The Design, University of Chicago. Administration and Potential Revenue of Tobacco Excises." In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, Niu, S. R., G. H. Yang, Z. M. Chen, J. L. Wang, G. H. Wang, X. Z. He, and U.K.: Oxford University Press. others. 1998. "Emerging Tobacco Hazards in China: 2. Early Mortality Results from a Prospective Study." British Medical Journal 317 (7170): Tauras, J. A. 1999. "The Transition to Smoking Cessation: Evidence from 1423­24. Multiple Failure Duration Analysis." NBER Working Paper 7412. Cambridge, MA: National Bureau of Economic Research. Novotny, T. E., J. C. Cohen, A. Yurekli, D. Sweaner, and J. de Beyer. 2000. "Smoking Cessation and Nicotine-Replacement Therapies." In Tobacco Tauras, J. A., and F. J. Chaloupka. 2003. "The Demand for Nicotine Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, Replacement Therapies." Nicotine and Tobacco Research 5 (2): U.K.: Oxford University Press. 237­43. O'Malley, P. M., J. G. Bachman, and L. D. Johnston. 1988. "Period, Age, and Tauras, J. A., F. J. Chaloupka, M. Farrelly, G. A. Giovino, M. Wakefield, Cohort Effects on Substance Use among Young Americans: A Decade L. D. Johnston, and others. 2005. "State Tobacco Control Spending of Change, 1976­86." American Journal of Public Health 78 (10): and Youth Smoking." American Journal of Public Health 95 (2): 1315­21. 338­44. 884 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others Taylor, A. L., F. J. Chaloupka, E. Guindon, and M. Corbett. 2000. "The ------. 2003. "The Costs of Benefits: Smoking and Health Care Impact of Trade Liberalization on Tobacco Consumption." In Tobacco Expenditures." American Journal of Health Promotion 18 (2): 123­24. Control in Developing Countries, ed. P. Jha and F. J. Chaloupka, 343­64. Warner, K. E., F. J. Chaloupka, P. J. Cook, W. G. Manning, J. P. Newhouse, Oxford, U.K.: Oxford University Press. T. E. Novotny, and others. 1995. "Criteria for Determining an Optimal Townsend, J. L. 1993."Policies to Halve Smoking Deaths." Addiction 88 (1): Cigarette Tax: The Economist's Perspective." Tobacco Control 4: 43­52. 380­86. Townsend, J. L., P. Roderick, and J. Cooper. 1998. "Cigarette Smoking by Wasserman, J., W. G. Manning, J. P. Newhouse, and J. D. Winkler. 1991. Socio-Economic Group, Sex, and Age: Effects of Price, Income, and "The Effects of Excise Taxes and Regulations on Cigarette Smoking." Health Publicity." British Medical Journal 309 (6959): 923­26. Journal of Health Economics 10 (1): 43­64. U.K. Department of Health. 1992. "Effect of Tobacco Advertising on Weinstein, N. D. 1998. "Accuracy of Smokers' Risk Perceptions." Annals of Tobacco Consumption: A Discussion Document Reviewing the Behavioral Medicine 20 (2): 135­40. Evidence." London: U.K. Department of Health, Economics and WHO (World Health Organization). 1997. Tobacco or Heath: A Global Operational Research Division. Status Report. Geneva: WHO. U.S. DHHS (United States Department of Health and Human Services). ------. 2002. "The European Report on Tobacco Control Policy." Paper 1994. Preventing Tobacco Use amongst Young People. A Report of the presented at the WHO European Ministerial Conference for a Surgeon General. Atlanta: U.S. DHHS, Public Health Service, Centers Tobacco-free Europe, Warsaw. Document EUR/01/5020906/8, WHO for Disease Control, Center for Chronic Disease Prevention and Health Regional Office for Europe, Copenhagen. Promotion, Office on Smoking and Health. Woolery, T., S. Asma, and D. Sharp. 2000. "Clean Indoor-Air Laws and ------. 2000. Reducing Tobacco Use: A Report of the Surgeon General. Youth Access." In Tobacco Control in Developing Countries, ed. P. Jha Atlanta: U.S. DHHS, Public Health Service, Centers for Disease and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Control, Center for Chronic Disease Prevention and Health Zatonski, W., and P. Jha. 2000. "The Health Transformation in Eastern Promotion, Office on Smoking and Health. Europe after 1990: A Second Look." Warsaw: Marie Skeodowska-Curie U.S. Federal Trade Commission. 2004. Cigarette Report for 2002. Cancer Center and Institute of Oncology. Washington, DC: U.S. Federal Trade Commission. http://www.ftc. gov/reports/cigarette/041022cigaretterpt.pdf. Warner, K. E. 1986. "Smoking and Health Implications of a Change in the Federal Cigarette Excise Tax." Journal of the American Medical Association 255 (8): 1028­32. Tobacco Addiction | 885 Chapter 47 Alcohol Jürgen Rehm, Dan Chisholm, Robin Room, and Alan D. Lopez This chapter provides an overview of epidemiology of alco- others 2004). As figure 47.1 shows, alcohol consumption is hol use and health consequences as well as introducing cost- linked to acute and long-term health and social consequences effectiveness interventions to reduce alcohol-related harm. through three intermediate mechanisms--toxic and beneficial biochemical effects, intoxication, and dependence (Babor and others 2003; Rehm, Room, Graham, and others 2003)--as EPIDEMIOLOGY OF ALCOHOL USE AND follows: ALCOHOL-RELATED DISEASE CONDITIONS · Toxic and beneficial biochemical effects. These effects of Alcoholic beverages and the problems they engender have alcohol consumption may influence chronic disease in been familiar fixtures in human societies since the beginning either beneficial or harmful ways. Accepted beneficial effects of recorded history. Because alcohol is causally related to include the influence of moderate drinking on coronary more than 60 International Classification of Diseases codes heart disease through reduction of plaque deposits in arter- (Rehm, Room, Graham, and others 2003), disease outcomes ies, protection against blood clot formation, and promotion are among the most important alcohol-related problems. of blood clot dissolution (Zakhari 1997). Examples of Depending on the pattern of consumption, alcohol is also harmful effects include increased risk for high blood pres- protective against diseases, most important among them, sure and for liver damage (Rehm, Room, Graham, and coronary heart disease (Rehm, Sempos, and Trevisan 2003). others 2003) and direct toxic effects on acinar cells trigger- However, the net effect is negative, and 4 percent of the global ing pancreatic damage (Apte, Wilson, and Korsten 1997) or burden of disease is attributable to alcohol, or about as much hormonal disturbances (Emanuele and Emanuele 1997). death and disability globally as is attributable to tobacco and These are just examples, because alcohol exposure is associ- hypertension (Ezzati and others 2002; WHO 2002). Alcohol ated with a multitude of toxic effects on different organs. thus constitutes a serious public health problem (Room, · Intoxication. Alcohol intoxication is a powerful mediator for Babor, and Rehm 2005). Evidence-based preventive measures acute health outcomes, such as accidental or intentional are available at both the individual and the population levels, injuries or deaths, although intoxication can also be impli- with alcohol taxes, restrictions on alcohol availability, and cated in chronic health and social problems and in certain drinking-and-driving countermeasures among the most forms of heart disease. The subjective feeling of intoxication effective policy options (Babor and others 2003). This chapter is mainly caused by the effects of alcohol on the central reviews the cost-effectiveness of different interventions in nervous system, and these effects are felt and can be meas- developing regions of the world. ured even at light to moderate consumption levels (Eckardt and others 1998). Dimensions of Alcohol Related to Disease · Dependence. Alcohol dependence is a clinical disorder in its The relationship between alcohol consumption and health and own right, but it is also a powerful mechanism sustaining social outcomes is complex and multidimensional (Rehm and alcohol consumption and mediating its impact on both 887 in terms of high-risk drinking occasions and also in terms of Patterns of Average drinking in public settings and the proportion of drinking that drinking volume occurs outside of meals (for further details, see Rehm and others 2004). Epidemiology of High-Risk Alcohol Use The intervention analyses presented in this chapter focus on Toxic and Intoxication average high-risk drinking, although patterns of drinking were beneficial Dependence biochemical also incorporated into the disease burden calculations. High- effects risk drinking is defined in sex-specific terms as drinking 20 grams per day or more of pure alcohol on average for females and 40 grams per day or more of pure alcohol on aver- age for males (a bottle of table wine contains about 70 grams of Accidents and Acute social Chronic social Chronic pure alcohol). This definition of high-risk drinking is fairly injuries and and disease (acute disease) psychological psychological standard in alcohol epidemiology and was first introduced by problems problems English and others (1995) on the basis of Australian guidelines. Source: Adapted from Babor and others 2003. Originally, English and others (1995) used two categories: haz- ardous drinking (defined as drinking between 20 and 40 grams Figure 47.1 Model of Alcohol Consumption, Intermediate Outcomes, per day of pure alcohol on average for females and between 40 and Long-Term Consequences and 60 grams per day of pure alcohol for males) and harmful drinking (defined as drinking 40 grams per day or more of pure alcohol on average for females and 60 grams per day or more of pure alcohol on average for males). These categories have been chronic and acute physiological and social consequences of used in almost every comprehensive meta-analysis on alcohol alcohol (Drummond 1990). In the quantitative analyses and disease since 1995 (see Rehm, Room, Graham, and others reported in this chapter, alcohol dependence--and alcohol- 2003 for an overview). However, critics asserted that the terms use disorders (AUDs) in general--will be considered only as hazardous drinking and harmful drinking were not neutral; a health outcome related to high-risk alcohol use. thus, the CRA uses drinking categories II and III, referring to the term high-risk drinking when both categories are consid- This chapter, including the section on the cost-effectiveness ered together. High-risk drinking thresholds differ by sex of interventions, focuses primarily on health consequences, because the risk for chronic disease is related to lower volumes although later it briefly discusses the social consequences of of drinking for women than for men; thus, the thresholds for high-risk drinking and recommended interventions. The epi- high-risk drinking were set to reflect an approximately similar demiological calculations are taken from Ezzati and others' risk of chronic disease. (2002) comparative risk analysis (CRA) and the World Health Table 47.1 shows the distribution of high-risk drinking by Organization (WHO) assessment of the global burden of dis- age and by World Bank region. The table excludes the Middle ease (WHO 2002). (For further information, see Mathers and East and North Africa because prevalence rates of high-risk others 2003; Rehm, Rehn, and others 2003; Rehm, Room, drinking are considerably lower than 1 percent and this situa- Graham, and others 2003; Rehm, Room, Monteiro and others tion is unlikely to change in the near future. 2003; Rehm and others 2004). The CRA defines alcohol expo- Calculating the burden of high-risk alcohol use that is sure using two measures: the average volume of alcohol con- avertable by means of effective interventions requires additional sumption and patterns of drinking (figure 47.1). It then relates epidemiological data--in particular, rates of incidence to and these exposure measures to disease outcomes. remission from high-risk alcohol use and the relative fatality of The average volume of consumption has been the conven- high-risk alcohol users compared with non-high-risk alcohol tional measure of exposure in alcohol epidemiology (Bruun users. We derived remission rates from studies of natural and others 1975) and has been linked to many disease cate- recovery from alcohol problems, which found an average of gories following the seminal work of English and others (1995; 10.9 years to remission (Sobell, Ellingstad, and Sobell 2000), see also Rehm, Room, Graham, and others 2003). Patterns of with an adjustment of plus 20 percent for older age groups and drinking have been linked mainly to two categories of disease minus 20 percent for younger age groups. We set the relative risk outcome: acute effects of alcohol (such as accidental and inten- of mortality for high-risk alcohol users age 15 to 44 at 2.5 and the tional injuries) and cardiovascular outcomes (mainly coronary relative risk for older age groups at 1.3 for men and 1.4 for women heart disease). The CRA defines patterns of drinking primarily (Gmel, Gutjahr, and Rehm 2003; Rehm, Gutjahr, and Gmel 888 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others Table 47.1 Prevalence of High-Risk Drinking by Gender, Age Group, and Region, 2000 (percentage of the population) Age group (years) Region Gender 15­29 30­44 45­59 60­69 70 Europe and Central Asia Male 20.8 18.7 21.4 15.2 8.1 Female 11.2 10.4 11.5 7.9 5.7 Latin America and the Caribbean Male 9.7 11.1 10.6 7.9 3.4 Female 6.8 7.5 6.5 5.8 3.1 Sub-Saharan Africa Male 10.4 14.3 12.9 11.3 8.4 Female 3.1 4.7 5.1 3.2 2.2 East Asia and the Pacific Male 6.2 7.5 7.1 6.5 5.0 Female 0.3 0.2 0.1 0.1 0.0 South Asia Male 0.8 2.5 0.3 0.1 0.0 Female 1.2 0.4 0.4 0.0 0.0 High-income countries Male 18.0 17.9 16.2 10.9 7.6 Female 10.9 8.7 9.8 6.8 5.4 Source: Authors' calculations based on Rehm, Rehn, and others 2003 and Rehm and others 2004. Note: The criteria for high-risk drinking were set sex specific (for details see text). 2001). Using WHO disease-modeling software, we derived an whereas many of these criteria are associated with high-risk internally consistent epidemiological profile of current high-risk alcohol use, no strict classificatory rule indicates that people alcohol use in each region, including specifications of incidence with AUDs are a subcategory of high-risk drinkers. and the relative risk of mortality,with currently observed rates of Second, the prevalence of AUDs is often derived from sur- prevalence, remission, and risk of mortality as inputs. A final veys, where the operationalization usually requires that three input parameter is the disability level for high-risk alcohol use, symptoms be present in a lifetime and at least one of these cri- which we estimated at 0.154 (where zero equals no disability); teria be present within the past 12 months (see, for example, this is a weighted average based on the severity breakdown of Demyttenaere and others 2004, table 2). Thus individuals may high-risk drinkers from the CRA (80 percent category II, or be categorized as alcohol dependent even if they are currently hazardous; 20 percent category III, or harmful). The preference abstaining from alcohol. values for these health states of 0.11 and 0.33, respectively, are Third, qualitative studies across a wide range of cultures derived from Stouthard, Essink-Bot, and Bonsel (2000). have found that the criteria used for diagnosing AUDs often have different meanings and implications in different cultural Relationship between High-Risk Drinking and AUDs settings (Room and others 1996; Schmidt and Room 1999). For instance, in the United States over the past decade, the level of Assessing the relationship between high-risk drinking and reported AUDs increased despite decreases in high-risk drink- AUDs is not a straightforward exercise. Even though high-risk ing (Grant and others 2004). This fact has been explained in drinking over a long period entails the risk of AUDs, that all terms of changes in drinking norms and social attitudes during people with AUDs are also high-risk drinkers does not auto- a period when the United States has become a "drier" culture. matically follow. First, neither the definition of alcohol depend- Thus, the measurement of AUDs is quite complex and cultur- ence nor WHO's (1993) definition of harmful use includes ally dependent. Moreover, AUDs are only one outcome of alco- actual consumption levels. An individual is considered depend- hol consumption and, in many parts of the world, not the most ent if at least three of the following criteria apply: important one. As a result, we decided to focus on high-risk · strong desire or compulsion to take the substance alcohol consumption rather than AUDs. · impaired control and physiological withdrawal if the sub- stance is reduced or ceased Relationship between Alcohol Use and · tolerance to the effects of the substance Disease Categories · preoccupation with use of the substance · persistent use despite clear evidence of harmful conse- The exact procedures for quantifying the risk of disease attrib- quences. utable to alcohol are described in detail elsewhere (Rehm, Room, Graham, and others 2003; Rehm and others 2004). By contrast, harmful alcohol use is defined as a pattern of For most chronic disease categories, investigators have derived use that is causing damage to physical or mental health. Thus, alcohol-attributable fractions of disease by combining Alcohol | 889 prevalence and relative risk estimates based on meta-analyses had. These differences cannot be explained by the overall vol- (Corrao and others 2000; English and others 1995; Gutjahr, ume of drinking, which actually tended to be greater for those Gmel, and Rehm 2001; Ridolfo and Stevenson 2001; Single in higher socioeconomic groups. Rather, the differences can be and others 1996, 1999). For depression, we drew alcohol- explained by the fact that more of the drinking of those in attributable fractions from mental health surveys, looking at lower socioeconomic status categories is in high-risk patterns; the rates of comorbidity and the order of onset of depression that is, depending on the use values for drinking in the culture, and alcohol disorders. For coronary heart disease, we modeled poor drinkers may see little point in wasting resources on the interaction of average volumes and patterns of drinking drinking that is not to intoxication. Poorer drinkers are also based on multilevel analyses that include temporal informa- likely to be less protected physically and socially from possible tion as covariates (Gmel, Rehm, and Frick 2003; Rehm and harm arising from drinking, such as injuries and chronic and others 2004). For the final estimates, we based alcohol- infectious diseases. Mäkelä (1999) finds that multiple dimen- attributable fractions on these multilevel results for all countries, sions of socioeconomic status are required to capture all the except for developed countries with relatively favorable drinking adverse interactions of socioeconomic status with alcohol- patterns (Australia, Japan, and countries in North America and related mortality. Western Europe),which are not discussed here because the focus A critical macroeconomic question is how a country's level is on developing countries. For injuries, we took a similar multi- of economic development is related to alcohol-related risks to level approach to quantify the interaction of the average volume health. The impact of alcohol on disease and mortality may be of consumption and patterns of drinking in determining more potent in countries with greater poverty and nutritional alcohol-attributable fractions (Rehm and others 2004). deficiencies (Isichei, Ikwuagu, and Egbuta 1993; Room and Thus the analysis includes the following major disease others 2002, 119­30). However, most of the risk relationships categories: between alcohol and disease have been derived from studies in established market economies, and the extent of systematic · chronic disease research is currently insufficient to allow quantification of this º cancer (mouth and oropharyngeal, esophageal, liver, phenomenon. As a result, the estimated disease burden cited female breast) here may be considered as a lower-bound estimate of the actual º neuropsychiatric diseases (AUDs, unipolar major depres- alcohol-attributable disease burden in developing countries. sion, epilepsy) º diabetes º cardiovascular diseases (hypertensive diseases, coronary BURDEN OF DISEASE RELATED TO HIGH-RISK heart disease, stroke) ALCOHOL USE º gastrointestinal diseases (cirrhosis of the liver) In the following sections, the procedures to estimate alcohol- º conditions arising during the perinatal period (low birthweight) related burden of disease are described, as well as the limita- · injury tions of the used approach. º unintentional injury (motor vehicle accidents, drowning, falls, poisonings, other unintentional injuries) Determining the Alcohol-Related Burden of Disease º intentional injury (self-inflicted injuries, homicide, other Table 47.2 breaks down alcohol-attributable disability-adjusted intentional injuries). life years (DALYs) by disease category and World Bank region using a constant 3 percent per year discount rate, but with no We did not include other disease categories that are clearly age weighting. Results differ from those of the CRA (Ezzati and alcohol-related, such as fetal alcohol syndrome, because the others 2002; Rehm and others 2004; WHO 2002) because of current analysis was based on the CRA and was, thus, limited the use of non-age-weighted DALYs.1 to the global burden-of-disease categories. Determining the Burden of Disease Related to High-Risk Alcohol Consumption Social Determinants of Exposure and Risk In determining the burden of disease related to high-risk alco- Alcohol-specific risks to health are in part determined and hol consumption, we first divided the burden of disease modified by social determinants. For example, Harrison and between chronic and acute disease. For chronic disease, we Gardiner (1999) find that for men age 25 to 69 in England and assume that almost the entire disease burden reported in the Wales in 1988­94, those in the lowest socioeconomic status CRA is associated with high-risk alcohol use. Indeed, the over- category, unskilled labor, had a 15-fold greater risk for alcohol- all disease burden in the CRA is an underestimate, because related mortality than professionals in the highest category drinking up to 20 grams per day of pure alcohol by females and 890 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others Table 47.2 Alcohol-Attributable DALYs by Disease Category and World Bank Region, 2001 (thousands of DALYs) Europe Latin America Middle East East Asia and and the Sub-Saharan and North and the South High-income Disease category Central Asia Caribbean Africa Africa Pacific Asia countries World Chronic disease Maternal and perinatal 12 7 39 1 2 29 6 105 conditions Cancer 526 296 635 25 2,820 189 1,103 5,594 Neuropsychiatric 2,159 3,315 1,035 89 4,726 1,444 4,752 17,600 Vascular 2,639 926 556 40 1,751 1,199 2,488 5,209 Other noncommunicable 1,175 739 504 27 997 306 1,153 5,126 diseases Subtotal chronic disease 6,511 5,283 2,769 182 10,296 3,167 4,526 33,634 Injury Unintentional 4,127 1,984 2,308 135 3,613 2,222 1,753 15,619 Intentional 1,822 1,872 1,074 9 927 567 571 6,755 Subtotal injury 5,949 3,856 3,382 144 4,540 2,789 2,324 22,374 Total DALYs attributable 12,460 9,139 6,151 326 14,836 5,956 6,850 56,008 to alcohol Total DALYs from all diseases 116,502 104,287 344,754 65,570 346,225 408,655 149,161 1,535,871 Proportion of DALYs 10.7 8.8 1.8 0.5 4.3 1.5 4.6 3.6 attributable to alcohol (percent) Source: Authors' calculations based on Rehm and others 2004 and WHO 2002. Note: Negative DALYs can occur because certain patterns of alcohol have cardio-protective effects. up to 40 grams per day of pure alcohol by males is globally could calculate the proportion of per capita consumption associated with a net beneficial effect in relation to chronic dis- related to high-risk drinking in each region, thereby determin- ease. However, this effect occurs mainly in countries with mod- ing the proportion of injury caused by high-risk drinking erate drinking patterns (Rehm, Sempos, and Trevisan 2003), (table 47.3). Together with our calculation of the chronic which tend to be high-income countries (Rehm, Rehn, and disease burden attributable to high-risk alcohol use, this per- others 2003). Although high-risk but regular drinking patterns centage enabled us to estimate the overall disease burden may also have some beneficial effects, such effects are not attributable to high-risk alcohol use: whereas 3.6 percent of the important in countries with binge drinking patterns. (For the global burden was attributable to alcohol drinking generally, association between alcohol and coronary heart disease, see 2.8 percent was attributable to high-risk drinking. McKee and Britton 1998; Puddey and others 1999; Rehm, Sempos, and Trevisan 2003; for consequences on modeling the Limitations of the CRA Approach regional burden of disease, see Rehm and others 2004.) The CRA's estimates of the global and regional alcohol-related For injuries, which are considered to be acute outcomes, we burden of disease are based on a number of assumptions, of started by separating out the proportion of injury not caused which the following are the most crucial: by high-risk drinking, which we accomplished by assuming that injuries are linearly related to per capita consumption · The estimates of per capita consumption and unrecorded (Rehm and others 2004).2 This assumption is probably conser- consumption for different countries do not contain sub- vative, because high-risk drinkers in countries with binge stantial measurement error. drinking patterns are likely to have more frequent and intensive · The distribution of consumption as derived from surveys is drinking occasions, and the risk of injury usually rises similar to actual distribution in the population. logarithmically with the amount of drinking on a specific occa- · The relationships between alcohol and chronic disease sion (see, for example, National Highway Traffic Safety derived from meta-analyses of cohort and case-control Administration 1992). Following this initial calculation, we studies are stable among countries and regions. Alcohol | 891 Table 47.3 DALYs Attributable to High-Risk Average Alcohol Consumption by Disease Category and Region, 2001 (thousands of DALYs) Europe and Latin America and Sub-Saharan East Asia and High-income Disease category Central Asia the Caribbean Africa the Pacific South Asia countries World Total chronic disease 6,510 5,283 2,770 10,296 3,167 4,526 33,634 Total injury 3,149 1,500 1,693 1,532 514 1,092 9,207 Total DALYs attributable 9,659 6,783 4,463 11,828 3,681 5,618 42,841 to high-risk alcohol consumption Total DALYs from 116,502 104,287 344,754 346,225 408,655 149,161 1,535,871 all diseases Proportion of DALYs 8.3 6.5 1.3 3.4 0.9 3.8 2.8 attributable to high-risk alcohol consumption Source: Authors' calculations based on Rehm and others 2004 and WHO 2002. Some evidence indicates that per capita consumption can be and cost-effectiveness of such interventions. Methods and reliably estimated, and information on this indicator is avail- analyses draw on Chisholm and others (2004), adjusted as nec- able for the vast majority of countries (Rehm, Rehn, and oth- essary to conform to the analytical standards of this volume, ers 2003). With respect to survey information, reliability and including the specification of all costs in U.S. dollars rather worldwide coverage are lower. However, because the overall than international dollars. volume of consumption and, thus, the average volume per capita are based on production and sales estimates, the measure of the volume of drinking overall can be considered reliable. Population Model These factors leave the stability of relationships between alco- We determined intervention effectiveness using a state transi- hol and chronic disease as the most crucial part of our esti- tion population model (Lauer and others 2003), which traces mates. Some indications suggest that relative risks may not be the development of a regional population taking into account the same in developing countries as in developed countries (for births, deaths, and the specified risk factor--in this case, high- example, for tobacco and lung cancers, see Liu and others risk alcohol use. In addition to population size and structure, 1998). Thus, the CRA's estimates may be biased, most likely the population model uses a number of epidemiological toward an overestimation of the impact of alcohol. parameters (incidence and prevalence, remission, and cause- One additional problem pertains to the usual epidemiolog- specific and residual rates of mortality) and assigns age- and ical approach as applied to alcohol. Most information about gender-specific health state valuations to both the disease in alcohol and chronic disease is derived from cohorts. Because question and to the nondiseased population. The output of the cohorts are frequently not representative of the population as a model is an estimate of the total healthy life years experienced whole, specific patterns of consumption such as binge drinking by the population over a lifetime period (100 years). are often not represented, and thus their influence cannot be We ran the model for a number of possible scenarios, analyzed (Rehm, Gmel, and others 2003). Unfortunately, the including no intervention at all (natural history), current inter- patterns most often missing are those that are the most detri- vention coverage, and scaled-up coverage of current and possi- mental with respect to health; thus, the impact of alcohol on ble new interventions. For the intervention scenarios, we used chronic diseases that are influenced by patterns of drinking an implementation period of 10 years for an intervention pro- other than average volumes is underestimated. gram (after which epidemiological rates return to their natural history levels), from which we derived the number of addition- al DALYs averted each year compared with the case for no INTERVENTIONS FOR REDUCING intervention at all. We discounted DALYs at 3 percent but did HIGH-RISK DRINKING not age weight them. The next two sections estimate the burden of disease attributa- ble to high-risk alcohol consumption that is currently being Effectiveness averted or could be averted by a range of personal and nonper- A number of interventions have been evaluated and shown sonal intervention strategies and calculate the expected costs to be effective in reducing alcohol use, yet their level of 892 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others implementation remains low in all but a handful of countries applied these estimated effects to the proportion of total deaths and their potential effect on population-level health has rarely and of years lived with a disability attributed to alcohol-related been assessed. By contrast, some interventions without clearly traffic accidents (table 47.4). established effects continue to be widely used, including, for example, mass media public information campaigns and Taxation on Alcoholic Beverages. Excise taxation on alcoholic school-based education aimed at reducing alcohol consump- beverages primarily affects the incidence of drinking through tion. Recent reviews of measures to reduce alcohol misuse have reduced consumption. Effects are measured in terms of price assessed the quality of the evidence for four types of interven- elasticity, which relates the change in consumption to the size tions specifically aimed at reducing high-risk alcohol use of the price increase (table 47.5). We derived price elasticities, (Babor and others 2003; Ludbrook and others 2002): adjusted downward by one-third to reflect possible reduced price responsiveness among high-risk drinkers, with reference · policy and legislative interventions, including taxation of to preferred type of alcoholic beverage (beer, wine, or spirits) alcohol sales, laws on drunk driving, restrictions on retail by region, built up from country-level data (WHO 2003b). outlets, and controls on advertising This downward adjustment of price elasticities for high-risk · measures to better enforce these interventions, such as ran- drinkers is a conservative approach; most of the literature dom breath testing of drivers found similar effects on high-risk and dependent drinkers as · mass media and other awareness campaigns on social users (Babor and others 2003; see also Farrell, · brief interventions with individual high-risk drinkers. Manning, and Finch 2003). Price elasticities ranged from 0.3 for the most preferred On the basis of these reviews, we included the following beverage category to 1.5 for the least preferred (Babor and strategies and intervention effects in our analysis: drinking- others 2003; Levy and Ornstein 1983). For a beer-drinking and-driving legislation and random breath testing, taxation region where wine is the second-most preferred beverage type, of alcoholic beverages, reduced hours of sale in retail outlets, for example, elasticities were set as follows: beer 0.3, wine and advertising bans (included as population-based interven- 1.0, distilled spirits 1.5. We performed sensitivity analysis tions) and so-called brief interventions (included as interven- around these elasticities. We evaluated three rates of excise tax tions aimed at personal behavior). We considered including on alcoholic beverages: the current rate of tax, a 25 percent one other intervention strategy--mass media or school-based increase over the current rate, and a 50 percent increase over the awareness campaigns--but omitted it in the final analysis on current rate. We adjusted estimated reductions in the incidence the grounds that evidence for its effectiveness was weak, both in of high-risk alcohol use by the observed or expected level of terms of methodological quality and in terms of its effect on unrecorded consumption resulting from illicit production and consumption (as opposed to transfer of information or knowl- smuggling (for instance, an estimated 35 percent of alcohol edge alone) (Babor and others 2003; Edwards and others 1994; consumption in Eastern Europe and Central Asia is unrecord- Foxcroft and others 2003; Foxcroft, Lister-Sharp, and Lowe ed, a proportion that was modeled to increase by 10 to 15 per- 1997; Ludbrook and others 2002). cent with the tax increases). In regions with rates of unrecorded consumption already greater than 50 percent (South Asia and Drunk-Driving Legislation and Random Breath Testing. Sub-Saharan Africa), tax increases can actually have a regressive Drunk-driving laws and reinforcement policies, such as impact on incidence if accompanied by a rise in the already high random breath testing of drivers, influence fatal and nonfatal level of unrecorded (and therefore untaxed) consumption. traffic injuries among both high-risk alcohol users and other members of the population, such as passengers and pedestri- Reduced Hours of Sale in Retail Outlets. Access to and avail- ans. We assessed two independent effects on alcohol-related ability of alcohol can be dramatically reduced by prohibition or traffic injuries, but note that evidence for these effects comes rationing, but implementing and sustaining such strategies from the developed countries, where road infrastructures and without adverse effects, such as black markets and poisonings driving patterns may differ significantly from those in the from home-produced alcohol, present considerable challenges. developing world. The first intervention was drunk-driving A more modest strategy is to reduce the hours of sale of retail laws, estimated to reduce traffic fatalities by 7 percent if widely outlets selling alcoholic beverages (for example, no sales for off- implemented across a region. The second was enforcement by premise consumption for a 24-hour period at the weekend), random breath testing, estimated to reduce fatalities by 6 to 10 which in Scandinavia has reduced consumption and alcohol- percent in regions partially implementing such a strategy and related harm (Leppänen 1979; Nordlund 1984; Norström and by 18 percent with wide implementation. The effect on non- Skog 2003). On the basis of these studies, we modeled a mod- fatal injuries was estimated to be a reduction of 15 percent est reduction of 1.5 to 3.0 percent in the incidence of high-risk (Peek-Asa 1999; Shults and others 2001). In each region, we drinking and 1.5 to 4.0 percent in alcohol-related traffic Alcohol | 893 Table 47.4 Effectiveness of Drinking-and-Driving Legislation and Its Enforcement (per 100,000 population) Effectiveness of Attributable fractions drinking-and-driving laws (per 100,000 deaths) and random breath testing Deaths Reduced Deaths attributed to Reduced years lost due WHO attributed to alcohol-related deaths to disability World Bank region subregion Sex traffic accidentsa traffic accidentsa (per 100,000) (per 100,000) Europe and Central Asia Europe B Male 1,473 657 141 77 Female 542 74 16 6 Europe C Male 2,197 1,396 299 193 Female 799 223 48 30 Latin America and Americas B Male 4,358 2,053 439 148 the Caribbean Female 1,514 220 47 12 Americas D Male 2,599 861 184 64 Female 1,093 101 22 6 Sub-Saharan Africa Africa D Male 2,159 417 89 43 Female 1,079 90 19 9 Africa E Male 2,075 803 172 107 Female 1,027 123 26 17 East Asia and the Pacific Southeast Asia B Male 7,809 1,993 427 164 Female 2,343 127 27 8 Western Pacific B Male 3,629 723 155 66 Female 1,790 157 34 12 South Asia Southeast Asia D Male 3,689 591 126 45 Female 1,451 53 11 3 Source: Deaths attributed to traffic accidents: WHO 2003a; deaths attributed to alcohol-related traffic accidents: Rehm and others 2004. B low child mortality, low adult mortality; C low child mortality, high adult mortality; D high child mortality, high adult mortality; E high child mortality, very high adult mortality. a. Percentages for all age groups combined shown here. fatalities, depending on the regional pattern of drinking, with Brief Interventions. We modeled brief interventions (such as the largest effects in regions with the highest levels of high-risk physician advice provided in primary health care settings), drinking occasions. which involve a small number of education sessions and psy- chosocial counseling, to influence the prevalence of high-risk Advertising Bans. Public health specialists are becoming drinking by increasing remission and reducing disability. increasingly interested in the effect of a comprehensive ban on Efficacy reviews of brief interventions reveal an estimated 13 alcohol advertising, including advertising on television and to 34 percent net reduction in consumption among high-risk through radio and billboards. However, available evidence drinkers (Higgins-Biddle and Babor 1996; Moyer and others from econometric studies suggests a modest effect on con- 2002; Whitlock and others 2004), which, if applied to the total sumption at best, even for a comprehensive ban, arguably population at risk, would reduce the overall prevalence of because of the continuing presence of other alcohol marketing high-risk drinking by 35 to 50 percent, equivalent to a 14 to strategies, such as product placement or event sponsorship 18 percent improvement in the rate of recovery over no treat- (Grube and Agostinelli 2000; Saffer 2000; Saffer and Dave ment at all. After taking into account adherence (70 percent) 2002). Here we consider the potential effects of a comprehen- and potential treatment coverage in the population (50 per- sive advertising ban (television, radio, and billboards) by mod- cent of high-risk drinkers), however, we estimated remission eling a 2 to 4 percent reduction in the incidence of high-risk rates to be between 4.9 and 6.4 percent higher than natural alcohol use, depending on regional drinking patterns. history rates. 894 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others Table 47.5 Effect of Taxation on the Incidence of High-Risk Alcohol Use Prevalence by preferred Rate of taxation beverage (percent) by preferred beverage (percent) Price increases (percent)b Elasticity Elasticity Elasticity Nonrecorded or Effect (percent)c World Bank WHO Most Next Least Most Next Least 0.3, most 1.0, next 1.5, least untaxed consumption region subregiona preferred preferred preferred preferred preferred preferred preferred preferred preferred (percent) Baseline Lowerd Uppere Europe and Europe B 0.45 0.30 0.25 0.29 0.13 0.12 (current rate) 0.04 0.08 0.11 0.34 (current rate) 0.05 0.03 0.06 Central Asia (spirits) (beer) (wine) 0.36 0.16 0.15 (25 percent increase) 0.05 0.09 0.13 0.37 (10 percent increase) 0.05 0.04 0.07 0.44 0.20 0.18 (50 percent increase) 0.06 0.11 0.15 0.39 (15 percent increase) 0.06 0.04 0.08 Europe C 0.68 0.21 0.11 0.65 0.13 0.25 (current rate) 0.08 0.08 0.20 0.36 (current rate) 0.06 0.04 0.08 (spirits) (beer) (wine) 0.81 0.16 0.31 (25 percent increase) 0.09 0.09 0.24 0.40 (10 percent increase) 0.06 0.05 0.09 0.98 0.20 0.38 (50 percent increase) 0.10 0.11 0.27 0.42 (15 percent increase) 0.07 0.05 0.09 Latin America Americas B 0.53 0.30 0.17 0.16 0.49 0.22 (current rate) 0.03 0.22 0.18 0.29 (current rate) 0.08 0.06 0.10 and the Caribbean (beer) (spirits) (wine) 0.20 0.61 0.28 (25 percent increase) 0.03 0.25 0.22 0.32 (10 percent increase) 0.09 0.06 0.12 0.24 0.74 0.33 (50 percent increase) 0.04 0.28 0.25 0.34 (15 percent increase) 0.10 0.07 0.13 Americas D 0.58 0.39 0.03 0.26 0.21 0.25 (current rate) 0.04 0.12 0.20 0.22 (current rate) 0.06 0.04 0.08 (spirits) (beer) (wine) 0.33 0.26 0.31 (25 percent increase) 0.05 0.14 0.24 0.24 (10 percent increase) 0.07 0.05 0.09 0.39 0.32 0.38 (50 percent increase) 0.06 0.16 0.27 0.25 (15 percent increase) 0.08 0.05 0.10 Sub-Saharan Africa D 0.79 0.16 0.05 0.36 0.41 0.35 (current rate) 0.05 0.19 0.26 0.77 (current rate) 0.02 0.01 0.03 Africa (beer) (spirits) (wine) 0.45 0.51 0.44 (25 percent increase) 0.06 0.23 0.30 0.85 (10 percent increase) 0.01 0.01 0.02 0.54 0.62 0.53 (50 percent increase) 0.07 0.25 0.34 0.89 (15 percent increase) 0.01 0.01 0.02 Africa E 0.49 0.30 0.21 0.28 0.50 0.38 (current rate) 0.04 0.22 0.28 0.47 (current rate) 0.08 0.06 0.10 (beer) (spirits) (wine) 0.35 0.63 0.48 (25 percent increase) 0.05 0.26 0.32 0.52 (10 percent increase) 0.08 0.06 0.11 0.42 0.75 0.57 (50 percent increase) 0.06 0.29 0.36 0.55 (15 percent increase) 0.09 0.06 0.11 East Asia and Southeast 0.88 0.12 0.00 0.30 0.40 0.00 (current rate) 0.05 0.19 0.00 0.36 (current rate) 0.04 0.03 0.05 the Pacific Asia B (spirits) (beer) (wine) 0.38 0.50 0.00 (25 percent increase) 0.05 0.22 0.00 0.39 (10 percent increase) 0.05 0.03 0.06 0.45 0.60 0.00 (50 percent increase) 0.06 0.25 0.00 0.41 (15 percent increase) 0.05 0.03 0.07 Western 0.88 0.11 0.01 0.17 0.09 0.11 (current rate) 0.03 0.06 0.10 0.27 (current rate) 0.02 0.02 0.03 Pacific B (spirits) (beer) (wine) 0.21 0.11 0.14 (25 percent increase) 0.04 0.07 0.12 0.32 (10 percent increase) 0.03 0.02 0.04 0.26 0.14 0.17 (50 percent increase) 0.04 0.08 0.14 0.31 (15 percent increase) 0.03 0.02 0.04 South Asia Southeast 0.89 0.11 0.00 0.40 0.25 0.00 (current rate) 0.06 0.13 0.00 0.79 (current rate) 0.01 0.01 0.02 Asia D (spirits) (beer) (wine) 0.50 0.31 0.00 (25 percent increase) 0.07 0.16 0.00 0.87 (10 percent increase) 0.01 0.01 0.01 0.60 0.38 0.00 (50 percent increase) 0.08 0.18 0.00 0.91 (15 percent increase) 0.01 0.01 0.01 Source: WHO 2003b. a. B low child mortality, low adult mortality; C low child mortality, high adult mortality; D high child mortality, high adult mortality; E high child mortality, very high adult mortality. b. Price rise caused by tax [percentage of tax/(1 percentage of tax)] elasticity 2/3 (high-risk drinkers less responsive). c. Effect sum of (prevalence price increase) for each beverage (1 percentage of unrecorded consumption). d. Lower-range elasticities 0.2, 0.7, 1.2. e. Upper-range elasticities 0.4, 1.3, 2.0. Costs males (the proportion for random breath testing rises to 80 to Costs covered in the analysis include program-level costs asso- 90 percent because of the higher proportion of deaths and ciated with running the intervention (such as administration, injuries attributed to traffic accidents among men). A clear dif- training, and media costs) and patient-level costs (such as costs ference is also apparent between regions with relatively high of primary care visits). Program-level costs include resource rates of high-risk alcohol use (that is, prevalence in the total inputs used in the production of an intervention at a level population greater than 5 percent) and regions with generally above that of the patient or providing facility, such as central low levels of high-risk drinking (that is, less than 2 percent). planning, policy, and administration functions, as well as As shown in table 47.6, in the three regions with a higher resources devoted to preventive programs, such as the enforce- prevalence of high-risk alcohol use--Europe and Central Asia, ment of drunk-driving legislation by police officers (Johns and Latin America and the Caribbean, and Sub-Saharan Africa-- others 2003). We derived estimated quantities of resources the most effective interventions were taxation and brief physi- required to implement each intervention for 10 years at the cian advice to individual high-risk drinkers, with each averting national, provincial, and district levels with reference to the more than 500 DALYs per million population per year. The region's prevailing characteristics--for example, the stability remaining control strategies--random breath testing, reduced and efficiency of tax systems, the volume of traffic (for breath access to alcoholic beverage retail outlets, and a comprehensive testing), and the strength of antidrinking sentiment as indicated advertising ban--mainly produced effects in the range of 200 by existing alcohol controls (advertising bans, restricted sales). to 400 DALYs averted per million population per year. In the In this analysis, patient-level resource inputs used in the provi- two regions with lower rates of high-risk drinking (particularly sion of a given health care intervention (for example, hospital among the female population), by contrast, the burden that is inpatient days, outpatient visits, medications, and laboratory avertable through taxation is very much reduced (10 to 100 tests) are relevant only to brief interventions. We estimated an DALYs averted per million population per year). In South Asia, average of four primary care visits per year for the intervention the most effective intervention is enforcement of drinking- itself, plus an additional 0.33 outpatient visits (20 percent and-driving laws by means of random breath testing, because 1.67 visits) and 0.25 inpatient days (5 percent 5 days) (see, of the higher rate of traffic-related injuries than elsewhere as for example, Fleming and others 2000). We applied these well as the low levels of high-risk drinking. patient-level resource inputs to the 50 percent of prevalent Population-Level Costs high-risk alcohol users in receipt of brief advice in year 1 and (because we model an enduring effect for 10 years) year 6 and Table 47.7 summarizes the costs and cost-effectiveness of each to the 50 percent of incident cases in years 2 to 5 and 7 to 10. intervention and of two combination strategies by region. The Note that, throughout, the costing does not include possible most costly interventions to implement in all regions were ran- offsetting revenues for the government, for instance, from dom breath testing and brief physician advice in primary care. drunk-driving convictions and, in particular, from the rev- The higher costs of brief advice stem from a combination of enues likely to result from increased alcohol taxes. patient-level costs in the provision of the intervention itself (an Unit costs and prices of program- and patient-level resource average annual cost of US$7 to US$20 per treated case), plus inputs include the salaries of central administrators; the capital program costs associated with administration and training pri- costs of vehicles, offices, and furniture; and the cost per outpa- mary care providers (15 to 40 percent of total costs). Random tient visit (see chapter 7 for an overview of the costing method- breath testing is also a relatively resource-intensive intervention ology, plus prices by World Bank region). All costs are to implement because of the need for regular sobriety check- expressed in U.S. dollars for 2001 and are discounted at an points administered by law enforcement officers. Other inter- annual rate of 3 percent. ventions, including taxation, had a per capita cost in the range US$0.02 to US$0.13, depending in part on the efficiency of the tax collection system and the degree of antidrinking sentiment. COST-EFFECTIVENESS OF INTERVENTIONS In the following section, we provide results relating to the Population-Level Cost-Effectiveness population-level health effects, costs, and cost-effectiveness of Compared with doing nothing, taxation is the most cost- the evidence-based interventions previously reviewed. effective population-level strategy in Europe and Central Asia, Latin America and the Caribbean, and Sub-Saharan Africa, the three regions with a relatively high prevalence of high-risk Population-Level Effects drinking (table 47.7). At the current rate of tax, for example, Except for random breath testing, two-thirds of the total pop- each DALY averted costs US$104 to US$225, equivalent to ulation-level health gain from these interventions was among 4,435 to 9,633 DALYs averted per US$1 million expenditure. 896 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others Table 47.6 Population-Level Effects of Interventions to Reduce High-Risk Alcohol Use by World Bank Region Europe and Latin America East Asia Coveragea Central and the Sub-Saharan and the (percent) Asia Caribbean Africa Pacific South Asia Burden of disease (DALYs/million population) 20,241 12,894 6,685 6,263 2,652 Total effect (DALYs averted/ million population/year) Excise tax on alcoholic beverages 0.95 685 586 697 83 13 (current situation) Excise tax on alcoholic beverages 0.95 756 654 724 96 10 (25 percent increase) Excise tax on alcoholic beverages 0.95 828 719 764 109 8 (50 percent increase) Reduced access to alcoholic beverage 0.95 441 287 386 203 32 retail outlets Comprehensive advertising ban 0.95 395 243 406 226 20 on alcohol Random breath testing of motor 0.80 284 307 197 181 125 vehicle drivers Brief advice to heavy drinkers by 0.50 1,328 713 539 362 80 a primary care physician Combination: highest tax 2,048 1,360 1,237 447 83 brief advice Combination: highest tax 2,551 1,784 1,715 790 210 advertising ban random breath testing brief advice Reduction in current burden (percent) Excise tax on alcoholic beverages 0.95 0.03 0.05 0.10 0.01 0.01 (current situation) Excise tax on alcoholic beverages 0.95 0.04 0.05 0.11 0.02 0.00 (25 percent increase) Excise tax on alcoholic beverages 0.95 0.04 0.06 0.11 0.02 0.00 (50 percent increase) Reduced access to alcoholic 0.95 0.02 0.02 0.06 0.03 0.01 beverage retail outlets Comprehensive advertising ban 0.95 0.02 0.02 0.06 0.04 0.01 on alcohol Random breath testing of motor 0.80 0.01 0.02 0.03 0.03 0.05 vehicle drivers Brief advice to heavy drinkers by 0.50 0.07 0.06 0.08 0.06 0.03 a primary care physician Combination: highest tax 0.10 0.11 0.19 0.07 0.03 brief advice Combination: highest tax advertising 0.13 0.14 0.26 0.13 0.08 ban random breath testing brief advice Source: Chisholm and others 2004. a. Refers to the modeled percentage of all high-risk drinkers exposed to the intervention. Advertising bans had a cost per unit of effect similar to that of DALY, approximately 500 to 1,000 DALYs averted per US$1 reduced access to sales outlets, US$134 to US$380, equivalent million dollars expenditure. Brief physician advice provided in to 2,631 to 7,442 averted DALYs per US$1 million dollars primary care settings had an average cost-effectiveness in the expenditure, whereas random breath testing had the highest range of US$204 to US$502 per DALY averted, or close to 2,000 estimated cost per DALY averted: US$973 to US$1,856 per to 5,000 averted DALYs for every US$1 million expenditure. Alcohol | 897 Table 47.7 Costs and Cost-Effectiveness of Interventions to Reduce High-Risk Alcohol Use by World Bank Region Coveragea Europe and Latin America Sub-Saharan East Asia and (percent) Central Asia and the Caribbean Africa the Pacific South Asia Total cost (US$ million/year/million population) Excise tax on alcoholic beverages (current situation) 0.95 0.10 0.13 0.07 0.04 0.04 Excise tax on alcoholic beverages (25 percent increase) 0.95 0.10 0.13 0.07 0.04 0.04 Excise tax on alcoholic beverages (50 percent increase) 0.95 0.10 0.13 0.07 0.04 0.04 Reduced access to alcoholic beverage retail outlets 0.95 0.10 0.10 0.06 0.03 0.03 Comprehensive advertising ban on alcohol 0.95 0.07 0.09 0.05 0.03 0.02 Random breath testing of motor vehicle drivers 0.80 0.53 0.47 0.19 0.18 0.07 Brief advice to heavy drinkers by a primary care physician 0.50 0.36 0.36 0.11 0.08 0.04 Combination: highest tax brief advice 0.44 0.48 0.18 0.12 0.07 Combination: highest tax advertising ban random 0.97 0.97 0.39 0.30 0.15 breath testing brief advice Cost-effectiveness relative to no intervention (US$/DALY averted) Excise tax on alcoholic beverages (current situation) 0.95 141 225 104 516 2,671 Excise tax on alcoholic beverages (25 percent increase) 0.95 127 202 100 447 3,654 Excise tax on alcoholic beverages (50 percent increase) 0.95 116 184 95 394 4,641 Reduced access to alcoholic beverage retail outlets 0.95 216 340 152 146 827 Comprehensive advertising ban on alcohol 0.95 185 380 134 123 1,123 Random breath testing of motor vehicle drivers 0.80 1,856 1,542 973 984 531 Brief advice to heavy drinkers by a primary care physician 0.50 270 502 204 224 462 Combination: highest tax brief advice 216 350 143 269 845 Combination: highest tax advertising ban 381 546 229 383 707 random breath testing brief advice DALYs averted/US$ million expenditure Excise tax on alcoholic beverages (current situation) 0.95 7,107 4,435 9,633 1,937 374 Excise tax on alcoholic beverages (25 percent increase) 0.95 7,847 4,953 10,007 2,239 274 Excise tax on alcoholic beverages (50 percent increase) 0.95 8,590 5,442 10,553 2,536 215 Reduced access to alcoholic beverage retail outlets 0.95 4,638 2,940 6,580 6,856 1,209 Comprehensive advertising ban on alcohol 0.95 5,417 2,631 7,442 8,139 891 Random breath testing of motor vehicle drivers 0.80 539 648 1,027 1,016 1,882 Brief advice to heavy drinkers by a primary care physician 0.50 3,705 1,992 4,891 4,460 2,163 Combination: highest tax brief advice 4,627 2,859 7,016 3,718 1,184 Combination: highest tax advertising ban random 2,621 1,833 4,364 2,612 1,415 breath testing brief advice Source: Chisholm and others 2004. a. Refers to the modeled percentage of all high-risk drinkers exposed to the intervention. Starting from the current situation in these regions, the physician advice in primary care. Even a multifaceted strategy most efficient strategies for reducing high-risk alcohol use made up of an increase in taxation plus full implementation of would be tax increases (additional gains are obtained at virtu- the other interventions considered here has a favorable ratio of ally no extra cost because the costs of tax administration and costs to health benefits. enforcement remain relatively constant whatever the rate of In East Asia and the Pacific and South Asia, the two regions tax), followed by the introduction or escalation of comprehen- with lower rates of high-risk alcohol use, a comparison of sive advertising bans on alcohol products, reduced access to intervention costs and effects to a no-intervention scenario retail outlets, and the provision of brief interventions such as reveals that current practice--namely, excise taxes on alcoholic 898 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others beverages--is not the most efficient response to the existing population-wide measures, such as taxation of alcoholic bever- burden of alcohol use. The reduced efficiency of taxation in ages. Of these, taxation has the most sizable and least resource- these lower-prevalence regions is related both to the distribu- intensive effect on reducing the avertable burden of high-risk tion of the fixed costs of administering and enforcing alcohol alcohol use. In regions where high-risk alcohol use represents tax legislation across a smaller target population of drinkers less of a public health burden, targeted approaches such as brief and to underlying drinking patterns: more than 85 percent of physician advice as well as other intervention strategies that all alcohol consumption falls into a single preferred drink cat- restrict the supply or promotion of alcoholic beverages appear egory, spirits, which therefore diminishes the scope for reduc- to be the most cost-effective mechanisms, although greater ing the consumption of less preferred but more elastic cate- empirical support for the efficacy of these interventions in gories of alcoholic beverages. In South Asia, targeted strategies these localities is clearly needed before considering their wide- such as brief physician advice and random breath testing have spread implementation. the lowest cost per DALY averted (around US$500), while tax- Even though sectoral cost-effectiveness analysis pursues a ation policies are the most expensive at more than US$2,500 societal perspective, considerable challenges remain in relation per DALY averted. In East Asia and the Pacific, the most cost- to the appropriate measurement of certain societal costs and effective interventions are brief physician advice, a comprehen- effects that fall outside the boundaries of the health system. sive ban on advertising, and reduced access to retail outlets Therefore, this analysis has not been able to successfully cap- (below US$250 per DALY averted). ture potential reductions in workforce and household produc- tivity losses among high-risk drinkers, nor does it incorporate the economic costs associated with alcohol-related crime, vio- Implications and Limitations of Sectoral lence, and harm reduction. It also does not value the time spent Cost-Effectiveness Analyses by patients and informal caregivers in seeking or providing care This cost-effectiveness analysis offers a new approach to gener- and support. Including these modest additional costs and sub- ating economic evidence that can inform public health policy stantial incremental effects is likely to improve the cost- on alcohol in a wide range of cultural and epidemiological effectiveness ratios of all interventions, but to a variable and settings (Chisholm and others 2004). Resulting estimates of currently unknown extent. cost-effectiveness can inform policy makers not only by deter- mining the efficiency of existing resource allocation and practices, but also by identifying priorities for future alcohol ECONOMIC BENEFITS OF INTERVENTIONS control strategies. Furthermore, use of a common methodology enables comparison with cost per DALY estimates for other By design, estimates of the burden of alcohol do not include risk factors or disease entities, which may constitute an impor- most social harm and harm to people other than the drinker; tant argument when considering priorities for the allocation however, the burden of social problems from drinking can be of scarce health care resources. However, the application of a at least as significant as the health burden. The burden attrib- broad sectoral approach using entire regions as the unit of utable to alcohol in the CRA estimates is actually a substantial analysis clearly limits the approach's use in specific country underestimate of the full harm alcohol imposes on human contexts, where demographic or epidemiological characteris- welfare. The estimates reported earlier reflect primarily the tics, as well as treatment costs and coverage, may not coincide chronic disease and injury effects of drinking. Because the CRA with estimates for the region as a whole. In addition, extrapo- focused on disease and disability, the estimates were not lation of the extent of intervention effects from relatively designed to take account of the social harm and problems that information-rich countries to other sociocultural settings are particular to alcohol and that result for the drinker and for lessens the precision of derived estimates of population-level others as a consequence of a person's drinking (Klingemann health gains. and Gmel 2001). These problems are quite prevalent in many Although an ongoing analytical step is to calibrate results populations (Room and others 2003) and are also affected by at the country level, the primary purpose and utility of the the interventions listed earlier. sectoral approach is to identify interventions that are clearly Some information on the relative burden of alcohol for cost-effective as opposed to those that clearly do not seem to social services versus health services is available for a handful of offer good value for money. In this respect, the primary con- societies. In an estimate of the staffing and service costs attrib- clusion to be drawn from the analysis is that in regions with utable to alcohol in different service systems in Scotland for high or moderate rates of high-risk alcohol use, a number of fiscal year 2001/02, for instance, health services accounted for intervention strategies can have a notable effect on population only 21 percent of the estimated costs, whereas social services health, including both individual-based interventions, such as accounted for 19 percent, and criminal justice and fire services brief physician advice at the primary care level, as well as accounted for 60 percent (Catalyst Health Consultants 2001, 3). Alcohol | 899 If those estimates are used as a rough gauge of the burden to the 1993 rate, and the rate rose again slightly in 1996 and society, the illness and disability burden of alcohol may thus 1997. Medical specialists in Mauritius agree that patients constitute half or less of the total burden when social problems with alcohol problems account for an increasing portion are also taken into consideration. of admissions in general medical wards and now represent Thus, policies that affect the levels of alcohol-related health between 40 and 50 percent of bed occupancy (Abdool and social harm not only are a matter of intervening to save 1998). people from the detrimental effects of their own behavior, but · Age-adjusted death rates per 100,000 population for chronic also potentially have a broader effect on the health and well- liver disease and cirrhosis rose from 32.8 for males and being of families and of associates of drinkers. This issue is 4.0 for females in 1993 to 42.7 for males and 5.3 for females especially relevant for women: even though men predominate in 1996 (WHO 1999, 2000). among high-risk drinkers worldwide (Rehm and others 2004; Room and others 2002), women bear much of the burden of Even though available statistics are limited, the reduction in harm from others' drinking, not only in such forms as domes- alcohol import taxes clearly had a substantial negative effect on tic violence, but also in such forms as diversion of family the health of Mauritians. Thus, the government's 1997 call for resources from greater needs. control measures for alcohol--specifically, new permits for licensed premises, increased excise duties on alcohol, and limi- tations on bars' opening hours--was not surprising. Alcohol IMPLEMENTATION OF CONTROL STRATEGIES: taxes were increased somewhat in the 1999/2000 budget (U.S. LESSONS OF EXPERIENCE Department of State 1999). However, an analysis by World Bank staff that did not take health effects into account called The following paragraphs provide a few concrete examples of for further reductions in maximum tariff rates, identifying interventions or policy changes that illustrate the actual imple- Mauritius as having an antitrade bias on the basis of the struc- mentation and effects of control strategies in developing soci- ture of its alcohol and tobacco taxes (Hinkle and Herrou- eties (the examples are taken from Room and others 2002). Aragon 2001). Wallace and Bird (2003) suggest the following general prin- Tax Rate Reduction and the Resulting Disease ciples for setting and collecting alcohol taxes in the context of Burden in Mauritius developing societies from the perspective of revenue genera- Mauritius, an island nation in the Indian Ocean, has a popula- tion rather than public health (see also Tax Policy Chief tion of about 1 million. These people are of Indian, African, Directorate 2002): European, and Chinese origin. By religious affiliation, 53 per- cent are Hindu, 29 percent are Christian, and 17 percent are · Countries around the world need revenues they can raise Muslim. Tourism is the third-ranked industry in terms of hard relatively efficiently, but this need is probably more critical currency earnings. In June 1994, the government drastically in the case of developing nations. That said, alcohol taxes are lowered customs duties on imported alcoholic beverages to probably a good bet for future revenues. 80 percent from rates that had ranged from 200 percent for · Excise taxes on alcohol should be set by alcohol content, wine to 600 percent for whisky and other spirits (Abdool 1998). rather than as a percentage of the price. The government made the change under pressure from the · Tax rates should be logically defined so that alcoholic bever- hotel industry, which claimed that tourists were not purchasing ages with similar alcohol content are treated similarly, with enough alcohol because of its high prices (Lee 2001). Other rea- stronger alcohol beverages taxed more heavily. sons given for the change were to reduce unofficial imports · Analyses of revenue-maximizing rates should be conducted from abroad and to make better, more refined alcoholic bever- to determine a range of tax rates that is likely to maximize ages available to the local population. Despite little evidence to government revenues. support the view, there were claims in the public discussion that · Tax systems should be designed to be as simple as possible better-quality alcohol would result in fewer health problems. to allow for the maximum efficiency of tax administration. The effects of the change were felt mainly by Mauritians rather than tourists, as follows: Reduced Access through Locational Prohibition in Brazil · Arrests for driving with blood alcohol over the legal limit The second example involves the institution of a new control made primarily in connection with traffic crashes increased on alcohol availability in an environment where it is likely to by 23 percent between 1993 and 1997. be combined with driving. Although we have modeled the · Admissions of alcoholism cases to the island's psychiatric effects of another, better studied availability control (namely, hospital shot up in 1994. The 1995 rate was more than twice closing on a weekend day), a wide variety of possible 900 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others measures is available to control the time and place of alcohol Drunk-Driving Enforcement in South Africa purchase or drinking (Babor and others 2003; Room and oth- No published studies are available of the implementation of ers 2002). Even though in this case the particular control was random breath testing in a developing country. However, some extremely limited in scope, it appears to have had measurable data are available on a campaign to increase drunk-driving effects. enforcement in South Africa, a strategy that has often shown Traffic deaths are an important source of mortality in Brazil, some effects, although weaker and less lasting than those of amounting to 3.6 percent of overall mortality. The few available random breath testing. studies suggest that alcohol plays a significant role in traffic The minister of finance launched a short-term campaign, casualties. For instance, one study in São Paulo found positive ARRIVE ALIVE, for the period October 1997 to January 1998, blood alcohol levels in 72 percent of pedestrian deaths and in response to the high rate of traffic fatalities and injuries. The 32 percent of driver and passenger deaths of persons age 13 and campaign's main aim was to mobilize all available traffic polic- older (Carlini-Cotrim and Chasin 2000). ing, control, and education resources to reduce traffic accidents In 1985, motivated by concern about alcohol and impaired on South African roads by at least 5 percent, especially in the driving and about the lax enforcement of drinking and driv- Western Cape, Gauteng, and KwaZulu Natal provinces, because ing laws, a conservative party politician from the state of São 75 percent of all accidents occurred in those provinces. The Paulo introduced legislation to prohibit alcohol sales in com- ARRIVE ALIVE campaign targeted, in turn, what were consid- mercial facilities that had access to state highways. Even ered the three critical factors having the greatest impact on though the bill passed in the legislature, its implementation injuries: failing to wear seat belts, drinking and driving, and was delayed by the state's alcohol producers and commercial speeding. Unofficially, the campaign came to be called "belts, and industrial federations, which claimed that the law would booze, and bats out of hell." be a barrier to improved facilities for travelers, would encour- As many of the parties interested in road safety as possible age people to carry bottles in their cars, and would restrict were involved, with funding drawn from a variety of govern- individual freedoms. Discussion in the press was also generally ment and business sources. The campaign included a number unsympathetic. In August 1988, however, a new state gover- of components particularly relevant to alcohol use. New equip- nor from the same party implemented the law. At that time, ment purchased by the provinces included alcohol screening the press was slightly more supportive. Since then, the law has devices, alcohol evidentiary units, and so-called booze buses been on the books, although site visits to restaurants and (vehicles containing all the technology needed to check breath snack bars along a state highway in 1997 suggested a low level and blood alcohol levels). Sentences were increased to under- of compliance. In 1995, another legislator from the same line the point that traffic violations are serious offenses, with a party proposed repealing the law on the grounds that no three-month suspension of a driver's license and an increased studies proved that it lowered traffic accidents. The repeal maximum fine for a first conviction for drunk driving and with passed the legislature without significant public debate, but license suspension for one to five years for second offenders. the state governor vetoed it. Undaunted, the same legislator Traffic supervisors underwent intensive training courses before then proposed a law to criminalize buying as well as selling the start of the campaign. alcohol along state highways. That law passed but has not yet Because the aim of the campaign included educating road been implemented. users, advertisements covering aspects of the campaign were A study by Carlini-Cotrim, Pinsky, and Serrano Barbosa run on the radio, on television, and in movie theaters through- (1998) assesses the effects of the intervention. Finding data for out the country. Supplements were published in national and a controlled study comparing traffic casualties on state high- provincial newspapers. Private companies, such as a supermar- ways with casualties on federal highways, which were unaffect- ket chain and an automobile manufacturer, also promoted ed by the law, proved impossible. The best data available were the campaign. A national transportation center, established to on crashes and crashes resulting in injuries per 10,000 vehicles collect and collate data from local and provincial authorities, traveling on three short highway systems administered by a pri- operated for 12 hours every day throughout the campaign. vate agency. Linear regressions on those data for 1983­93 Traffic authorities staffed an additional 80 roadside communi- showed that the law had made a significant difference in the cations points, and at selected points on certain routes, road number of accidents resulting in injuries on all three roads and signs were erected and updated to display the percentage of a significant difference in all accidents on two of the roads. A speed limit and drinking-and-driving violations and the rate of separate analysis on estimated accidents and accidents with seat belt use in that area. injuries per 10,000 vehicles in two geographic areas of the state A total of 776 enforcement points were set up on 195 strate- did not show significant effects of the law. Overall, the analyses gic routes in the selected provinces. Posters, pamphlets, key do provide some support for the law having a beneficial effect rings, and license decals were produced for distribution and on the rate of traffic casualties. Alcohol | 901 display at roadblocks in the three provinces. Between October brief intervention techniques. The project's hypothesis was 1, 1997, and January 17, 1998, 6,674 notices of prosecution that the amount of change in alcohol consumption over a were issued for alcohol-related traffic offenses, 83 percent of nine-month period would be proportional to the intensity of which were issued in the intervention provinces. the intervention provided by a trained primary health care Comparison studies showed a decrease in the drinking rate professional. The results showed a significant effect of inter- of drivers in the three provinces, whereas the other six ventions on both consumption and intensity of drinking provinces, as a group, showed an increase. KwaZulu Natal had among males, but the intensity of the intervention was not the lowest drinking rate of all drivers throughout the campaign related to the amount of change in drinking behavior; 5 min- (3 to 7 percent), and the Western Cape had the most dramatic utes of simple advice turned out to be as effective as 20 min- decrease (from 12.0 to 9.3 percent in October). Except for in utes of brief counseling (Babor and Grant 1992). The female Gauteng, the drinking rates for pedestrians decreased from sample was too small for the results to attain significance, and more than 15 percent to less than 7 percent. Overall, during the the intervention did not significantly affect men's frequency of months targeted, drinking-and-driving rates decreased by 2 to dependence symptoms, problems related to alcohol, or con- 4 percent, as measured by breath testing. The total number of cern expressed by others (WHO Brief Intervention Study crashes decreased by 8 percent, and fatalities dropped by 9 per- Group 1996). cent. The ratio of benefits to costs for the intervention was esti- The findings suggest that in a population of high-risk mated as 4 to 1, based on an investment in the campaign of drinkers, behavior change is more a function of motivational R50 million, or about US$4.4 million at 2002 rates (ARRIVE factors and social influence than of the moderation skills and ALIVE Campaign 2000). social learning techniques that behavioral self-control training Despite the potential inconvenience of roadblocks and other packages typically use. Changes in drinking were not attribut- enforcement activities, the public generally perceived the cam- able solely to the small number of patients who achieved an paign positively. The liquor retail and hospitality industries abstinence goal, nor to the small number who gave up daily or complained about decreased sales, and tow truck operators almost daily drinking. Rather, changes seem to have been dis- complained about reduced business. tributed across a broad spectrum of the drinkers who reduced Even though driver behavior improved during the focus their consumption by small, but clinically meaningful, months, violations often increased after the focus was changed, amounts. for example, from drunk driving to seat belt use. This finding emphasizes the need for sustained enforcement as opposed to ad hoc campaigns. (This example was summarized from ARRIVE RESEARCH AND DEVELOPMENT AGENDA ALIVE Campaign 2000 and Cerff and Plüddemann 1998.) Research and development needs in the area of alcohol con- sumption are large and multidimensional. The work reported Implementation of Brief Interventions in Several in this chapter represents best estimates from the available data, Developing Countries some of which were developed to fill the needs of the analysis; In the first phase of the WHO Collaborative Project on however, we cite few figures for the developing world for which Identification and Management of Alcohol Related Problems we can say that the underlying data are so good that they could (Saunders and Aasland 1987), a screening measure suitable for not usefully be improved. Nevertheless, more and better data use in both developing and developed countries--the alcohol- are available on alcohol than on many other health topics. use disorders identification test--was developed to identify The health and social burdens of alcohol are clearly people at risk for alcohol problems among those attending pri- extremely large in most developing societies. Thus, the most mary health care services. In the second phase, a multicenter urgent focus should be on development and evaluation proj- clinical trial of brief intervention procedures designed to ects to study the outcomes of various policy and program reduce the health risks associated with hazardous alcohol use interventions. The projects must necessarily be attuned to what was carried out in primary health care settings in Australia, is politically feasible in a particular time and place. They are Bulgaria, Costa Rica, Kenya, Mexico, Norway, the Soviet Union, likely to include natural experiment studies, where the research the United Kingdom, the United States, and Zimbabwe (Babor tracks the effects of changes that governments undertake, and others 1994). whether those changes are expected to increase or to decrease The project's aims were to study the influence of simple the extent of alcohol problems. Where possible, the projects advice and brief counseling, to examine the moderating role of should include experimental and quasi-experimental studies, reduced consumption on the prevention of alcohol-related whereby the effects of a change at intervention sites are studied problems, and to evaluate the cross-national generalizability of in comparison to outcomes at control sites, with random 902 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others assignment where possible. Costing data should be included to CONCLUSION permit cost-effectiveness analysis. Also important are process studies--that is, research on The burden of disease attributable to alcohol in the developing how policy makers decide on policy changes, how they world is considerable, and the social harm not accounted for in implement them, and what the reactions and sequelae are. For this analysis increases the costs. However, known interventions example, deciding to introduce a new alcohol tax may be the can reduce the burden by up to 25 percent, depending on the easiest part of an initiative, but actually implementing it in a region of the world. Compared with other interventions in the developing society with a great deal of unrecorded alcohol in health care field, these interventions are quite cost-effective, but the informal market and with poorly guarded borders may be given the nature of many of the interventions, caution is needed. much more difficult. Currently, no international mechanism or In particular, the following recommendations can be given: nexus exists for developing and disseminating practical knowl- edge about implementing effective alcohol control strategies · Interventions and research about their effectiveness are between developing countries. based mostly on experiences from established market At this time, nearly all studies of alcohol interventions come economies; thus, the levels of effectiveness estimated in our from a limited range of developed countries. Extending knowl- analysis should be treated as broad indications. Depending edge and experience in and between developing societies is on actual methods of implementation, individual interven- urgently needed. tions could be more or less effective. A secondary need, but one that is also important, is to · Interventions should ideally be modeled on the basis of the extend the epidemiological database in developing societies on specific environment (that is, countries or provinces) and levels and patterns of drinking and on the health and social on the harm distribution in the respective environment, consequences of drinking. To this end, better estimation of including social harm. unrecorded alcohol consumption is needed. Which dimensions · General principles, such as restricting access to alcohol, of drinking patterns matter for what kinds of outcomes needs should be attuned to local cultures and traditions when to be studied in the context of different kinds of developing interventions are formulated. societies. Studies of the effects that the interaction of drinking · Population measures must take into account the complex levels and patterns with poverty and social exclusion have on interplay of public opinion and balance the interests of dif- the extent of alcohol-related problems are also necessary. ferent groups and stakeholders with conflicting values. One Because most of our knowledge about the health effects of of these stakeholders is, of course, the alcohol industry. drinking concerns mortality, studies of alcohol's role in various kinds of morbidity should be emphasized. Another area where If policy makers keep these principles in mind, reducing the data are lacking is the social harm arising from drinking, for alcohol-related health burden could be one of the most cost- which we cannot presently make the kinds of estimates that are effective targets of population-level health programs in devel- possible to make for harm to health. Developing and reaching oping countries. This target is even more attractive because the consensus on ways to measure the social harm caused by drink- measures discussed will also reduce the alcohol-related social ing is a substantial agenda for both the developed and the burden, thereby further contributing to development. developing world. Developing the epidemiological database can provide clues NOTES to etiology to be pursued further by biomedical and social 1. The global burden of disease attributable to alcohol is 4.0 percent researchers and, thus, offers hope for the development of new using age-weighted DALYs and 3.6 percent using non-age-weighted treatments or preventive interventions. It can provide informa- DALYs. This difference can be explained by the many alcohol-attributable tion on the distribution of drinking patterns and problems in outcomes occurring during adolescence and young adulthood, when age subpopulations that can be used to guide targeting and pre- weights are higher. 2. The CRA defined per capita consumption as average consumption of vention and treatment priorities. However, from a short-term pure alcohol per person 15 years old or older. policy perspective, the most important function of developing the epidemiological database in a particular country may be REFERENCES providing a base for creating political will for action. For exam- ple, the development of devices to measure blood and breath Abdool, R. 1998. "Alcohol Policy and Problems in Mauritius." Paper pre- alcohol and the collection of data on drinking and driving that pared for the World Health Organization Alcohol Policy in Developing Societies Project, World Health Organization, Geneva. they made possible were prerequisites for developing the polit- Apte, M. V., J. S. Wilson, and M. A. Korsten. 1997. "Alcohol-Related ical will and support for implementing drinking-and-driving Pancreatic Damage: Mechanisms and Treatment." Alcohol Health and countermeasures in industrial countries. Research World 21 (1): 13­20. Alcohol | 903 ARRIVE ALIVE Campaign. 2000. "ARRIVE ALIVE Safety Campaign." "Selected Major Risk Factors and Global and Regional Burden of Pretoria: Department of Transport. http://www.transport.gov.za/ Disease." Lancet 360: 1347­60. projects/arrive/presentation/slide41.htm. Farrell, S., W. G. Manning, and M. D. Finch. 2003. "Alcohol Dependence Babor, T., F. Caetano, S. Casswell, G. Edwards, N. Giesbrecht, K. Graham, and the Price of Alcoholic Beverages." Journal of Health Economics and others. 2003. Alcohol: No Ordinary Commodity--A Consumer's 22 (1): 117­47. Guide to Public Policy. Oxford, U.K.: Oxford University Press. Fleming, M. F., M. P. Mundt, M. T. French, L. B. Manwell, E. A. Stauffacher, Babor, T. F., and M. Grant, eds. 1992. Project on Identification and K. L. Barry, and others. 2000. "Benefit-Cost Analysis of Brief Physician Management of Alcohol-Related Problems: Report on Phase II--A Advice with Problem Drinkers in Primary Care Settings." Medical Care Randomized Clinical Trial of Brief Interventions in Primary Health Care. 38: 7­18. Geneva: World Health Organization, Program on Substance Abuse. Foxcroft, D. R., D. Ireland, D. J. Lister-Sharp, G. Lowe, and R. Breen. 2003. Babor, T. F., M. Grant, W. Acuda, F. H. Burns, C. Campillo, F. K. Del Boco, "Longer-Term Primary Prevention for Alcohol Misuse in Young and others. 1994. "Randomized Clinical Trial of Brief Interventions People: A Systematic Review." Addiction 98: 397­411. in Primary Health Care: Summary of a WHO Project (with Foxcroft, D. R., D. Lister-Sharp, and G. Lowe. 1997. "Alcohol Misuse Commentaries and a Response)." Addiction 89: 657­78. Prevention for Young People: A Systematic Review Reveals Bruun, K., G. Edwards, M. Lumio, K. Mäkelä, L. Pan, R. E. Popham, Methodological Concerns and Lack of Reliable Evidence of and others. 1975. Alcohol Control Policies in Public Health Perspective. Effectiveness." Addiction 92: 531­38. Helsinki: Finnish Foundation for Alcohol Studies. Gmel, G., E. Gutjahr, and J. Rehm. 2003. "How Stable Is the Risk Curve Carlini-Cotrim, B., and A. A. da M. Chasin. 2000. "Blood Alcohol Content between Alcohol and All-Cause Mortality and What Factors Influence and Death from Fatal Injury: A Study in the Metropolitan Area of São the Shape? A Precision-Weighted Hierarchical Meta-Analysis." Paulo, Brazil." Journal of Psychoactive Drugs 32: 269­75. European Journal of Epidemiology 18 (7): 631­42. Carlini-Cotrim, B., I. Pinsky, and M. T. Serrano Barbosa. 1998. Alcohol Gmel, G., J. Rehm, and U. Frick. 2003. "Trinkmuster, Pro-Kopf-Konsum Availability Restrictions in Developing Societies: The Case of São Paulo von Alkohol und koronare Mortalität." Sucht 49 (2): 95­104. Highways, Brazil. Report prepared for the WHO Alcohol Policy in Grant, B. F., D. A. Dawson, F. S. Stinson, S. P. Chou, M. C. Dufour, R. P. Developing Societies Project, World Health Organization, Geneva. Pickering, and others. 2004. "The 12-Month Prevalence and Trends in Catalyst Health Consultants. 2001. Alcohol Misuse in Scotland: Trends and DSM-IV Alcohol Abuse and Dependence: United States, 1991­1992 Costs--Final Report. Northwort, U.K.: Catalyst Health Consultants. and 2001­2002." Drug and Alcohol Dependence 11 (3): 223­34. http://www.scotland.gov.uk/health/alcoholproblems/docs/trco.pdf. Grube, J., and G. Agostinelli. 2000. Alcohol Advertising and Alcohol Cerff, P., and A. Plüddemann. 1998. "Brief on the ARRIVE ALIVE Consumption: A Review of Recent Research. Berkeley, CA: Prevention Campaign." Prepared for the Alcohol Policy in Developing Societies Research Center. project. Cape Town, South Africa: Medical Research Council, Gutjahr, E., G. Gmel, and J. Rehm. 2001. "Relation between Average Urbanisation, and Health Research Programme. Alcohol Consumption and Disease: An Overview." European Addiction Chisholm, D., J. Rehm, M. van Ommeren, and M. Monteiro. 2004. Research 7 (3): 117­27. "Reducing the Global Burden of Hazardous Alcohol Use: A Harrison, L., and E. Gardiner. 1999. "Do the Rich Really Die Young? Comparative Cost-Effectiveness Analysis." Journal of Studies on Alcohol Alcohol-Related Mortality and Social Class in Great Britain, 1988­94." 65 (6): 782­93. Addiction 94: 1871­80. Corrao, G., L. Rubbiati, V. Bagnardi, A. Zambon, and K. Poikolainen. 2000. Higgins-Biddle, J. C., and T. F. Babor. 1996. Reducing Risky Drinking. "Alcohol and Coronary Heart Disease: A Meta-Analysis." Addiction 94 Report prepared for the Robert Wood Johnson Foundation, University (10):1501­23. of Connecticut Health Center, Farmington. Demyttenaere, K., R. Bruffaerts, J. Posada-Villa, I. Gasquet, V. Kovess, J. P. Hinkle, L. E., and A. Herrou-Aragon. 2001. "How Far Did Africa's First Lepine, and others. 2004. "Prevalence, Severity, and Unmet Need for Generation Trade Reforms Go?" World Bank, Washington, DC. Treatment of Mental Disorders in the World Health Organization http://www.uesiglo21.edu.ar/pdfs%20dpto%20economia/ES?002? World Mental Health Surveys." Journal of the American Medical ECO.pdf. Association 291 (21): 2581­90. Isichei, H. U., P. U. Ikwuagu, and J. O. Egbuta. 1993."Pattern of Alcoholism Drummond, D. C. 1990. "The Relationship between Alcohol Dependence in Jos, Nigeria, and Castrop-Rauxel, West Germany: A Comparative and Alcohol-Related Problems in a Clinical Population." British Journal Study." In Epidemiology and Control of Substance Use in Nigeria, ed. I. S. of Addiction 85 (3): 357­66. Obot, 123­27. Jos, Nigeria: Centre for Research and Information on Eckardt, M. J., S. E. File, G. L. Gessa, K. A. Grant, C. Guerri, P. L. Hoffman, Substance Abuse. and others. 1998. "Effects of Moderate Alcohol Consumption in the Johns, B., R. Baltussen, T. Adam, and R. Hutubessy. 2003. "Programme Central Nervous System." Alcoholism: Clinical and Experimental Costs in the Economic Evaluation of Health Interventions." Research 22 (5): 998­1040. Cost-Effectiveness and Resource Allocation 1: 1. http://www Edwards, G., P. Anderson, T. F. Babor, S. Casswell, R. Ferrence, N. .resourceallocation.com. Giesbrecht, and others. 1994. Alcohol Policy and the Public Good. Klingemann, H., and G. Gmel, eds. 2001. Mapping the Social Consequences Oxford, U.K.: Oxford University Press. of Alcohol Consumption. Dordrecht, Netherlands: Kluwer Academic Emanuele, N., and M. A. Emanuele. 1997."The Endocrine System: Alcohol Publishers. Alters Critical Hormonal Balance." Alcohol Health and Research World Lauer, J. A., C. J. L. Murray, K. Roehrich, and H. Wirth. 2003. "PopMod: A 21 (1): 53­64. Longitudinal Population Model with Two Interacting Disease States." English, D. R., C. D. J. Holman, E. Milne, M. Winter, G. K. Hulse, G. Codde, Cost Effectiveness and Resource Allocation 1: 6. http://www and others. 1995. The Quantification of Drug Caused Morbidity and .resourceallocation.com. Morality in Australia. Canberra: Commonwealth Department of Lee, V. J. 2001. Tourism and Alcohol in the Developing World: Potential Human Services and Health. Effects on Alcohol Policies and Local Drinking Problems. Stockholm: Ezzati, M., A. D. Lopez, A. Rodgers, S. Vander Horn, C. J. L. Murray, and Stockholm University, Centre for Social Research on Alcohol and the Comparative Risk Assessment Collaborating Group. 2002. Drugs. 904 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others Leppänen, K. 1979. "Valtakunnallisen lauantaisulkemiskokoeilun vaiku- Factors, ed. M. Ezzati, A. D. Lopez, A. Rodgers, and C. J. L. Murray, tuksista alkoholijuomien myyntiin" (Effects of National Saturday 959­1108. Geneva: World Health Organization. Closing Experiment on Alcohol Sales). Alkoholipolitikka 44: 20­21. Rehm, J., C. T. Sempos, and M. Trevisan. 2003."Average Volume of Alcohol Levy, F., and S. I. Ornstein. 1983. "Price and Income Elasticities and the Consumption, Patterns of Drinking, and Risk of Coronary Heart Demand for Alcoholic Beverages." In Recent Developments in Disease: A Review." Journal of Cardiovascular Risk 10 (1): 15­20. Alcoholism, ed. M. Galanter, 303­45. New York: Plenum. Ridolfo, B., and C. Stevenson. 2001. The Quantification of Drug-Caused Liu, B. Q., R. Peto, Z. M. Chen, J. Boreham, Y. P. Wu, J. Y. Li, and others. Mortality and Morbidity in Australia 1998. Canberra: Australian 1998. "Emerging Tobacco Hazards in China: 1. Retrospective Institute of Health and Welfare. Proportional Mortality Study of One Million Deaths." British Medical Room, R., T. Babor, and J. Rehm. 2005. "Alcohol and Public Health: A Journal 317 (7170): 1411­22. Review." Lancet 365 (February 5): 519­30. Ludbrook, A, C. Godfrey, L. Wyness, S. Parrot, S. Haw, M. Napper, and Room, R., K. Graham, J. Rehm, D. Jernigan, and M. Monteiro. 2003. others. 2002. "Effective and Cost-Effective Measures to Reduce Alcohol "Drinking and Its Burden in a Global Perspective: Policy Considerations Misuse in Scotland: A Literature Review." Aberdeen, U.K.: Health and Options." European Addiction Research 9 (4): 165­75. Economics Research Unit. Room, R., A. Janca, L. A. Bennett, L. Schmidt, N. Sartorius, and others. Mäkelä, P. 1999. "Alcohol-Related Mortality as a Function of Socio- 1996. "WHO Cross-Cultural Applicability Research on Diagnosis Economic Status." Addiction 94: 867­86. and Assessment of Substance Use Disorders: An Overview of Methods Mathers, C. D., C. Bernard, K. Moesgaard Iburg, M. Inoue, D. Ma Fat, K. and Selected Results." Addiction 91 (2): 199­220. Shibuya, and others. 2003. "Global Burden of Disease in 2002: Data Room, R., D. Jernigan, B. Carlini-Marlatt, O. Gureje, K. Mäkelä, M. Sources, Methods and Results." Global Programme on Evidence Marshall, and others. 2002. Alcohol in Developing Societies: A Public for Health Policy Discussion Paper 54, Geneva, World Health Health Approach. Helsinki: Finnish Foundation for Alcohol Studies. Organization. Saffer, H. 2000. "Alcohol Consumption and Alcohol Advertising Bans." McKee, M., and A. Britton. 1998. "The Positive Relationship between NBER Working Paper 7758, National Bureau of Economic Research, Alcohol and Heart Disease in Eastern Europe: Potential Physiological Cambridge, MA. Mechanisms." Journal Royal Society Medicine 91: 402­7. Saffer, H., and D. Dave. 2002. "Alcohol Consumption and Alcohol Moyer, A., J. W. Finney, C. E. Swearingen, and P. Vergun. 2002. "Brief Advertising Bans." Applied Economics 34 (11): 1325­34. Interventions for Alcohol Problems: A Meta-Analytic Review of Controlled Investigations in Treatment-Seeking and Non-Treatment- Saunders, J. B., and O. G. Aasland. 1987. WHO Collaborative Project on Seeking Populations." Addiction 97: 279­92. Identification and Treatment of Persons with Harmful Alcohol Consumption. WHO/MNH/DAT 86.3. Geneva: World Health National Highway Traffic Safety Administration. 1992. "Driving under the Organization. Influence: A Report to Congress on Alcohol Limits." Washington, DC: U.S. Department of Transportation. Schmidt, L., and R. Room. 1999. "Cross-Cultural Applicability in International Classifications and Research on Alcohol Dependence." Nordlund, S. 1984. "Effekten av lørdagsstengningen ved Vinmonopolets Journal of Studies on Alcohol 60: 448­62. butikker" (Effects of Saturday Closing of the Wine/Liquor Monopoly Outlets). Alkoholpolitik--Tidsskrift for Nordisk Alkoholforskning 1: Shults, R. A, R. W. Elder, D. A. Sleet, J. L. Nichols, M. O. Alao, V. G. 221­29. Carande-Kulis, and others. 2001. "Reviews of Evidence Regarding Norström, T., and O. J. Skog. 2003. "Saturday Opening of Alcohol Retail Interventions to Reduce Alcohol-Impaired Driving." American Journal Shops in Sweden: An Impact Analysis." Journal of Studies on Alcohol 64: of Preventive Medicine 21 (Suppl. 4): 66­88. 393­401. Single, E., L. Robson, J. Rehm, and X. Xie. 1999. "Morbidity and Mortality Peek-Asa, C. 1999. "The Effect of Random Alcohol Screening in Reducing Attributable to Alcohol, Tobacco, and Illicit Drug Use in Canada." Motor Vehicle Crash Injuries." American Journal of Preventive Medicine American Journal of Public Health 89 (3): 385­90. 16 (1 Suppl.): 57­67. Single, E., L. Robson, X. Xie, and J. Rehm. 1996. The Cost of Substance Puddey, I. B., V. Rakic, S. B. Dimmitt, and L. J. Beilin. 1999. "Influence Abuse in Canada. Ottawa: Canadian Centre on Substance Abuse. of Drinking on Cardiovascular Disease and Cardiovascular Risk Sobell, L. C., T. P. Ellingstad, and M. B. Sobell. 2000. "Natural Recovery Factors--A Review." Addiction 94: 649­63. from Alcohol and Drug Problems: Methodological Review of the Rehm, J., G. Gmel, C. T. Sempos, and M. Trevisan. 2003. "Alcohol-Related Research with Suggestions for Future Directions." Addiction 95: Mortality and Morbidity." Alcohol Research and Health 27 (1): 39­51. 749­64. Rehm, J., E. Gutjahr, and G. Gmel. 2001."Alcohol and All-Cause Mortality: Stouthard, M. E., M. L. Essink-Bot, and G. L. Bonsel. 2000. "On Behalf of A Pooled Analysis." Contemporary Drug Problems 28: 337­61. the Dutch Disability Weights Group: Disability Weights for Diseases-- A Modified Protocol and Results for a Western European Region." Rehm, J., N. Rehn, R. Room, M. Monteiro, G. Gmel, D. Jernigan, and U. European Journal of Public Health 10: 24­30. Frick. 2003. "The Global Distribution of Average Volume of Alcohol Consumption and Patterns of Drinking." European Addiction Research Tax Policy Chief Directorate. 2002. The Taxation of Alcoholic Beverages in 9 (4): 147­56. South Africa and Its Impact on the Consumption Levels of Alcoholic Beverages. Pretoria: National Treasury. Rehm, J., R. Room, K. Graham, M. Monteiro, G. Gmel, and C. T. Sempos. 2003. "The Relationship of Average Volume of Alcohol Consumption U.S. Department of State. 1999. "Mauritius: 1999­2000 Budget Increased." and Patterns of Drinking to Burden of Disease: An Overview." U.S. Department of State, Washington, DC. http://www.tradeport.org/ Addiction 98 (10): 1209­28. ts/countries/mauritius/mrr/mark0003.html. Rehm, J., R. Room, M. Monteiro, G. Gmel, K. Graham, N. Rehn, and oth- Wallace, S., and R. Bird. 2003. "Taxing Alcohol in Africa: Reflections from ers. 2003. "Alcohol as a Risk Factor for Global Burden of Disease." International Experience." Paper presented at the South Africa European Addiction Research 9 (4):157­64. Conference on Excise Taxation, June 11­13, Gauteng, South Africa. http://www.iticnet.org/030707-PRESENTATION_Bird.pdf. Rehm, J., R. Room, M. Monteiro, G. Gmel, K. Graham, T. Rehn, and oth- ers. 2004. "Alcohol." In Comparative Quantification of Health Risks: Whitlock, E. P., M. R. Polen, C. A. Green, T. Orleans, and J. Klein. 2004. Global and Regional Burden of Disease Due to Selected Major Risk "Behavioral Counseling Interventions in Primary Care to Reduce Alcohol | 905 Risky/Harmful Alcohol Use by Adults: A Summary of the Evidence for ------. 2003a. "Burden of Disease Project." WHO, Geneva. http:// the U.S. Preventive Services Task Force." Annals of Internal Medicine www.who.int/evidence/bod. 140: 557­68. ------. 2003b. "WHO Global Alcohol Database." WHO, Geneva. WHO (World Health Organization). 1993. The ICD-10 Classification http://www3.who.int/whosis/alcohol. of Mental and Behavioral Disorders: Diagnostic Criteria for Research. WHO Brief Intervention Study Group. 1996. "A Cross-National Trial of Geneva: WHO. Brief Interventions with High Risk Drinkers." American Journal of ------. 1999. Global Status Report on Alcohol. WHO/HSC/SAB/99.11. Public Health 86: 948­55. Geneva: WHO, Substance Abuse Department. Zakhari, S. 1997. "Alcohol and the Cardiovascular System: Molecular ------. 2000. 1997­1999 World Health Statistics Annual. Geneva: WHO. Mechanisms for Beneficial and Harmful Action." Alcohol Health and http://www.who.int/whosis/. Research World 21 (1): 21­29. ------. 2002. World Health Report 2002: Reducing Risks, Promoting Healthy Life. Geneva: WHO. 906 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others Chapter 48 Illicit Opiate Abuse Wayne Hall, Chris Doran, Louisa Degenhardt, and Donald Shepard Illicit drugs are those banned by international drug control disease control priorities for illicit opioid dependence, because treaties. They include cannabis products (for example, mari- dependent users account for most of the illicit opioids con- juana, hashish, and bhang); stimulant drugs (such as cocaine sumed and experience most of the harm such dependence and methamphetamine); so-called dance-party drugs (such as causes (W. Hall, Degenhardt, and Lynskey 1999). 3, 4-methylenedioxymethamphetamine, also known as ecstasy or MDMA); and illicit opioids (for instance, heroin and opium) and diverted pharmaceutical opioids (such as NATURE, CAUSES, AND HEALTH CONSEQUENCES buprenorphine, methadone, and morphine) (see annex 48.A). OF ILLICIT OPIOID USE Worldwide, 185 million people were estimated to have used illicit drugs during 1998­2002 (UNODC 2004; UNODCCP Before considering interventions, we briefly summarize what is 2002). Cannabis was the most widely used illicit drug, with known about the antecedents, causes, and health consequences 146.2 million users in 2002, or 3.7 percent of the global of illicit opioid use. population over age 15. The stimulant drugs were the next most widely used illicit drugs: 29.6 million people worldwide used amphetamines; 13.3 million used cocaine; and 8.3 million Antecedents of Heroin Use used ecstasy. An estimated 15.3 million, or 0.4 percent of the Law enforcement efforts to reduce the availability of heroin world population age 15 to 64, used illicit opioids; more than aim to increase its price, deter illicit drug use, and promote half used heroin and the remainder used opium or diverted social values that discourage heroin use (Fergusson, Horwood, pharmaceutical opioids. Illicit opioids continue to be the major and Lynskey 1998; Hawkins, Catalano, and Miller 1992; illicit drug problem in most regions of the world in terms of Newcomb and Bentler 1988). These gains may be at the cost of impact on public health and public order (UNODC 2004). increasing harm among the minority who use opioids despite Even though cannabis use accounts for about 80 percent of the prohibition--for example, by encouraging injecting use as illicit drug use worldwide, the mortality and morbidity attrib- the most efficient way to use an expensive drug and increasing utable to its use are not well understood, even in developed needle sharing because clean injecting equipment is not freely countries (W. Hall and Pacula 2003; Macleod and others 2004; available (Rhodes and others 2003; Strathdee and others 2003). WHO Programme on Substance Abuse 1997). The same is true Two aspects of the family environment are associated with of the morbidity and mortality attributable to cocaine and increased rates of both licit and illicit drug use in young people amphetamine-type stimulants (Macleod and others 2004). in developed countries. The first is exposure to a disadvantaged Dance-party drugs have been used for too short a time in most home environment, with parental conflict and poor discipline developed societies to enable a good assessment of their poten- and supervision; the second is exposure to parents' and sib- tial for harm (Boot, McGregor, and Hall 2000; Macleod and lings' use of alcohol and other drugs (Hawkins, Catalano, and others 2004). The remainder of this chapter is concerned with Miller 1992). In developed countries, children who perform 907 poorly in school because of impulsive or problem behavior and 2. withdrawal, as manifested by either of the following: those who are early users of alcohol and other drugs are most a. the characteristic withdrawal syndrome for the likely to use illicit opioids (Fergusson, Horwood, and Swain- substance Campbell 2002). Affiliation with drug-using peers is a risk fac- b. the same (or closely related) substance is taken to relieve tor for drug use that operates independently of individual and or avoid withdrawal symptoms; 3. the substance is often taken in larger amounts or over a family risk factors (Fergusson, Horwood, and Lynskey 1998; longer period than was intended; Hawkins, Catalano, and Miller 1992). 4. there is a persistent desire or unsuccessful efforts to cut down or control substance use; 5. a great deal of time is spent in activities necessary to obtain Health Consequences of Heroin Use the substance (e.g., visiting multiple doctors, driving long The following sections describe the major health consequences distances), use the substance (e.g., chain smoking), or of heroin use. They include dependence, increased mortality recover from its effects; and morbidity attributable to drug overdoses, and bloodborne 6. important social, occupational, or recreational activities are viruses. given up or reduced because of substance use; 7. the substance use is continued despite knowledge of having Heroin Dependence. In household surveys, 1 to 2 percent of a persistent or recurrent physical or psychological problem adults in Australia, the United States, and Europe report using that is likely to have been caused or exacerbated by the substance. heroin at some time in their lives (Australian Institute of Health and Welfare 1999; EMCDDA 2002; SAMHSA 2002). The highest rates are typically among adults age 20 to 29. Self- Indirect estimation methods suggest that in Australia, the reported heroin use in population surveys probably underesti- United Kingdom, and the European Union fewer than 1 per- mates rates of use because heroin users are undersampled and cent of adults age 15 to 54 are heroin dependent (EMCDDA those who are sampled underreport their use (W. Hall, Lynskey, 2002; W. Hall and others 2000). Research in the United States and Degenhardt 1999). indicates that dependent heroin users who seek treatment or In developed countries, one in four of those who report who come to the attention of the legal system may use heroin heroin use become dependent on it (Anthony, Warner, and for decades (Goldstein and Herrera 1995; Hser, Anglin, and Kessler 1994). People who are heroin dependent continue to Powers 1993), with periods of use punctuated by abstinence use heroin in the face of problems that they know (or believe) (Bruneau and others 2004; Galai and others 2003), drug treat- to be caused by its use. These problems include being arrested ment, and imprisonment (Gerstein and Harwood 1990). When or imprisoned, having interpersonal and family problems, periods of abstinence are included, dependent heroin users use catching infectious diseases, and suffering from drug over- heroin daily for 40 to 60 percent of the 20 years that they typi- doses. Many heroin users who seek treatment have typically cally are addicts (Ball, Shaffer, and Nurco 1983; Maddux and been daily heroin injectors, although in Europe (EMCDDA Desmond 1992). 2002), North America (Office of National Drug Control Policy Illicit opioid use increased in Asia, Europe, and Oceania 2001), and parts of Asia, illicit opioid users also smoke or and, to a lesser extent, in Africa and South America in the "chase" the drug (inhale the fumes released when heroin is 1990s, but it has stabilized or declined since 2000 (UNODC heated) (UNODC 2004). 2004). Most illicit opioid users (7.8 million) live in Asian coun- The American Psychiatric Association defines drug depend- tries that surround the major opium-producing countries, ence as "a cluster of cognitive, behavioral, and physiologic Afghanistan and Myanmar. Europe accounts for about 25 per- symptoms indicating that the individual continues use of the cent of illicit opioid use (4 million users or 0.8 percent of the substance despite significant substance-related problems" adult population age 15 to 64). Two-thirds of users are in (American Psychiatric Association 1994, 176). In the fourth Eastern Europe, which reported large increases in illicit opioid edition of the association's Diagnostic and Statistical Manual of use during the second half of the 1990s (Atlani and others Mental Disorders (1994,), a diagnosis of substance dependence 2000; Hamers and Downs 2003; Kelly and Amirkhanian 2003; requires that three or more of the following occur together: Rhodes and others 1999; Uuskula and others 2002). Illicit opioid use stabilized in much of Asia between 2000 and At any time in the same 12-month period: 2002 (UNODC 2004) as a result of decreased opium production after the rapid expansion during the 1990s (Dorabjee and 1. tolerance, as defined by either of the following: a. need for markedly increased amounts of the substance Samson 2000; Reid and Crofts 2000). After 2000, India and to achieve intoxication or desired effect Pakistan reported stabilizing rates of illicit opioid use but b. markedly diminished effect with continued use of the increased injection of pharmaceutical opiates (Ahmed and same amount of the substance; others 2003; Dorabjee and Samson 2000; Strathdee and others 908 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others 2003). China has reported a steady rate of growth in illicit opiate account for 25 to 33 percent of deaths of young adult males use in its southern and northern provinces (Beyrer 2003; Beyrer (EMCDDA 2002). and others 2000; Yu and others 1998) and a 15-fold increase in the number of registered opioid addicts between 1990 and 2002, Economic Costs of Illicit Opioid Use. In Canada, Xie and bringing the total to about 1 million (UNODC 2004). others (1996) calculate the costs of illicit drugs as 0.2 percent Oceania experienced a marked rise in heroin use in the late of gross domestic product (GDP). In Australia, Collins and 1990s, largely driven by a dramatic increase in the availability of Lapsley (1996) estimate the economic costs of illicit drug abuse heroin in Australia (Darke, Topp, and others 2002; W. Hall, at 2 percent of GDP. Degenhardt, and Lynskey 1999). In late 2000, an abrupt heroin shortage resulted in a large reduction in fatal and nonfatal overdoses (Day and others 2004; Degenhardt, Day, and Hall CONTRIBUTION OF OPIOID DEPENDENCE 2004). TO THE GLOBAL BURDEN OF DISEASE Mortality, Morbidity, and Heroin Dependence. In developed Degenhardt, Hall, and others (2004) estimate the contribution countries, dependent heroin users have an increased risk of of illicit opioid dependence to the global burden of disease premature death from drug overdoses, violence, suicide, using data on deaths caused by opioid and other drug over- and alcohol-related causes (Darke and Ross 2002; Goldstein doses, suicides and accidents, and HIV/AIDS. When estimates and Herrera 1995; Vlahov and others 2004). Heroin users of morbidity attributable to illicit drug use were added in, illicit treated before the HIV epidemic were 13 times more likely to opioid use accounted for 0.7 percent of global disability- die prematurely than their peers (Hulse and others 1999), with adjusted life years (DALYs) in 2000 (WHO 2003). opioid overdose the most frequent cause of death (W. Hall, These estimates suggest that illicit opioid use is a significant Degenhardt, and Lynskey 1999). In countries with a high global cause of premature mortality and disability among prevalence of HIV infection, AIDS is a major cause of prema- young adults. Even so, they probably underestimate the disease ture death among drug users (EMCDDA 2002; UNAIDS and burden attributable to illicit opioids, because they omit differ- WHO 2002). Fatal opioid overdose deaths increased in many ences across subregions in the quality of data on causes of mor- developed countries during the 1990s before declining after tality and estimates of mortality and morbidity attributable to 2000 (UNODC 2004). hepatitis and violence (Degenhardt, Hall, and others 2004). In parts of Asia, Eastern Europe, and the United States, the sharing of contaminated injecting equipment accounts for a substantial proportion of new HIV infections (EMCDDA INTERVENTIONS FOR ILLICIT OPIOID 2002; UNAIDS and WHO 2002; UNODC 2004). Injecting DEPENDENCE opioid use has been a major driver of HIV epidemics in China (Yu and others 1998), Myanmar (Beyrer and others 2000), the Methods adopted to control the problems arising from illicit Russian Federation and former Soviet republics (Hamers and opioid dependence include source-country control; interdic- Downs 2003), and Vietnam (Beyrer and others 2000; Hien tion of supply into end-use countries; enforcement by the and others 2001). police force and the criminal justice system of legal prohibi- The prevalence of infection with hepatitis B and C viruses tions on the supply, possession, and use of opioids; treatment among injecting drug users is greater than 60 percent in of those who are opioid dependent, both voluntarily and under Australia (National Centre in HIV Epidemiology and Clinical legal coercion from the criminal justice system; school-based Research 1998), Canada (Fischer and others 2004), China and mass media preventive educational programs; and regula- (Ruan and others 2004), the United States (Fuller and others tory policies restricting the prescription of opioids (Manski, 2004), and the European Union (EMCDDA 2002). Chronic Pepper, and Petrie 2001). infection occurs in 75 percent of infections, and 3 to 11 percent of chronic hepatitis C virus carriers develop liver cirrhosis within 20 years (Hepatitis C Virus Projections Working Group Prevention of Heroin Use 1998). Countries use a variety of interventions in attempts to prevent Heroin-related deaths primarily occur among young adults the initiation of use of illicit drugs such as cannabis (Manski, and account for a large number of life years lost in developed Pepper, and Petrie 2001; Spooner and Hall 2002), in the belief societies. In Australia in 1996, for example, such deaths that early initiation of cannabis use leads to an increased accounted for 2.2 percent of life years lost, with each death risk of using illicit opioids (Fergusson, Horwood, and accounting for 22 years of life lost (Mathers, Vos, and Swain-Campbell 2002). These interventions include legal pro- Stephenson 1999). In Scotland and Spain, opiate-related deaths hibitions on the manufacture, sale, and use of opioid drugs Illicit Opiate Abuse | 909 for nonmedical purposes; enforcement of these sanctions by others 2000; Kimber and others 2003). Supervised injecting law enforcement officials by means of fines and imprisonment; facilities have been introduced in Germany, the Netherlands, and enforcement of restrictions on medically prescribed opi- and Switzerland (Dolan and others 2000; Kimber and others oids to prevent their diversion (Manski, Pepper, and Petrie 2003), but their effect on overdose deaths has not been rigor- 2001). Preventive measures also include mass media and ously evaluated to date. A supervised injecting facility was eval- school-based educational campaigns about the health risks of uated in Australia, but the evaluation was limited by the con- opioid and other illicit drug use (Spooner and Hall 2002). It is current onset of a heroin shortage that resulted in a 40 percent unclear how effective these interventions are in preventing decline in overdose deaths (Kaldor and others 2003). cannabis use and even less clear whether they reduce the initia- A fourth strategy is to increase methadone maintenance tion of opioids (Caulkins and others 1999; Manski, Pepper, and among older, high-risk opioid-dependent people, because indi- Petrie 2001). viduals enrolled in methadone maintenance treatment (MMT) The most popular interventions against illicit opioid use in are substantially less likely to suffer from a fatal overdose many developed societies have been the interdiction of drug (Caplehorn and others 1994; Gearing and Schweitzer 1974; supply and the enforcement of legal sanctions against the pos- Langendam and others 2001). session, use, and sale of opioid drugs (Manski, Pepper, and Petrie 2001). As a consequence, imprisonment is the most com- mon intervention to which many illicit opioid users have been Treatment Interventions for Dependent Opioid Users exposed (Gerstein and Harwood 1990). In Asia and Eastern The range of treatment interventions includes voluntary pro- Europe, high rates of imprisonment of drug users have been a grams such as detoxification, abstinence-oriented treatments, factor in HIV transmission, because drug users engage in high- and oral Methadone maintenance treatment, as well as invol- risk injecting while imprisoned (Beyrer and others 2000). untary options imposed by criminal justice systems. Detoxification. Detoxification is supervised withdrawal from Interventions to Reduce Heroin-Related Harm a drug of dependence that attempts to minimize withdrawal The most effective intervention to reduce bloodborne virus symptoms. It is not a treatment for heroin dependence; it infection arising from illicit injecting of opioids and other provides a respite from opioid use and may be a prelude to drugs is the provision of clean injecting equipment to reduce abstinence-based treatment (Mattick and Hall 1996). users' risks of contracting or transmitting bloodborne viruses. Naltrexone is a longer-acting opiate antagonist than nalox- This intervention has been widely supported in most developed one; it can be used to accelerate the opioid withdrawal process. countries, but it has been incompletely adopted in developing Ultra-rapid opioid detoxification accelerates withdrawal by countries that have problems with the concept of facilitating giving the patient naltrexone under general anesthetic. There is the injection of illicit drugs (UNAIDS and WHO 2002). Vac- no evidence that accelerated withdrawal in itself reduces the cinations are available against hepatitis B but not hepatitis C. high rate of relapse to heroin use in the absence of further These important interventions are covered in chapter 18. treatment (W. Hall and Mattick 2000). A number of strategies can potentially reduce deaths from opioid overdoses (Darke and Hall 2003; Sporer 2003). First, Abstinence-Oriented Treatments. Abstinence-oriented treat- injecting drug users can be educated about the dangers of com- ments aim to achieve enduring abstinence from all opioid bining the use of opioids with alcohol and benzodiazepines drugs by providing some type of intervention after withdrawal (McGregor and others 2001), both of which heighten the risk to reduce the high rate of relapse to opioids (Mattick and Hall of a fatal opioid overdose (Darke and Zador 1996; Warner- 1996). The interventions may include social and psychological Smith and others 2001). Heroin users also need to be discour- support only or such support supplemented by pharmacologi- aged from injecting in the streets or alone, thereby denying cal methods. themselves assistance in the event of an overdose. These inter- Residential treatment in therapeutic communities and out- ventions have yet to be evaluated. patient drug counseling may entail encouraging patients to A second strategy is to encourage drug users who witness become involved in self-help groups such as Narcotics overdoses to seek medical assistance and to use simple resusci- Anonymous. These approaches share a commitment to achiev- tation techniques until help arrives. A more controversial ing abstinence from all opioids, using group and psychological option is to distribute the opioid antagonist naloxone to high- interventions to help dependent heroin users remain abstinent. risk heroin users (Darke and Hall 1997; Strang and others Therapeutic communities and drug counseling are usually pro- 1996). Neither of these interventions has been evaluated. vided through specialist addiction or mental health services. A third strategy is to provide supervised injecting facilities The former are residential, and the latter are provided on an in areas with high rates of injecting opioid use (Dolan and outpatient basis. 910 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others No randomized controlled trials of therapeutic communi- of injected heroin, allowing the individual to take advantage of ties or outpatient drug counseling have been carried out. psychotherapeutic and rehabilitative services. Observational studies in the United Kingdom (Gossop, Every one of the small number of randomized controlled Marsden, and Stewart 1998; Gossop and others 1997) and the trials of MMT compared with placebo or no treatment has United States (Hubbard and others 1989; Simpson and Sells produced positive results (W. Hall, Ward, and Mattick 1998; 1982) have found that therapeutic communities and drug Mattick and others 2003). Large observational studies show counseling were less successful than MMT in attracting and that MMT decreases heroin use and criminal activity and retaining dependent heroin users, but they substantially reduces HIV transmission while patients remain in treatment reduced heroin use and crime among those who remained in (Gerstein and Harwood 1990; Simpson and Sells 1990; Ward, treatment for at least three months (Gerstein and Harwood Hall, and Mattick 1998). MMT is the best-supported form of 1990; Gossop, Marsden, and Stewart 1998; Gossop and others opioid maintenance treatment (Farre and others 2002; Marsch 1997). Some evidence indicated that therapeutic communities 1998; Mattick and others 2003). may be more effective if they are used in combination with Buprenorphine is a mixed agonist-antagonist that also legal coercion to ensure that heroin users are retained in treat- blocks the effects of heroin. When given in high doses, its ment long enough to benefit from it (Gerstein and Harwood effects can last for up to three days, while its antagonist effects 1990). substantially reduce overdose and abuse (Oliveto and Kosten Recovering drug users run Narcotics Anonymous groups 1997; Ward, Hall, and Mattick 1998). Meta-analyses have found using an adaptation of the 12-step philosophy of Alcoholics that buprenorphine is effective in the treatment of heroin Anonymous. Some individuals use these groups as their sole dependence (Mattick and others 2003) and is of equivalent effi- form of support for abstinence, whereas for others these groups cacy to MMT when delivered in primary health care and spe- complement therapeutic communities that are based on the cialist treatment settings in Australia (Gibson and others 2003). same principles. Such groups are usually not open to people Bammer and others (2003) have proposed injectable heroin who are in opioid substitution treatment programs. maintenance as a way of attracting into treatment those heroin The most extensive research on self-help has been in the users who are not interested in or have failed to respond to treatment of alcohol dependence. Treated alcoholics who par- MMT. This method has recently been evaluated in the ticipate in Alcoholics Anonymous groups have higher rates of Netherlands (Central Committee on the Treatment of Heroin abstinence than those who do not (see, for example, Tonigan, Addicts 2002) and Switzerland (Perneger and others 1998; Connors, and Miller 2003; Tonigan, Toscova, and Miller 1996). Uchtenhagen, Gutzwiller, and Dobler-Mikola 1998). Perneger The good outcome in those who attend Alcoholics Anonymous and others (1998) report a randomized controlled trial of meetings may reflect the self-selection of motivated partici- injectable heroin maintenance in people who had failed at pants into self-help groups. Recent studies that have attempted MMT. Stabilizing and safely maintaining heroin addicts on to control for this possibility using sophisticated statistical injectable heroin (self-administered on-site in a comprehensive methods have produced mixed results, with some showing the health and social service) proved feasible for six months and persistence of an effect of self-help after correction (Tonigan, substantially improved their health and social well-being. The Connors, and Miller 2003) while others do not (Fortney and Swiss trials showed that it was possible to maintain opioid others 1998). addicts on injectable heroin for up to two years (Rehm and Shepard and others (forthcoming) evaluate the effect of self- others 2001; Uchtenhagen, Gutzwiller, and Dobler-Mikola help participation on substance abuse 24 months after treat- 1998). A recent randomized controlled trial in the Netherlands ment for members of a mixed population of substance abusers (Central Committee on the Treatment of Heroin Addicts 2002) treated at two treatment facilities in the United States, some of confirms the findings of Perneger and others (1998). whom had problems with heroin. They find that participation in self-help groups was associated with longer abstinence from Criminal Justice Interventions for Dependent Illicit Opioid all drugs. Correction for self-selection did not eliminate the Users. The most common intervention for illicit opioid association in one treatment setting, but it made the results dependence in most developed societies is imprisonment much more equivocal in the other. (EMCDDA 2003; Gerstein and Harwood 1990). Imprisonment is not intended to be a health intervention. Nonetheless, it is an Oral Methadone Maintenance Treatment. This treatment ineffective way of reducing opioid dependence, when judged substitutes a long-acting, orally administered opioid for the by the high recidivism in longitudinal studies of dependent shorter-acting heroin, with the aim of stabilizing dependent heroin users (see, for example, Hser, Anglin, and Powers 1993; heroin users so that they are amenable to rehabilitation (Marsh Manski, Pepper, and Petrie 2001). and others 1990; Ward, Hall, and Mattick 1998). When given in Legally coerced treatment is treatment that is legally forced high or blockade doses, methadone blocks the euphoric effects on those who have been charged with or convicted of an Illicit Opiate Abuse | 911 offense to which their drug dependence has contributed Detoxification. The National Evaluation of Pharma- (W. Hall 1997). It is most often provided as an alternative to cotherapies for Opioid Dependence Project in Australia con- imprisonment, under the threat of imprisonment if the person ducted a cost-effectiveness analysis of five interventions: fails to comply with the treatment (W. Hall 1997; Manski, Pepper, and Petrie 2001; Spooner, Hall, and Mattick 2001). Its · naltrexone-induced rapid opioid detoxification under major justification is that it is an effective way of treating anesthesia offenders' drug dependence that reduces the likelihood of their · naltrexone-induced rapid opioid detoxification under offending again (Gerstein and Harwood 1990). A consensus sedation view prepared for the World Health Organization (WHO) · conventional inpatient detoxification (Porter, Arif, and Curran 1986) was that compulsory treatment · conventional outpatient detoxification was legally and ethically justified only if the rights of the · buprenorphine outpatient detoxification. individuals were protected by due process and if the treatment provided was effective and humane. A successful outcome was defined as achieving abstinence Research into the effectiveness of legally coerced treatment from heroin for one week (Mattick and others 2001). for opioid dependence has been limited to observational studies Rapid detoxification under sedation was the most cost- (W. Hall 1997; Manski, Pepper, and Petrie 2001; Wild, Roberts, effective method of detoxification (US$2,355 for one week of and Cooper 2002). Anglin's (1988) quasi-experimental studies abstinence) and conventional outpatient detoxification the least of the California Civil Addict Program provide the strongest cost-effective (US$12,031). Rapid detoxification under anesthe- evidence of efficacy. These studies compared heroin-dependent sia achieved high rates of abstinence in the first week, but its offenders who entered the program between 1962 and 1964 expense reduced its cost-effectiveness (Mattick and others with a group of similar offenders who went through the crimi- 2001). nal justice system during the same period. They found that com- Doran and others (2003) compared the cost-effectiveness of pulsory hospital treatment followed by close supervision in the detoxification from heroin using buprenorphine in a specialist community substantially reduced heroin use and crime. Australian clinic and in a shared care setting. They conducted The effectiveness of less coercive forms of treatment has a randomized controlled trial with 115 heroin-dependent been supported by analyses of the effectiveness of community- patients receiving a five-day treatment regime of buprenor- based treatment provided while on probation or parole phine. The specialist clinic was a community-based treatment (Hubbard and others 1989; Simpson and others 1986). These agency in Sydney. Shared care involved treatment by a general studies showed that individuals who entered community-based practitioner, supplemented by weekend dispensing and some therapeutic communities and drug-free outpatient counseling counseling at the specialist clinic. They estimate that buprenor- under legal pressure did as well as those who did so voluntarily phine detoxification in the shared care setting was US$17 more (Hubbard and others 1988; Simpson and Friend 1988). The expensive per patient than the costs of treatment at the clinic recent creation of specialized drug courts in the United States to (US$236 per patient). process those arrested for drug-related offenses awaits rigorous evaluation (Belenko 2002; Manski, Pepper, and Petrie 2001). Drug-Free Treatment. The limited economic evaluations of Legally coerced MMT is also effective. The strongest drug-free treatment have used data from observational studies evidence comes from a study in which drug offenders were ran- of treatment outcomes in samples of patients who have mixed domly assigned to parole with and without community-based substance abuse problems that include opioids. For example, MMT (Dole and others 1969). This study showed a greater Shepard, Larson, and Hoffmann (1999) calculate a range of reduction in heroin use and lower rates of incarceration among estimated costs for achieving an abstinent year in 408 patients those enrolled in MMT in the year following their release from at two different treatment facilities in the United States. The prison. These findings are supported by observational studies cost-effectiveness depended on the severity of the problem and that found no major differences in response to MMT between the intensiveness and cost of the intervention. For outpatients those who enrolled under legal coercion and those who did not with the least severe drug problems, the cost of an abstinent (Anglin, Brecht, and Maddahain 1989; Brecht, Anglin, and year was US$7,000, whereas the same outcome in patients with Wang 1993; Hubbard and others 1988). more severe problems who received long-term residential treat- ment cost US$20,000. Economic Evaluations of Interventions for Illicit Shepard and others (forthcoming) use these data to esti- Opioid Dependence mate the cost-effectiveness of involvement in mutual self- The few published economic evaluations of treatment inter- help groups, such as Alcoholics Anonymous and Narcotics ventions for illicit opioid dependence indicate varying levels of Anonymous, in sustaining abstinence for up to 24 months after cost-effectiveness. treatment. They find a positive association between self-help 912 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others involvement and abstinence 12 and 24 months after treat- The costs of injectable heroin maintenance in the Dutch ment. Applying statistical methods to correct for the effects of study was between US$18,015 and US$23,243 per patient per self-selection into self-help, they find that in a Veterans year (Bammer and others 2003). Most of the costs arose from Administration hospital, the effects of self-help on abstinence the supervision of heroin use and the security required to pre- persisted after the statistical correction, but at the other site, the vent the diversion of heroin to the black market. Injectable results depended on the method of analysis that was used. They heroin maintenance needs to produce substantially greater estimate the cost of achieving an abstinent year by means of benefits for each participant than MMT to make it as cost- self-help in the year following treatment at US$13,000, all of effective as MMT. that due to the costs that participants incurred in attending a group. Economic Modeling of the Cost-Effectiveness of Opioid Maintenance Treatment. Barnett (1999), using data on the Oral Opioid Maintenance Treatment. Goldschmidt's (1976) efficacy of MMT in reducing mortality derived from economic evaluation of MMT found that it was as effective Gronbladh, Ohlund, and Gunne's (1990) Swedish study and as a therapeutic community intervention and twice as cost- U.S. cost data, estimated that MMT saved an additional year of effective. Cartwright's (2000) review of the literature since 1976 life at a cost of US$5,900. Barnett, Zaric, and Brandeau (2001), identified a number of studies, all of which reported positive using a similar approach, estimated that the use of buprenor- benefit-cost ratios for MMT. phine by patients who would not use methadone would cost Gerstein, Harwood, and Suter's (1994) California Drug and less than US$45,000 per quality-adjusted life year. Overall, Alcohol Treatment Assessment study is the most comprehen- however, they found that BMT was much less effective and sive cost-benefit analysis carried out to date. The authors exam- more costly than MMT. Zaric, Barnett, and Brandeau (2000) ine the effects of treatment--residential programs, outpatient assessed the economic benefits of using MMT to reduce HIV programs, and methadone programs--on alcohol and drug transmission in heroin users. They found that for heroin users use, criminal activity, health and health care utilization, and living in a community with a high prevalence of HIV infection, source of income. For each treatment modality, they found that expanding MMT use produced an additional year of quality- the benefits during the first year of treatment significantly adjusted life at a cost of US$8,200. exceeded the cost of delivering the care. The benefit-cost ratio was 4.8 for residential treatment and 11.0 and 12.6 for outpa- Comparing the Cost-Effectiveness of Different Interventions tients and discharged methadone participants, respectively. Comparative cost-effectiveness analyses of these interventions Doran and others (2003) compared the cost-effectiveness face major obstacles because the small number of published of buprenorphine and methadone treatment for opioid studies used different methods to cost interventions and differ- dependence. In a randomized controlled trial, 405 subjects ent endpoints to assess the outcome of treatment. The follow- were randomly assigned to each treatment at one of three ing list, therefore, only ranks treatment interventions in the specialist outpatient drug treatment centers. The study found approximate order of their cost-effectiveness. We believe that that treatment with methadone was less expensive and more estimates of their likely contribution to DALYs worldwide effective than treatment with buprenorphine, but the differ- would be too speculative. ence in cost (US$143 per additional heroin-free day gained) had a wide range of uncertainty around it ( US$1,469 to · Detoxification. Buprenorphine and supervised naltrexone- US$1,284). accelerated withdrawal delivered on an outpatient basis are The National Evaluation of Pharmacotherapies for Opioid the most efficient and effective ways to achieve withdrawal Dependence Project also provided a cost-effectiveness analysis from opioids. of methadone, buprenorphine, LAAM (levo-alpha-acetyl- · Self-help groups. These groups provide the simplest form of methadol), and naltrexone maintenance treatments (Mattick postwithdrawal support for enduring abstinence and are and others 2001). The daily costs of these maintenance treat- also a low-cost intervention, because patients bear most of ments were similar for methadone and LAAM, but naltrexone the costs; however, they have a low rate of uptake, and their was slightly more expensive. Buprenorphine maintenance effectiveness is only modest. treatment (BMT) was more expensive, but its cost-efficiency · Oral opioid agonist maintenance treatment. This form of could have been improved to make its cost similar to that for treatment is the most widely used intervention for illicit the other treatments. MMT was the most cost-effective treat- opioid dependence in developed societies. It has a better ment for opioid dependence because it achieved one of the uptake than other interventions, and it is moderately effec- highest rates of retention in treatment among the four phar- tive under the usual delivery conditions. macotherapies examined. Naltrexone treatment was the least · Drug-free residential treatment. This form of treatment has a cost-effective. relatively low rate of treatment uptake and is costly because Illicit Opiate Abuse | 913 of its residential character and the need for intensive staff- mortality by 25 percent. In the sensitivity analysis, we varied patient interaction. It is effective for the minority of people the reduction from 15 to 35 percent (using the confidence who are retained in treatment long enough to benefit from intervals around the estimated reduction). We assumed that it (usually three months). Retention in treatment may be the reduction in mortality associated with BMT was 20 per- improved if patients enter treatment under some form of cent, which we varied in the sensitivity analysis from 10 to legal coercion. 30 percent. Finally, we assumed that those who were alive and · Naltrexone maintenance treatment. This form of treatment in treatment experienced a 25 percent reduction in disability, has not been rigorously evaluated. consistent with the Dutch disability weights. · Injectable opioid maintenance. This intervention is a more The third step was to estimate the burden for those not expensive variant of agonist maintenance treatment that has treated. For those users not in treatment, we calculated DALYs been used for patients with more severe cases of dependency using the original mortality rates. but for whom retention and treatment outcomes have been The fourth step was to estimate the total avertable burden good. from treatment with methadone or buprenorphine by (a) adding the results of the second and third steps, the revised DALYs for Calculation of the Averted, Avertable, those in treatment, and the residual for those not in treatment and Unavertable Burden and (b) subtracting those figures from the base case estimates. Assuming that the disease burden from opioid dependence is The fifth step was to cost the interventions using data on potentially avertable, we used the following approach to esti- MMT and BMT from Doran and others (2003). They estimated mate the avoidable burden of opioid dependence. We initially the cost of MMT at $A 1,415 and of BMT at $A 1,729 for six modeled the avertable burden using MMT and used this model months of treatment.We converted these estimates into U.S.dol- for BMT. The first step was to establish the base case for opioid lars and multiplied them by two to provide yearly estimates dependence using 2002 as the baseline year. We established the of treatment costs of US$1,732 for MMT and US$2,117 for BMT. model of the base case for opioid dependence for regions and We applied relative price weights for each region using the subregions according to WHO country classifications. We used Western Pacific as the reference case (1.00). We calculated the population estimates for each region for those age 15 to 59, the relative price weights for each cost type using data provided by age range in which heroin dependence is most prevalent. We the World Bank. The prices are a reflection of the public health incorporated Degenhardt, Hall, and others' (2004, table 13.1) systems in each region, and as far as possible they reflect the figures for the prevalence of opioid use by region, assuming opportunity cost of health care resources in these regions. that the prevalence was 30 percent higher among male users than female users. Results. Our results are presented in table 48.1. We explored We obtained population-attributable fractions related to various combinations of coverage and reductions in mortality opioid dependence from the editors of this volume. We used for MMT and BMT. For each intervention, as coverage and nine relevant WHO categories to estimate the burden of dis- reductions in mortality increased, the number of DALYs averted ease attributable to opioid dependence--namely HIV/AIDS, increased. The wide discrepancies in DALYs averted within drug-use disorders, road traffic accidents, poisonings, falls, regions primarily reflect differences in population-attributable fires, drownings, other unintentional injuries, and self-inflicted fractions for HIV/AIDS. Costs increased as a consequence of injuries. increased coverage for both interventions, whereas results for We calculated the mortality rate for opioid deaths by divid- cost-effectiveness differ by both intervention and mortality. ing the number of deaths by the estimated number of users. We The cost-effectiveness analysis suggests that for MMT (with took estimates of years of life lost (YLLs) and years lived with a coverage of 25, 50, or 75 percent and reductions in mortality of disability (YLDs), by gender, for each region from data 35 percent) the cost in international dollars per DALY averted obtained from the editors of this volume. We then used those ranges from a low of $128 in Africa, with high child and adult estimates to calculate the DALYs for male users, female users, mortality where the prevalence of illicit opioid dependence is and all users (YLL YLD DALY). We discounted the YLLs, low (0.01 percent), to a high of $3,726 in Eastern Europe, with YLDs, and DALYs using a 3 percent discount rate. low child and adult mortality where the prevalence of illicit The second step was to estimate the avertable burden by opioid dependence is high (0.55 percent). Across all the treatment with methadone or buprenorphine. Using the popu- regions, the average cost-effectiveness ratio for MMT (with 25, lation and prevalence data, we assumed, in the first instance, 50, and 75 percent coverage and 35 percent reduction in mor- that 50 percent of those dependent on opioids entered treat- tality) is estimated at $2,236 per DALY averted. ment. In the sensitivity analysis, we varied this proportion from 25 to 75 percent coverage. On the basis of Caplehorn and Assessment. The results shown in table 48.1 provide a first others' (1994) meta-analysis, we assumed that MMT reduced approximation of the potential avertable burden in DALYs if 914 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others Table 48.1 Cost-Effectiveness Results Total effect (DALYs averted per 1 million population) Eastern The Mediter- Southeast Western Africa Americas ranean Europe Asia Pacific Coverage Mortality Treatment (%) (%) AFR-D AFR-E AMR-A AMR-B AMR-D EMR-A EMR-D EUR-A EUR-B EUR-C SEAR-B SEAR-D WPR-A WPR-B MMT 25 15 125 79 153 107 158 179 105 117 48 198 63 48 39 26 MMT 50 15 251 158 306 214 316 358 210 234 96 397 126 97 77 53 MMT 75 15 376 237 459 321 474 538 315 352 144 595 190 145 116 79 MMT 25 25 150 81 184 121 173 217 151 141 59 264 93 70 51 35 MMT 50 25 300 163 369 243 347 435 303 283 117 527 185 140 102 70 MMT 75 25 450 244 553 364 520 652 454 424 176 791 278 211 152 105 MMT 25 35 174 84 216 136 189 256 198 165 69 329 122 92 63 43 MMT 50 35 349 167 432 272 378 511 396 331 139 657 244 184 126 87 MMT 75 35 523 251 648 408 566 767 594 496 208 986 367 276 189 130 BMT 25 10 113 78 137 100 150 160 82 105 43 166 48 38 32 22 BMT 50 10 226 156 274 199 301 320 163 210 85 331 97 75 65 44 BMT 75 10 339 234 412 299 451 480 245 315 128 497 145 113 97 67 BMT 25 20 138 80 169 114 166 198 128 129 53 231 78 59 45 31 BMT 50 20 275 160 337 228 332 397 256 258 107 462 156 119 89 61 BMT 75 20 413 240 506 342 497 595 384 388 160 693 234 178 134 92 BMT 25 30 162 82 200 129 181 237 175 153 64 296 107 81 57 39 BMT 50 30 324 165 400 258 362 473 350 307 128 592 215 162 114 78 BMT 75 30 487 247 601 386 543 710 524 460 192 888 322 243 171 117 Total costs (US$ per 1 million population) MMT 25 15, 25, 35 0.10 0.01 0.25 0.06 0.12 0.95 0.65 0.20 0.16 0.35 0.06 0.19 0.07 0.03 MMT 50 15, 25, 35 0.19 0.02 0.50 0.11 0.24 1.90 1.30 0.40 0.32 0.71 0.11 0.39 0.13 0.07 MMT 75 15, 25, 35 0.29 0.03 0.74 0.17 0.36 2.86 1.95 0.60 0.49 1.06 0.17 0.58 0.20 0.10 BMT 25 10, 20, 30 0.12 0.01 0.30 0.07 0.15 1.16 0.80 0.24 0.20 0.43 0.07 0.24 0.08 0.04 BMT 50 10, 20, 30 0.24 0.03 0.60 0.14 0.29 2.33 1.59 0.49 0.40 0.86 0.14 0.47 0.16 0.08 BMT 75 10, 20, 30 0.35 0.04 0.91 0.20 0.44 3.49 2.39 0.73 0.59 1.29 0.20 0.71 0.24 0.12 Cost-effectiveness (US$ per DALY averted) MMT 25, 50, 75 15 768 136 1,618 520 755 5,315 6,213 1,711 3,379 1,782 875 3,984 1,716 1,284 MMT 25, 50, 75 25 643 132 1,342 458 688 4,381 4,300 1,419 2,764 1,341 597 2,749 1,301 974 MMT 25, 50, 75 35 552 128 1,146 408 632 3,726 3,288 1,212 2,339 1,074 453 2,099 1,048 784 BMT 25, 50, 75 10 1,041 168 2,204 682 969 7,269 9,764 2,329 4,646 2,606 1,396 6,277 2,493 1,867 BMT 25, 50, 75 20 855 164 1,793 595 880 5,869 6,210 1,895 3,716 1,869 867 3,975 1,809 1,354 BMT 25, 50, 75 30 726 159 1,510 527 805 4,921 4,553 1,598 3,096 1,458 629 2,909 1,419 1,062 DALYs averted per US$1 million spent MMT 25, 50, 75 15 1,302 7,363 618 1,922 1,325 188 161 585 296 561 1,142 251 583 779 MMT 25, 50, 75 25 1,556 7,575 745 2,185 1,453 228 233 705 362 746 1,676 364 768 1,027 MMT 25, 50, 75 35 1,811 7,787 873 2,448 1,582 268 304 825 428 931 2,210 476 954 1,275 BMT 25, 50, 75 10 961 5,939 454 1,465 1,032 138 102 429 215 384 717 159 401 536 BMT 25, 50, 75 15 1,170 6,112 558 1,681 1,137 170 161 528 269 535 1,153 252 553 739 BMT 25, 50, 75 20 1,378 6,286 662 1,896 1,242 203 220 626 323 686 1,590 344 705 942 Illicit Opiate Abuse | 915 MMT and BMT were applied to 50 percent of the opioid- Third, in societies with a sizable illicit opioid dependence dependent population in each region. Because the methods and problem, cultural attitudes and beliefs will affect societal data used to estimate avertable DALYs are subject to certain responses, especially attitudes toward illicit opioid use and limitations, those results should be considered preliminary. dependence (Gerstein and Harwood 1990). A critical determi- nant of the social response will be the relative dominance of moral and medical understandings of drug dependence in gen- RELEVANCE TO DEVELOPING COUNTRIES eral and opioid dependence in particular. A moral model of addiction sees addiction as largely a voluntary behavior, in Much of the epidemiological research on illicit opioid depend- which case it is seen as an excuse for bad behavior that allows ence, its disease burden, and its societal harm comes from drug users to continue to take drugs without assuming respon- Australasia, Europe, and the United States. The major excep- sibility for their conduct (Szasz 1985). In this view, drug users tion is research on the role of injecting drug use in HIV trans- who offend against the criminal code should be imprisoned mission in developing countries (see, for example, Beyrer and (Szasz 1985). This model is the dominant one in many devel- others 2000; Yu and others 1998). In addition, research on the oped societies, which imprison drug users at high rates without effectiveness and cost-effectiveness of interventions for illicit any effect on the prevalence of drug abuse. Countries that opioid dependence has been conducted primarily in developed adopt punitive policies toward drug users are reluctant to countries (Ward, Hall, and Mattick 1998), with the exception of embrace harm reduction measures, such as needle and syringe studies of the effectiveness of methadone treatment in Hong programs and opioid maintenance treatment (Ainsworth, Kong, China (see, for instance, Newman and Whitehill 1979), Beyrer, and Soucat 2003). A medical model of addiction, by and in Thailand (Vanichseni and others 1991), both of which contrast, recognizes that dependent opioid users require spe- showed comparable effectiveness to that found in developed cific treatment if the sufferer is to become and remain abstinent countries (W. Hall, Ward, and Mattick 1998). (see, for example, Leshner 1997). Translating findings on interventions for opioid depend- These competing views will affect the societal acceptability ence in developed countries into disease control priorities for of opioid maintenance and abstinence-oriented approaches to opioid dependence in developing countries presents three the treatment of opioid dependence (Cohen 2003). Those who major challenges. First, countries differ in the scale of illicit have a moral view of addiction will tend to prefer drug-free and opioid use and in the resulting disease burden. This variation self-help approaches toward treatment. Supporters of medical reflects the effects of differences in the prevalence of injecting models of addiction will favor some form of opioid substitu- and noninjecting opioid users; the dependent opioid users' tion treatment and the provision of clean needles and syringes access to treatment and health services for overdoses, blood- to reduce the transmission of bloodborne viruses by injecting borne viruses, and other complications of drug use; the access opioid and other drug users. Stronger advocacy by interna- to needle and syringe programs; the extent to which illicit tional organizations and agencies is needed for the adoption of opioid use is concentrated in socially disadvantaged minority such harm reduction measures as needle and syringe programs groups; and the capacity of public health services to monitor and agonist substitution programs. and respond to emerging infectious disease and drug-use epi- demics. The burden is likely to be greatest in settings where the RESEARCH AND DEVELOPMENT primary route of administration is injecting and where public and personal health services are poorly developed, as appears to Two main areas are important for research and development. be the case in Asia and in Eastern Europe. First, better estimates are needed of the prevalence of illicit opi- Second, societal wealth and health care infrastructure oid dependence and prospective studies of the morbidity and affect the capacity of developing societies to treat illicit opioid mortality that it causes in both developed and developing dependence.A country's capacity to provide opioid substitution countries. These estimates are especially needed in countries treatment will be affected by the cost of oral opioid drugs, such where illicit opioid use is high because of their proximity to as methadone, LAAM, and buprenorphine, and the existence of source countries. Second, we need evaluations of the effective- specialist drug treatment centers; trained medical, nursing, and ness and cost-effectiveness of self-help, drug-free, and oral opi- pharmacy staff; and a drug regulatory system, which are oid substitution treatment in developing countries. A priority required so as to deliver opioid substitution treatment safely should be the identification of safe, innovative, and less expen- and effectively. Few developing countries possess this infra- sive ways of effectively delivering culturally acceptable forms of structure. However, examples exist of apparently successful drug opioid maintenance treatments in developing countries. This substitution programs, using such tools as sublingual buprenor- effort may require experimentation with a range of substitute phine, that have been conducted with minimal resources in opioids, such as buprenorphine, and cheaper options, such as extremely poor settings (Crofts and others 1998). codeine and opium tincture. 916 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others CONCLUSIONS: PROMISES AND PITFALLS and the United States, with the highest rates in Denmark, France and the United Kingdom (EMCDDA 2002; W. Hall and Illicit opioid use, especially injecting use, contributes to prema- Pacula 2003). The limited data from developing countries ture mortality and morbidity in many developed and develop- suggest that, with some exceptions (for example, Jamaica and ing societies. Fatal overdoses and HIV/AIDS resulting from the South Africa), rates of cannabis use are lower in Africa, Asia, sharing of dirty injecting equipment are major contributors the Caribbean, and South America than they are in Europe to mortality and morbidity, and the economic costs of illicit and in English-speaking countries (W. Hall, Johnston, and opioid dependence are substantial. Illicit opioid dependence Donnelly 1999). generates substantial externalities that are not included in Surveys in the United States have found long waves of burden-of-disease estimates, principally law enforcement costs cannabis use among young people since 1975. Cannabis use incurred in handling drug dealing and property crime. increased during the 1970s to peak in 1979, before declining The most popular interventions for illicit opioid depend- steadily between 1980 and 1991. Use rose sharply in 1992 and ence in many developed societies have been law enforcement increased throughout the 1990s, before leveling off in the late efforts to interdict the drug supply and enforce legal sanctions 1990s (Johnston, O'Malley, and Bachman 1994a, 1994b). There against the use of opioid drugs. One consequence of this strat- was also a rise in cannabis use during the early 1990s in egy has been that most illicit opioid users have been exposed Australia, Canada, and some European countries (W. Hall and to the least effective intervention: imprisonment for drug or Pacula 2003). property offenses. Prisons rarely take the opportunity to treat The natural history of cannabis use in the United States dependence using opioid maintenance or to reduce the harm typically begins in the mid to late teens and reaches its maxi- caused by illicit opioid use by providing access to clean inject- mum in the early 20s before declining in the mid to late 20s. ing equipment. Only a minority of young adults continue to use cannabis into In treatment settings, the most popular interventions have their 30s (Bachman and others 1997; Chen and Kandel 1995). been detoxification (which is not a treatment but a prelude to Getting married and having children substantially reduces rates treatment) and drug-free treatment (which is the least attrac- of cannabis use (Bachman and others 1997). tive and the least effective in retaining opioid-dependent peo- Cannibis use can have several adverse health effects, as dis- ple in treatment). Opioid agonist maintenance treatment has cussed below. been ambivalently supported in many developed societies despite its being the treatment for which there is the best evi- Acute Effects of Cannabis Use. The most frequent unpleasant dence of effectiveness, safety, and cost-effectiveness. The range effects of cannabis use are anxiety and panic reactions, which of opioid agonists available for maintenance treatment is most often occur in users who are unfamiliar with the drug's increasing. A number of developed countries have approved effects. Psychotic symptoms such as delusions and hallucina- the use of BMT, which the limited data suggest may be approx- tions may be experienced following very high doses. There are imately equivalent to MMT in efficacy and cost-effectiveness. no cases of fatal cannabis poisoning in the medical literature, Opioid antagonists have a niche role in the treatment of opioid and the fatal dose in humans is likely to exceed what recre- dependence because of poor compliance and an increased ational users are able to ingest (W. Hall and Pacula 2003). risk of overdose on return to heroin use. Their efficacy may Cannabis intoxication impairs a wide range of cognitive and improve with the development of long-acting injectable forms behavioral functions that are involved in driving an automobile of the drug. or operating machinery (Beardsley and Kelly 1999; Jaffe 1985). It has been difficult to determine whether these impairments increase the risk of being involved in motor vehicle accidents ANNEX 48.A: PREVALENCE OF USE, ADVERSE (Smiley 1999). Studies of the effect of cannabis on driving per- HEALTH EFFECTS OF AND INTERVENTIONS formance on the road have found only modest impairments, FOR CANNABIS, COCAINE, AMPHETAMINES, because cannabis-intoxicated drivers drive more slowly and AND MDMA USE AND DEPENDENCE take fewer risks than drivers intoxicated by alcohol (Smiley 1999). Cannabis Cannabinoids are found in the blood of substantial propor- Cannabis is the most widely used illicit drug globally, with tions of persons killed in motor vehicle accidents (Bates and about 150 million users, or 3.7 percent of the world's popula- Blakely 1999; Chesher 1995; Walsh and Mann 1999), but these tion age 15 and older (UNODCCP 2003). Patterns of cannabis findings have been difficult to evaluate because they have not use have been most extensively studied in Australia, Canada, distinguished between past and recent cannabis use the United States, and Europe (W. Hall and Pacula 2003). (Ramaekers and others 2004). More recent research using bet- Europe generally has lower rates of use than Australia, Canada, ter indicators of recent cannabis use has found a dose-response Illicit Opiate Abuse | 917 relationship between cannabis and risk of motor vehicle disease, and coronary atherosclerosis (Chesher and Hall 1999; crashes (Ramaekers and others 2004). Cannabis used in com- Sidney 2002). One controlled study suggests that cannabis use bination with alcohol substantially increases risk of accidents can precipitate heart attacks in middle-aged cannabis users (Bates and Blakely 1999; Ramaekers and others 2004). who have atherosclerosis in the heart, brain, and peripheral blood vessels (Mittleman and others 2001). Health Effects of Chronic Cannabis Use. Cannabis smoke is Regular cannabis smoking impairs the functioning of the a potential cause of cancer because it contains many of the large airways and causes chronic bronchitis (Tashkin 1999; same carcinogenic substances as cigarette smoke (Marselos and Taylor and others 2002). Given that tobacco and cannabis Karamanakos 1999). Cancers have been reported in the aerodi- smoke contain similar carcinogenic substances, it is likely that gestive tracts of young adults who were daily cannabis smokers chronic cannabis smoking increases the risks of respiratory (W. Hall and MacPhee 2002), and a case-control study has cancer (Tashkin 1999). found an association between cannabis smoking and head and neck cancer (Zhang and others 1999). A prospective cohort Psychological Effects of Chronic Cannabis Use. Psychological study of 64,000 adults did not find any increase in rates of head effects of chronic cannabis use can include a dependence and neck or respiratory cancers (Sidney and others 1997). syndrome, cognitive effects, and psychotic disorders. Further studies are needed to clarify the issue. Three studies of different types of cancer have reported Dependence Syndrome A cannabis dependence syndrome an association with maternal cannabis use during pregnancy occurs in heavy chronic users of cannabis (American (W. Hall and MacPhee 2002). There have not been any increases Psychiatric Association 1994). Regular cannabis users develop in the rates of these cancers that parallel increases in rates of tolerance to THC. Some experience withdrawal symptoms on cannabis use (W. Hall and MacPhee 2002). cessation of use (Kouri and Pope 2000), and some report prob- High doses of cannabinoids impair cell-mediated and lems controlling their cannabis use (W. Hall and Pacula 2003). humoral immunity and reduce resistance to infection by The risk of dependence is about 1 in 10 among those who ever bacteria and viruses in rodents (Klein 1999). Cannabis smoke use the drug, between 1 in 5 and 1 in 3 among those who use impairs the functioning of alveolar macrophages, the first line cannabis more than a few times, and about 1 in 2 among daily of the body's immune defense system in the lungs. The doses users (W. Hall and Pacula 2003). that produce these effects have been very high, and extrapola- tion to the doses used by humans is complicated by the fact that Cognitive Effects Long-term daily cannabis use does not tolerance to these effects develops (Hollister 1992). There is as severely impair cognitive function, but it may more subtly yet no epidemiological evidence that rates of infectious disease impair memory, attention, and the ability to integrate complex are higher among chronic heavy cannabis users. Several large information (Solowij 1998; Solowij and others 2002). It remains prospective studies of HIV-positive homosexual men have not uncertain whether these effects are due to the cumulative effect found that cannabis use makes it more likely that HIV-positive of regular cannabis use on cannabinoid receptors in the brain or men develop AIDS (W. Hall and Pacula 2003). whether they are residual effects of THC that will disappear after Chronic administration of tetrahydrocannabinol (THC) an extended period of abstinence (W. Hall and Pacula 2003). disrupts male and female reproductive systems in animals, reducing testosterone secretion and sperm production, motil- Psychotic Disorders There is now good evidence that chronic ity, and viability in males and disrupting ovulation in females cannabis use may precipitate psychosis in vulnerable individu- (Brown and Dobs 2002). It is uncertain whether cannabis use als (see, for example, Arseneault and others 2002; van Os and has these effects in humans because of the limited research on others 2002; Zammit and others 2002). It is less likely that human males and females (Murphy 1999). cannabis use can cause psychosis de novo, because the inci- The use of cannabis during pregnancy is associated with dence of schizophrenia has either remained stable or declined smaller birthweight (English and others 1997; Fergusson, while cannabis use has increased among young adults Horwood, and Northstone 2002), but it does not appear to (Degenhardt, Hall, and Lynskey 2003). increase the risk of birth defects (W. Hall and Pacula 2003). In Effects of Cannabis Use on Adolescents. Cannabis use has a some studies, infants exposed to cannabis during pregnancy number of effects on adolescents. show behavioral and developmental effects during the first few months after birth; these effects are smaller than those seen Gateway Hypothesis Adolescents in developed societies typi- after tobacco use during pregnancy (Fried and Smith 2001). cally use alcohol and tobacco before using cannabis, which in The changes that cannabis smoking causes in heart rate turn, they use before using hallucinogens, amphetamines, and blood pressure are unlikely to harm healthy young adults, heroin, and cocaine (Kandel 2002). Generally, the earlier the but they may harm patients with hypertension, cerebrovascular age of first use and the greater the involvement with any drug 918 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others in the sequence, the more likely a young person is to use the Cocaine next drug in the sequence (Kandel 2002). The role played by After cannabis, cocaine is one of the most widely used illicit cannabis in this sequence remains controversial (W. Hall and drugs in developed and developing societies. Some 14 million Lynskey forthcoming; W. Hall and Pacula 2003). people were estimated to have used cocaine globally in 2003, The simplest hypothesis is that cannabis use has a pharma- with demand for treatment second only to heroin (UNODCCP cological effect that increases the risk of using drugs later in the 2003). The highest rates of reported cocaine use--and the best sequence. Equally plausible hypotheses are that it is due to a data on trends in cocaine use--come from the United States, combination of (a) early recruitment into cannabis use of the world's largest cocaine market. Rates of cocaine use in the nonconforming and deviant adolescents who are likely to use United States increased from the mid 1970s until 1985, when alcohol, tobacco, and illicit drugs; (b) a shared genetic vulnera- 5.7 million Americans age 12 and older reported using cocaine bility to dependence on alcohol, tobacco, and cannabis; and in the preceding month. Rates of cocaine use in the preceding (c) socialization of cannabis users within an illicit drug­using month have declined steadily since 1985. In 2000, 11.2 percent subculture, which increases the opportunity, and encourage- of Americans over age 12 reported that they had used cocaine ment to use other illicit drugs (W. Hall and Pacula 2003). at some time in their lives, and 0.4 percent (800,000 people) reported weekly cocaine use (SAMHSA 2001). Among young Adolescent Psychosocial Outcomes Cannabis use is associated U.S. adults age 18 to 25, lifetime prevalence was 14.9 percent in with early withdrawal from high school, early family formation, 2001, rising slightly to 15.4 percent in 2002 (SAMHSA 2003). poor mental health, and involvement in drug-related crime. In In 2002, annual prevalence figures from student surveys were the case of each of these outcomes, the strong associations in 15 percent lower than 1998 figures and 60 percent lower than cross-sectional data are more modest when account is taken of 1985 figures (UNODCCP 2003). A more recent study of U.S. the fact that cannabis users show characteristics before they use adults age 35 years found that 6 percent of men and 3 percent cannabis that predict these outcomes. For example, they have of women had used cocaine within the preceding 12 months lower academic aspirations and poorer school performance than (Merline and others 2004). peers who do not use cannabis (Lynskey and Hall 2000; Macleod The reported prevalence of cocaine use in other developed and others 2004). Nonetheless, the evidence increasingly sug- societies is much lower than that in the United States. In gests that regular cannabis use adds to the risk of these outcomes Europe, for example, rates of lifetime cocaine use range from in adolescents already at risk (W. Hall and Pacula 2003). 0.5 percent to 5 percent (EMCDDA 2003), compared with 12.3 percent among American adults in 2001 (SAMHSA 2001). Interventions for Cannabis Dependence. Although many Rates of cocaine use in Australia resemble those in Europe, with dependent cannabis users may succeed in quitting without 4.3 percent of adults reporting lifetime use (Darke and others professional help, some are unable to stop on their own and 2000). will need assistance to do so. There has not been a great deal of The prevalence of cocaine use is likely to be lower in devel- research on pharmacological treatments for cannabis depend- oping societies, but the poor quality of the available data makes ence, although a recent study trialed divalproex sodium with it difficult to be sure (UNDCP 1997). There probably has been promising results (Levin and others 2004). Limited research an increase in cocaine use in some developing countries in exists on the effectiveness of different types of psychosocial recent years, but it is difficult to estimate the size of the increase treatments for dependent cannabis use (Budney and others (United Nations Commission on Narcotic Drugs 2000). The 2000; Copeland and others 2001; Stephens, Roffman, and region with the highest rates of cocaine use among developing Simpson 1994). These approaches have involved short-term societies is likely to be Central and South America. The botan- cognitive behavioral treatments modeled on similar treatments ical source is indigenous to the region and has traditionally for alcohol dependence, usually given in three to six sessions on been used by local populations. Moreover, several nations in an outpatient basis. Central and South America have a history of production and In all of these studies, rates of abstinence at the end of export to global markets. Recent reports indicate that cocaine treatment have been modest (20 to 40 percent), and subse- abuse is increasing in South America (UNODCCP 2003), and quent high rates of relapse mean that rates of abstinence after a recent household survey on drug abuse in São Paulo, Brazil, 12 months have been very modest (Budney and Moore 2002). estimated cocaine prevalence at 2.1 percent (Galduroz and Nonetheless, treatment does substantially reduce cannabis use others 2003). and problems. These outcomes are not very different from those observed in the treatment for alcohol and other forms of drug dependence (Budney and Moore 2002). Much more Adverse Health Effects of Cocaine. Most cocaine use is infre- research is needed before sensible advice can be given about the quent; regular cocaine use (monthly or more frequently) can best ways to achieve abstinence from cannabis. be a major public health problem. Regular cocaine users who Illicit Opiate Abuse | 919 inject cocaine or smoke crack cocaine are especially likely to Previous studies have documented a variety of neuropsycho- develop dependence and to experience problems related to logical effects of cocaine use, including deficits in memory and their cocaine use (Platt 1997). In the United States, it has been problem solving (Beatty and others 1995; Hoff and others estimated that one in six of those who ever use cocaine become 1996; O'Malley and others 1992). More recently, a twin study dependent on the drug (Anthony, Warner, and Kessler 1994). indicated that cocaine may lead to impaired attention and High rates of cocaine dependence are found among people motor skills up to one year after the conclusion of heavy use treated for alcohol and drug problems and among arrestees in (Toomey and others 2003). the United States (Anglin and Perrochet 1998). The method by which cocaine is administered can result in In large doses, cocaine may be harmful in both cocaine- adverse health effects (Platt 1997). Snorting cocaine through naive and cocaine-tolerant individuals (Platt 1997; Vasica and the nose can lead to rhinitis, damage to the nasal septum, and Tennant 2002). The vasoconstrictor effects of cocaine in large loss of the sense of smell. Smoking cocaine can lead to respira- doses place great strains on a number of the body's physiolog- tory problems, and injecting cocaine leads to the risks of infec- ical systems (McCann and Ricaurte 2000). Effects on the car- tions and bloodborne viruses associated with all injecting drug diovascular system can result in a range of difficulties, from use. chest pain to fatal cardiac arrests (Lange and Hillis 2001). Users who inject cocaine, either on its own or in combina- Neurological problems include cerebral vascular accidents such tion with heroin ("speedballs"), inject much more frequently as strokes or seizures. Other effects of cocaine can include gas- than other injecting drug users and, as a consequence, engage trointestinal problems such as vomiting, colitis, and bowel in more needle sharing, take more sexual risks, and have infarction and respiratory problems such as asthma, respiratory higher rates of HIV infection (Chaisson and others 1989; collapse, pulmonary edema, and bronchitis. Hyperthermia Schoenbaum and others 1989; van Beek, Dwyer, and Malcolm may occur because of the increased metabolism, peripheral 2001). Associations between cocaine use and HIV risk-taking vasoconstriction, and inability of the thalamus to control have been reported in Europe (Torrens and others 1991), body temperature (Crandall, Vongpatanasin, and Victor 2002). Australia (Darke and others 1992), and the United States Obstetric complications can include irregularities in placental (Chaisson and others 1989). Recent Australian research blood flow, premature labor, and low neonate birthweight has indicated that injecting cocaine users report more prob- (Majewska 1996; Platt 1997; Vasica and Tennant 2002). lems related to injecting drug use--such as vascular problems, Adverse health effects from cocaine are potentially fatal and abscesses, and infections--than other injecting drug users can occur among healthy users irrespective of cocaine dose and (Darke, Kaye, and Topp 2002). frequency of use (Lange and Hillis 2001; Vasica and Tennant The link between cocaine use and HIV risk is not restricted 2002). Although the likelihood of health problems may to those who inject cocaine. Crack smoking has been linked to increase with dosage and frequency of use, there is wide indi- higher levels of needle risk, sexual risk taking, and HIV infec- vidual variation in reactions to cocaine and, therefore, no spe- tion (Chaisson and others 1989; Chirgwin and others 1991; cific combination of conditions under which adverse health Desjalais and others 1992; Grella, Anglin, and Wugalter 1995). effects can be predicted. There is no antidote to cocaine over- Two mechanisms probably underlie the relationship between dose as there is for an overdose of heroin (Platt 1997). cocaine use and HIV infection. First, the short half-life of The impact of cocaine on mental health is also complex. cocaine promotes a much higher frequency of injecting by Although cocaine can produce feelings of pleasure, it may also users than that seen in heroin injectors. Second, cocaine itself result in negative psychological symptoms such as anxiety, disinhibits and stimulates users, encouraging them to take depression, paranoia, hallucinations, and agitation (American greater risks with sexual activity and needle use (Darke and Psychiatric Association 1994). Regular cocaine users experience others 2000). high rates of psychiatric disorders. In the United States, regular Cocaine is associated with a risk of intentional injuries and cocaine users report high rates of anxiety and affective disor- injuries in general. A recent review reported that 28.7 percent of ders (Gawin and Ellinwood 1988; Platt 1997). The repeated use people with intentional injuries and 4.5 percent of injured driv- of large doses of cocaine can also produce a paranoid psychosis ers tested positive for cocaine (Macdonald and others 2003). (Majewska 1996; Manschreck and others 1988; Platt 1997; Satel Users are also at risk of death from an accidental overdose of and Edell 1991). People who are acutely intoxicated by cocaine cocaine. A recent study of accidental deaths from drug overdose can become violent, especially those who develop a paranoid in New York between 1990 and 1998 found that 70 percent of psychosis (Platt 1997). deaths were caused by cocaine, often in combination with opi- Animal studies suggest that cocaine use may be neurotoxic ates (Coffin and others 2003). The causes of cocaine-related in large doses--that is, it can produce permanent changes in deaths are usually related to cardiovascular complications the brain and neurotransmitter systems (Majewska 1996; Platt (Vasica and Tennant 2002), but death may also be due to brain 1997). It is unclear whether use is also neurotoxic in humans. hemorrhage, stroke, and kidney failure (Brands, Sproule, and 920 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others Marshman 1998). Injection of cocaine is most likely to cause an there is evidence of previous or consequent symptoms of overdose, followed by smoking it, with intranasal use involving depression. Other antidepressants have been used with mixed the least risk (Pottieger and others 1992). results: imipramine and trazodone have been found to have Much less is known about nonfatal cocaine overdose. A more adverse effects than desipramine, and fluoxetine has not study in Miami, Florida, found that 40 percent of users had been found to be effective (Mendelson and Mellon 1996). A overdosed on cocaine at least once (Pottieger and others 1992). recent systematic review found no current evidence to support More recently, a study in Brazil found that 20 percent of users the use of antidepressants in the treatment of cocaine depend- had experienced an overdose, with 50 percent knowing some- ence (Lima and others 2003). one who had died from an overdose (Mesquita and others Dopamimetic drugs have also been used to treat cocaine 2001). A study in Sydney, Australia, found that 17 percent of dependence; such treatments are based on the action of cocaine injecting cocaine users and 6 percent of noninjecting cocaine to block reuptake of dopamine. Unfortunately, although some of users had ever overdosed, with 9 percent and 3 percent, respec- these drugs are relatively effective, they also result in quite tively, overdosing in the preceding 12 months (Kaye and Darke severe adverse effects (Mendelson and Mellon 1996). Current 2003). Frequency of cocaine use, severity of dependence, and evidence does not support the clinical use of dopamine ago- route of administration did not predict an overdose, support- nists for cocaine dependence (Soares and others 2003). Opioid ing the view that cocaine overdose is an unpredictable event. antagonists (for example, naltrexone) or opioid mixed agonist- antagonists (such as buprenorphine) have been explored, on Interventions for Cocaine Dependence. Efforts at interven- the basis that cocaine dependence may be accompanied by tion have included pharmacological treatments as well as psy- dependence on opiates. Although there have been problems chotherapy and cognitive behavioral therapy. with compliance with naltrexone therapy (National Research Council Committee on Clinical Evaluation of Narcotic Pharmacological Interventions Despite much research effort Antagonists 1978), buprenorphine has shown promising pre- there are no effective pharmacological treatments for cocaine clinical and clinical trial results (Kosten, Kleber, and Morgan dependence (Kreek 1997; McCance 1997; Mendelson and 1989). Other promising directions include cannabinoid recep- Mellon 1996; Nunes 1997; Silva de Lima and others 2002; van tor antagonists and cortisol synthesis inhibitors (van den Brink den Brink and van Ree 2003). Attempts have been made to and van Ree 2003) and vaccination against the effects of develop longer-acting agonist drugs that act on the same cocaine (Kantak 2003), but there is as yet no evidence on the molecular targets as cocaine without producing its euphoric effectiveness of any of these interventions. effects (for example, methylphenidate) (Kreek 1997) or that Acupuncture has also been used to treat cocaine depend- block its rewarding and euphoric effects (McCance 1997). ence. Auricular acupuncture is frequently used, but the small There has also been a search for drugs that indirectly change number of trials that have been conducted have not provided the effects that cocaine has on the brain by acting on other sufficient evidence of effectiveness (van den Brink and van Ree neurotransmitter systems, such as the serotonergic system 2003). (for example, fluoxetine) (McCance 1997). None of these approaches has produced an effective pharmacotherapy for Psychotherapy and Cognitive Behavioral Therapy The lack of cocaine dependence (Lima and others 2003; Platt 1997; Soares evidence for pharmacological therapy means that treatment and others 2003). for cocaine dependence currently relies on cognitive behavior Development of pharmacological therapies for cocaine therapies combined with contingency management strategies. dependence and their evaluation is complicated by the multiple Unfortunately, psychosocial treatments for cocaine dependence interactive processes that may have contributed--for example, are also of limited effectiveness. Treatments such as therapeutic coexisting substance abuse or mental health issues (Mendelson communities, cognitive behavioral treatments, contingency and Mellon 1996). Many of the approaches to the treatment of management, and 12 step­based self-help approaches benefit cocaine dependence have also been used in treating patients cocaine-dependent people by reducing their rates of cocaine with alcoholism and other substance abuse disorders. use and improving their health and well-being, but rates of A number of drugs have been used to treat cocaine based on relapse to cocaine use after treatment remain high (Platt 1997). their relevance to the symptoms of cocaine dependence (Silva de Mendelson and Mellon (1996) conclude that there are no Lima and others 2002; van den Brink and van Ree 2003). The specific cognitive or behavioral interventions that are uniquely frequency of depressive symptoms has led to the exploration of effective in treating cocaine dependence. However, some success the effectiveness of antidepressant drugs. Desipramine has been has been demonstrated with incentive-based programs in which used with mixed effectiveness for cocaine detoxification and the rewards are provided for urine samples that are free of cocaine, maintenance of abstinence (Covi and others 1994; Gawin, although there is doubt about whether results are sustained Kleber, and Byck 1989), but it appears to be most effective when (Roozen and others 2004). Such programs are generally more Illicit Opiate Abuse | 921 effective when the patient's family and friends are involved for stimulant drugs such as hallucinogens and cocaine (Darke (Higgins and others 1994). Petry and others (2004) suggested and Hall 1995; Hando and Hall 1994; Vincent and others 1998). that contingency management was effective in reducing Globally, Europe is the main center of amphetamine produc- cocaine use in a community-based treatment setting. They tion, particularly Belgium, the Netherlands, and Poland, with found that the benefits of treatment depended on the magni- production increasing in Eastern Europe (UNODCCP 2003). tude of reward, with those earning up to US$240 obtaining bet- Half of all Western European countries reported an increase in ter results than those earning up to US$80. They suggested that amphetamine abuse in 2000, but in 2001 the figure fell to 33 per- this form of intervention may work best for people with more cent (UNODCCP 2003). Lifetime use of amphetamines is severe dependence on cocaine. reported to be between 0.5 percent and 6 percent among A multicenter investigation examining the efficacy of four European Union countries, with the exception of the United psychosocial treatments for cocaine-dependent patients con- Kingdom, where the figure is 11 percent. Denmark and Norway cluded that individual drug counseling in combination with also have relatively higher rates of use (EMCDDA 2003). group drug counseling showed the most promise for effective treatment of cocaine dependence over two forms of traditional Adverse Health Effects of Amphetamine Use. Amphetamine psychotherapy (Crits-Christoph and others 1999). Community users who inject the drug are at high risk of bloodborne infec- reinforcement involving an intensive, biopsychosocial, multi- tions through needle sharing. Amphetamine users are as likely faceted approach to lifestyle change has shown positive effects as opioid users to share injection equipment (Darke, Ross, over four to six weeks and has the advantage of being tailored Cohen, and others 1995; Darke, Ross, and Hall 1995; W. Hall, to individual goals (Roozen and others 2004). Bell, and Carless 1993; Hando and Hall 1994; Kaye and Darke The few studies of the long-term effects of treatment have not 2000; Loxley and Marsh 1991). In addition, the youth of shown particularly encouraging results.A one-year follow-up of amphetamine users places them at risk of sexual transmission the U.S. Drug Abuse Treatment Outcome Studies reported that of diseases such as HIV and hepatitis B virus (although not reductions in the use of cocaine in the year following treatment hepatitis C). Primary amphetamine users have been demon- were associated with longer duration of treatment, particularly strated to be a sexually active group, and small proportions six months or more in long-term residential or outpatient treat- engage in paid sex to support their drug use (Darke, Ross, ments (Hubbard, Craddock, and Anderson 2003). A five-year Cohen, and others 1995; Hando and Hall 1994). Among gay national follow-up study of 45 U.S. treatment programs found and bisexual men, amphetamines may be used to enhance sex- that only 33 percent of the sample had highly favorable out- ual encounters, which may lead to unprotected anal intercourse comes (Flynn and others 2003). and increased risk of HIV infection (Urbina and Jones 2004). High-dose amphetamine use, especially by injection, can result in a schizophreniform paranoid psychosis, associated Amphetamines with loosening of associations, delusions, and hallucinations According to WHO, amphetamines and methamphetamines (Gawin and Ellinwood 1988; Jaffe 1985). The psychosis could are the most widely abused illicit drugs after cannabis, with an be reproduced by the injection of large doses in addicts (Bell estimated 35 million users worldwide (Rawson, Anglin, and 1973) and by the repeated administration of large doses to nor- Ling 2002). mal volunteers (Angrist and others 1974). In Australia, the lifetime prevalence of amphetamine use is High proportions of regular amphetamine injectors describe between 6 and 8 percent in the general population, making symptoms of anxiety, panic attacks, paranoia, and depression. amphetamines the most commonly used illicit drug after The emergence of such symptoms is associated with injecting cannabis during that period (Makkai and McAllister 1998). In the drugs, greater frequency of use, and dependence on amphet- 1998, the lifetime prevalence of amphetamine use was highest amines (W. Hall and others 1996; McKetin and Mattick 1997, (25 percent) among male users age 20 to 29. 1998). Recent evidence also suggests that women may experi- The use of amphetamines is generally less frequent than that ence more emotional effects of amphetamine intoxication than of opioids (Darke and Hall 1995; Darke, Kaye, and Ross 1999; W. men and higher rates of anorexia nervosa than women without Hall, Bell, and Carless 1993; Hando, Topp, and Hall 1997; amphetamine disorders (Holdcraft and Iacono 2004). Vincent and others 1998). This pattern is no doubt due to the In sufficiently high doses, amphetamines can be lethal physical and psychological toll taken by regular amphetamine (Derlet and others 1989). However, the risk is low compared use. Although such use is less frequent overall, however, there is with the high risks of overdose associated with central nervous widespread bingeing on amphetamines, with frequent use over system depressants such as heroin. Typically, amphetamine- several consecutive days, which may be followed by benzodi- related deaths are associated with the effects of amphetamines azepine use to "come down." Polydrug use is particularly com- on the cardiovascular system--for example, cardiac failure and mon among amphetamine users,who show a marked preference cerebral vascular accidents (Mattick and Darke 1995). 922 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others There is evidence that amphetamines are neurotoxic that those who received more Matrix treatment had better (Robinson and Becker 1986). Evidence from animal studies abstinence rates than those who had less treatment but that indicates that heavy amphetamine use results in dopaminergic desipramine had no effect on treatment outcome. depletion (Ellison 1992; Fields and others 1991). The few stud- J. Hall and others (1999) conducted an evaluation of the ies of the neuropsychological effects of amphetamine abuse effectiveness of the Iowa Case Management Project. The proj- report findings similar to those found with cocaine abuse. ect was designed to supplement interventions provided by a Deficits in memory and attention have been attributed to drug abuse treatment agency and is a comprehensive social amphetamine use (McKetin and Mattick 1997, 1998). More work intervention, including outreach activities and provision recently, a twin study indicated that amphetamine abuse might of limited emergency funds. The results of the evaluation lead to impaired attention and motor skills up to one year after showed that comprehensive case management was effective in the conclusion of heavy use (Toomey and others 2003). improving employment status among amphetamine users sub- sequent to treatment. There was an almost significant lower Interventions for Amphetamine Dependence. Treatment for incidence of depression among those who received the pro- methamphetamine abuse has been a relatively recent develop- gram compared with controls. Drug use decreased significantly ment and has generally been based on previous treatments for for clients in both control and program conditions. cocaine abuse (Huber and others 1997). Cretzmeyer and oth- More recently, an Australian study evaluated the effective- ers (2003) reviewed treatments for methamphetamine abuse, ness of brief cognitive-behavioral interventions among regular noting that there has been little research on the effectiveness users of amphetamines (Baker, Boggs, and Lewin 2001). The of drug treatment, probably because many amphetamine researchers found a clinically significant reduction in daily users use multiple drugs. The combination of methampheta- amphetamine use among the intervention groups compared mine use with use of marijuana or other sedating drugs indi- with controls and concluded that further studies of brief cates that effective treatments need to address the use of mul- cognitive-behavioral interventions are feasible and warranted. tiple drugs. A Cochrane Review concluded that evidence for Although some promising interventions have been identified success in treatment of amphetamine dependence is very lim- to assist methamphetamine abusers, no single treatment ited, with no pharmacological treatment demonstrated to be option has yet been established as better than any other in a effective (Srisurapanont, Jarusuraisin, and Kittirattanapaiboon randomized controlled trial (Cretzmeyer and others 2003). 2003). An early study explored the use of aversion therapy in a Methylenedioxymethamphetamine multimodal treatment program using educational groups, Methylenedioxymethamphetamine is more widely known individual counseling, occasional family counseling, and after- as ecstasy or MDMA. In Australia, the lifetime prevalence of care planning. The intervention paired an aversive stimulus MDMA use increased from 1 percent of the population in 1988 (either chemical or electrical) with the act of using metham- to 4.6 percent (about one in 20 persons ) in 1998, with 2.3 per- phetamines. Cocaine use was also treated in this way. After cent reporting MDMA use in the preceding 12 months (Topp 12 months, 53 percent of patients were abstinent and the and others 1998). In 2001, 6.1 percent of Australians age 14 researchers noted that their results were promising, despite a years or older reported lifetime use of MDMA, with 2.9 percent number of limitations to the study (Frawley and Smith 1992). reporting use within the preceding year (Degenhardt, Barker, An intervention combining imipramine, a tricyclic antide- and Topp 2004). Rates of use are generally higher among males pressant, with intensive group counseling has been evaluated than females (3.1 percent versus 1.5 percent). MDMA use in with cocaine and methamphetamine abusers. Patients received the preceding 12 months is most common among those age either a low or higher dose (as needed) of imipramine, as well 20 to 29 (5 percent of females and 12 percent of males) (Topp as intensive group counseling and access to medical and psy- and others 1998). chiatric care. Those who received the higher dose stayed in The availability of MDMA has also increased, as indicated by treatment longer, but the results did not support the use of the proportion of the population who have been offered MDMA imipramine for methamphetamine abuse (Galloway and (from 4 percent in 1988 to 7 percent in 1991) (Makkai and others 1994). McAllister 1998), with 14 percent of those age 14 to 29 reporting The Matrix Program for methamphetamine and cocaine that they had been offered MDMA in the preceding year. abusers has also been evaluated. The Matrix Program uses a Research suggests that the pattern of MDMA use changed cognitive behavioral approach with an emphasis on relapse during the 1990s (Topp and others 1998). Users of MDMA are prevention (Huber and others 1997). The study evaluated the commencing use at a younger age, and they appear to be using effectiveness of three conditions: Matrix treatment alone, larger doses more frequently. The incidence of bingeing on Matrix treatment plus desipramine, and Matrix treatment plus MDMA appears to have increased, as does the prevalence of the placebo (Shoptaw and others 1994). The researchers concluded parenteral use of this drug. The increase in the use of MDMA Illicit Opiate Abuse | 923 by injection has been noted among surveys of MDMA users A recent U.K. study of 430 regular users of MDMA reported and of injecting drug users generally. that 83 percent of participants reported low mood and 80 per- An examination of trends in the United States suggested cent experienced impaired concentration. Long-term effects that, although the use of MDMA has increased over time, its of MDMA included the development of tolerance to MDMA prevalence is significantly less than that of other drugs of abuse (59 percent), impaired ability to concentrate (38 percent), and (Yacoubian 2003b). A study of 14,520 U.S. college students depression (37 percent) (Verheyden and others 2003). indicated 6 percent lifetime use of MDMA, 3 percent within the Physical symptoms that were perceived as being due to preceding 12 months, and 1 percent within the preceding MDMA use alone (Topp and others 1998) included an inability 30 days. Those who had used MDMA in the preceding to urinate, blurred vision, vomiting, numbness or tingling, loss 12 months were more likely to be white and a member of a fra- of sexual urge, and hot and cold flushes. As with amphetamines, ternity or sorority and to have used a range of other drugs the use of MDMA to facilitate sexual encounters may lead to (Yacoubian 2003a). Rates of use are much higher in surveys of risky sexual behavior and risk of sexually transmitted infections club attendees. A recent U.S. survey found 86 percent reporting such as HIV. Studies of gay and bisexual men have found an lifetime use, 51 percent 30-day use, and 30 percent use within association between MDMA use and high-risk sexual behavior the preceding 2 days (Yacoubian and others 2003). (Urbina and Jones 2004). Abuse of MDMA had showed signs of decreasing in Western MDMA has been implicated in a growing number of deaths, Europe but has recently shown signs of increase (UNODCCP both in Australia and in other countries (Henry, Jeffreys, and 2003). Although MDMA use appears to be still diffusing, in Dawling 1992; Solowij 1993; White, Bochner, and Irvine 1997). 2003 only four countries (Ireland, the Netherlands, Spain, and Although the reasons for extreme reactions have yet to be the United Kingdom) reported a rate of more than 3 percent clearly determined, deaths have most often been attributed to use among young adults in the preceding 12 months hyperthermia when MDMA was used at dance venues. A com- (EMCDDA 2003). In the United States, use declined in 2002 for bination of sustained exertion, high ambient temperatures, and the first time, but it increased in other regions, particularly the inadequate fluid replacement appears to compound the effect Caribbean, parts of South America, Oceania, Southeast Asia, of MDMA on thermoregulatory mechanisms, causing a rapid the Near East, and southern Africa (UNODCCP 2003). and fatal rise in body temperature (Topp and others 1998). Lifetime experience of MDMA is reported to range from Some deaths have been attributed to excessive water consump- 0.5 percent to 5 percent in European Union countries, with use tion, which causes cerebral edema (Cook 1996; Matthai and more common in the Netherlands (EMCDDA 2003). others 1996). Population survey findings from New Zealand reported an increase in the preceding-year use of MDMA from 1.5 percent REFERENCES in 1998 to 3.4 percent in 2001. The increase was particularly Ahmed, M. A., T. Zafar, H. Brahmbhatt, G. Imam, S. ul Hassan, J. C. Bareta, evident among young men age 20 to 24 (from 4.3 percent to and S. A. Strathdee. 2003. "HIV/AIDS Risk Behaviors and Correlates of 12.5 percent) (Wilkins and others 2003). Injection Drug Use among Drug Users in Pakistan." Journal of Urban Health 80 (2): 321­29. Adverse Health Effects of MDMA. Early studies of MDMA Ainsworth, M., C. Beyrer, and A. Soucat. 2003. "AIDS and Public Policy: The Lessons and Challenges of `Success' in Thailand." Health Policy use in Australia and the United States documented relatively 64 (1): 13­37. few problems associated with the drug's use (Beck 1990; Beck American Psychiatric Association. 1994. Diagnostic and Statistical Manual and Rosenbaum 1994; Downing 1986; Solowij, Hall, and of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Lee 1992). A survey of 100 MDMA users (Solowij, Hall, and Association. Lee 1992) found that the most common adverse effects were Anglin, M. D. 1988. "The Efficacy of Civil Commitment in Treating Narcotic Drug Addiction." In Compulsory Treatment of Drug Abuse: the side effects of acute use, such as appetite loss, dry mouth, Research and Clinical Practice, ed. C. G. Leukefeld and F. M. Tims, 8­34. palpitations, and bruxism (teeth grinding). Among the few Rockville, MD: National Institute on Drug Abuse. heavy users in the study, only two reported feeling dependent Anglin, M. D., M. L. Brecht, and E. Maddahain. 1989. "Pre-treatment on the drug. Characteristics and Treatment Performance of Legally Coerced versus Voluntary Methadone Maintenance Admissions." Criminology 27 (3): With a change in the pattern of MDMA use in Australia, 537­57. there has been an increase in the MDMA-related harms Anglin, M. D., and B. Perrochet. 1998. "Drug Use and Crime: A Historical reported (Topp and others 1998). Some of the acute physical Review of Research Conducted by the UCLA Drug Abuse Research and psychological adverse effects that MDMA users have Center." Substance Use and Misuse 33 (9): 1871­914. attributed to the use of this drug include energy loss, irritabil- Angrist, B., G. Sathananthan, S. Wilk, and S. Gershon. 1974. "Amphetamine Psychosis: Behavioural and Biochemical Aspects." ity, muscular aches, insomnia, and depression. More chronic Journal of Psychiatric Research 11: 13­23. adverse effects were also reported, including weight loss, Anthony, J. C., L. Warner, and R. Kessler. 1994. "Comparative depression, energy loss, insomnia, anxiety, and teeth problems. Epidemiology of Dependence on Tobacco, Alcohol, Controlled 924 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others Substances, and Inhalants: Basic Findings from the National Brecht, M. L., M. D. Anglin, and J. C. Wang. 1993. "Treatment Effectiveness Comorbidity Survey." Experimental and Clinical Psychopharmacology for Legally Coerced versus Voluntary Methadone Maintenance 2 (3): 244­68. Clients." American Journal of Drug and Alcohol Abuse 19 (1): 89­106. Arseneault, L., M. Cannon, R. Poulton, R. Murray, A. Caspi, and T. E. Brown, T. T., and A. S. Dobs. 2002. "Endocrine Effects of Marijuana." Moffitt. 2002. "Cannabis Use in Adolescence and Risk for Adult Journal of Clinical Pharmacology 42 (Suppl. 11): 90S­96S. Psychosis: Longitudinal Prospective Study." British Medical Journal Bruneau, J., S. B. Brogly, M. W. Tyndall, F. Lamothe, and E. L. Franco. 2004. 325 (7374): 1212­13. "Intensity of Drug Injection as a Determinant of Sustained Injection Atlani, L., M. Carael, J. B. Brunet, T. Frasca, and N. Chaika. 2000. "Social Cessation among Chronic Drug Users: The Interface with Social Change and HIV in the Former USSR: The Making of a New Factors and Service Utilization." Addiction 99 (6): 727­37. Epidemic." Social Science and Medicine 50 (11): 1547­56. Budney, A. J., S. T. Higgins, K. J. Radonovich, and P. L. Novy. 2000. "Adding Australian Institute of Health and Welfare. 1999. "1998 National Drug Voucher-Based Incentives to Coping Skills and Motivational Strategy Household Survey: First Results." Drug Statistics Series 1, Enhancement Improves Outcomes during Treatment for Marijuana Australian Institute of Health and Welfare, Canberra. Dependence." Journal of Consulting and Clinical Psychology 68 (6): Bachman, J. G., K. N. Wadsworth, P. M. O'Malley, L. D. Johnston, and 1051­61. J. Schulenberg. 1997. Smoking, Drinking, and Drug Use in Young Budney, A. J., and B. A. Moore. 2002. "Development and Consequences Adulthood: The Impacts of New Freedoms and New Responsibilities. of Cannabis Dependence." Journal of Clinical Pharmacology 42 Mahwah, NJ: Lawrence Erlbaum. (Suppl. 11): 28S­33S. Baker, A., T. G. Boggs, and T. J. Lewin. 2001. "Randomized Controlled Trial Caplehorn, J. R., S. Dalton, M. C. Cluff, and A. M. Petrenas. 1994. of Brief Cognitive-Behavioral Interventions among Regular Users of "Retention in Methadone Maintenance and Heroin Addicts' Risk of Amphetamine." Addiction 96: 1279­87. Death." Addiction 89 (2): 203­9. Ball, J. C., J. W. Shaffer, and D. N. Nurco. 1983. "The Day-to-Day Cartwright, W. S. 2000."Cost-Benefit Analysis of Drug Treatment Services: Criminality of Heroin Addicts in Baltimore: A Study in the Continuity Review of the Literature." Journal of Mental Health Policy and of Offence Rates." Drug and Alcohol Dependence 12 (2): 119­42. Economics 3 (1): 11­26. Bammer, G., W. van den Brink, P. Gschwend, V. Hendriks, and J. Rehm. Caulkins, J. P., C. P. Rydell, S. M. S. Everingham, J. R. Chiesa, and S. 2003."What Can the Swiss and Dutch Trials Tell Us about the Potential Bushway. 1999. An Ounce of Prevention, a Pound of Uncertainty: The Risks Associated with Heroin Prescribing?" Drug and Alcohol Review Cost-Effectiveness of School-Based Drug Prevention Programs. Santa 22 (3): 363­71. Monica, CA: Rand. Barnett, P. G. 1999. "The Cost-Effectiveness of Methadone Maintenance as Central Committee on the Treatment of Heroin Addicts. 2002. Medical a Health Care Intervention." Addiction 94 (4): 479­88. Co-prescription of Heroin: Two Randomized Controlled Trials. Utrecht, Barnett, P. G., G. S. Zaric, and M. L. Brandeau. 2001. "The Cost- Netherlands: Central Committee on the Treatment of Heroin Addicts. Effectiveness of Buprenorphine Maintenance Therapy for Opiate Chaisson, R. E., P. Bacchetti, D. Osmond, B. Brodie, M. A. Sande, and Addiction in the United States." Addiction 96 (9): 1267­78. A. R. Moss. 1989."Cocaine Use and HIV Infection in Intravenous Drug Bates, M. N., and T. A. Blakely. 1999. "Role of Cannabis in Motor Vehicle Users in San Francisco." Journal of the American Medical Association Crashes." Epidemiologic Reviews 21: 222­32. 261 (4): 561­65. Beardsley, P., and T. Kelly. 1999. "Acute Effects of Cannabis on Human Chen, K., and D. B. Kandel. 1995. "The Natural History of Drug Use from Behavior and Central Nervous System Functions." In The Health Effects Adolescence to the Mid-Thirties in a General Population Sample." of Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, and R. Smart, American Journal of Public Health 85 (1): 41­47. 127­265. Toronto, ON: Centre for Addiction and Mental Health. Chesher, G. 1995. "Cannabis and Road Safety: An Outline of Research Beatty, W. W., V. M. Katzung, V. J. Moreland, and S. J. Nixon. 1995. Studies to Examine the Effects of Cannabis on Driving Skills and "Neuropsychological Performance of Recently Abstinent Alcoholics Actual Driving Performance." In The Effects of Drugs (Other Than and Cocaine Abusers." Drug and Alcohol Dependence 37: 247­53. Alcohol) on Road Safety, ed. Parliament of Victoria Road Safety Beck, J. 1990. "The Public Health Implications of MDMA Use." In Ecstasy: Committee, 67­96. Melbourne, Australia: Road Safety Committee. The Clinical, Pharmacological, and Neurotoxicological Effects of the Drug Chesher, G., and W. D. Hall. 1999. "Effects of Cannabis on the MDMA, ed. S. J. Peroutka. Boston: Kluwer. Cardiovascular and Gastrointestinal Systems." In The Health Effects of Beck, J., and M. Rosenbaum. 1994. Pursuit of Ecstasy: The MDMA Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, and R. Smart, 435­58. Experience. Albany: State University of New York Press. Toronto, ON: Centre for Addiction and Mental Health. Belenko, S. 2002. "The Challenges of Conducting Research in Drug Chirgwin, K., J. A. DeHovitz, S. Dillon, and W. M. McCormack. 1991."HIV Treatment Court Settings." Substance Use and Misuse 37 (12­13): Infection, Genital Ulcer Disease, and Crack Cocaine Use among 1635­64. Patients Attending a Clinic for Sexually Transmitted Diseases." Bell, D. S. 1973. "The Experimental Reproduction of Amphetamine American Journal of Public Health 81 (12): 1576­69. Psychosis." Archives of General Psychiatry 29 (1): 35­40. Coffin, P. O., S. Galea, J. Ahern, A. C. Leon, D. Vlahov, and K. Tardiff. 2003. Beyrer, C. 2003. "Hidden Epidemic of Sexually Transmitted Diseases in "Opiate, Cocaine and Alcohol Combinations in Accidental Drug China: Crisis and Opportunity." Journal of the American Medical Overdose Deaths in New York City, 1990­1998." Addiction 98: 739­47. Association 289 (10): 1303­5. Cohen, J. 2003. "Asia: The Next Frontier for HIV/AIDS." Science 301 Beyrer, C., M. H. Razak, K. Lisam, J. Chen, W. Lui, and X. F. Yu. 2000. (5640): 1650­63. "Overland Heroin Trafficking Routes and HIV-1 Spread in South and Collins, D., and H. Lapsley. 1996. The Social Costs of Drug Abuse in South-East Asia." AIDS 14 (1): 75­83. Australia in 1988 and 1992. Canberra: Australian Government Boot, B., I. McGregor, and W. D. Hall. 2000. "MDMA (Ecstasy) Publishing Service. Neurotoxicity: Assessing and Communicating the Risks." Lancet 355 Cook, T. M. 1996. "Cerebral Oedema after MDMA (`Ecstasy') and (9217): 1818­21. Unrestricted Water Intake." British Medical Journal 313: 689. Brands, B., B. Sproule, and J. Marshman. 1998. Drugs and Drug Abuse. Copeland, J., W. Swift, R. Roffman, and R. Stephens. 2001. "A Randomized 3rd ed. Toronto, ON: Addiction Research Foundation. Controlled Trial of Brief Cognitive-Behavioral Interventions for Illicit Opiate Abuse | 925 Cannabis Use Disorder." Journal of Substance Abuse Treatment 21 (2): Degenhardt, L., W. D. Hall, and M. Lynskey. 2003. "Testing Hypotheses 55­64. about the Relationship between Cannabis Use and Psychosis." Drug Covi, L., I. D. Montoya, J. Hess, and N. Kreiter. 1994. "Double-Blind and Alcohol Dependence 71 (1): 37­48. Comparison of Desipramine and Placebo for Treatment of Cocaine Degenhardt, L., W. D. Hall, M. Warner-Smith, and M. Lynskey. 2004."Illicit Dependence." Clinical Pharmacology and Therapeutics 55: 132. Drug Use." In Comparative Risk Assessment, vol. 1, ed. M. Ezzati, Crandall, C. G., W. Vongpatanasin, and R. G. Victor. 2002. "Mechanism of A. Lopez, and C. Murray, 1109­76. Geneva: World Health Organization. Cocaine-Induced Hyperthermia in Humans." Annals of Internal Derlet, R. W., P. Rice, B. Z. Horowitz, and R. V. Lord. 1989. "Amphetamine Medicine 136: 785­91. Toxicity: Experience with 127 Cases." Journal of Emergency Medicine 7 Cretzmeyer, M., M. V. Sarrazin, D. L. Huber, R. I. Block, and J. A. Hall. (2): 157­61. 2003. "Treatment of Methamphetamine Abuse: Research Findings and Desjalais, D. C., J. Wenston, S. R. Friedman, J. L. Sotheran, R. Maslansky, Clinical Directions." Journal of Substance Abuse Treatment 24 (3): and M. Marmor. 1992. "Crack Cocaine Use in a Cohort of Methadone 267­77. Maintenance Patients." Journal of Substance Abuse Treatment 9: 319­25. Crits-Christoph, P., L. Siqueland, J. Blaine, A. Frank, L. Luborsky, Dolan, K., J. Kimber, C. Fry, J. Fitzgerald, D. MacDonald, and L. S. Onken, and others. 1999. "Psychosocial Treatments for Cocaine F. Trautmann. 2000. "Drug Consumption Facilities in Europe and the Dependence: National Institute on Drug Abuse Collaborative Cocaine Establishment of Supervised Injecting Centres in Australia." Drug and Treatment Study." Archives of General Psychiatry 56 (6): 493­502. Alcohol Review 19 (3): 337­46. Crofts, N., G. Costigan, P. Narayanan, J. Gray, J. Dorabjee, B. Langkham, Dole, V. P., J. W. Robinson, J. Oracca, E. Towns, P. Searcy, and E. Caine. and others. 1998. "Harm Reduction in Asia: A Successful Response to 1969. "Methadone Treatment of Randomly Selected Addicts." Hidden Epidemics--The Asian Harm Reduction Network." AIDS 12 New England Journal of Medicine 280 (25): 1372­75. (Suppl. B): S109­15. Dorabjee, J., and L. Samson. 2000. "A Multi-Centre Rapid Assessment of Darke, S., A. Baker, J. Dixon, A. Wodak, and N. Heather. 1992. "Drug Use Injecting Drug Use in India." International Journal of Drug Policy 11 and HIV Risk-Taking Behaviour among Clients in Methadone (1­2): 99­112. Maintenance Treatment." Drug and Alcohol Dependence 29: 263­68. Doran, C. M., M. Shanahan, R. P. Mattick, R. Ali, J. White, and J. Bell. 2003. Darke, S., and W. D. Hall. 1995. "Levels and Correlates of Polydrug Use "Buprenorphine versus Methadone Maintenance: A Cost-Effectiveness among Heroin Users and Regular Amphetamine Users." Drug and Analysis." Drug and Alcohol Dependence 71 (3): 295­302. Alcohol Dependence 39: 231­35. Downing, J. 1986. "The Psychological and Physiological Effects of MDMA ------. 1997. "The Distribution of Naloxone to Heroin Users." Addiction on Normal Volunteers." Journal of Psychoactive Drugs 18: 335­40. 92 (9): 1195­99. Ellison, G. 1992. "Continuous Amphetamine and Cocaine Have Similar ------. 2003. "Heroin Overdose: Research and Evidence-Based Neurotoxic Effects in Lateral Habenular and Fasciculus Retroflexus." Intervention." Journal of Urban Health 80 (2): 189­200. Brain Research 598: 353­56. Darke, S., S. Kaye, and J. Ross. 1999. "Transitions between the Injection of EMCDDA (European Monitoring Centre for Drugs and Drug Addiction. Heroin and Amphetamines." Addiction 94: 1803­11. 2002. Annual Report on the State of the Drugs Problem in the European Darke, S., S. Kaye, and L. Topp. 2002. "Cocaine Use in New South Wales, Union, 2001. Lisbon: EMCDDA. Australia, 1996­2000: 5-Year Monitoring of Trends in Price, Purity, Availability, and Use from the Illicit Drug Reporting System." Drug and ------. 2003. Annual Report 2003: The State of the Drugs Problem in the Alcohol Dependence 6: 81­88. European Union and Norway. Lisbon: EMCDDA. Darke, S., and J. Ross. 2002. "Suicide among Heroin Users: Rates, Risk English, D., G. Hulse, E. Milne, C. Holman, and C. Bower. 1997. "Maternal Factors, and Methods." Addiction 97 (11): 1383­94. Cannabis Use and Birth Weight: A Meta-Analysis." Addiction 92: 1553­60. Darke, S., J. Ross, J. Cohen, J. Hando, and W. D. Hall. 1995. "Injecting and Sexual Risk-Taking Behavior among Regular Amphetamine Users." Farre, M.,A. Mas, M. Torrens,V. Moreno, and J. Cami. 2002."Retention Rate AIDS Care 7: 17­24. and Illicit Opioid Use during Methadone Maintenance Interventions: A Meta-Analysis." Drug and Alcohol Dependence 65 (3): 283­90. Darke, S., J. Ross, and W. D. Hall. 1995. "Benzodiazepine Use among Injecting Heroin Users." Medical Journal of Australia 162: 645­47. Fergusson, D. M., L. J. Horwood, and M. Lynskey. 1998. "Child and Adolescent Psychiatric Disorders." In Mental Health in New Zealand Darke, S., J. Ross, J. Hando, W. D. Hall, and L. Degenhardt. 2000. Illicit from a Public Health Perspective, eds. P. Ellis and S. Collings, 136­63. Drug Use in Australia: Epidemiology, Use Patterns, and Associated Harm. Wellington: Ministry of Health. National Drug Strategy Monograph 43. Canberra: Department of Health and Aged Care. Fergusson, D. M., L. J. Horwood, and K. Northstone. 2002. "Maternal Use Darke, S., I. Topp, H. Kaye, and W. Hall. 2002. "Heroin Use in New South of Cannabis and Pregnancy Outcome." British Journal of Obstetrics and Wales, Australia, 1996­2000: Five-Year Monitoring of Trends in Price, Gynaecology 109 (1): 21­27. Purity, Availability, and Use from the Illicit Drug Reporting System Fergusson, D. M., L. J. Horwood, and N. Swain-Campbell. 2002."Cannabis (IDRS)." Addiction 97 (2): 179­86. Use and Psychosocial Adjustment in Adolescence and Young Darke, S., and D. Zador. 1996. "Fatal Heroin `Overdose': A Review." Adulthood." Addiction 97 (9): 1123­35. Addiction 91 (12): 1765­72. Fields, J. Z., L. Wichlinski, G. E. Drucker, K. Engh, and J. H. Gordon. 1991. Day, C., L. Degenhardt, S. Gilmour, and W. D. Hall. 2004. "Effects of "Long-Lasting Dopamine Receptor Up-Regulation in Amphetamine- Reduction in Heroin Supply on Injecting Drug Use: Analysis of Data Treated Rats Following Amphetamine Neurotoxicity." Pharmacology, from Needle and Syringe Programmes." British Medical Journal 329 Biochemistry, and Behavior 40 (4): 881­86. (7463): 428­29. Fischer, B., E. Haydon, J. Rehm, M. Krajden, and J. Reimer. 2004."Injection Degenhardt, L., B. Barker, and L. Topp. 2004. "Patterns of Ecstasy Use in Drug Use and the Hepatitis C Virus: Considerations for a Targeted Australia: Findings from a National Household Survey." Addiction Treatment Approach--The Case Study of Canada." Journal of Urban 99 (2): 187­95. Health 81 (3): 428­47. Degenhardt, L., C. Day, and W. D. Hall, eds. 2004. The Causes, Course, and Flynn, P. M., G. W. Joe, K. M. Broome, D. D. Simpson, and B. S. Brown. Consequences of the Heroin Shortage in Australia. Adelaide, Australia: 2003. "Looking Back on Cocaine Dependence: Reasons for Recovery." National Drug Law Enforcement Research Fund. American Journal on Addictions 12 (5): 398­411. 926 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others Fortney, J., B. Booth, M. Zhang, J. Humphrey, and E. Wiseman. 1998. Hall, J., M. Vaughan, T. Vaughn, R. I. Block, D. L. Huber, and A. Schut. "Controlling for Selection Bias in the Evaluation of Alcoholics 1999. "Iowa Case Management for Rural Drug Abuse: Preliminary Anonymous as Aftercare Treatment." Journal of Studies on Alcohol Results." Case Management Journal 1: 232­43. 59 (6): 690­707. Hall, W. D. 1997. "The Role of Legal Coercion in the Treatment of Frawley, P., and J. L. Smith. 1992. "One-Year Follow-up after Multimodal Offenders with Alcohol and Heroin Problems." Australian and Inpatient Treatment for Cocaine and Methamphetamine New Zealand Journal of Criminology 30 (2): 103­20. Dependencies." Journal of Substance Abuse Treatment 9: 271­86. Hall, W. D., J. Bell, and J. Carless. 1993. "Crime and Drug Use among Fried, P. A., and A. R. Smith. 2001. "A Literature Review of the Conse- Applicants for Methadone Maintenance." Drug and Alcohol quences of Prenatal Marihuana Exposure: An Emerging Theme of a Dependence 31: 123­29. Deficiency in Aspects of Executive Function." Neurotoxicology and Hall, W. D., L. J. Degenhardt, and M. T. Lynskey. 1999. "Opioid Overdose Teratology 23 (1): 1­11. Mortality in Australia, 1964­1997: Birth-Cohort Trends." Medical Fuller, C. M., D. C. Ompad, S. Galea, Y. Wu, B. Koblin, and D. Vlahov. 2004. Journal of Australia 171 (1): 34­37. "Hepatitis C Incidence: A Comparison between Injection and Noninjection Drug Users in New York City." Journal of Urban Health Hall, W. D., J. Hando, S. Darke, and J. Ross. 1996. "Psychological Morbidity 81 (1): 20­24. and Route of Administration among Amphetamine Users in Sydney, Australia." Addiction 91 (1): 81­87. Galai, N., M. Safaeian, D. Vlahov, A. Bolotin, and D. D. Celentano. 2003. "Longitudinal Patterns of Drug Injection Behavior in the Alive Study Hall, W. D., L. Johnston, and N. Donnelly. 1999. "Epidemiology of Cohort, 1988­2000: Description and Determinants." American Journal Cannabis Use and Its Consequences." In The Health Effects of Cannabis, of Epidemiology 158 (7): 695­704. ed. H. Kalant, W. Corrigal, W. D. Hall, and R. Smart, 69­125. Toronto, ON: Centre for Addiction and Mental Health. Galduroz, J. C., A. R. Noto, S. A. Nappo, and E. L. Carlini. 2003. "First Household Survey on Drug Abuse in São Paulo, Brazil, 1999: Principal Hall, W. D., and M. Lynskey. Forthcoming. "Testing Hypotheses about the Findings." São Paulo Medical Journal 121 (6): 231­37. Relationship between Cannabis Use and the Use of Other Illicit Drugs." Drug and Alcohol Review. Galloway, G., J. A. Newmeyer, T. Knapp, S. Stalcup, and D. Smith. 1994. "Imipramine for the Treatment of Cocaine and Methamphetamine Hall, W. D., M. Lynskey, and L. Degenhardt. 1999. Heroin Use in Australia: Dependence." Journal of Addictive Diseases 13 (4): 201­16. Its Impact on Public Health and Public Order. NDARC Monograph 42. Gawin, F. H., and E. H. Ellinwood Jr. 1988."Cocaine and Other Stimulants. Sydney, Australia: National Drug and Alcohol Research Centre. Actions, Abuse, and Treatment." New England Journal of Medicine 318 Hall, W. D., and D. MacPhee. 2002. "Cannabis Use and Cancer." Addiction (18): 1173­82. 97 (3): 243­47. Gawin, F. H., H. D. Kleber, and R. Byck. 1989. "Desipramine Facilitation Hall, W. D., and R. P. Mattick. 2000. "Is Ultra-Rapid Opioid Detoxification of Initial Cocaine Abstinence."Archives of General Psychiatry 46: 117­21. a Viable Option in the Treatment of Opioid Dependence?" CNS Drugs Gearing, F. R., and M. D. Schweitzer. 1974. "An Epidemiologic Evaluation 14 (4): 251­55. of Long-Term Methadone Maintenance Treatment for Heroin Hall, W. D., and R. L. Pacula. 2003. Cannabis Use and Dependence: Public Addiction." American Journal of Epidemiology 100 (2): 101­12. Health and Public Policy. Melbourne, Ausralia: Cambridge University Gerstein, D., and H. Harwood. 1990. Treating Drug Problems. Vol. 1 of A Press. Study of Effectiveness and Financing of Public and Private Drug Hall, W. D., J. E. Ross, M. T. Lynskey, M. G. Law, and L. J. Degenhardt. 2000. Treatment Systems. Washington, DC: National Academy Press. "How Many Dependent Heroin Users Are There in Australia?" Medical Gerstein, D., H. Harwood, and N. Suter. 1994. Evaluating Recovery Services: Journal of Australia 173 (10): 528­31. The California Drug and Alcohol Treatment Assessment. Sacramento: Hall, W. D., J. Ward, and R. Mattick. 1998. "The Effectiveness of State of California Health and Welfare Agency, Department of Alcohol Methadone Maintenance Treatment 1: Heroin Use and Crime." In and Drug Programs. Methadone Maintenance Treatment and Other Opioid Replacement Gibson, A. E., C. M. Doran, J. R. Bell, A. Ryan, and N. Lintzeris. 2003. "A Therapies, ed. J. Ward, R. Mattick, and W. D. Hall, 17­57. Amsterdam: Comparison of Buprenorphine Treatment in Clinic and Primary Care Harwood Academic. Settings: A Randomised Trial." Medical Journal of Australia 179 (1): Hamers, F. F., and A. M. Downs. 2003. "HIV in Central and Eastern 38­42. Europe." Lancet 361 (9362): 1035­44. Goldschmidt, P. G. 1976. "A Cost-Effectiveness Model for Evaluating Health Care Programs: Application to Drug Abuse Treatment." Inquiry Hando, J., and W. D. Hall. 1994. "HIV Risk-Taking Behavior among 13 (1): 29­47. Amphetamine Users in Sydney, Australia." Addiction 89 (1): 79­85. Goldstein, A., and J. Herrera. 1995. "Heroin Addicts and Methadone Hando, J., L. Topp, and W. D. Hall. 1997. "Amphetamine-Related Harms Treatment in Albuquerque: A 22-Year Follow-up." Drug and Alcohol and Treatment Preferences of Regular Amphetamine Users in Sydney, Dependence 40 (2): 139­50. Australia." Drug and Alcohol Dependence 46 (1-2): 105­13. Gossop, M., J. Marsden, and D. Stewart. 1998. NTORS at One Year: Hawkins, J. D., R. F. Catalano, and J. Y. Miller. 1992. "Risk and Protective Changes in Substance Use, Health, and Criminal Behaviour One Year Factors for Alcohol and Other Drug Problems in Adolescence and after Intake. London: Department of Health. Early Adulthood: Implications for Substance Abuse Prevention." Psychological Bulletin 112 (1): 64­105. Gossop, M., J. Marsden, D. Stewart, C. Edwards, P. Lehmann, A. Wilson, and G. Segar. 1997. "The National Treatment Outcome Research Study Henry, J. A., K. L. Jeffreys, and S. Dawling. 1992. "Toxicity and Deaths from in the United Kingdom: Six-Month Follow-up Outcomes." Psychology 3,4-Methylenedioxymethamphetamine (`Ecstasy')." Lancet 340: 384­87. of Addictive Behaviors 11 (4): 324­37. Hepatitis C Virus Projections Working Group. 1998. Estimates and Grella, C. E., M. D. Anglin, and S. E. Wugalter. 1995."Cocaine and Crack Use Projections of the Hepatitis C Virus Epidemic in Australia. Sydney, and HIV Risk Behaviors among High-Risk Methadone Maintenance Australia: National Centre in HIV Epidemiology and Clinical Research. Clients." Drug and Alcohol Dependence 37 (1): 15­21. Hien, N. T., L. T. Giang, P. N. Binh, W. Deville, E. J. van Ameijden, and Gronbladh, L., L. Ohlund, and L. Gunne. 1990. "Mortality in Heroin I. Wolffers. 2001. "Risk Factors of HIV Infection and Needle Sharing Addiction: Impact of Methadone Treatment." Acta Psychiatrica among Injecting Drug Users in Ho Chi Minh City, Vietnam." Journal of Scandinavica 82 (3): 223­27. Substance Abuse 13 (1­2): 45­58. Illicit Opiate Abuse | 927 Higgins, S. T., A. J. Budney, W. K. Bickel, F. E. Foerg, R. Donham, and Kimber, J., K. Dolan, I. van Beek, D. Hedrich, and H. Zurhold. 2003. "Drug G. J. Badger. 1994. "Incentives Improve Outcome in Outpatient Consumption Facilities: An Update since 2000." Drug and Alcohol Behavioral Treatment of Cocaine Dependence." Archives of General Review 22 (2): 227­33. Psychiatry 51 (7): 568­76. Klein, T. 1999. "Cannabis and Immunity." In The Health Effects of Hoff, A. L., H. Riordan, L. Morris, V. Cestaro, M. Wieneke, R. Alpert, and Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, and R. Smart, 347­73. G. J. Wang. 1996. "Effects of Crack Cocaine on Neurocognitive Toronto, ON: Centre for Addiction and Mental Health. Function." Psychiatry Research 60: 167­76. Kosten, T. R., H. D. Kleber, and C. Morgan. 1989. "Role of Opioid Holdcraft, L. C., and W. G. Iacono. 2004. "Cross-Generational Effects on Antagonists in Treating Intravenous Cocaine Abuse." Life Science 44: Gender Differences in Psychoactive Drug Abuse and Dependence." 887­92. Drug and Alcohol Dependence 74 (2): 147­58. Kouri, E. M., and H. G. Pope. 2000. "Abstinence Symptoms during Hollister, L. 1992. "Marijuana and Immunity." Journal of Psychoactive Withdrawal from Chronic Marijuana Use." Experimental and Clinical Drugs 24: 159­64. Psychopharmacology 8 (4): 483­92. Hser, Y. I., D. Anglin, and K. Powers. 1993. "A 24-Year Follow-up of Kreek, M. J. 1997. "Opiate and Cocaine Addictions: Challenge for Pharma- California Narcotics Addicts." Archives of General Psychiatry 50 (7): cotherapies." Pharmacology, Biochemistry, and Behavior 57 (3): 551­69. 577­84. Lange, R. A., and L. D. Hillis. 2001. "Cardiovascular Complications of Hubbard, R. L., J. J. Collins, J. V. Rachal, and E. R. Cavanaugh. 1988. "The Cocaine Use." New England Journal of Medicine 345: 351­58. Criminal Justice Client in Drug Abuse Treatment." In Compulsory Langendam, M. W., G. H. van Brussel, R. A. Coutinho, and E. J. van Treatment of Drug Abuse: Research and Clinical Practice, ed. Ameijden. 2001. "The Impact of Harm-Reduction-Based Methadone C. G. Leukefeld and F. M. Tims, 57­80. Rockville, MD: National Treatment on Mortality among Heroin Users." American Journal of Institute on Drug Abuse. Public Health 91 (5): 774­80. Hubbard, R. L., S. G. Craddock, and J. Anderson. 2003. "Overview of Leshner, A. I. 1997. "Addiction Is a Brain Disease, and It Matters." Science 5-Year Followup Outcomes in the Drug Abuse Treatment 278 (5335): 45­47. Outcome Studies (Datos)." Journal of Substance Abuse Treatment 25 Levin, F. R., D. McDowell, S. M. Evans, E. Nunes, E. Akerele, S. Donovan, (3): 125­34. and S. K. Vosburg. 2004. "Pharmacotherapy for Marijuana Hubbard, R. L., M. Marsden, J. V. Rachal, H. Harwood, E. Cavanaugh, and Dependence: A Double-Blind, Placebo-Controlled Pilot Study of H. Ginzburg. 1989. Drug Abuse Treatment: A National Study of Divalproex Sodium." American Journal on Addictions 13 (1): 21­32. Effectiveness. Chapel Hill: University of North Carolina Press. Lima, M. S., A. A. Reisser, B. G. Soares, and M. Farrell. 2003. "Anti- Huber, A., W. Ling, S. Shoptaw, V. Gulati, P. Brethen, and R. Rawson. 1997. depressants for Cocaine Dependence." Cochrane Database of "Integrating Treatments for Methamphetamine Abuse: A Psychosocial Systematic Reviews (2): CD002950. [PMID: 12804445]. Perspective." Journal of Addictive Diseases 16 (4): 41­50. Loxley, W., and A. Marsh. 1991. "Nodding and Speeding: Age and Hulse, G. K., D. R. English, E. Milne, and C. D. Holman. 1999. "The Injecting Drug Use in Perth." National Centre for Research into the Quantification of Mortality Resulting from the Regular Use of Illicit Prevention of Drug Abuse, Curtin University of Technology, Perth, Opiates." Addiction 94 (2): 221­29. Australia. Jaffe, J. 1985. "Drug Addiction and Drug Abuse." In The Pharmacological Lynskey, M., and W. D. Hall. 2000. "The Effects of Adolescent Basis of Therapeutics, eds. A. Gilman, L. Goodman and F. Murad, Cannabis Use on Educational Attainment: A Review." Addiction 96 (3): 532­81. New York: Macmillan. 433­43. Johnston, L. D., P. M. O'Malley, and J. G. Bachman. 1994a. National Survey Macdonald, S., K. Anglin-Bodrug, R. E. Mann, P. Erickson, A. Hathaway, Results on Drug Use from the Monitoring the Future Study, 1975­1993: M. Chipman, and M. Rylett. 2003."Injury Risk Associated with Cannabis College Students and Young Adults. Rockville, MD: National Institute on and Cocaine Use." Drug and Alcohol Dependence 72 (2): 99­115. Drug Abuse. Macleod, J., R. Oakes,A. Copello, I. Crome, M. Egger, M. Hickman, and oth- ------. 1994b. National Survey Results on Drug Use from the Monitoring ers. 2004. "Psychological and Social Sequelae of Cannabis and Other the Future Study, 1975­1993: Secondary School Students. Rockville, MD: Drug Use by Young People: A Systematic Review of Longitudinal, National Institute on Drug Abuse,. General Population Studies." Lancet 363 (9421): 1579­88. Kaldor, J., H. Lapsley, R. P. Mattick, D. Weatherburn, and A. Wilson. 2003. Maddux, J. F., and D. P. Desmond. 1992. "Methadone Maintenance and Final Report on the Evaluation of the Sydney Medically Supervised Recovery from Opioid Dependence." American Journal of Drug and Injecting Centre. Sydney, Australia: Medically Supervised Injecting Alcohol Abuse 18 (1): 63­74. Centre Evaluation Committee. Majewska, M. D., ed. 1996. Neurotoxicity and Neuropathology Associated Kandel, D. B., ed. 2002. Stages and Pathways of Drug Involvement: with Cocaine Abuse. NIDA Research Monograph 163. Rockville, MD: Examining the Gateway Hypothesis. New York: Cambridge University U.S. Department of Health and Human Services. Press. Makkai, T., and I. McAllister. 1998. Patterns of Drug Use in Australia, Kantak, K. M. 2003. "Vaccines against Drugs of Abuse: A Viable Treatment 1985­95. Canberra: Australian Government Publishing Service. Option?" Drugs 63 (4): 341­52. Manschreck, T. C., J.A. Laughery, C. C.Weisstein, D.Allen, B. Humblestone, Kaye, S., and S. Darke. 2000. "A Comparison of the Harms Associated with M. Neville, and others. 1988. "Characteristics of Freebase Cocaine the Injection of Heroin and Amphetamines." Drug and Alcohol Psychosis." Yale Journal of Biology and Medicine 61 (2): 115­22. Dependence 58 (1­2): 189­95. Manski, C. F., J. V. Pepper, and C. V. Petrie, eds. 2001. Informing America's ------. 2003. "Non-Fatal Cocaine Overdose and Other Adverse Events Policy on Illegal Drugs: What We Don't Know Keeps Hurting Us. among Injecting and Non-Injecting Cocaine Users." NDARC Technical Washington, DC: National Academy Press. Report 170, National Drug and Alcohol Research Centre, University of Marsch, L. A. 1998. "The Efficacy of Methadone Maintenance New South Wales, Sydney, Australia. Interventions in Reducing Illicit Opiate Use, HIV Risk Behavior, and Kelly, J. A., and Y. A. Amirkhanian. 2003. "The Newest Epidemic: A Review Criminality: A Meta-Analysis." Addiction 93 (4): 515­32. of HIV/AIDS in Central and Eastern Europe." International Journal of Marselos, M., and P. Karamanakos. 1999. "Mutagenicity, Developmental STD and AIDS 14 (6): 361­71. Toxicity and Carcinogeneity of Cannabis." Addiction Biology 4 (1): 5­12. 928 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others Marsh, K., G. Joe, D. Simpson, and W. Lehman. 1990. "Treatment History." Newcomb, M. D., and P. M. Bentler. 1988. Consequences of Adolescent Drug In Opioid Addiction and Treatment: A 12-Year Follow-Up, eds. D. Use: Impact on the Lives of Young Adults. Thousand Oaks, CA: Sage. Simpson and S. Sells, 137­56. Malabar, FL: Krieger. Newman, R. G., and W. B. Whitehill. 1979. "Double-Blind Comparison of Mathers, C., T. Vos, and C. Stephenson. 1999. The Burden of Disease and Methadone and Placebo Maintenance Treatments of Narcotic Addicts Injury in Australia. Canberra: Australian Institute of Health and Welfare. in Hong Kong." Lancet 2 (8141): 485­88. Matthai, S. M., J. A. Sills, D. C. Davidson, and D. Alexandrou. 1996. Nunes, E. V. 1997. Methodologic Recommendations for Cocaine Abuse "Cerebral Oedema after Ingestion of MDMA (`Ecstasy') and Clinical Trials: A Clinician-Researcher's Perspective. In Medication Unrestricted Intake of Water." British Medical Journal 312: 1359. Development for the Treatment of Cocaine Dependence: Issues in Clinical Mattick, R. P., and S. Darke. 1995. "Drug Replacement Treatments: Is Efficacy Trials, NIDA Research Monograph 175, 73­95. Rockville, MD: Amphetamine Substitution a Horse of a Different Color?" Drug and U.S. Department of Health and Human Services. Alcohol Review 14: 389­94. Office of National Drug Control Policy. 2001. Pulse Check: Trends in Drug Mattick, R. P., E. Digiusto, C. M. Moran, S. O'Brien, M. Shanahan, J. Abuse, November 2001. Washington, DC: Executive Office of the Kimber, N. Henderson, C. Breen, J. Shearer, J. Gates, A. Shakeshaft, President, Office of National Drug Control Policy. NEPOD Trial Investigators. 2001. "National Evaluation of Oliveto, A., and T. Kosten. 1997. "Buprenorphine." In New Treatments for Pharmacotherapies for Opioid Dependence (NEPOD)." Canberra: Opioid Dependence, eds. S. Stine and T. Kosten, 25­67. New York: Commonwealth Department of Health and Ageing. Guilford Press. Mattick, R. P., and W. D. Hall. 1996. "Are Detoxification Programmes O'Malley, S., M. Adamse, R. K. Heaton, and F. H. Gawin. 1992. Effective?" Lancet 347 (8994): 97­100. "Neuropsychological Impairment in Chronic Cocaine Abusers." Mattick, R. P., J. Kimber, C. Breen, and M. Davoli. 2003. "Buprenorphine American Journal of Drug and Alcohol Abuse 18: 131­44. Maintenance versus Placebo or Methadone Maintenance for Opioid Perneger, T. V., F. Giner, M. del Rio, and A. Mino. 1998. "Randomised Trial Dependence." Cochrane Database Systematic Review (2) CD002207 of Heroin Maintenance Programme for Addicts Who Fail in [PMID: 12804429]. Conventional Drug Treatments." British Medical Journal 317 (7150): McCance, E. F. 1997. "Overview of Potential Treatment Medications for 13­8. Cocaine Dependence." In Medication Development for the Treatment of Petry, N. M., J. Tedford, M. Austin, C. Nich, K. M. Carroll, and Cocaine Dependence: Issues in Clinical Efficacy Trials, NIDA Research B. J. Rounsaville. 2004. "Prize Reinforcement Contingency Manage- Monograph 175, 36­72. Rockville, MD: U.S. Department of Health ment for Treating Cocaine Users: How Low Can We Go, and with and Human Services. Whom?" Addiction 99 (3): 349­60. McCann, U. D., and G. A. Ricaurte. 2000. "Drug Abuse and Dependence: Platt, J. J. 1997. Cocaine Addiction: Theory, Research and Treatment. Hazards and Consequences of Heroin, Cocaine, and Amphetamines." Cambridge, MA: Harvard University Press. Current Opinion in Psychiatry 13: 321­25. Porter, L., A. Arif, and W. Curran. 1986. The Law and the Treatment of Drug McGregor, C., R. Ali, P. Christie, and S. Darke. 2001. "Overdose among and Alcohol Dependent Persons: A Comparative Study of Existing Heroin Users: Evaluation of an Intervention in South Australia." Legislation. Geneva: World Health Organization. Addiction Research 9 (5): 481­501. Pottieger, A. E., P. A. Tressell, J. A. Inciardi, and T. A. Rosales. 1992. McKetin, R., and R. P. Mattick. 1997. "Attention and Memory in Illicit "Cocaine Use Patterns and Overdose." Journal of Psychoactive Drugs 24: Amphetamine Users." Drug and Alcohol Dependence 48 (3): 235­42. 399­410. ------. 1998. "Attention and Memory in Illicit Amphetamine Users: Ramaekers, J. G., G. Berghaus, M. van Laar, and O. H. Drummer. 2004. Comparison with Non-Drug-Using Controls." Drug and Alcohol "Dose Related Risk of Motor Vehicle Crashes after Cannabis Use." Drug Dependence 50 (2): 181­4. and Alcohol Dependence 73 (2): 109­19. Mendelson, J. H., and N. K. Mellon. 1996. "Management of Cocaine Abuse Rawson, R., M. Anglin, and W. Ling. 2002. "Will the Methamphetamine and Dependence." New England Journal of Medicine 334 (15): 965­72. Problem Go Away?" Journal of Addictive Diseases 21: 5­19. Merline, A. C., P. M. O'Malley, J. E. Schulenberg, J. G. Bachman, and Rehm, J., P. Gschwend, T. Steffen, F. Gutzwiller, A. Dobler-Mikola, and L. D. Johnston. 2004. "Substance Use among Adults 35 Years of Age: A. Uchtenhagen. 2001. "Feasibility, Safety, and Efficacy of Injectable Prevalence, Adulthood Predictors, and Impact of Adolescent Substance Heroin Prescription for Refractory Opioid Addicts: A Follow-Up Use." American Journal of Public Health 94 (1): 96­102. Study." Lancet 358 (9291): 1417­23. Mesquita, F., A. Kral, A. Reingold, I. Haddad, M. Sanches, G. Turienzo, and Reid, G., and N. Crofts. 2000. "Rapid Assessment of Drug Use and HIV others. 2001."Overdoses among Cocaine Users in Brazil." Addiction 96: Vulnerability in South-East and East Asia." International Journal of 1809­13. Drug Policy 11 (1­2): 113­24. Mittleman, M. A., R. A. Lewis, M. Maclure, J. B. Sherwood, and J. E. Muller. Rhodes, T., A. Ball, G. V. Stimson, Y. Kobyshcha, C. Fitch, V. Pokrovsky, and 2001. "Triggering Myocardial Infarction by Marijuana." Circulation others. 1999. "HIV Infection Associated with Drug Injecting in the 103: 2805­9. Newly Independent States, Eastern Europe: The Social and Economic Murphy, L. 1999. "Cannabis Effects on Endocrine and Reproductive Context of Epidemics." Addiction 94 (9): 1323­36. Function." In The Health Effects of Cannabis, ed. H. Kalant, Rhodes, T., L. Mikhailova, A. Sarang, C. M. Lowndes, A. Rylkov, W. Corrigall, W. D. Hall, and R. Smart, 375­400. Toronto, ON: Centre M. Khutorskoy, and A. Renton. 2003. "Situational Factors Influencing for Addiction and Mental Health. Drug Injecting, Risk Reduction, and Syringe Exchange in Togliatti City, National Centre in HIV Epidemiology and Clinical Research. 1998. Russian Federation: A Qualitative Study of Micro Risk Environment." HIV/AIDS and Related Diseases in Australia: Annual Surveillance Social Science and Medicine 57 (1): 39­54. Report 1998. Sydney, Australia: National Centre in HIV Epidemiology Robinson, T. E., and J. B. Becker. 1986. "Enduring Changes in Brain and and Clinical Research. Behavior Produced by Chronic Amphetamine Administration: A National Research Council Committee on Clinical Evaluation of Narcotic Review and Evaluation of Animal Models of Amphetamine Psychosis." Antagonists. 1978. "Clinical Evaluation of Naltrexone Treatment of Brain Research 396 (2): 157­98. Opiate-Dependent Individuals." Archives of General Psychiatry 35: Roozen, H. G., J. J. Boulogne, M. W. van Tulder, W. van den Brink, C. A. De 355­40. Jong, and A. J. Kerkhof. 2004. "A Systematic Review of the Effectiveness Illicit Opiate Abuse | 929 of the Community Reinforcement Approach in Alcohol, Cocaine, and ------. 1998. Cannabis and Cognitive Functioning. Cambridge, U.K.: Opioid Addiction." Drug and Alcohol Dependence 74 (1): 1­13. Cambridge University Press. Ruan, Y. H., K. X. Hong, S. Z. Liu, Y. X. He, F. Zhou, G. M. Qin, and others. Solowij, N., W. D. Hall, and N. Lee. 1992. "Recreational MDMA Use in 2004. "Community-Based Survey of HCV and HIV Coinfection in Sydney: A Profile of `Ecstasy' Users and Their Experiences with the Injection Drug Abusers in Sichuan Province of China." World Journal Drug." British Journal of Addiction 87: 1161­72. of Gastroenterology 10 (11): 1589­93. Solowij, N., R. S. Stephens, R. A. Roffman, T. Babor, R. Kadden, M. Miller, SAMHSA (Substance Abuse and Mental Health Services Adminstration). and others. 2002. "Cognitive Functioning of Long-Term Heavy 2001. Summary of Findings from the 2000 National Household Survey on Cannabis Users Seeking Treatment." Journal of the American Medical Drug Abuse. Rockville, MD: Office of Applied Statistics, SAMHSA. Association 287 (9): 1123­31. ------. 2002. Results from the 2001 Household Survey on Drug Abuse. Spooner, C. and W. D. Hall. 2002. "Public Policy and the Prevention of Vol. 1 of Summary of National Findings. Rockville, MD: Office of Substance Use Disorders."Current Opinion in Psychiatry 15 (3): 235­39. Applied Statistics, SAMHSA. Spooner, C., W. D. Hall, and R. P. Mattick. 2001."An Overview of Diversion ------. 2003. Overview of Findings from the 2002 National Survey on Drug Strategies for Australian Drug-Related Offenders." Drug and Alcohol Use and Health Office. Rockville, MD: Office of Applied Statistics, Review 20 (3): 281­94. SAMHSA. Sporer, K. A. 2003. "Strategies for Preventing Heroin Overdose." British Satel, S. L., and W. S. Edell. 1991."Cocaine-Induced Paranoia and Psychosis Medical Journal 326 (7386): 442­44. Proneness." American Journal of Psychiatry 148 (12): 1708­11. Srisurapanont, M., N. Jarusuraisin, and P. Kittirattanapaiboon. 2003. Schoenbaum, E. E., D. Hartel, P. A. Selwyn, R. S. Klein, K. Davenny, "Treatment for Amphetamine Dependence and Abuse." Cochrane M. Rogers, and others. 1989. "Risk Factors for Human Database of Systematic Reviews (4): CD003022 [PMID: 11687171]. Immunodeficiency Virus Infection in Intravenous Drug Users." New Stephens, R. S., R. A. Roffman, and E. E. Simpson. 1994. "Treating Adult England Journal of Medicine 321 (13): 874­49. Marijuana Dependence--A Test of the Relapse Prevention Model." Shepard, D., M. J. Larson, and N. G. Hoffmann. 1999. "Cost-Effectiveness Journal of Consulting and Clinical Psychology 62 (1): 92­99. of Substance Abuse Services: Implications for Public Policy." Strang, J., S. Darke, W. D. Hall, M. Farrell, and R. Ali. 1996. "Heroin Psychiatric Clinics of North America 22 (2): 385­400. Overdose: The Case for Take-Home Naloxone: Home-Based Supplies Shepard, D., G. Strickler, J. McKay, D. Bury-Maynard, H. Yeom, C. Love, of Naloxone Would Save Lives." British Medical Journal 312 (7044): and others. Forthcoming. "Cost-Effectiveness of Self-Help for 1435­36. Controlling Substance Use: Controlling for Self-Selection." Strathdee, S. A., T. Zafar, H. Brahmbhatt, A. Baksh, and S. ul Hassan. 2003. Shoptaw, S., R. Rawson, M. McCann, and J. Obert. 1994. "The "Rise in Needle Sharing among Injection Drug Users in Pakistan dur- Matrix Model of Outpatient Stimulant Abuse Treatment: Evidence ing the Afghanistan War." Drug and Alcohol Dependence 71 (1): 17­24. of Efficacy." In Experimental Therapeutics in Addiction Medicine, ed. S. Magura and S. Rosenblum, 129­41. Binghamton, NY: Haworth Szasz, T. 1985. Ceremonial Chemistry: The Ritual Persecution of Drugs, Press. Addicts, and Pushers. Holmes Beach, FL: Learning Publications. Sidney, S. 2002. "Cardiovascular Consequences of Marijuana Use." Journal Tashkin, D. P. 1999."Effects of Cannabis on the Respiratory System." In The of Clinical Pharmacology 42 (11 Suppl.): 64S­70S. Health Effects of Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, and R. Smart,311­45.Toronto,ON: Centre forAddiction and Mental Health. Sidney, S., J. E. Beck, I. S. Tekawa, C. P. Quesenberry, and G. D. Friedman. 1997. "Marijuana Use and Mortality." American Journal of Public Taylor, D. R., D. M. Fergusson, B. J. Milne, L. J. Horwood, T. E. Moffitt, Health 87 (4): 585­90. M. R. Sears, and R. Poulton. 2002. "A Longitudinal Study of the Effects of Tobacco and Cannabis Exposure on Lung Function in Young Silva de Lima, M., B. Garcia de Oliveira Soares, A. Alves Pereira Reisser, and Adults." Addiction 97 (8): 1055­61. M. Farrell. 2002. "Pharmacological Treatment of Cocaine Dependence: A Systematic Review." Addiction 97: 931­49. Tonigan, J., R. Toscova, and W. Miller. 1996."Meta-Analysis of the Literature on Alcoholics Anonymous: Sample and Study Characteristics Moderate Simpson, D. S., and H. J. Friend. 1988. "Legal Status and Long-Term Findings." Journal of Studies on Alcohol 57 (1): 65­72. Outcomes for Addicts in the DARP Followup Project." In Compulsory Treatment of Drug Abuse: Research and Clinical Practice, ed. Tonigan, J. S., G. J. Connors, and W. R. Miller. 2003. "Participation and C. G. Leukefeld and F. M. Tims, 81­96. Rockville, MD: National Involvement in Alcoholics Anonymous." In Treatment Matching in Institute on Drug Abuse. Alcoholism, ed. T. F. Babor and F. K. Del Boca, 184­204. Cambridge, UK: Cambridge University Press. Simpson, D. S., G. W. Joe, W. E. K. Lehman, and S. B. Sells. 1986."Addiction Careers: Etiology, Treatment, and 12-Year Follow-up Outcomes." Toomey, R., M. J. Lyons, S. A. Eisen, H. Xian, S. Chantarujikapong, Journal of Drug Issues 16 (1): 107­21. L. J. Seidman, and others. 2003. "A Twin Study of the Neuropsychological Consequences of Stimulant Abuse." Archives of Simpson, D. S., and S. Sells. 1982. "Effectiveness of Treatment for Drug General Psychiatry 60 (3): 303­10. Abuse: An Overview of the DARP Research Program." Advances in Alcohol and Substance Abuse 2 (1): 7­29. Topp, L., J. Hando, L. Degenhardt, P. Dillon, A. Roche, and N. Solowij. 1998. Ecstasy Use in Australia. NDARC Monograph 39. Sydney, ------, eds. 1990. Opioid Addiction and Treatment: A 12-Year Follow-up. Australia: National Drug and Alcohol Research Centre, University of Malabar, FL: Krieger. New South Wales. Smiley, A. 1999. "Marijuana: On Road and Driving Simulator Studies." In The Health Effects of Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, Torrens, M., L. San, J. M. Peri, and J. M. Olle. 1991. "Cocaine Abuse and R. Smart, 171­91. Toronto, ON: Centre for Addiction and Mental among Heroin Addicts in Spain." Drug and Alcohol Dependence 27 (1): Health. 29­34. Uchtenhagen, A., F. Gutzwiller, and A. Dobler-Mikola. 1998. Medical Soares, B. G., M. S. Lima, A. A. Reisser, and M. Farrell. 2003. "Dopamine Prescription of Narcotics Research Programme: Final Report of the Agonists for Cocaine Dependence." Cochrane Database of Systematic Principal Investigators. Zurich, Switzerland: Institut für Sozial- und Reviews (2): CD003352 [PMID: 12804461]. Präventivmedizin der Universität Zurich. Solowij, N. 1993. "Ecstasy (3,4-Methalenedioxymethamphetamine)." UNAIDS (Joint United Nations Programme on HIV/AIDS) and WHO Current Opinion in Psychiatry 6: 411­15. (World Health Organization). 2002. AIDS Epidemic Update December 930 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others 2002. Geneva: Joint United Nations Programme on HIV/AIDS and Ward, J., W. D. Hall, and R. P. Mattick. 1998. Methadone Maintenance World Health Organization. Treatment and Other Opioid Replacement Therapies. Amsterdam: UNDCP (United Nations Drug Control Programme). 1997. World Drug Harwood Academic. Report. Oxford, U.K.: Oxford University Press. Warner-Smith, M., S. Darke, M. Lynskey, and W. D. Hall. 2001. "Heroin United Nations Commission on Narcotic Drugs. 2000. "World Situation Overdose: Causes and Consequences." Addiction 96 (8): 1113­25. with Regard to Drug Abuse, in Particular among Children and Youth." White, J. M., F. Bochner, and R. J. Irvine. 1997. "The Agony of `Ecstasy': Vienna, United Nations Commission on Narcotic Drugs. How Can We Avoid More `Ecstasy'-Related Deaths?" Medical Journal of UNODC (United Nations Office on Drugs and Crime). 2004. World Drug Australia 166: 117­18. Report. Vienna: UNODC. WHO (World Health Organization). 2003. The World Health Report 2003: UNODCCP (United Nations Office for Drug Control and Crime Shaping the Future. Geneva: WHO. Prevention). 2002. Global Illicit Drug Trends, 2002. New York: WHO (World Health Organization) Programme on Substance Abuse. UNODCCP. 1997. Cannabis: A Health Perspective and Research Agenda. Geneva: ------. 2003. Global Illicit Drug Trends, 2003. New York: UNODCCP. WHO, Division of Mental Health and Prevention of Substance Abuse. Urbina, A., and K. Jones. 2004. "Crystal Methamphetamine, Its Analogues, Wild, T. C., A. B. Roberts, and E. L. Cooper. 2002. "Compulsory Substance and HIV Infection: Medical and PsychiatricAspects of a New Epidemic." Abuse Treatment: An Overview of Recent Findings and Issues." Clinical Infectious Diseases 38 (6): 890­4. European Addiction Research 8 (2): 84­93. Uuskula, A., A. Kalikova, K. Zilmer, L. Tammai, and J. DeHovitz. 2002. Wilkins, C., K. Bhatta, M. Pledger, and S. Casswell. 2003. "Ecstasy Use in "The Role of Injection Drug Use in the Emergence of Human New Zealand: Findings from the 1998 and 2001 National Drug Immunodeficiency Virus Infection in Estonia." International Journal of Surveys." New Zealand Medical Journal 116 (1171): U383. Infectious Diseases 6 (1): 23­27. Xie, X., J. Rehm, E. Single, and L. Robson. 1996. The Economic Costs of van Beek, I., R. Dwyer, and A. Malcolm. 2001."Cocaine Injecting: The Sharp Alcohol, Tobacco, and Illicit Drug Abuse in Ontario: 1992. Toronto, ON: End of Drug-Related Harm!" Drug and Alcohol Review 20: 333­42. Addiction Research Foundation. van den Brink,W.,and J.M.van Ree.2003."Pharmacological Treatments for Yacoubian, G. S. Jr. 2003a. "Correlates of Ecstasy Use among Students Heroin and Cocaine Addiction." European Neuropsychopharmacology Surveyed through the 1997 College Alcohol Study." Journal of Drug 13 (6): 476­87. Education 33 (1): 61­69. Vanichseni, S., B. Wongsuwan, K. Choopanya, and K. Wongpanich. 1991. ------. 2003b. "Tracking Ecstasy Trends in the United States with Data "A Controlled Trial of Methadone Maintenance in a Population of from Three National Drug Surveillance Systems." Journal of Drug Intravenous Drug Users in Bangkok: Implications for Prevention of Education 33 (3): 245­58. HIV." International Journal of the Addictions 26 (12): 1313­20. van Os, J., M. Bak, M. Hanssen, R. V. Bijl, R. de Graaf, and H. Verdoux. Yacoubian, G. S. Jr., C. Boyle, C. A. Harding, and E. A. Loftus. 2003. "It's a 2002. "Cannabis Use and Psychosis: A Longitudinal Population-Based Rave New World: Estimating the Prevalence and Perceived Harm of Study." American Journal of Epidemiology 156 (4): 319­27. Ecstasy and Other Drug Use among Club Rave Attendees." Journal of Drug Education 33(2): 187­96. Vasica, G., and C. C. Tennant. 2002. "Cocaine Use and Cardiovascular Complications." Medical Journal of Australia 177 (5): 260­62. Yu, X. F., J. Chen, Y. Shao, C. Beyrer, and S. Lai. 1998. "Two Subtypes of HIV-1 among Injection-Drug Users in Southern China." Lancet 351 Verheyden, S. L., J. A. Henry, and H. V. Curran. 2003. "Acute, Sub-Acute, (9111): 1250. and Long-Term Subjective Consequences of `Ecstasy' (MDMA) Consumption in 430 Regular Users." Human Psychopharmacology Zammit, S., P. Allebeck, S. Andreasson, I. Lundberg, and G. Lewis. 2002. 18 (7): 507­17. "Self Reported Cannabis Use as a Risk Factor for Schizophrenia in Swedish Conscripts of 1969: Historical Cohort Study." British Medical Vincent, N., J. Shoobridge, A. Ask, S. Allsop, and R. Ali. 1998. "Physical and Journal 325 (7374): 1199­201. Mental Health Problems in Amphetamine Users from Metropolitan Adelaide, Australia." Drug and Alcohol Review 17: 187­95. Zaric, G. S., P. G. Barnett, and M. L. Brandeau. 2000. "HIV Transmission and the Cost-Effectiveness of Methadone Maintenance." American Vlahov, D., C. L. Wang, N. Galai, J. Bareta, S. H. Mehta, S. A. Strathdee, and Journal of Public Health 90 (7): 1100­11. K. E. Nelson. 2004. "Mortality Risk among New Onset Injection Drug Users." Addiction 99 (8): 946­54. Zhang, Z. F., H. Morgenstern, M. R. Spitz, D. P. Tashkin, G. P. Yu, J. R. Marshall, and others. 1999. "Marijuana Use and Increased Risk of Walsh, G. W., and R. E. Mann. 1999. "On the High Road: Driving under Squamous Cell Carcinoma of the Head and Neck." Cancer the Influence of Cannabis in Ontario." Canadian Journal of Public Epidemiology Biomarkers and Prevention 8 (12): 1071­78. Health-Revue Canadienne De Santé Publique 90 (4): 260­63. Illicit Opiate Abuse | 931 Chapter 49 Learning and Developmental Disabilities Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, Karin B. Nelson, Geoffrey C. Solarsh, Nicole Bellows, Richard M. Scheffler, and Karen J. Hofman Learning and developmental disabilities (LDDs) include func- · Lifelong duration. By definition, LDDs have an early onset, tional limitations that manifest in infancy or childhood as a with the causes frequently occurring in the prenatal period. result of disorders of or injuries to the developing nervous These effects are typically lifelong,affecting learning and other system (Institute of Medicine Committee on Nervous System neurological functions, educational achievement, quality of Disorders in Developing Countries 2001). These limitations life, earning potential, and productivity across the life span. range from mild to severe and can affect cognition, mobility, · Costs. The extensive costs include the direct costs of acute hearing, vision, speech, and behavior. The known causes of LDD care, outpatient health care services, long-term care, rehabil- are numerous and include genetic factors, nutritional factors, itation, and special education, as well as the indirect costs of infections, toxic exposures, trauma, perinatal factors, and multi- morbidity and increased mortality (Waitzman, Romano, and factorial conditions (table 49.1). Selected causes of LDD that are Scheffler 1994). Additionally, the costs and effects extend not addressed in detail in this chapter are described in box 49.1. beyond the individuals affected to include entire families. Although information on the prevalence and impact of Health, careers and employment of parents, family dispos- disabilities in low- and middle-income countries (LMICs) is able income, health and adaptation of siblings, and family scarce, five considerations support the conclusion that LDDs interaction are adversely affected when a family member has are a public health priority in LMICs today: an LDD (Stein and Jessop 2003). It is difficult to comprehend the extent of these effects, just as it is difficult to measure · Prevalence. Although each individual cause is relatively rare, them and develop economic models that account for them. taken together, LDD affects a large proportion of children. In · Education and work. As societies and economies become high-income countries, 10 to 20 percent of children have an increasingly information-oriented and dependent on edu- LDD (Benedict and Farel 2003). With improvements in child cated and literate workers, the impact of disabilities affect- survival in LMICs, it is not known whether the prevalence of ing cognition and learning becomes greater (Institute of disabilities among children is increasing, as has been seen in Medicine Committee on Nervous System Disorders in wealthier countries (Winter and others 2002), but the few Developing Countries 2001). data available from LMICs suggest that the prevalence of · Proven interventions. The prospects for preventing LDD and specific causes and types of LDD may be even higher than in for improving outcomes are considerable and can be high-income countries. Examples include cognitive disabili- achieved, to some extent, by implementing interventions ties associated with prenatal iodine deficiency, brain infec- that have been shown to be effective and cost-effective else- tions, and blindness associated with vitamin A deficiency where but that are not being implemented in LMICs. (Durkin 2002). The prevalence of childhood disabilities in LMICs is not well established, but it is likely higher than in This chapter provides an overview of the range of interven- high-income countries. tions likely to improve child development and educational 933 Table 49.1 Categories of Causes of LDD Category Examples Genetic Chromosomal Down syndrome, chromosomal rearrangements Segmental autosomal syndromes Prader-Willi syndrome, Angelman syndrome Sex-linked, single gene Fragile X syndrome, Rett syndrome Autosomal recessive Phenylketonuria, Tay­Sachs disease Autosomal dominant Neurocutaneous syndromes, such as neurofibromatosis Multifactorial Genetic and nutritional Neural tube defects Nutritional Prenatal: maternal iodine deficiency Developmental iodine deficiency disorder Childhood: vitamin A deficiency Xerophthalmia, night blindness Infections Prenatal or perinatal Toxoplasmosis, rubella, cytomegalovirus, herpes, gonorrhea, syphilis, group B streptococcus, chlamydia, trichomonas vaginalis, bacterial vaginosis, herpes simplex virus, HIV Postnatal or childhood Encephalitis, meningitis, varicella, cerebral malaria, polio, trachoma, otitis media Toxic exposures Prenatal Alcohol, lead, mercury, antimicrobials (such as sulfonamides, isoniazid, ribavirin), anticonvulsants (such as phenytoin, carbamazepine), and other drugs (such as accutane, thalidomide) Postnatal or childhood Lead, mercury Other maternal disorders Thyroid disease Cerebral palsy Other perinatal complications Brain injuries associated with Cerebral palsy, cognitive disabilities, seizure disorders premature birth, birth asphyxia Injury Traumatic brain injuries and other Cognitive, motor, speech, vision, hearing, seizure, and behavioral disabilities disabling injuries from vehicle crashes, child abuse and neglect, falls, burns, warfare, and so forth Poverty, economic disadvantage Social and cognitive deprivation Mild mental retardation Unknown LDD of unknown cause outcomes for children in LMICs. Evidence of cost-effectiveness risk factors. Attempts have been made to estimate the DALYs is considered in some detail for three selected interventions. An associated with specific causes of LDDs. For example, it is esti- overview of other key risk factors and conditions that result in mated that 9.8 million DALYs, or nearly 1 percent of the global LDD is provided. A research agenda is outlined for advancing burden of disease, are due to one relatively minor form of LDD, knowledge of how to prioritize cost-effective interventions and namely, mild mental retardation (MR) caused by lead ingestion how best to devote resources for the prevention of LDD in from environmental sources (Fewtrell and others 2004). Since LMICs. only a small fraction--probably much less than 10 percent--of LDD worldwide can be attributed to lead-induced mild MR, LDD AND THE GLOBAL BURDEN OF DISEASE this estimate suggests that LDD as a whole must account for a large proportion, perhaps more than 10 percent of the global Estimates for disability-adjusted life years (DALYs) (Mathers burden of disease. Where DALY estimates are available, we use 2006) are not available to convey the full range of LDDs or their them as a basis for economic analysis to estimate the costs of 934 | Disease Control Priorities in Developing Countries | Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, and others Box 49.1 Interventions for the Prevention of Childhood Neurological Disabilities Attention Deficit Hyperactivity Disorder for a person with autistic disorder exceeded UK £12.4 mil- Attention deficit hyperactivity disorder (ADHD) is the lion, with most of the expense related to living support most common neurological disorder in children in the and daily activities. United States, with an estimated prevalence of 3 to 11 per- cent. The prevalence is not known in LMICs, but as school- Infection ing increasingly becomes the norm, ADHD is likely to Numerous prenatal, perinatal, and postnatal infections become more obvious. The burden of ADHD in settings of can damage the developing nervous system or sensory large class sizes will likely pose an increasing challenge. In pathways and cause long-term disabilities in children. The addition to its major impact on school performance, relative contribution of these infections to the burden of ADHD affects family relationships and social competence, LDD is likely to vary by country. It will be influenced by with lasting consequences. Children with ADHD are also at overall infant mortality, postneonatal contribution to higher risk for injury, depression, and substance abuse. infant mortality, and regional difference in the distribu- Worldwide, with the growing use in school settings of stim- tion of the infections known to be associated with neuro- ulants to control this chronic disorder, the impact on health logical sequelae during different periods in the early life care costs is potentially huge. Although there are a paucity cycle. A few of the most important infections that may of data on this topic, in one study, the cost of medicating result in LDD include the following: children for ADHD was close to an average of US$500 or more per child per year, and this figure is considered a sub- · Congenital rubella (chapter 20). This disease is a major stantial underestimate (Chan, Zhan, and Homer 2002). global cause of preventable hearing impairment, blind- ness, and intellectual disability. The incidence of con- genital rubella syndrome has been variably set at 0.5 to Autism Spectrum Disorders 2.2 out of every 1,000 live births in LMICs during epi- All autism spectrum disorders (ASDs) are characterized demics, which occur every four to seven years (Cutts by varying degrees of impairment in communication and others 1997). Though some LMICs have set elimi- skills and social interactions and in restricted, repetitive nation goals and vaccination has been noted to be cost- patterns of behavior or interests. Although only 50 per- effective, only 28 percent of LMICs routinely vaccinate cent of children in the United States with ASDs are diag- against rubella (Robertson and others 1997). nosed before six years of age, this group of disorders can · HIV/AIDS infection (chapter 18). Neurological prob- reliably be diagnosed by three years of age and in some lems in HIV-infected children vary in different parts of cases by as early as 18 months. ASDs range from a severe the world but may be as high as 40 to 50 percent (Bobat form called autistic disorder to a milder form known as and others 1998). The developmental trajectory of Asperger syndrome. Prevalence studies of ASDs in Asia, infected children is confounded by maternal, social, Europe, and North America estimate that 2 to 6 out of and biological risk factors during pregnancy and early every 1,000 children have an ASD. Screening instruments childhood. Maternal substance and drug abuse, more using responses from children and parents are available. common in HIV-infected women, have an independent Evidence indicates that early intervention (ideally in opti- adverse effect on brain growth and neurodevelopmen- mal educational settings for at least two years during pre- tal outcome. Low birthweight and prematurity, pov- school) results in improved outcomes. Individuals with erty, protein calorie malnutrition, and micronutrient ASDs generally respond well to highly structured, special- deficiencies--more frequently seen in HIV-infected ized programs. A variety of medications is used to treat children and particularly in LMICs--may similarly associated depression, anxiety, ADHD, seizures, and other compromise early child development (Brouwers and behavioral symptoms. Adults with severe ASDs require others 1996). intensive and constant supervision. Little information is · Malaria (chapter 21). In Sub-Saharan Africa, malaria is available regarding the parental and service costs of ASDs. the leading cause of childhood mortality and morbidity. In a 2001 study in the United Kingdom, the lifetime cost Cerebral malaria is a well-known complication and Learning and Developmental Disabilities | 935 Box 49.1 (Continued) may result in neurological sequelae in survivors, con- screening tools for high-risk drinkers, who include women tributing significantly to the burden of LDD. in prisons, drug rehabilitation centers, hospital emergency · Bacterial meningitis (chapter 20). This disease results facilities, and sexually transmitted disease clinics (Sokol, in long-term sequelae for many children, including Delaney-Black, and Nordstrom 2003). Little is known approximately 40 percent of children who survive about the costs around the world. Annual costs for all indi- Haemophilus influenza meningitis, 50 percent who viduals with fetal alcohol syndrome in the United States survive pneumococcal meningitis, and 10 percent during 1998 was estimated at US$4 billion, with lifetime who survive meningococcal meningitis. Cost-effective care per person, for individuals requiring such care, at immunization can prevent meningitis from all these US$1.4 million (Lupton, Burd, and Harwood 2004). causes. Environmental Exposures Children are more susceptible to environmental factors, Alcohol including unsafe home environments, road traffic, and Prenatal alcohol exposure resulting in fetal alcohol syn- chemicals (see chapters 42 and 43). Even in high-income drome may be the most common single preventable cause countries in Europe, mild MR resulting from lead expo- of MR worldwide (Viljoen 1999), but substantial chal- sure accounted for 4.4 percent of DALYs among children lenges remain in diagnosing and preventing this disorder zero to four years of age. Legislative efforts are under way (see chapter 47). In addition to growth retardation and to eliminate lead from gasoline and other environmental congenital heart disease, effects include ADHD, memory sources of lead exposure in LMICs (Khan and Khan 1999; deficits, and mood disorders. Adults continue to have Alliance to End Childhood Lead Poisoning 2002). In the 0 attention and social difficulties and higher rates of alcohol, to 19 years age group, injuries from all causes accounted nicotine, and drug dependence. Children exposed to even for 19 percent of DALYs. The poor and vulnerable road small amounts of alcohol (half a drink per day) in utero users--pedestrians, cyclists, and motorcyclists--bear the have poor outcomes, suggesting that abstinence should be greatest burden of road injuries. Nearly 25 percent of all recommended during conception and throughout preg- nonfatally injured victims requiring hospitalization sus- nancy (Sokol, Delaney-Black, and Nordstrom 2003). tain a traumatic brain injury as a result of motor vehicle Although tools are available to help providers identify crashes (Peden and others 2004). Although the effective- women who consume alcohol, detection of maternal alco- ness of bicycle helmets for road safety is high, their use in hol exposure is a challenge. The overall rate of fetal alco- LMICs is low (Thompson, Rivara, and Thompson 1999). hol syndrome for LMICs has been placed at 1 to 4.8 out of Interventions aimed at reducing children's exposure to every 1,000 population (Sampson and others 1997) and is environmental factors and injuries could result in sub- higher among low socioeconomic populations and sub- stantial gains. Targeted action by region, even within a populations with particularly high alcohol intakes. If single country, is likely to prove most productive (Valent individuals with the full spectrum of fetal alcohol and others 2004). syndrome­related effects are included, this rate may be as high as 1 in every 100 births. A prevalence rate of 40.5 to 46.4 out of every 1,000 children in South Africa, the high- Nutritional Deficiency est rate worldwide, is attributable to particular historical Iodine deficiency from inadequate quantities of iodine in and social conditions (May and others 2000). soil, water, and food affects 13 percent of the world's pop- Public health measures to prevent prenatal alcohol ulation, and an additional 30 percent are at risk (see chap- exposure have had limited success, and rates have not ter 28). Maternal iodine deficiency during pregnancy may changed over the past decade in the United States (Floyd result in an average loss of 15 IQ points in offspring, mak- and Sidhu 2004). These measures include putting warning ing it a leading preventable cause of MR. Iodine deficiency labels on alcoholic beverages and broadcasting public can be prevented with adequate consumption of iodized messages about alcohol dangers during pregnancy. salt, which is now consumed by about 70 percent of Improved outcomes might result from targeting the use of households worldwide. 936 | Disease Control Priorities in Developing Countries | Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, and others prevention of LDD. In this chapter, we estimated only costs of by dietary restrictions to prevent damage to the nervous sys- the interventions for Down syndrome (DS), neural tube defects tem and effective emergency medical care for head injury. (NTDs), and congenital hypothyroidism. · Tertiary prevention refers to rehabilitation and special edu- cational services to mitigate disability and improve func- IMPAIRMENT, DISABILITY, AND PARTICIPATION tional and participatory or social outcomes once disability has occurred. Quantifying the impacts of LDD and their preventive interven- tions is complicated by the fact that these disorders can exist UNINTENDED CONSEQUENCES OF SUCCESSFUL and be measured at multiple levels, including three levels OR PARTIALLY SUCCESSFUL INTERVENTIONS distinguished by the World Health Organization (WHO) in International Classification of Functioning, Disability, and Interventions to reduce mortality and morbidity may be fol- Health (WHO 2001): lowed by increases in the prevalence of LDD. Examples include the following: · impairment, which refers to physiological or psychological defects or abnormalities, such as failure of the neural tube to · Improved survival of very low birthweight infants at high close risk for LDD may cause the prevalence of disability in the · function or disability, which refers to the ability of an indi- population to increase at the same time that it increases the vidual to perform a task, such as walking, seeing, hearing, absolute number of survivors without disabilities. learning language, and reading · Rubella vaccination programs with less than optimal cover- · participation, which refers to the degree to which an indi- age will prevent infections in those vaccinated but leave vidual participates in school, employment, social role, and unvaccinated girls at risk for acquiring rubella infection dur- recreational activities. ing their childbearing years (rather than during childhood, as might be expected in the absence of a vaccination pro- A given impairment may be associated with a range of func- gram), thereby increasing the risk of congenital rubella tional outcomes. Some but not all of these may be recognized infection and disability in the population. as disability. Disability is context specific and may vary from · Newborn screening and treatment for phenylketonuria in culture to culture. For example, conditions such as dyslexia, infancy and childhood prevent MR, but phenylalanine attention deficit and hyperactivity disorder (ADHD), and mild dietary restriction for women with phenylketonuria during MR may be especially disabling in school but not as noticeable their childbearing years is essential to prevent prenatal neu- in nonacademic settings and environments where schooling is rological damage and MR in their offspring. optional. Environmental factors and social stigma may deter- mine the participation of people with disabilities more than do the functional deficits themselves. Some interventions may be OTHER FACTORS LEADING TO INCREASES designed to enhance participation (for example, ramps, acces- IN MEASURED PREVALENCE sible public toilets, inclusive education), whereas others may target impairment and disability (for example, nutritional for- Progress in the field of LDD may result in increases in the rec- tification, surgery, rehabilitation, special education, newborn ognized prevalence of disability and in social and economic screening, and early treatment). costs, as in the following examples: · Increased availability of services may increase the number of THREE LEVELS OF PREVENTION children with recognized disabilities. Just as it is ethically problematic to screen for disorders for which no services Prevention of LDD involves primary, secondary, and tertiary can be offered, expansion of case finding becomes justified prevention activities: and ethically demanded as services become available, with · Primary prevention includes efforts to control the underly- the potential result of increasing the measured prevalence of ing cause or condition that results in disability. Examples disability. include (a) maternal antiretroviral therapy to reduce the · As educational expectations and awareness of LDD increase, risk of mother-to-child transmission of HIV and (b) fortifi- the prevalence of recognized disability may increase. cation of the food supply to prevent birth defects such as spina bifida and iodine deficiency disorders. In consideration of these trends and relationships between · Secondary prevention aims at preventing an existing illness or public health advances and increases in disability, it may not be injury from progressing to long-term disability. Examples realistic to expect short-term control of disability or cost sav- include newborn screening for metabolic disorders followed ings following interventions that reduce mortality, even if those Learning and Developmental Disabilities | 937 Conception Nutritional Alcohol Nutrition supplementation Genetics Infections Treatment of infections Drugs Hypertension Prenatal screening Positron emission and obstetric care Smoking tomography Delivery Low Congenital birthweight abnormali- Meningitis ties Head injury Perinatal Congenital asphyxia infections Diarrhea Neonatal screening HIV/AIDS Immunization Pneumonia Growth monitoring and nutrition education Undernutrition Developmental screening Poverty Deworming and micronutrient supplementation Poor maternal Maternal education education and support Early child Inadequate stimulation and stimulation education Social welfare support Abnormal Interventions development and Risk factors disability Source: Authors. Figure 49.1 Causal Pathways for LDD interventions have a net positive effect on public health. The prevention of disability. The vertical axis distinguishes four costs of disability and its prevention may increase initially in levels of evidence for cost-effectiveness: the wake of interventions that successfully reduce mortality. · evidence available for LMICs Figure 49.1 summarizes the causal pathways and potential · evidence available for high-income countries only interventions for the prevention of LDD. · evidence for cost-effectiveness not available, but cost- effectiveness can be estimated from existing data · evidence not available, but potential for benefits exists. INTERVENTIONS IN LOW- AND The literature indicates that the economic outcomes of a MIDDLE-INCOME COUNTRIES given intervention may vary widely for two reasons: Numerous interventions are effective in preventing LDD. · Variations exist across populations, even within the same Table 49.2 provides a summary of these interventions classified country, in the prevalence of the disorder, the cost of health on two axes. The horizontal axis distinguishes whether the care, and the existing infrastructure available to implement intervention would accomplish primary, secondary, or tertiary the intervention. 938 | Disease Control Priorities in Developing Countries | Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, and others Table 49.2 Classification of Interventions to Prevent LDD Tertiary (rehabilitation or Primary (prevention of Secondary (prevention of prevention of further condition that can lead to disability once condition disability once disability disability) has occurred) has occurred) Evidence for cost-effectiveness Food fortification (folic acid and iodinea) available for LMICs Rubella vaccinea Hemophilus vaccine Removal of lead from paint and fuel Vitamin A supplementation (vision) Measles vaccine Evidence for cost-effectiveness Prenatal screening for DS available for high-income and prevention of DS births countries only Newborn screening for metabolic disorders followed by interventions to prevent disability Evidence for cost-effectiveness Malaria preventionb Early detection and care Special education not available, but cost-effectiveness of neonatal jaundice Prosthetics can be estimated from existing data Management of malaria Braille Treatment for otitis media Sign language Prevention and treatment of neonatal Occupational, physical, complications through emergency and speech therapies obstetric and pediatric services Surgery Eyeglasses Residential care Hearing aids Assistive devices Detection and treatment of maternal thyroid disorders Evidence for cost-effectiveness Fetal alcoholism prevention Dehydration/diarrhea treatment Community-based not available, but potential for Trauma prevention (bicycle Postnatal combined cognitive rehabilitation benefits exists helmets, burns) stimulation and nutritional intervention Prevention of shaken baby Therapeutic stimulants for syndrome and child abuse treatment of ADHD Note: Italicized text represents somewhat detailed consideration of cost-effectiveness included in this chapter. a. Covered in chapter 56, but chapter emphasis is not on implications for preventing developmental disabilities. b. Covered in detail in chapter 21. · Differences between studies exist in analytical methods Too little is known about the fourth type of intervention, used, such as the willingness to pay versus the human community-based rehabilitation, to evaluate it. There is a capital approach to valuation, and in cost categories, such as paucity of knowledge and a history of failed interventions whether to include parental time costs. Though these differ- for the prevention of premature birth and the disabilities ences make cross-population comparisons difficult, the associated with premature birth. overall evidence of cost-effectiveness is demonstrated by repeated findings that the benefits of a particular interven- Neural Tube Defects: Burden and Cost-Effectiveness tion outweigh the costs in a number of different settings. of Folic Acid Fortification Current evidence suggests that three interventions are cost NTDs, which are the most common malformation of the cen- saving: folic acid fortification to prevent NTDs, prenatal tral nervous system, result from failure of the neural tube to screening and selective pregnancy termination to prevent DS, close during the first month of pregnancy. Anencephaly typi- and neonatal screening and treatment for congenital hypothy- cally results in pregnancy loss, stillbirth, or neonatal death. roidism (CH). Spina bifida (open spine defect) is associated with a range of Learning and Developmental Disabilities | 939 functional deficits (requiring multiple surgical and rehabilita- · Grains from large mills are relatively cheap to fortify; more tive interventions), including paralysis of the lower extremities resources are required to fortify grains milled in smaller and often primary enuresis and cognitive disabilities. Large neighborhood mills. geographic variations in the prevalence of NTDs exist both · The amount of folate consumed by different populations in within and between countries. The burden of disease is highest the absence of fortification varies. in South Asia and lowest in LMICs of Europe and Central Asia. · Prevalence of NTDs varies across populations, and the cost- Similarly, deaths from NTDs are high in South Asia but lowest effectiveness increases with prevalence. in high-income countries. Estimates suggest that almost all NTD disease burden is concentrated in the age group zero to Costs of food fortification may be lower in high-income four years (Mathers 2006). countries, where most people consume cereals processed in a few large mills, equipment for fortification is likely to be in Folic Acid. Folate is a vitamin that occurs naturally in green place, and quality assurance is facilitated. In contrast, mills in leafy vegetables, legumes, citrus, and other foods. Folic acid (FA) LMICs lack fortification equipment and capital, and running is an easily absorbed synthetic form of folate that can be deliv- costs are higher in the short run. ered as a dietary supplement or through FA fortification of flour or other common staple foods. NTDs can be reduced by 70 per- Costs of Food Fortification For optimal daily consumption, cent if women consume 400 g of FA daily around the time of the actual level of food fortification (defined as g of FA per conception and until closure of the neural tube. At a population 100 grams of the food item) should be adjusted for storage and level, either supplementation or fortification of the food supply other losses so that a daily dose of 400 g is achieved. Food items is necessary to ensure that 400 g of FA is consumed at the crit- that should be fortified depend on specific dietary habits. ical period of fetal development, as this dose is higher than can Staples such as rice and flour are obvious choices; salt, sugar, reasonably be consumed by relying on naturally occurring bread, milk, and edible oils are promising candidates. There are folate in foods. Fortification is much more likely than supple- economies of scale in FA fortification.It can be and usually is car- mentation to reach the population at risk because the benefit of ried out in conjunction with other forms of fortification, such as enhanced FA intake occurs early, typically before the pregnancy iron, iodine, and vitamin A fortification. Many food items are is recognized. Fortification is of particular value to women who already fortified in high-income countries. Other factors to be may not receive prenatal care until the third trimester. considered in the choice of food for fortification are items that This section considers only evidence of cost-effectiveness of are centrally processed and allow for quality control.Soy sauce in FA fortification in LMICs with respect to the benefit of pre- China is an example: it is consumed on a daily basis by 70 percent venting NTDs. Additional health benefits can be expected with of the population and is prepared in a few large factories. respect to stroke, heart disease, and cancer. The recommended fortification level is thus 240 g per 100 grams of the staple food. This fortification rate is assumed Cost-Effectiveness of Folic Acid Fortification A cost-benefit for all regional strata where the per capita staple consumption analysis of grain fortification in the United States (Romano and per day is less than 300 grams. Wheat, rice, maize, or a combina- others 1995) included costs related to the addition of FA to tion of these foods is the staple in most countries. The food, to annual testing and surveillance, to a one-time packag- recommended level of FA fortification varies from 150 g to ing change, and to potential (though not substantiated) adverse 240 g per 100 grams of cereal. So that women receive a daily health effects associated with undiagnosed vitamin B12 defi- dose of 400 g,the target cereals for fortification should be those ciency. Benefits included avoided costs of NTDs, such as mor- for which daily per capita consumption is at least 200 grams. In tality costs (particularly for anencephaly) and costs of caring Sub-Saharan Africa,daily per capita cereal consumption exceeds for those with spina bifida. The benefits of fortification out- 200 grams only if wheat, rice, and maize are considered together. weighed costs with cost-benefit ratios of 1 to 4.3 for low-level Quality assurance is done through analytic testing of forti- fortification and 1 to 6.1 for high-level fortification. fied products to confirm FA levels. Quality assurance costs in Cost-effectiveness relative to status quo of FA fortification the United States are estimated at US$0.64 cents per ton of for- depends on several factors: tified grain in quality assurance costs. The costs of FA fortification include the cost of FA, setup, · Costs of food fortification depend on the types and quanti- and analytic testing. The analysis is done using two different ty of food that are fortified and the level of fortification. cost estimates: US$0.15 and US$0.50 per ton of grain fortified. · The proportion of the target population reached by the The cost of FA determines the cost of premix added to the fortified food is important since, in most LMICs, many peo- flour. FA is almost never added alone; usually FA, iron, zinc, ple consume food produced on their own farms or within and niacin are added in combination. The material cost of FA their villages. alone is about US$0.10 to US$0.20 per metric ton of milled 940 | Disease Control Priorities in Developing Countries | Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, and others wheat. However, a more realistic cost for the premix (including common genetic cause of mental retardation. Identifying a other supplements) is about US$0.50 per metric ton of milled fetus with DS before birth and giving parents the option to ter- wheat. This higher estimate is conservative and does not minate the pregnancy early can help decrease the burden of the account for the health benefits from the other supplements. disease on families and society. During counseling, parents may Either way, the per capita costs are only a few cents in each receive information about the consequences of DS, which will region. The low per capita cost in high-income countries of allow them to make an informed decision about the best care US$0.009 assumes that 80 percent of the cereal supply is forti- for the newborn or about termination of the pregnancy. fied. In South Asia, where NTDs have the highest burden, the Prenatal screening services provide an opportunity to pro- per capita cost is estimated at US$0.067 (Bagriansky n.d.). foundly reduce the impact of MR. The cost-effectiveness of prenatal screening for DS is based on two parameters: efficacy Benefits of Folic Acid Fortification The cost-effectiveness of FA (by assessing the false positive rate of screening procedures and fortification in terms of its cost per DALY and per death averted the number of fetal losses caused by screening) and financial assumes that the fortification strategy will reduce the incidence costs (costs of screening per DS pregnancy averted). On the of NTDs by 50 percent. The costs are relatively high because of basis of the evidence, the best screening method is proposed, the high cost of FA. Even a few cents per capita becomes expen- and sensitivity of the parameters of interest to the LMIC is sive if the per capita prevalence of NTD is very low. tested. No formal comparisons are made between the costs of screening and care for a person with DS. The purpose of this Other Costs and Benefits The benefits of FA fortification analysis is to suggest the most cost-effective way of screening outweigh the costs. The benefits estimated here are conserva- that provides families with information about the health of the tive for three reasons: child; it is not a cost-benefit analysis of whether a couple should terminate a pregnancy. · Strokes and coronary deaths are also prevented by FA forti- fication and occur more frequently than NTDs. Burden. DS is caused by trisomy of chromosome 21--an extra · The percentage of NTDs that can be prevented by FA chromosome rather than the usual diploid form--and is a fortification may be greater than 50 percent, because up to major cause of severe MR (IQ less than 50 with substantial 70 percent of NTDs can be prevented by 400 g of pericon- deficits in adaptive behavior). The incidence of DS is higher ceptional FA daily. than the birth prevalence because many fetuses are sponta- · These estimates do not take account of the costs of clinical neously miscarried and, in some cases, selectively terminated. care and management for complications when NTDs are In the absence of prenatal screening and intervention, most not prevented. DS conceptions (71 percent) result in spontaneous abortion; another 3 percent result in stillbirth, and 26 percent result in live Interventions to Prevent Disability Caused birth with subsequent LDD (Kline, Stein, and Susser 1989). by Down Syndrome Because the incidence of DS cannot be determined without Screening programs are critical public health interventions that doing surveillance of all conceptions, the frequency of DS is use universal or targeted screening tests to identify potential typically measured in terms of prevalence per 1,000 live births causes or cases of LDD, including DS. rather than in terms of incidence. Thus, the population preva- lence of DS varies depending on the maternal age structure Prenatal and Neonatal Screening. Prenatal screening for (steep increase after age 35 years) as well as the availability and genetic abnormalities allows parents to determine whether to use of prenatal diagnosis followed by selective termination. continue with an affected pregnancy, whereas neonatal screen- Estimates from 10 LMICs range widely, from 0.1 out of every ing's fundamental purpose is to improve the infant's prognosis 1,000 live births in Indonesia to 4.4 out of every 1,000 live births through early diagnosis and treatment. in Pakistan (Institute of Medicine 2003). Most studies, in both A number of LDDs have been screened for in high-income high-income countries and LMICs, show DS birth prevalence countries since the 1960s, and researchers have conducted eco- in the range of 1.0 to 1.6 out of every 1,000 births. The birth nomic evaluations of these screening programs, including those prevalence of DS is likely higher in LMICs because of a higher for Tay-Sachs disease carriers, DS (Cusick and others 2003), proportion of births among women over age 35 (11 to 15 per- sickle cell disease (Panepinto and others 2000), phenylke- cent) relative to that in high-income countries (5 to 9 percent) tonuria (Lord and others 1999), and several other inborn errors (Kline, Stein, and Susser 1989) and possibly because of differen- of metabolism (Insinga, Laessig, and Hoffman 2002). tial access to prenatal screening for chromosomal abnormalities. Estimates for Prenatal Screening, Diagnosis, and Selective Life Expectancy and Quality of Life. Life expectancy for chil- Pregnancy Termination for Down Syndrome. DS is the most dren with DS is substantially lower than that of the general Learning and Developmental Disabilities | 941 Table 49.3 Distribution of DALYs Lost to and Deaths Caused medical care and community services, as well as on sustained by Down Syndrome, by World Bank Region, 2002 support from family members. It is also important to note that dollar costs of care for a DS child in LMICs would be much lower Region DALYs Deaths than such costs in high-income countries because of lower East Asia and the Pacific 4,101,694 1,328 prices as well as lower treatment intensity. For example, in some Europe and Central Asia 507,723 652 countries, congenital heart disease, which affects 40 to 60 per- High income countries 199,215 2,113 cent of DS children, cannot be treated effectively. This lack of Latin America and the Caribbean 214,346 1,979 treatment will lower costs of care as well as life expectancy, and Middle East and North Africa 347,898 1,311 cost estimates will vary for each individual area or region. South Asia 2,005,766 11,336 Sub-Saharan Africa 478,851 4,967 Cost-Effectiveness of Prenatal Screening, Diagnosis, and Selective Pregnancy Termination. Prenatal screening can be Source: Mathers and others 2006. implemented to allow selective termination of DS pregnancies and prevention of disability related to DS in the population. population. Congenital heart disease occurs in 40 to 60 percent This intervention raises ethical, social, and cultural concerns of children with DS and accounts for 30 to 35 percent of for some individuals and populations that may preclude its deaths. Survival and life expectancy of children with DS have applicability. increased dramatically: In a 1940­60 birth cohort in England, A screening program incorporating maternal serum triple only 50 percent of infants with DS survived beyond age two. By screening in all pregnant women, regardless of maternal age, comparison, in 1981­85, 90 percent survived beyond age five yields an excellent DS detection rate and is associated with a low (McGrother and Marshall 1990). Table 49.3 describes the esti- false-positive rate (Wald and others 2003). DS pregnancies yield mated total deaths caused by DS by region, as well as the esti- lower levels of alpha-fetoprotein and unconjugated estriol but mated total DALYs lost. have elevated levels of human chorionic gonadotropin com- DS is always associated with cognitive impairment. pared with other pregnancies. Ultrasound evaluation of the Disability can range from mild to profound, and most children fetus neck thickness improves screening sensitivity. It is also are affected moderately (IQ 40­55). Early intervention and useful when used in conjunction with serum screening (Wald therapy can improve functional outcomes. Of children with DS, and others 2003). A positive screening result is followed by diag- 60 to 80 percent have hearing loss, and approximately 70 per- nosis using amniocentesis or chorionic villus sampling (CVS). cent have ophthalmologic problems. As life expectancy of DS Although both diagnostic procedures are guided by ultra- individuals has increased, many grow to adulthood and face an sound to reduce risk, they are invasive, are more expensive than increased risk of early onset Alzheimer's disease, cataracts, hear- the screening procedure, and carry a small risk of miscarriage ing loss, hypothyroidism, and degenerative vascular disease. of an unaffected pregnancy. Thus, only a select group screening positive for possible trisomy 21 are offered the invasive diag- Costs of Care. Based on 1988 data, the estimated incremental nostic procedures. Amniocentesis, which involves the aspira- lifetime economic costs of DS are US$410,000 per case or tion of amniotic fluid, is performed between the 14th and 16th US$647,709 in 2004 dollars (Waitzman, Romano, and Scheffler weeks of pregnancy. CVS involves aspiration of villi and can be 1994). In another study, the estimate of per capita incremental performed between the 10th and 12th weeks of pregnancy. costs of DS, converted to 2004 dollars, include net medical costs Although CVS can be performed earlier in the pregnancy, of US$168,567, developmental services costs of US$80,530, amniocentesis is easier to perform and is more widely used in special education costs of US$171,593, and total costs of the second trimester. Following diagnostic confirmation of DS, US$420,690 (Waitzman, Romano, and Scheffler 1994). parents are provided with genetic counseling and the option of An estimate of lifetime costs per live born baby with DS-- terminating the pregnancy. including education, health, and lost productivity costs-- Although DS risk increases with maternal age, most births ranged from US$137,000 in 1990 to US$515,000 in 1993 occur in younger women and, therefore, two-thirds of all DS (Gilbert and others 2001). Net savings using the annual pro- births occur in younger mothers (Ross and Elias 1997). If pre- gram of screening, diagnosis, and selective termination was natal diagnosis is available only for mothers 35 years or older, estimated to be US$885, with costs of US$446,000 per 10,000 only 33 percent of DS births will be detected. Studies demon- pregnancies for a program that detects and prevents 9.7 DS strate that heavy reliance on maternal age to screen for DS may births per year and a lower bound estimate of US$137,000 of not be desirable in LMICs. Maternal age factor is not so useful potential lifetime costs per 9.7 births prevented. in settings where early marriage and motherhood are the norm The increased life span of individuals with DS and accompa- and most DS pregnancies involve mothers younger than 35 nying age-associated morbidity impose heavy demands on (Gupta and others 2001). Therefore, maternal serum screening 942 | Disease Control Priorities in Developing Countries | Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, and others of all pregnant women is important in preventing DS births averted (Institute of Medicine 2003). The goal is to minimize and achieving cost-effectiveness (Wald and others 2003). this ratio. The efficacy of prenatal screening varies with preva- lence, and the primary determinant of variations in prevalence Procedure Costs Genetic screening and counseling services are of DS is the age structure of women giving birth. The preva- expensive. Even after initial high fixed costs to establish prena- lence of DS and the efficacy of prenatal screening increase with tal screening services, provision of high-quality services the percentage of births to mothers over the age of 35. In this requires staff training, equipment, and laboratory mainte- analysis, a 90 percent rate of selective termination is used nance. A recent report suggests establishing genetic screening (Waitzman, Romano, and Scheffler 1994). On this basis, the services when other public health interventions have reduced number of fetal losses per DS birth avoided varies from 7.13 the infant mortality rate to the range of 20 to 40 out of every (for 1 in 10,000 prevalence) to 0.16 (for 44 in 10,000 preva- 1,000 live births (Institute of Medicine 2003). Above this level, lence). Therefore, in countries with low prevalence of DS, such other public health interventions may have greater benefits. as Indonesia, more unaffected fetuses are lost than DS births The breakdown of tasks is as follows: averted because of screening. In areas where the ratio of unaf- fected fetal losses to DS births avoided is above 1, the efficacy of · screening costs, which consist of laboratory expenses (con- screening for DS is questionable. sumables and staff); informing women of results (by mail Because of higher loss rates for CVS, we use a 1.5 percent if negative, by phone if positive); service costs (processing fetal loss rate in our sensitivity analysis (Lippman and others results and monitoring the service); training in ultrasound 1992). Other costs not considered in this study are the psycho- measurement of neck skin translucency; and overhead logical effects of a positive test on the parents, anxiety that may expenses persist from a false-positive test, and potential complications · diagnostic costs, which comprise counseling before CVS or resulting from pregnancy termination. Complications from amniocentesis, equipment and staff for these procedures, termination may vary (Stray-Pedersen and others 1991) and laboratory expenses (consumables and staff), and overhead may not be the same in LMICs, which should be taken into expenses account. The sensitivity rate for the triple serum test followed · costs of termination of selected pregnancies, which include by the amniocentesis is 62.3 percent in the clinical trials surgical dilation, evacuation (11 to 13 weeks), or medical (Vintzileos and others 2000), and the uptake of amniocentesis termination with mifepristone (after 13 weeks). is 90 percent for affected mothers and 80 percent for unaffected mothers (Waitzman, Romano, and Scheffler 1994). We assume We assume infrastructure exists for prenatal screening, diag- the false-positive rate of 5 percent. The false-positive rate nosis, and intervention. We use the following costs: triple serum affects the probability of losing an unaffected fetus as a result of test, US$70; amniocentesis, US$1,200; genetic counseling, invasive testing that follows serum screening. US$100; and termination of pregnancy, US$2,000. These cost Financial cost-effectiveness is defined as the screening costs estimates have been widely used in the literature (Cusick and per DS birth averted. It is presented in table 49.4. Cost- others 2003). However, the medical costs can be significantly effectiveness is the highest in countries with high birth preva- lower in LMICs and will also vary across and within countries. lence of DS, given that women have access and receive prenatal care. Costs of prenatal screening and termination per DS birth Cost-Effectiveness and Efficacy We assume that 100 percent of averted vary from US$1,497,390 in Indonesia (for 1 in 10,000 women attend a prenatal clinic between 10 and 14 weeks of ges- prevalence) to US$37,185 in Pakistan (44 in 10,000 prevalence). tation and are offered tests in the first trimester,or between 15 and A similar relationship is seen between prevalence and cost per 19 weeks for the tests in the second trimester.We discuss the effect DALY. In our analysis, we use costs data that are based on esti- of low uptake of prenatal care and its effect on cost-effectiveness mates from developed countries. Because costs of care will vary of prenatal screening programs in our sensitivity analysis. widely across and within countries, cost estimates should be In terms of economic considerations, it is desirable to bal- done for individual regions. Lower costs of care will reduce cost- ance the probability of the birth of a DS child with the risk of effectiveness of prenatal screening for DS. However, even after procedure-related miscarriage. Sensitivity of prenatal screening the cost adjustment, it is unlikely that the benefits will complete- and the false-positive rates vary widely, depending on the ly go away, because of the large difference between a relatively method used. The risk of procedure-related miscarriage can cheap screening program and high burden of disease of DS. vary from 0.04 to 0.8 percent (Nyberg and others 1998). We use the conservative fetal loss rate of 0.9 percent (Gilbert and Sensitivity Analysis The results of the analysis above depend others 2001) for both procedures. on assumptions that may not hold in some LMICs. For exam- Efficacy of prenatal screening is defined as the number of ple, if many women accept screening but few decide to have an unaffected fetuses lost due to prenatal testing per each DS birth amniocentesis, cost-effectiveness will be adversely affected. The Learning and Developmental Disabilities | 943 Table 49.4 Financial Cost-Effectiveness and Efficacy of Prenatal Screening and Pregnancy Termination for the Prevention of Down Syndrome Births DS births per 100,000 Cost per Cost per Cost of Cost per Fetal population DS 100,000 DS birth detection and DS birth losses per Representative (birth births population detected DS births termination avoided US$ per Unaffected DS birth country prevalence) detected (US$) (US$) prevented (US$) (US$) DALY fetal losses prevented East Asia and the Pacific Indonesia 10 5.61 7,546,188 1,345,851 5.05 7,556,281 1,497,390 14.88 36.0 7.13 Europe and Central Asia Hungary 56 31.40 7,574,655 241,237 28.26 7,631,174 270,041 38.31 35.98 1.27 High-income countries Canada 120.79 67.73 7,614,750 112,433 60.95 7,736,658 126,926 36.09 35.96 0.59 Latin America and the Caribbean Argentina 160 89.71 7,639,014 85,150 80.74 7,800,496 96,612 22.50 35.94 0.45 Middle East and North Africa Israel 100 56.07 7,601,884 135,579 50.46 7,702,810 152,643 22.14 35.96 0.71 South Asia Pakistan 440 246.71 7,812,290 3,1666 222.04 8,256,364 37,185 4.12 35.84 0.16 Sub-Saharan Africa South Africa 210 117.75 7,669,956 65,139 105.9 7,881,901 74,377 16.46 35.92 0.34 public health benefits of screening for DS in socioeconomically Because fetal losses following CVS are often higher than deprived areas are small because of low uptake of amniocente- those for amniocentesis, efficacy analysis should be conducted sis (Ford and others 1998). With lower uptake rates of amnio- assuming a 1.5 percent fetal loss risk attributable to invasive centesis, both efficacy and financial cost-effectiveness are testing in the first trimester. With higher fetal losses, the efficacy adversely affected as a result of low detection rates, and the of the prenatal screening is adversely affected, although the number of unaffected fetal losses decreases. It is also important cost-effectiveness will not change. to note that, in some countries, many women may not have In addition, assuming a higher false-positive rate of 8.3 per- access to prenatal care or may not seek prenatal care and pre- cent increases the number of invasive tests on unaffected moth- natal testing. In such areas, programs that try to reduce DS ers and the number of unaffected fetal losses, thus adversely prevalence will have limited success, especially if a population affecting the efficacy of the prenatal testing (Vintzileos and at greater risk of DS is not tested. others 2000). Cost-effectiveness is often measured per DS birth averted The analysis presented above is limited to an evaluation of since reduction in DS prevalence is the ultimate goal of prena- the cost-effectiveness of prenatal screening for DS only. Some tal testing. In many cultures, an abortion is not an acceptable serum markers (for example, alpha-fetoprotein) will identify option. Acceptance of elective termination of pregnancy may other abnormalities, the benefits of which are not included in also vary across ethnicities and other subgroups within a given this analysis. country. A study in California found the uptake of termination following the DS diagnosis varied from 47.5 percent for Equity and Access The desirable policy is that women of sim- Hispanics to 65.8 percent for whites and 70.8 percent for Asians ilar risk for DS have equal access to diagnostic tests. With lim- (Cunningham and Tompkinison 1999). If few families decide ited access to prenatal care, the introduction of the screening to terminate pregnancy to avoid having a child with severe dis- programs can have small public health effects. Although the ability, cost-effectiveness per DS birth averted will be adversely approach used in cost-effectiveness analysis is optimization of affected, and the screening program may fail to reduce the birth societal net benefit, the policies to be recommended for the prevalence of DS. If a large percentage of families are opposed prevention of disability must also consider individuals' free- to induced abortion of fetal DS, the uptake of amniocentesis dom in decision making at each step of the prenatal diagnosis. also will be low. Successful policies need to be based on cost-effectiveness 944 | Disease Control Priorities in Developing Countries | Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, and others estimates that take into account the needs, sensitivities, and val- normal brain development, and little or no thyroid hormone in ues of individuals and cultures (Institute of Medicine 2003). the neonatal period results in damage to the nervous system. Various causes of anatomical maldevelopment of the thyroid gland are responsible for CH, and several genes have been Interventions to Prevent Disability Caused implicated. With biochemical newborn screening (best con- by Congenital Hypothyroidism ducted in centralized regional laboratories) using dried blood For CH, like DS, screening programs are critical public health spots and diagnosis in the first few weeks of life, MR is avoid- interventions. able. Without appropriate treatment, two-thirds of patients with CH have low IQ, and 30 percent experience severe or pro- Neonatal Screening in Low- and Middle-Income Countries. found cognitive disability (Beaulieu 1994). Even with appro- When considering the costs and benefits associated with a CH priate treatment, some subtle intellectual impairment and screening program, one must first have an estimate of how behavior deficits may still occur--the mean IQ may be approx- prevalent CH is in the population so that the avoided costs imately 10 points lower than that of the general population associated with disability can be calculated. It is important to (Tillotson and others 1994). In the United States, infants are note that in high-income countries and in several middle- screened as newborns and again at two to six weeks of age to income countries screening is usually done for a series of con- detect missed cases. For optimal outcomes, lifelong treatment ditions rather than for a single disorder. This fact is likely to with thyroid hormone is required, with subsequent monitoring affect the cost. In several of these conditions, the treatment and adjustments recommended every 3 to 12 months until includes dietary modification as well as costly prepared foods growth and puberty are complete. Many females born with CH and formulas. Policies in countries where this type of screening are now reaching childbearing age and require increased occurs include labeling of food to alert potentially vulnerable dosages of thyroid hormone during pregnancy for optimal consumers. neuropsychological outcome in their offspring. Several studies have examined the prevalence of CH in spe- cific populations, with substantially varying results. A review Costs of Care. Estimated lifetime costs of care for the child of 13 studies reporting findings on CH prevalence identified with CH include the following (Barden and Kessel 1984): through individual screening programs found the lowest rate to be 1 case of CH per over 6,000 screened in Thailand · Institutional care. At the time of the study, 15 percent of con- (Wasant, Liammongkolkul, and Srisawat 1999). Contrasting genital hypothyroid individuals were institutionalized from this is the highest rate reported: 1 case in 1,000 screened in age 5 to 70. Pakistan (Lakhani and others 1989). Prevalence can vary not · Foster care. About 25 percent of congenital hypothyroid only from one country to the next, but also within countries, cases received foster care from age 5 to 20. depending on different analyses or subpopulations within one · Residential care. Such care was provided for 40 percent of country. These variations demonstrate the need for identifying affected cases. the appropriate population in order to conduct economic eval- · Special education expenditures. Such expenses varied with uations of screening interventions. the level of MR (15 percent severe, 25 percent moderate, and According to three cost-benefit analyses of CH screening 40 percent mild). (Layde, Von Allmen, and Oakley 1979; Barden and Kessel 1984), the benefits included savings from institutionalization, In 2004, estimated lifetime costs of CH care is US$191,000, special education, medical care, lost parent and child produc- with a 6 percent discount rate. This estimate of the financial tivity, and slightly decreased life expectancy. The costs included costs of care for an affected person is fairly conservative; it does those of the screening program as well as the cost of treating not take into account lost productivity of the person with CH, detected cases. Overall, CH screening programs are substantially a potential loss of income attributable to the time inputs of the cost saving, with a cost-benefit ratio as high as 1 to 8.9 in high- family members who are taking care of the affected person, or income countries (Dhondt and others 1991). Such savings have effects on quality of life. not yet been evaluated in LMICs. Because the treatment is inexpensive and highly effective, it is anticipated that CH Cost-Effectiveness of Neonatal Screening. Table 49.5 presents screening would also be substantially cost saving in LMICs. cost-effectiveness analysis of neonatal screening for representa- tive countries in the World Bank regions. Screening costs Burden. Congenital hypothyroidism is a common cause of include blood sample collection, laboratory costs, discounted MR that can be prevented by newborn screening and treat- lifetime treatment cost, and costs of care for those missed ment. By the end of the 1970s, neonatal screening programs by the screening. Specimen collection and laboratory had been established in many regions of Canada, Europe, costs (Barden and Kessel 1984; OTA 1988) constitute (in Japan, and the United States. Thyroid hormone is required for 2004 dollars) US$989,000 and US$969,000, respectively, per Learning and Developmental Disabilities | 945 Table 49.5 Cost-Effectiveness of Neonatal Screening for Congenital Hypothyroidism by World Bank Region Program costs Cost per Cost CH births per CH for screening disability without Cost Representative 100,000 (birth births and treatment averted testing savings country prevalence) detected (US$) (US$) (US$) (US$) East Asia and the Pacific Thailand 23.94 22.74 2,236,661 98,366 4,342,987 2,106,326 Europe and Central Asia Estonia 34.97 33.22 2,407,937 72,492 6,344,406 3,936,468 High-income countries United States 25.00 23.75 2,253,200 94,872 4,536,250 2,283,050 Latin America and the Caribbean Mexico 40.7 38.67 2,496,991 64,580 7,385,022 4,888,032 Middle East and North Africa Saudi Arabia 36.25 34.43 2,427,813 70,509 6,576,658 4,148,845 South Asia Pakistan 100.00 95.00 3,417,801 35,977 18,145,000 14,727,199 Sub-Saharan Africa South Africa 24.13 22.93 2,239,768 97,686 4,379,292 2,139,524 100,000 children tested. Lifetime discounted (at 6 percent) the cost-effectiveness analysis, because the difference between treatment costs are US$6,292.64 in 2004 dollars (Barden and program and treatment expenditures and lifetime costs will Kessel 1984). Analysis of costs and benefits in table 49.5 shows remain even after we scale the medical costs. that, although screening for the population as a whole requires The analysis presented above is limited to an evaluation of considerable investment and infrastructure, the burden from the cost-effectiveness of neonatal screening for CH. For mini- the disorder is high and treatment is cheap. Screening all new- mal extra cost, collected blood samples for CH can also be used borns is beneficial compared with the high costs of lifelong care to identify other inherited disorders, including phenylke- for the affected individuals. Cost savings are positive for all rep- tonuria, maple syrup urine disease, and other inborn errors of resentative countries despite high variance in the prevalence of metabolism. Without the benefits of early detection and treat- CH. Even for a low birth prevalence estimate of 4 out of every ment for these conditions, the result is severe MR. 100,000 in Thailand (Wasant, Liammongkolkul, and Srisawat 1999), the cost savings would be US$106,326. Effectiveness of the newborn screening in identifying the Community-Based Rehabilitation affected infants depends on the ability of the screening pro- Community-based rehabilitation is a set of low-cost approaches gram to collect blood samples from all infants in the first week to providing rehabilitation services such as physical therapy, and to perform tests in time to initiate treatment. This effort occupational therapy, prosthetics, and assistive devices for may be difficult in some settings, where infants are born at people with disabilities in developing countries. Such rehabili- home or released on the first day after birth and do not have tation also aims to minimize stigmatization of people with dis- contact with the health care system in the first month of their abilities and to support inclusive education and integration of lives (Sack, Feldman, and Kaiserman 1998). The wider the cov- people with disabilities into society. WHO and other organiza- erage of the screening program, the higher will be the cost sav- tions have actively promoted community-based models for ings of screening. Also, follow-up screening for those infants providing rehabilitation services--including services for chil- who test as false negative will increase sensitivity to 100 percent dren with developmental disabilities--in resource-poor set- and improve cost-effectiveness of the program. tings (Institute of Medicine Committee on Nervous System In our cost-effectiveness analysis, we assumed the lifetime Disorders in Developing Countries 2001). care and treatment costs to be similar to those estimated for the Although 80 percent or more of the world's people with United States. However, medical costs may vary significantly disabilities live in developing countries, only 2 percent have among and within countries. Such variation is unlikely to alter access to rehabilitation services (WHO Community-Based 946 | Disease Control Priorities in Developing Countries | Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, and others Rehabilitation 1982). If families of people with disabilities are Efficacy. Other attempts to evaluate community-based reha- taken into account, the number of people experiencing the bilitation in different settings and for a range of outcomes effects of disability is estimated to be up to 25 percent of include one for preschool children with disabilities in rural the world population. Community-based rehabilitation is Guyana. The children showed significant improvement after six designed to expand access to rehabilitation services in poor and months in the program, and noticeable improvements in the rural areas by providing training in rehabilitation techniques to attitudes of parents and others toward children with disabilities individuals with disabilities, their family members, and others were seen (O'Toole 1988). Community-based rehabilitation in their communities. It also attempts to change negative atti- programs in the Philippines and Zimbabwe found gains in tudes toward disability and to remove barriers in the physical activities of daily living and communication as well as higher environment that prevent people with disabilities from fully rates of starting school and employment after six months in the participating in society. program (Lagerkvist 1992). Similarly, people with disabilities participating in a community-based rehabilitation program in Botswana showed high levels of independence in activities of Costs. In the community-based rehabilitation model, com- daily living; 20 percent of adults were working, and most munity interventions are shifted from institutions and centers school-age children were attending class (Lundgren-Lindquist to the homes and communities of people with disabilities and and Nordholm 1996). are carried out largely by family members and volunteers. By Some have questioned the efficacy of community-based using volunteer workers and the existing infrastructure in the rehabilitation and its ability to expand access for people with communities, this form of rehabilitation minimizes costs of disabilities. Many people with disabilities do not patronize such delivering services and is assumed to be cost-effective relative programs, and many who do try leave dissatisfied (Kassah 1998). to the alternative institution-based rehabilitation (Institute After initial diagnosis, only half of the identified individuals of Medicine Committee on Nervous System Disorders in with disabilities continued (Rottier and others 1993). Developing Countries 2001; WHO Community-Based Community-based rehabilitation programs also face the diffi- Rehabilitation 1982; Lagerkvist 1992). Institutional care has culty of working in diverse communities with unique cultural, higher costs because it relies on paid staff, medical equipment, religious, economic, and social conditions, making it difficult for building maintenance, and medical costs. Some advocate for a single model to meet the needs for rehabilitation services in provision of institutional, center-based, medical, and commu- developing countries (Crishna 1999). nity-based approaches in a complementary fashion. In a Zimbabwean community-based rehabilitation project two- thirds of the patients were referred to hospitals or clinics Prevention of Premature Birth (Rottier and others 1993). Annual costs for training workshops Premature birth--after 20 weeks but before 37 weeks--is a and salaries of rehabilitation workers amounted to US$60,000 powerful predictor of death and disability. Infants born before to treat 1,614 individuals with disabilities. 28 weeks of gestation have a 50-fold increased risk of cerebral Little information is available about the full costs of com- palsy (Drummond and Colver 2002) and heightened risk of sen- munity-based rehabilitation and how they vary across disabili- sory, cognitive, and educational impairment (Taylor and others ties, age groups, and societies. The cost-effectiveness of such 2004). The rate of preterm births in the United State has rehabilitation or whether its costs are lower than alternative increased steadily since 1990 (MacDorman and others 2002). rehabilitation models is unknown. It is usually implemented in Survival of infants born before 34 weeks requires intensive and settings where no other rehabilitation models exist. The costs to expensive medical care (Gilbert, Nesbitt, and Danielsen 2003), consumers in terms of their efforts, time, and money may be and the global survival rate differs depending on neonatal care substantial (Thomas and Thomas 1998). No formal estimates availability (Lorenz and others 2001).Infection or inflammation are available of time costs and opportunity costs to family of the placenta is common in preterm pregnancies (Goldenberg members involved in community-based rehabilitation. and Culhane 2003), and many deaths attributed to asphyxia may Meeting the needs of a family member with a disability may be caused by maternal, placental, or neonatal infection. The prohibit or disrupt labor force participation of the caregiver cerebral palsy rate is significantly higher in premature infants and reduce family income. This need for caregiving may espe- whose births are monitored electronically (Shy and others cially affect women (Giacaman 2001). The effectiveness of 1990).With the exception of magnesium sulfate administered to community-based rehabilitation in improving functional out- women in preterm labor (Crowther and others 2003), no spe- comes for children with cerebral palsy in Bangladesh showed cific interventions result in decreased cerebral palsy among pre- no improvement, but researchers unexpectedly found a signif- mature infants (Crowther and Henderson-Smart 2004). icant increase in reported stress and symptoms of depression in Dietary supplements might decrease the frequency of low- the mothers of children in the community-based rehabilitation birthweight births and perhaps the frequency of marked intervention group (McConachie and others 2000). prematurity. In Bangladesh, where the rate of preterm labor Learning and Developmental Disabilities | 947 was high, women who went into labor before term were older, · methodological and prevalence studies to ensure that the shorter, thinner, less educated, and more disadvantaged eco- impacts of LDD are effectively measured by DALYs or other nomically, with closer spacing of births (Begum, Buckshe, and international indicators Pande 2003). Deaths attributable to prematurity in LMICs are · cost-effectiveness of interventions to prevent specific nutri- seldom due to poor management and are largely related to tional, infectious, genetic, and other causes of LDD poor health facilities (Pattinson 2004). · impact on child development of multiple insults and risk factors especially common in LMICs, such as neurotoxic exposures, trauma, infectious disease or malnutrition, Electronic Fetal Monitoring in Labor poverty, maternal illiteracy, and other social factors For decades most cerebral palsy and a major share of MR, · health services research related to access to prenatal care and epilepsy, and learning and behavioral disorders of childhood prenatal and newborn screening and evaluation of compo- were considered to be due to deprivation of oxygen supply to nents of the public health system that might impair or the fetus during birth. Recent research confirms that only a enhance integration of services for patients with LDD minority of cerebral palsy cases, as well as associated MR and · prevalence of ADHD and a cost-benefit analysis of the use seizures, are related to markers of birth asphyxia (Nelson and of psychotropic medications Ellenberg 1986; Torfs and others 1990). Low Apgar scores, the · prevalence and costs of autism spectrum disorders need for respiratory support, and neonatal seizures are more · strategies to improve interventions for the prevention of commonly due to etiologies other than asphyxia, most notably fetal alcohol syndrome and to develop effective intervention intrauterine exposure to infection (Wu and others 2003). So programs for children affected by prenatal alcohol exposure that medical workers could recognize the onset of asphyxia and · evaluation of criteria for newborn screening and effects of "rescue" the fetus, continuous electronic fetal monitoring new technology on measured incidence, costs, and system (EFM) of the fetal heart rate during labor was introduced in effectiveness the 1970s. This intervention was disseminated before being · evaluation of financing of successful newborn screening and tested in randomized trials to compare continuous electronic treatment programs monitoring with intermittent observation by stethoscope (aus- · model systems of care for individuals diagnosed through cultation). Since the introduction of EFM, fetal death in labor newborn screening from infancy to adulthood. has decreased, the cesarean section rate has quadrupled (Natale and Dodman 2003), and the rate of cerebral palsy has remained steady. Accordingly, EFM cannot be recommended SUMMARY for use in LMICs. Intermittent auscultation with a stethoscope appears to be the appropriate way to monitor fetal status Many potential interventions exist for the prevention of LDDs, during labor. and relatively few are being implemented for the benefit of chil- dren in LMICs. The following three interventions are effective and cost-effective in preventing LDD: RESEARCH AGENDA FOR PREVENTION OF DISABILITIES IN LOW- AND MIDDLE-INCOME · Folic acid fortification of the food supply can reduce the COUNTRIES occurrence of NTDs by 50 percent or more. This interven- tion was found to be highly cost-effective in the United Research needed as a basis for developing policies and inter- States; however, in low-income countries, high capital and ventions to prevent LDD in low- and middle-income countries running costs may compromise cost-effectiveness, at least in includes basic research, epidemiology, and evaluations of early the short run. interventions, clinical treatments, prevention strategies, and · Prenatal screening and selective pregnancy termination to health services that are culturally appropriate and feasible. prevent DS are highly cost-effective under some conditions Suggested research priorities include the following: but raise ethical, social, and cultural concerns that may preclude their applicability in some LMICs. Screening is not · etiology and prevention of adverse pregnancy outcomes only expensive; it also has negative health outcomes: the associated with LDD, such as low birthweight, preterm false-positive rates and the subsequent anxiety, a risk of birth, intrauterine growth retardation, and related factors miscarriage of an unaffected pregnancy, and the resulting · community-based rehabilitation, including effectiveness, potential complications from pregnancy termination. cost-effectiveness, and social effects of different models for Another concern is that, where access to prenatal care is lim- providing rehabilitation services and special education to ited, the potential for public health benefits of prenatal children with LDD in LMICs screening will be small. 948 | Disease Control Priorities in Developing Countries | Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, and others · Neonatal screening and treatment for CH is highly cost- Barden, H. S., and R. Kessel. 1984."The Costs and Benefits of Screening for effective in developed countries, where it provides a low- Congenital Hypothyroidism in Wisconsin." Social Biology 31 (3­4): 185­200. cost strategy for preventing MR. For minimal extra cost, col- Beaulieu, M. D. 1994. "Screening for Congenital Hypothyroidism." lected blood samples from newborns can also be used to In Canadian Guide to Clinical Preventive Health Care, ed. Canadian identify and prevent the disabling effects of other inborn Task Force on the Periodic Health Examination. Ottawa: Health errors of metabolism, such as phenylketonuria and maple Canada. syrup urine disease. However, when only a part of the new- Begum, F., K. Buckshe, and J. N. Pande. 2003."Risk Factors Associated with born population is reached by screening, high costs will be Preterm Labour." Bangladesh Medical Research Council Bulletin 29 (2): 59­66. incurred to care for those missed by the screening, thereby Benedict, R. E., and A. M. Farel. 2003. "Identifying Children in Need of reducing the cost-benefit ratio. Ancillary and Enabling Services: A Population Approach." Social Science and Medicine 57 (11): 2035­47. For another type of intervention considered, community- Bobat, R., D. Moodley, A. Coutsoudis, H. Coovadia, and E. Gouws. 1998. based rehabilitation, costs and benefits have not been quantified "The Early Natural History of Vertically Transmitted HIV-1 Infection in African Children from Durban, South Africa." Annals of Tropical sufficiently to allow evaluation. Such rehabilitation is designed Paediatrics 18 (3): 187­96. to expand access to services in poor and rural areas, to change Brouwers, P., C. Decarli, M. P. Heyes, H. A. Moss, P. L. Wolters, G. Tudor- negative attitudes toward disability, to lower the costs of deliv- Williams, and others. 1996. "Neurobehavioral Manifestations of ering services, and to enhance the participation of persons with Symptomatic HIV-1 Disease in Children: Can Nutritional Factors Play disabilities in society. The benefits of community-based reha- a Role?" Journal of Nutrition 126 (Suppl. 10): S2651­62. bilitation may come at a high cost in terms of time and financial Chan, E., C. Zhan, and C. J. Homer. 2002. "Health Care Use and Costs for Children with Attention-Deficit/Hyperactivity Disorder: National resources of family members. Estimates from the Medical Expenditure Panel Survey." Archives of Another intervention, electronic fetal monitoring in labor, Pediatrics and Adolescent Medicine 156 (5): 504­11. has been shown to be unsuccessful in preventing childhood Crishna, B. 1999. "What Is Community-Based Rehabilitation? A View neurological disability associated with premature birth: the risk from Experience." Child: Care, Health, and Development 25 (1): 27­35. of cerebral palsy was significantly higher in infants delivered Crowther, C. A., and D. J. Henderson-Smart. 2004. "Vitamin K Prior to using EFM. Consequently, this intervention is not recom- Preterm Birth for Preventing Neonatal Periventricular Haemorrhage." Cochrane Database of Systematic Reviews (4). mended for use during labor. Crowther, C. A., J. E. Hiller, L. W. Doyle, and R. R. Haslam. 2003. "Effect DALY estimates are not available to convey the full range of of Magnesium Sulfate Given for Neuroprotection before Preterm LDDs or their risk factors. However, available data are consis- Birth: A Randomized Controlled Trial." Journal of the American tent with the possibility that these disabilities account for a Medical Association 290 (20): 2669­76. large proportion of the global burden of disease. Quantifying Cunningham, G. C., and D. G. Tompkinison. 1999."Cost and Effectiveness the impacts of LDDs and their preventive interventions is com- of the California Triple Marker Prenatal Screening Program." Genetics in Medicine 1 (5): 199­206. plicated by the fact that these disorders can exist at multiple Cusick, W., P. Buchanan, T. W. Hallahan, D. A. Krantz, J. W. Larsen Jr., and levels and that disability is context-specific, with impacts that J. N. Macri. 2003. "Combined First-Trimester versus Second-Trimester may vary across cultures. Several research priorities for Serum Screening for Down Syndrome: A Cost Analysis." American improving knowledge and developing policies and interven- Journal of Obstetrics and Gynecology 188 (3): 745­51. tions to prevent LDD in LMICs are suggested. Cutts, F. T., S. E. Robertson, J. L. Diaz-Ortega, and R. Samuel. 1997. "Control of Rubella and Congenital Rubella Syndrome (CRS) in Developing Countries, Part 1: Burden of Disease from CRS." Bulletin of the World Health Organization 75 (1): 55­68. ACKNOWLEDGMENTS Dhondt, J. L., J. P. Farriaux, J. C. Sailly, and T. Lebrun. 1991. "Economic Evaluation of Cost-Benefit Ratio of Neonatal Screening Procedure for The authors thank Emmy Cauthen at the Fogarty International Phenylketonuria and Hypothyroidism." Journal of Inherited Metabolic Disease 14 (4): 633­39. Center for her support. Drummond, P. M., and A. F. Colver. 2002. "Analysis by Gestational Age of Cerebral Palsy in Singleton Births in North-East England 1970­94." Paediatric and Perinatal Epidemiology 16 (2): 172­80. Durkin, M. 2002. "The Epidemiology of Developmental Disabilities in REFERENCES Low-Income Countries." Mental Retardation and Developmental Disabilities Research Reviews 8 (3): 206­11. Alliance to End Childhood Lead Poisoning. 2002. "The Global Lead Initiative: A Proposed Outcome for the World Summit on Sustainable Fewtrell, L. J., A. Pruss-Ustun, P. Landrigan, and J. L. Ayuso-Mateos. 2004. Development."Alliance to End Childhood Lead Poisoning,Washington, "Estimating the Global Burden of Disease of Mild Mental Retardation DC. http://www.globalleadnet.org/pdf/GlobalLeadInitiative.pdf. and Cardiovascular Diseases from Environmental Lead Exposure." Environmental Research 94 (2): 120­33. Bagriansky, J. No date. "What Are the Costs for the Premixes?" http://www.sph.emory.edu/wheatflour/Comm/Resource/CDs/Bali/ Floyd, R. L., and J. S. Sidhu. 2004. "Monitoring Prenatal Alcohol Bali_files/v3_slide0089.htm. Exposure." American Journal of Medical Genetics 127C (1): 3­9. Learning and Developmental Disabilities | 949 Ford, C., A. J. Moore, P. A. Jordan, W. A. Bartlett, M. P. Wyldes, A. F. Jones, Lupton, C., L. Burd, and R. Harwood. 2004. "Cost of Fetal Alcohol and W. E. MacKenzie. 1998. "The Value of Screening for Down's Spectrum Disorders." American Journal of Medical Genetics 127C (1): Syndrome in a Socioeconomically Deprived Area with a High Ethnic 42­50. Population." British Journal of Obstetrics and Gynaecology 105 (8): MacDorman, M. F., A. M. Minino, D. M. Strobino, and B. Guyer. 2002. 855­59. "Annual Summary of Vital Statistics--2001." Pediatrics 110 (6): Giacaman, R. 2001."A Community of Citizens: Disability Rehabilitation in 1037­52. the Palestinian Transition to Statehood." Disability and Rehabilitation Mathers, C. D., A. D. Lopez, and C. J. L. Murray. "The Burden of Disease 23 (14): 639­44. and Mortality by Condition: Data, Methods, and Results for 2001." In Gilbert, R. E., C. Augood, R. Gupta, A. E. Ades, S. Logan, M. Sculpher, and Global Burden of Disease and Risk Factors, eds. A. D. Lopez, C. D. J. H. van der Meulen. 2001. "Screening for Down's Syndrome: Effects, Mathers, M. Ezzati, D. T. Jamison, and C. J. L. Murray. New York: Safety, and Cost Effectiveness of First and Second Trimester Strategies." Oxford University Press. British Medical Journal 323 (7310): 423­25. May, P. A., L. Brooke, J. P. Gossage, J. Croxford, C. Adnams, K. L. Jones, and Gilbert, W. M., T. S. Nesbitt, and B. Danielsen. 2003. "The Cost of others. 2000. "Epidemiology of Fetal Alcohol Syndrome in a South Prematurity: Quantification by Gestational Age and Birth Weight." African Community in the Western Cape Province." American Journal Obstetrics and Gynecology 102 (3): 488­92. of Public Health 90 (12): 1905­12. Goldenberg, R. L., and J. F. Culhane. 2003."Infection as a Cause of Preterm McConachie, H., S. Huq, S. Munir, S. Ferdous, S. Zaman, and N. Z. Khan. Birth." Clinics in Perinatology 30 (4): 677­700. 2000. "A Randomized Controlled Trial of Alternative Modes of Service Gupta, R., R. D. Thomas, V. Sreenivas, S. Walter, and J. M. Puliyel. 2001. Provision to Young Children with Cerebral Palsy in Bangladesh." "Ultrasonographic Femur-Tibial Length Ratio: A Marker of Down Journal of Pediatrics 137 (6): 769­76. Syndrome from the Late Second Trimester." American Journal of McGrother, C. W., and B. Marshall. 1990. "Recent Trends in Incidence, Perinatology 18 (4): 217­24. Morbidity and Survival in Down's Syndrome." Journal of Mental Insinga, R. P., R. H. Laessig, and G. L. Hoffman. 2002. "Newborn Screening Deficiency Research 34 (Part 1): 49­57. with Tandem Mass Spectrometry: Examining Its Cost-Effectiveness in Natale, R., and N. Dodman. 2003. "Birth Can Be a Hazardous Journey: the Wisconsin Newborn Screening Panel." Journal of Pediatrics 141 (4): Electronic Fetal Monitoring Does Not Help." Journal of Obstetrics and 524­31. Gynaecology Canada 25 (12): 1007­9. Institute of Medicine. 2003. Reducing Birth Defects: Meeting the Challenge Nelson, K. B., and J. H. Ellenberg. 1986. "Antecedents of Cerebral Palsy: in the Developing World. Washington, DC: National Academies Press. Multivariate Analysis of Risk." New England Journal of Medicine 315 Institute of Medicine Committee on Nervous System Disorders in (2): 81­86. Developing Countries. 2001. Neurological, Psychiatric, and Develop- Nyberg, D. A., D. A. Luthy, R. G. Resta, B. C. Nyberg, and M. A. Williams. mental Disorders: Meeting the Challenges in the Developing World. 1998. "Age-Adjusted Ultrasound Risk Assessment for Fetal Down's Washington, DC: National Academies Press. Syndrome during the Second Trimester: Description of the Method Kassah, A. K. 1998. "Community-Based Rehabilitation and Stigma and Analysis of 142 Cases." Ultrasound Obstetrics and Gynecology Management by Physically Disabled People in Ghana." Disability and 12 (1): 8­14. Rehabilitation 20 (2): 66­73. O'Toole, B. 1988. "A Community-Based Rehabilitation Programme for Khan, N. Z., and A. H. Khan. 1999. "Lead Poisoning and Psychomotor Pre-school Disabled Children in Guyana." International Journal of Delay in Bangladeshi Children." Lancet 353 (9154): 754. Rehabilitation Research 11 (4): 323­34. Kline, J. K., Z. Stein, and M. Susser. 1989. Conception to Birth: Epidemiology Panepinto, J. A., D. Magid, M. J. Rewers, and P. A. Lane. 2000. "Universal of Prenatal Development. New York: Oxford University Press. versus Targeted Screening of Infants for Sickle Cell Disease: A Cost- Effectiveness Analysis." Journal of Pediatrics 136 (2): 201­8. Lagerkvist, B. 1992. "Community-Based Rehabilitation--Outcome for the Disabled in the Philippines and Zimbabwe." Disability and Pattinson, R. C. 2004. "Are Deaths Due to Prematurity Avoidable in Rehabilitation 14 (1): 44­50. Developing Countries?" Tropical Doctor 34 (1): 7­10. Lakhani, M., M. Khurshid, S. H. Naqvi, and M. Akber. 1989. "Neonatal Peden, M., R. Scurfiled, D. Sleet, D. Mohan, A. Hyder, E. Jarawan, and Screening for Congenital Hypothyroidism in Pakistan." Journal of the C. Mather, eds. 2004. World Report on Road Traffic Injury Prevention. Pakistan Medical Association 39 (11): 282­84. Geneva: World Health Organization. Layde, P. M., S. D. Von Allmen, and G. P. Oakley Jr. 1979. "Congenital Robertson, S. E., F. T. Cutts, R. Samuel, and J. L. Diaz-Ortega. 1997. Hypothyroidism Control Programs: A Cost-Benefit Analysis." Journal "Control of Rubella and Congenital Rubella Syndrome (CRS) in of the American Medical Association 241 (21): 2290­92. Developing Countries, Part 2: Vaccination against Rubella." Bulletin of the World Health Organization 75 (1): 69­80. Lippman, A., D. J. Tomkins, J. Shime, and J. L. Hamerton. 1992. "Canadian Multicentre Randomized Clinical Trial of Chorion Villus Sampling Romano, P. S., N. J. Waitzman, R. M. Scheffler, and R. D. Pi. 1995. "Folic and Amniocentesis: Final Report." Prenatal Diagnosis 12 (5): 385­408. Acid Fortification of Grain: An Economic Analysis." American Journal of Public Health 85 (5): 667­76. Lord, J., M. J. Thomason, P. Littlejohns, R. A. Chalmers, M. D. Bain, G. M. Addison, and others. 1999. "Secondary Analysis of Economic Data: Ross, H. L., and S. Elias. 1997. "Maternal Serum Screening for Fetal Genetic A Review of Cost-Benefit Studies of Neonatal Screening for Disorders." Obstetrics and Gynecology Clinics of North America 24 (1): Phenylketonuria." Journal of Epidemiology and Community Health 33­47. 53 (3): 179­86. Rottier, M. J. N., R. W. Broer, A. Vermeer, and H. J. M. Finkenflügel. 1993. Lorenz, J. M., N. Paneth, J. R. Jetton, L. den Ouden, and J. E. Tyson. 2001. "A Study of Follow Up of Clients in Community-Based Rehabilitation "Comparison of Management Strategies for Extreme Prematurity Projects in Zimbabwe." Journal of Rehabilitation Sciences 6 (2): 35­41. in New Jersey and the Netherlands: Outcomes and Resource Sack, J., I. Feldman, and I. Kaiserman. 1998. "Congenital Hypothyroidism Expenditure." Pediatrics 108 (6): 1269­74. Screening in the West Bank: A Test Case for Screening in Developing Lundgren-Lindquist, B., and L. A. Nordholm. 1996. "The Impact of Regions." Hormone Research 50 (3): 151­54. Community-Based Rehabilitation as Perceived by Disabled People in a Sampson, P. D., A. P. Streissguth, F. L. Bookstein, R. E. Little, S. K. Clarren, Village in Botswana." Disability and Rehabilitation 18 (7): 329­34. P. Dehaene, and others. 1997. "Incidence of Fetal Alcohol Syndrome 950 | Disease Control Priorities in Developing Countries | Maureen S. Durkin, Helen Schneider, Vikram S. Pathania, and others and Prevalence of Alcohol-Related Neurodevelopmental Disorder." Environmental Factors and Injury among Children and Adolescents in Teratology 56 (5): 317­26. Europe." Lancet 363 (9426): 2032­39. Shy, K. K., D. A. Luthy, F. C. Bennett, M. Whitfield, E. B. Larson, G. van Belle, Viljoen, D. 1999. "Fetal Alcohol Syndrome." South African Medical Journal and others. 1990. "Effects of Electronic Fetal-Heart-Rate Monitoring, 89 (9): 958­60. as Compared with Periodic Auscultation, on the Neurologic Vintzileos, A. M., C. V. Ananth, J. C. Smulian, D. L. Day-Salvatore, Development of Premature Infants." New England Journal of Medicine T. Beazoglou, and R. A. Knuppel. 2000. "Cost-Benefit Analysis of 322 (9): 588­93. Prenatal Diagnosis for Down Syndrome Using the British or the Sokol, R. J., V. Delaney-Black, and B. Nordstrom. 2003. "Fetal Alcohol American Approach." Obstetrics and Gynecology 95 (4): 577­83. Spectrum Disorder." Journal of the American Medical Association Waitzman, N. J., P. S. Romano, and R. M. Scheffler. 1994. "Estimates of the 290 (22): 2996­99. Economic Costs of Birth Defects." Inquiry 31 (2): 188­205. Stein, R. E., and D. J. Jessop. 2003. "The Impact on Family Scale Revisited: Wald, N. J., C. Rodeck, A. K. Hackshaw, J. Walters, L. Chitty, and A. M. Further Psychometric Data." Journal of Developmental and Behavioral Mackinson. 2003. "First and Second Trimester Screening for Down's Pediatrics 24 (1): 9­16. Syndrome: The Results of the Serum, Urine and Ultrasound Screening Stray-Pedersen, B., J. Biornstad, M. Dahl, T. Bergan, G. Aanestad, Study (SURUSS)." Health Technology Assessment 7 (11): 1­77. L. Kristiansen, and K. Hansen. 1991. "Induced Abortion: Wasant, P., S. Liammongkolkul, and C. Srisawat. 1999. "Neonatal Microbiological Screening and Medical Complications." Infection Screening for Congenital Hypothyroidism and Phenylketonuria at 19 (5): 305­8. Siriraj Hospital, Mahidol University, Bangkok, Thailand--A Pilot Taylor, H. G., N. M. Minich, N. Klein, and M. Hack. 2004. "Longitudinal Study." Southeast Asian Journal of Tropical Medicine and Public Health Outcomes of Very Low Birth Weight: Neuropsychological Findings." 30 (Suppl. 2): 33­37. Journal of the International Neuropsychological Society 10 (2): 149­63. WHO (World Health Organization). 2001. International Classification of Thomas, M., and M. J. Thomas. 1998. "Controversies on Some Conceptual Functioning, Disability and Health. Geneva: WHO. http://www3. issues in Community-Based Rehabilitation." Asia Pacific Disability who.int/icf/icftemplate.cfm?myurl=homepage.html&mytitle=Home Rehabilitation Journal 9 (1): 12­14. %20Page. Thompson, D. C., F. P. Rivara, and R. Thompson. 1999. "Helmets for WHO Community-Based Rehabilitation. 1982. "Report of a WHO Preventing Head and Facial Injuries in Bicyclists." Cochrane Database Interregional Consultation." Colombo, Sri Lanka. of Systematic Reviews (4) CD001855. Winter, S., A. Autry, C. Boyle, and M. Yeargin-Allsopp. 2002. "Trends in the Tillotson, S. L., P. W. Fuggle, I. Smith, A. E. Ades, and D. B. Grant. 1994. Prevalence of Cerebral Palsy in a Population-Based Study." Pediatrics "Relation between Biochemical Severity and Intelligence in Early 110 (6): 1220­25. Treated Congenital Hypothyroidism: A Threshold Effect." British Wu, Y. W., G. J. Escobar, J. K. Grether, L. A. Croen, J. D. Greene, and T. B. Medical Journal 309 (6952): 440­45. Newman. 2003. "Chorioamnionitis and Cerebral Palsy in Term and Torfs, C. P., B. van den Berg, F. W. Oechsli, and S. Cummins. 1990. Near-Term Infants." Journal of the American Medical Association "Prenatal and Perinatal Factors in the Etiology of Cerebral Palsy." 290 (20): 2677­84. Journal of Pediatrics 116 (4): 615­19. Valent, F., D. Little, R. Bertollini, L. E. Nemer, F. Barbone, and G. Tamburlini. 2004. "Burden of Disease Attributable to Selected Learning and Developmental Disabilities | 951 Chapter 50 Loss of Vision and Hearing Joseph Cook, Kevin D. Frick, Rob Baltussen, Serge Resnikoff, Andrew Smith, Jeffrey Mecaskey, and Peter Kilima Although the loss of vision and hearing has multiple causes, account uncorrected refractive errors, but this change has not and the burden of these diseases is complex, three major points yet been approved. emerge from the outset: The major causes of adult-onset blindness are cataract (47.8 percent), glaucoma (12.3 percent), macular degeneration · Impairments of the essential senses of vision and hearing (8.7 percent), diabetic retinopathy (4.8 percent), trachoma contribute to early demise and are important causes of mor- (3.6 percent), and onchocerciasis (0.8 percent). Uncorrected bidity for individuals who are blind or deaf. refractive errors are also a major cause of morbidity related to · Cost-effective interventions are available to address several vision, but this type of disability is not included in the global causes of these burdens now. burden of disease by definition. It has been estimated to be on · The number of cost-effectiveness analyses of interventions the order of 15 percent of the total blind population and could to preserve hearing or vision in developing countries is quite add 50 percent to the low-vision population. However, there limited. are no published data to do more than speculate. The major causes of childhood vision loss have marked Table 50.1 summarizes the conditions causing the sensory regional variations. They include vitamin A deficiency (xeroph- deficits, the proposed interventions and sites of delivery, and thalmia) and ophthalmia neonatorum in low-income coun- the cost and effectiveness of these interventions to the extent of tries, retinopathy of prematurity and hereditary conditions in current knowledge. Earlier work by Evans and others (1996) in middle-income countries, and congenital cataract and glau- Myanmar does not appear because the cost data are quite old coma everywhere. Table 50.3 shows the estimated number of and because the cost-effectiveness data were in dollars per case blind persons worldwide in 2002. of blindness averted rather than dollars per disability-adjusted Vision loss is chronic and, almost invariably, without remis- life year (DALY) averted, which the latest information provides. sion. The extent of morbidity is related to the level of alteration of vision function. However, 80 percent of cases are avoidable, either through treatment (cataract and refractive errors) or NATURE, CAUSES, AND EPIDEMIOLOGY through primary prevention (onchocerciasis, trachoma, glau- OF VISION LOSS coma, and diabetic retinopathy). Strictly speaking, blindness attributable to glaucoma and diabetic retinopathy can be pre- Table 50.2 provides definitions of visual impairment, vented. However, prevention depends on the availability of a blindness, and low vision according to the International simple, cheap, and efficacious diagnostic test and rigorous Classification of Diseases, Injuries, and Causes of Death (WHO treatment. These are not readily amenable to public health 1993). At this time, the World Health Organization (WHO) is programs even in the most technologically advanced countries, considering changing the classification in order to take into especially in the case of glaucoma. 953 Table 50.1 Cost and Effectiveness Data for Vision and Hearing Care Interventions Incremental cost- Cost data effectiveness data Condition Intervention (2004 US$) Effectiveness data (2004 US$/DALY averted) Trachoma Trichiasis surgery 7.14 per village-based surgerya 77 percent cure rate over two yearsb 4­82c Tetracycline -- 51 percent cure rate in children at 9,600d six months following treatmentd Azithromycin -- 88 percent cure rate in children at 4,100d six months following treatmente Cataract Extracapsular surgery -- -- 200 (low- and middle- income countries); 2,400 (high-income countries)b Onchocerciasis Ivermectin -- -- 40f Source: Authors. -- not available. a. Frick, Keuffel, and Bowman 2001. b. Baltussen, Sylla, and Mariotti 2004. c. Baltussen and others (2005). Cost-effectiveness calculations are based on data from Frick and others (2001) for mass treatment of children only, not greater efficacy reported by Bowman and others (2000) and Solomon and others (2004) for mass treatment of entire communities. The greater efficacy reported in mass treatment of entire communities may lead to better cost-effectiveness. d. Bowman and others 2000. e. Reacher and others 1992. f. Waters, Rehwinkel, and Burnham 2004. Table 50.2 Definitions of Visual Impairment Levels Table 50.3 Number of Blind Worldwide in 2002 from Various Conditions Degree of Visual impairment impairment Definition categories Condition Number blind (millions) Low vision Visual acuity of less than 6/18 1 and 2 Cataract 17.6 (Snellen 20/70) but equal to or Glaucoma 4.5 better than 3/60 (20/400) Age-related macular degeneration 3.2 in the better eye with best possible correction Corneal opacity 1.9 Diabetic retinopathy 1.8 Blindness Visual acuity of less than 3/60 3, 4, and 5 (20/400) or corresponding visual Childhood blindness 1.4 field loss of less than Trachoma 1.3 10 degrees in the better eye with Onchocerciasis 0.3 best possible correction Other causes 4.8 Visual Blindness as well as low vision 1, 2, 3, 4, and 5 Total 36.8 impairment Sources: Pascolini and others 2004; Resnikoff and others 2004. Source: Authors, based on current international definitions by WHO 1993. Burden of Loss for Vision and Risk Factors (ROP) is an important cause in middle-income countries The risk factors for loss of vision are age, gender, poverty, and (Gilbert and Foster 2001). Unfortunately, screening for ROP in poor access to health care. The overall prevalence of vision loss, preterm infants, as well as the organization and provision of which mainly affects the population above age 40, is a function low-vision services, is a tertiary-level function (requiring a well- of age. It is estimated that more than 82.2 percent of all blind equipped clinic or hospital with the most modern facilities), and individuals are 50 or older. Increasing life expectancy results in no data on cost-effectiveness of interventions are available. a growing number of cases of age-related blindness (for exam- More disease-specific factors are poor hygiene, overcrowding, ple, cataract, glaucoma, macular degeneration). Among the 50 ultraviolet radiation, diabetes mellitus, drugs, micronutrient and older age group, cigarette smoking is a clear risk factor for deficiency, heredity and ethnic background, and consanguinity. both cataract and macular degeneration. Childhood vision loss Estimates of the global burden of visual impairment in 2002 represents approximately 4 percent of the total number of visu- were updated using the most recent available data on blindness ally impaired. However, it is the second largest cause of "blind- and low vision (Pascolini and others 2004). The global number person years," following cataract. Retinopathy of prematurity of people who are visually impaired is in excess of 161 million, 954 | Disease Control Priorities in Developing Countries | Joseph Cook, Kevin D. Frick, Rob Baltussen, and others of whom 36.8 million are blind (Resnikoff and others 2004). Several population-based surveys reported higher risk of Because the international definition refers to the best-corrected mortality among people with visual impairment. Relative risk visual acuity (table 50.2), these figures actually underestimate of mortality among blind and low-vision people varied from the magnitude of the global burden of the visual impairment, 1.5 to 4.1 and from 1.1 to 1.6, respectively. In industrial coun- especially in developing countries, where most of the refractive tries, the relative risk of mortality varied from 1.6 to 2.0. The errors are not corrected (Dandona and Dandona 2003; Fotouhi effect may vary by gender (Lee and others 2002; Taylor and oth- and others 2004; Naidoo and others 2004). A WHO working ers 1991). The link between visual impairment and mortality group has recommended the use of the more accurate "pre- remains poorly understood and cannot be attributed to known senting vision," recognizing that many people do not have their associations with underlying disease. best-corrected vision. This recommendation is under review The burden from visual impairment accounts for approxi- and, if approved, would substantially increase the estimates of mately 3 percent of the total global burden of disease and 9 per- the burden of disease attributable to impaired vision. cent of total years lived with disability in 2001. Table 50.4 shows The number of women with visual impairment, as estimated the global burden by vision-related cause in DALYs. Globally, from the available studies, is higher than that of men, even after half of the burden from visual impairment is due to cataract. adjustment for age. Female-to-male prevalence ratios indicate The burden of visual impairment is not distributed uni- that women are more likely to have a visual impairment than formly throughout the world; the least developed regions carry men in every region of the world: the ratios from past studies the largest share, as shown by World Bank region in table 50.5. range between 1.5 to 1 and 2.2 to 1. (Resnikoff and others Local and in-country variations, as well as regional variations, 2004). The major reported reason is women's reduced access to are related to the following factors: eye care services. Higher exposure to risk factors also con- tributes in the case of trachoma. (Abou-Gareeb and others · Epidemiology of cause (for example, onchocerciasis, 2001; Nirmalan, Padmavathi, and Thulasiraj 2003). trachoma). Table 50.4 Global Burden of Visual Impairment, by Major Cause, 2002 Visual impairment Blindness Low vision Percentage Percentage (thousands (thousands Thousands of total of total Condition of DALYs) of DALYs) of DALYs YLDs DALYs Cataract 8,798 15,053 25,251 4.5 1.7 Glaucoma 1,202 2,442 3,866 0.7 0.3 Trachoma 1,403 772 2,329 0.4 0.2 Onchocerciasis 203 146 484 0.1 0.0 Other 4,657 8,814 14,191 2.5 1.0 Total 16,263 27,227 46,121 8.2 3.2 Source: Pascolini and others 2004; Resnikoff and others 2004. YLDs years of life lived with disability. Table 50.5 Global Burden of Visual Impairment, by World Bank Region, 2002 (thousands of DALYs) East Asia and Europe and Latin America and Middle East and Condition the Pacific Central Asia the Caribbean North Africa South Asia Sub-Saharan Africa Worldwide Cataract 6,141 239 956 934 10,259 5,369 23,898 Glaucoma 1,184 168 165 401 566 1,009 3,493 Trachoma 410 0 102 201 226 1,272 2,211 Onchocerciasis 0 0 1 23 0 458 482 Other 5,821 903 1,031 971 2,447 1,046 12,219 Total 13,556 1,310 2,255 2,530 13,458 9,154 42,303 Source: Pascolini and others 2004; calculated from Resnikoff and others 2004. Loss of Vision and Hearing | 955 · Socioeconomic patterns (poverty and socioeconomic depri- morbidity). Surgery is in fact tertiary prevention--that is, vation), an essential element in most causes. repairing (and halting further) damage. The SAFE strategy has · Access to adequate eye care.Uneven access to good-quality eye not been subjected to a comprehensive economic evaluation, care (for example, for cataract, glaucoma, diabetic retinopa- but some cost-effectiveness information is available regarding thy) results from such factors as distance, affordability, and the antibiotic and surgery components. culture. Lack of resources is only part of the problem; existing The initial trachoma infection can be effectively treated with facilities are sometimes underused. antibiotics, either through mass treatment of all children below 10 years of age or through targeted treatment of infected chil- Interventions dren and household members. A work by Baltussen and others Not all causes of visual impairment can be addressed using (2005) for trachoma-endemic areas in the world--similar to public health types of interventions. Cataract, trachoma, child- studies of cataract control surgery--reveals that interventions hood blindness, and onchocerciasis are discussed below. using antibiotics cost between US$4,000 and US$220,000 per DALY averted for all regions studied. Targeted treatment with Cataract. Surgery to remove the opacified lens is the only antibiotics (be it on the basis of azithromycin or tetracycline) is effective treatment for cataracts. Neither diet nor medications not cost-effective, and mass treatment of all children (not have been shown to stop cataract formation. There are several entire communities) is cost-effective only when azithromycin is possible approaches for the surgical extraction of cataracts. donated. In Myanmar, tetracycline has been shown to be mod- Intracapsular cataract extraction using aphakic glasses is a erately cost-effective (Evans and others 1996). Cost-effective- technique by which the whole lens is removed from the eye. ness studies are not available on mass treatment of entire com- After surgery, special eyeglasses are provided to patients to munities, the approach now most commonly in use with restore sight. A disadvantage of this intervention is the non- donated azithromycin. Recent studies by Solomon and others compliance of people who need to wear glasses, which has been (2004) report a 70 percent fall in prevalence in an area in found to be between 18 and 73 percent. Although this behavior Tanzania; moreover, the total community burden of ocular may be characterized as noncompliance, it must be said that Chlamydia trachomatis infection (measured by polymerase some programs do not provide glasses or provide aphakic chain reaction) fell to 8.7 percent of pretreatment levels at six glasses of inferior quality. Also, replacing needed aphakic months after treatment. Additionally, Chidambaram and glasses is impossible for some patients. It is also true that apha- others (2004) have demonstrated that, after mass azithromycin kic glasses cause tremendous distortions in vision, thus impair- treatment of a population in Ethiopia, an indirect protective ing compliance. effect occurred among untreated children who resided in vil- In extracapsular cataract extraction with implantation of a lages in which most individuals had been treated. As noted in posterior chamber intraocular lens, the lens and the front por- table 50.1, greater efficacy of azithromycin than that used to tion of the capsule are removed and then replaced with an arti- calculate cost-effectiveness of mass treatment of children alone ficial lens. Baltussen, Sylla, and Mariotti (2004) have evaluated may lead to better cost-effectiveness than shown in the table. To work on the cost-effectiveness of cataract surgery. That work date, if governments purchased the drug, mass distribution (done by WHO regions rather than by World Bank regions) would be excessively expensive from a societal perspective. showed that both intracapsular and extracapsular surgeries are However, from the perspective of the governments of countries cost-effective ways to reduce the impact of cataract blindness. in which azithromycin (donated by Pfizer Inc. through the However, extracapsular surgery is both less costly and more International Trachoma Initiative) is being distributed, mass effective than intracapsular surgery and can therefore be con- distribution appears to be relatively cost-effective. sidered the best choice for cataract control. Its cost-effective- Trichiasis scarring is amenable to surgical repair. To date, the ness ratios are below US$200 per DALY averted in low- and cost-effectiveness analyses that have been done suggest that middle-income countries and below US$2,400 in high-income surgery is not particularly expensive per case of blindness pre- countries. vented, assuming that the eyelid correction prevents blindness Trachoma. WHO recommends an integrative approach to tra- and that the individuals with operated trichiasis are not more choma control through its SAFE strategy (surgery, antibiotics likely to be affected by other conditions (for example, dry eye) to control the infection, facial cleanliness, and environmental that might lead to corneal opacification. Baltussen and others improvements). The facial cleanliness and environmental (2005) suggest that trichiasis surgery--with cost-effectiveness improvements are preventive public health measures aimed at ranging between approximately US$4 and US$82 per DALY reducing the incidence of infection. Antibiotic treatment, espe- averted across trachoma-endemic areas--would be even more cially when given on a mass or community basis, is both pri- cost-effective than cataract surgery. mary prevention (reducing transmission in the community) From these cost-effectiveness evaluations, one could con- and secondary prevention (treating active infection to avoid clude that it is best simply to correct lid damage attributable to 956 | Disease Control Priorities in Developing Countries | Joseph Cook, Kevin D. Frick, Rob Baltussen, and others trachoma. Surgery (tertiary prevention) would then remain a tribution. A coordination group of nongovernmental organi- low but continuing cost. These evaluations do not, of course, zations is working closely with all three onchocerciasis control take into account the possibility of eliminating this blinding programs and with national counterparts in virtually all disease. The implementation of the full SAFE strategy includes endemic countries. If present efforts in endemic countries are primary, secondary, and tertiary prevention, and although successfully completed, the disease will be brought under con- more costly at the outset, it could eliminate infection, pain, and trol by 2010. blindness (and the need and cost of lid surgery) into the future. The WHO Alliance for the Global Elimination of Trachoma (GET 2020) was established in 1997 to support the work of a NATURE, CAUSES, AND EPIDEMIOLOGY broad spectrum of collaborating international organizations, OF HEARING LOSS nongovernmental development organizations, and founda- In this chapter, the term hearing loss, used by itself, denotes any tions in implementing the SAFE strategy. Kumaresan and or all levels of severity of hearing difficulty. These levels of Mecaskey (2003) report that 10 countries have initiated tra- hearing impairment comprise mild (26­40 decibel hearing choma elimination programs using donated azithromycin, and level, dBHL), moderate (41­60 dBHL), severe (61­80 dBHL), many more programs are expected. They make the point that and profound (81 dBHL or greater). The term deafness denotes the promise of elimination provides the justification for invest- profound hearing impairment (WHO 1991, 1997). Disabling ing in trachoma control. hearing impairment in adults is defined as "a permanent unaided hearing threshold level for the better ear of 41 dB or Childhood Blindness. In 1993, WHO estimated that as many greater; for this purpose, the hearing threshold level is to be as 13.8 million children have some degree of eye damage taken as the better ear average hearing threshold level for the because of vitamin A deficiency; however, the number of chil- four frequencies 0.5, 1, 2, and 4 kHz." Disabling hearing dren with actual blindness is much lower--less than 500,000 in impairment in children under the age of 15 years is defined as 1992. Recent WHO studies (Resnikoff and others 2004) include a permanent, unaided hearing threshold level for the better ear vitamin A deficiency among causes of childhood blindness. of 31 dB or greater; for this purpose, the hearing threshold level Cost-effectiveness studies of vitamin A supplementation, dis- is to be taken as the better ear average hearing threshold level cussed in chapter 56, focus only on deaths averted unrelated to for the four frequencies 0.5, 1, 2, and 4 kHz. blindness, but this public health intervention appears to be Mathers and others (2003) estimate that in 2002, 255 million cost-effective. people worldwide had disabling hearing loss (moderate or worse hearing loss in the better ear). Those 192 million people Onchocerciasis. Onchocerciasis, or "river blindness," is with adult-onset loss (age 20 years and above) and 63 million endemic in 28 countries in tropical Africa, where 99 percent of people with childhood-onset loss make up almost 4.1 percent of infected people live. Isolated foci of infection also occur in the world's population and just over 40 percent of all people Latin America (six countries) and Yemen. Although it accounts globally with hearing loss of any severity. The prevalence rates for only 0.8 percent of world blindness (Resnikoff and others of adult-onset hearing loss were estimated by subtracting the 2004), the distribution of ivermectin, given at no cost by prevalence rate for childhood onset (estimated in terms of Merck, has so far proved successful in drastically reducing this prevalence in ages around 15 to 19). Numbers with childhood- cause of blindness. Additionally, patients suffer severe skin onset hearing loss by cause have so far not been estimated sepa- lesions and pruritus, also remedied by the annual dosing with rately but are included among sequelae of other diseases (for ivermectin. Studies have shown that the cost per DALY averted example, infectious diseases such as meningitis, otitis media, is as little as US$40 when adjusted for inflation (Waters, congenital conditions). It has been estimated that at least Rehwinkel, and Burnham 2004). 50 percent of the burden of hearing loss could be prevented by During the past 25 years considerable progress has been primary, secondary, and tertiary preventive measures (Brobby made by the Onchocerciasis Control Program in West Africa, 1989; WHO 1991). both through control of the black-fly vector (insecticide spraying) and through the distribution of ivermectin. This success, expressed in health, economic, and development Causes and Characteristics terms, was the motivating rationale for the launching in Hearing loss is grouped according to International Classi- December 1995 of a new program, the African Program for fication of Diseases and Related Health Problems, 10th revision, Onchocerciasis Control. The objective is to create, by 2007, version for 2003 (ICD-10) into conductive and sensorineural sustainable community-directed distribution systems using loss and other hearing loss, ICD-10 codes 90­91 (WHO 2003). ivermectin. In Latin America, the Onchocerciasis Elimination The main causes are shown in table 50.6 according to the pro- Program in the Americas is successfully using ivermectin dis- portion that these contribute to the total burden (WHO 1986). Loss of Vision and Hearing | 957 Table 50.6 Main Causes of Hearing Loss, by Proportion of the effect of noise (Mizoue, Miyamoto, and Shimizu 2003). Total Burden Other risk factors include poverty, poor access to health care, poor hygiene, and overcrowding, all of which can lead to upper High proportion Moderate proportion Low proportion respiratory tract infections, otitis media, and other infections Genetic causes Excessive noise Nutritionally related that may cause hearing loss, such as measles and meningitis. Otitis media Ototoxic drugs and chemicals Trauma related Detailed risk factors and indicators have been developed for Presbycusis Prenatal and perinatal problems Menière's disease neonates and infants (Joint Committee on Infant Hearing 2000); Infectious causes Tumors these include conditions that should require admission to a Wax and foreign bodies Cerebrovascular disease neonatal intensive care unit, stigmata of syndromes causing hearing loss, positive family history, craniofacial anomalies, cer- Source: WHO 1986. tain in utero and post-natal infections (cytomegalovirus,herpes, rubella, syphilis, toxoplasmosis, meningitis), hyperbilirubine- Chronic otitis media (COM, as in ICD-10 codes H65­H67) mia, conditions requiring prolonged mechanical ventilation or includes chronic suppurative otitis media and otitis media with oxygenation, persistent otitis media with effusion, and others. effusion. These forms of otitis media, together with some other Ototoxic medications, low birth-weight, and low Apgar middle ear diseases, such as perforation of the tympanic mem- scores have also been cited as risk factors for neonates (Vohr and brane, cholesteatoma, and otosclerosis, are the major causes of others 2000). Offspring of consanguineous marriages have a conductive hearing loss. In most WHO estimates of the burden significantly higher incidence of autosomal recessive diseases, of otitis media, the data are not disaggregated into acute and including hearing impairment. Such diseases are an important chronic otitis media. cause in communities where consanguinity is common (Shahin Hearing loss is a chronic and often lifelong disability that, and others 2002; Zakzouk 2000). Certain ethnic groups (First depending on the severity and frequencies affected, can cause Nations peoples such as Inuit and North American Indians, as profound damage to the development of speech, language, and well as Australian Aboriginal people) appear to be at higher risk cognitive skills in children, especially if commencing prelin- of developing COM (WHO 1998). gually. That damage, in turn, affects the child's progress in school and, later, his or her ability to obtain, keep, and perform an occu- Age, Geographic, and Gender Burdens pation. For all ages and for both sexes, it causes difficulties with interpersonal communication and leads to significant individual The prevalence of disabling hearing impairment that increases social problems, especially isolation and stigmatization.All these markedly with age is mainly related to the effect of presbycusis. difficulties are much magnified in developing countries, where The current shortage of data, particularly in developing coun- there are generally limited services, few trained staff members, tries, prevents accurate assessment of the global distribution of and little awareness about how to deal with these difficulties. the burden and causes. In addition to its individual effects, hearing loss substan- Male-to-female ratios of age-standardized adult-onset tially affects social and economic development in communities prevalence rates were found to be greater than 1 in most stud- and countries. Ruben (2000), taking into account rehabilita- ies in all WHO regions (Mathers, Smith, and Concha 2005). tion, special education, and loss of employment, estimated the This finding may be related to occupational noise-induced cost to the U.S. economy in 1999 of communication disorders hearing loss, which differentially affects men. (hearing, voice, speech, and language disorders) at between US$176 billion and US$212 billion (2004 dollars; 2.5­3 percent Mortality of the gross national product of the United States in that year). Barnett and Franks (1999) have found evidence that adults Hearing loss accounted for about one-third of the prevalence with postlingual onset of deafness have higher mortality than of these communication disorders. nondeaf adults. A 10-year longitudinal analysis of participants (age 55 to 74 years) in the U.S. National Health and Nutrition Risk Factors Examination Survey I found that, at baseline, hearing loss pre- Occupations exposed to high levels of noise or ototoxic chemi- dicts mortality; relative risk 1.17 (Mui and others 1998). cals are also at risk, and noise exposure potentiates chemical oto- Other studies have reported that the association disappears toxicity in some cases (Fechter 1995; Morata 1998). Certain after controlling for age, and in any case, any relationship that lifestyles (for example, use of personal stereos, noisy toys, fire- may exist is too small to appear in published WHO estimates of crackers) and hobbies (for example, hunting) are also linked to deaths by cause (WHO 2004a, annex table 2) and by years of levels of noise exposure that can cause hearing loss (Berglund life lost, or YLLs (Mathers, Smith, and Concha 2005), in any and others 2000; Goelzer, Hansen, and Sehrndt 2001). Smoking region. A small number of deaths (4,000 globally in 2002) are may be a risk factor for high-frequency hearing loss, adding to recorded for otitis media (WHO 2004a), but these deaths are 958 | Disease Control Priorities in Developing Countries | Joseph Cook, Kevin D. Frick, Rob Baltussen, and others mainly due to infective complications and, hence, are not starts later in life and for which later surveillance is needed directly caused by hearing loss. (Grote 2000). No publications were found that have addressed the DALY burden that might be avoided by implementing Years Lived with Disability and DALYs neonatal hearing screening. Data on years of life lived with disability (YLDs) and DALYs are A recent WHO meeting of experts on noise-induced hearing available only for adult-onset hearing loss. The disease model loss (WHO 1998) concluded that exposure to excessive noise is used, the assumptions and methods used for calculation, and the major avoidable cause of permanent hearing impairment the disability weights are described elsewhere (Mathers, Smith, worldwide. They agreed that, in developing countries, occupa- and Concha 2005). Total global YLDs for adult-onset hearing tional noise and urban environmental noise are increasing risk loss in 2001 are estimated to be 25.87 million, or 4.7 percent of factors for hearing impairment. Experts attending the meeting total YLDs attributable to all causes, which makes hearing loss recommended that all countries implement national programs a leading cause of YLDs. Because YLLs are taken to be zero for for prevention of noise-induced hearing loss, including effec- all regions, the DALY figures are identical to the YLD figures. tive hearing conservation. However, there are no published The most comprehensive data available are for all adult-onset reports yet on the effectiveness of such programs in developing hearing loss (WHO 2004b; Mathers, Smith, and Concha 2003). countries. The United States has produced a guide to hearing Fewer data on the burden are available at present for childhood conservation programs in the workplace (Franks, Stephenson, hearing loss and specific causes. and Merry 1996). It advises how to appraise programs by assess- ing the completeness of their components and by evaluating Interventions both the individual audiometric data for threshold shift and the Effective interventions include screening programs, education, group data for variability compared with a nonexposed popu- surgery, medications, and assistive devices. lation. Even in developed countries, there have been few, if any, clinical trials and little convincing evidence of the efficacy of Population-Based Interventions. Neonatal or early infant occupational hearing conservation programs (Dobie 1995). hearing screening is important because early identification of hearing loss (before 6 months of age, with early intervention) Personal Services. Chronic suppurative otitis media is one of is associated with significantly better language development the most common causes of hearing impairment in developing and may lead to better school and occupational performance countries. Opportunities for prevention arise at all levels of than that of children identified after 6 months with early inter- national health systems, particularly in the community and at vention (Keren and others 2002; Yoshinaga-Itano and others the primary level through primary ear and hearing care (PEHC) 1998). Implementation of neonatal hearing screening raises (WHO 1998). Appropriate health promotion measures include from 20 to 80 percent the numbers of children with normal breastfeeding, immunization, adequate nutrition, personal development of language, compared with children whose hear- hygiene, improved housing, reduced overcrowding, and ade- ing loss is detected later (Yoshinaga-Itano and Gravel 2001). quate access to clean water. Primary health care workers can be Early identification of hearing impairment can reduce the given appropriate training and basic equipment for early detec- median age of identification of hearing impairment from tion and management of chronic suppurative otitis media, but between 12 and 18 months to 6 months or less. Universal the effectiveness and cost-effectiveness of this intervention in neonatal hearing screening is highly sensitive, but depending developing countries has not yet been assessed. on the test method used, it may result in many false positives Although WHO does not currently recommend treating (which may increase parental anxiety and lead to unnecessary what is commonly called chronic middle ear infection with follow-up tests and interventions). It has a low positive predic- antibiotics at the primary level (WHO 2000), evidence suggests tive value. Some screening protocols may decrease false- that antibiotics, especially topical quinolones, are more effec- positive rates (Kennedy and others 2000). Universal neonatal tive and cost-effective than ear toilet alone (Acuin, Smith, and hearing screening has been endorsed in developed countries Mackenzie 2000). WHO is reviewing these recommendations (Joint Committee on Infant Hearing 2000), although some (WHO 2004b). New methods of delivery of effective but experts urge caution (Paradise 1999); however, it is expensive to expensive topical antibiotics may lower the cost in poor com- implement and, for most developing countries, is not yet an munities, but treatment failure may be due to a high reinfec- option. Hearing screening targeted at high-risk neonates is tion rate attributable to poor hygienic conditions. To be generally used in developing countries that do any type of effective as public health measures, interventions need to be neonatal screening, but screening may fail to detect 50 percent implemented on a large scale, with good coverage of the targeted or more of cases of impairment (Lutman and Grandori 1999). population (van Hasselt and van Kregten 2002). Ear surgery Neonatal screening programs will not detect the 10 to 20 per- plays an essential part in the prevention of further hearing cent of cases of permanent childhood hearing impairment that impairment and, sometimes, in the improvement of hearing. Loss of Vision and Hearing | 959 Services at the secondary level of intervention include pro- Intervention Cost and Cost-Effectiveness. All the data on the vision of hearing aids in developing countries, which should costs and cost-effectiveness of interventions related to hearing assign priority to children with moderate or severe hearing loss (including school-age screening, treatment of COM, surgi- loss, followed by adults (Arslan and Genovese 1996; WHO cal interventions, hearing aids, and cochlear implants) come 2004c). However, even though globally about 6 million hearing from developed countries. Although they can be summarized aids are dispensed annually (WHO 1999), there have been no quite readily, it is not clear whether and how they relate to the published randomized, controlled trials of the effectiveness of costs that would be experienced in developing countries. hearing aids in reducing hearing disability in developing coun- tries and few trials in developed countries. RESEARCH AND DEVELOPMENT AGENDA A randomized trial of amplification in 194 U.S. veterans showed significant improvements in communication, cogni- The public health research and development agenda for con- tion, and social and emotional function, plus significant allevia- trolling and reducing the burden of disease related to the loss tion of depression, with hearing aids compared with controls of sight and hearing should include the following: (Mulrow and others 1990). No significant differences were observed in clinical effectiveness and cost-effectiveness between · further population-based studies on the magnitude, causes, newer hearing aids that use digital signal processing and those and distribution of the burden that do not--in particular, analog-based aids (Parving 2003; · economic analysis,especially on cost-effectiveness (for exam- Taylor, Paisley, and Davis 2001). Digital signal processing aids ple,cost-effectiveness of each of the components of the SAFE are not affordable for most people in developing countries. strategy in trachoma control) Over-the-counter hearing aids that can be purchased and used · research to develop eye and hearing care systems without prior training are commonly available in some devel- · operational research on eye and hearing care delivery (par- oping countries. Those aids were found not to meet the pre- ticularly for cataract, diabetic retinopathy, and affordable scription gain requirements of the majority of elderly clients hearing aids in underserved areas) who usually purchased them (Cheng and McPherson 2000). · clinical and field trials on interventions: pneumococcal and Learning to use a hearing aid and developing "hearing tac- meningitis vaccines, treatment for chronic suppurative otitis tics"are also important. Random assignment to a course for new media, primary care of ears and hearing, and prevention of hearing aid users significantly reduced the handicap compared noise damage. with controls not assigned (Beynon,Thornton,and Poole 1997). Lack of compliance in use is a substantial problem everywhere Basic scientific research, particularly for age-related macular among elderly and child users, including in developing coun- degeneration, must move forward, as it is doing in the indus- tries (Furuta and Yoshino 1998; Sorri, Luotonen, and Laitakari trial countries, where this disease constitutes a major burden 1984). Thus, measuring coverage without taking into account and where highly developed research establishments exist. actual usage is not enough to assess alleviation of the burden. Cochlear implants are provided to children and adults with CONCLUSIONS: PROMISES AND PITFALLS severe and profound bilateral deafness on the basis that known short-term outcomes in auditory receptive skills (Richter and With what we now know about some of the cost-effective inter- others 2002) will translate through various medium-term out- ventions cited above, we could make significant reductions in comes into greater social independence and quality of life (the the burden of disease related to loss of vision. Although wait- social and quality outcomes have not yet been tested in a trial ing for someone to have a condition and then remedying the or observational study) (Summerfield and Marshall 1999). situation is not a particularly common "public health" recom- Cochlear implantation is beneficial in prelingually and postlin- mendation, given the costs of and knowledge of prevention gually deaf children (Makhdoum, Snik, and van den Broek at this point, we can strongly recommend surgery both for 1997) and, when accompanied by aural (re)habilitation, leads cataract (the primary option) and for trachoma (apparently a to higher rates of mainstream placement in schools and lower better use of resources than mass treatment with antibiotics-- dependence on special education support services (Francis and even if not acceptable on a humanitarian basis). For example, others 1999). Multichannel implants are superior to single- clearing the backlog of cataract surgery globally could reduce channel implants (Cohen, Waltzman, and Fisher 1993) and are the DALYs associated with vision loss by more than half. more beneficial when implanted in young children (Richter Hearing loss interventions have only begun to demonstrate and others 2002). There has been no economic analysis of their potential effectiveness in developing countries, and no cochlear implants in developing countries, and such interven- cost work has been done in these settings. Furthermore, tions are currently not a priority in most parts of the develop- although the means to reduce the burden of adult-onset ing world (Berruecos 2000; WHO 2004c; Zeng 1995). hearing loss are not as straightforward nor as easily applied, 960 | Disease Control Priorities in Developing Countries | Joseph Cook, Kevin D. Frick, Rob Baltussen, and others eliminating adult hearing loss would avoid slightly more YLDs Dobie, R. A. 1995."Prevention of Noise-Induced Hearing Loss." Archives of than eliminating the cataract surgery backlog. The data suggest Otolaryngology--Head and Neck Surgery 121 (4): 385­91. that these interventions (particularly cataract surgery) are rela- Evans, T. G., M. K. Ranson, T. A. Kyaw, and C. K. Ko. 1996. "Cost Effectiveness and Cost Utility of Preventing Trachomatous Visual tively cost-effective, but a lack of political will, a failure to rec- Impairment: Lessons from 30 Years of Trachoma Control in Burma. ognize that steps can be taken now, insufficient capacity within British Journal of Ophthalmology 80 (10): 880­89. ministries of health to carry out the known beneficial interven- Fechter, L. D. 1995. "Combined Effects of Noise and Chemicals." tions, and, finally, a lack of equipment or funding for the pro- Occupational Medicine 10 (3): 609­21. grams still remain barriers to alleviating disabilities related to Fotouhi, A., H. Hashemi, K. Mohammad, and K. H. Jalali. 2004. "The Prevalence and Causes of Visual Impairment in Tehran: The Tehran vision and hearing loss. Eye Study." British Journal of Ophthalmology 88 (6): 740­45. Francis, H. W., M. E. Koch, J. R. Wyatt, and J. K. Niparko. 1999. "Trends in REFERENCES Educational Placement and Cost-Benefit Considerations in Children with Cochlear Implants." Archives of Otolaryngology--Head and Neck Surgery 125 (5): 499­505. Abou-Gareeb, I., S. Lewallen, K. Bassett, and P. Courtright. 2001. "Gender and Blindness: A Meta-Analysis of Population-Based Prevalence Franks, J. R., M. R. Stephenson, and C. J. Merry, eds. 1996. Preventing Surveys." Ophthalmic Epidemiology 8 (1): 39­56. Occupational Hearing Loss. NIOSH Publication 96-110. Cincinnati, OH: U.S. Department of Health and Human Services, National Acuin, J., A. Smith, and I. Mackenzie. 1998. "Interventions for Chronic Institute for Occupational Safety and Health. http://www.cdc.gov/ Suppurative Otitis Media." Cochrane Database of Systematic niosh/96-110.html. Reviews (2): CD000473.http://www.cochrane.org/cochrane/revabstr/ AB000473.htm. Frick, K. D., E. L. Keuffel, and R. J. Bowman. 2001. "Epidemiological, Demographic, and Economic Analyses: Measurement of the Value of Arslan E., and E. Genovese. 1996. "Hearing Aid Systems in Undeveloped, Trichiasis Surgery in The Gambia."Ophthalmic Epidemiology 8 (2­3): Developed, and Industrialized Countries. Scandinavian Audiology 42: 191­201. (Suppl.) 35­39. Frick, K. D., T. M. Lietman, S. O. Holm, H. C. Jha, J. S. Chaudhary, and Baltussen, R., M. Sylla, K. Frick, and S. Mariotti. 2005. "Cost-Effectiveness R. C. Bhatta. 2001. "Cost-Effectiveness of Trachoma Control Measures: of Trachoma Control in Seven World Regions." Ophthalmic Comparing Targeted Household Treatment and Mass Treatment of Epidemiology 12 (2): 91­101. Children." Bulletin of the World Health Organization 79 (3): 201­7. Baltussen, R., M. Sylla, and S. Mariotti. 2004. "Cost-Effectiveness of Furuta, H., and T. Yoshino. 1998. "The Present Situation of the Use of Cataract Surgery: A Global and Regional Analysis." Bulletin of the Hearing Aids in Rural Areas of Sri Lanka: Problems and Future Pros- World Health Organization 82 (5): 338­45. pects." International Journal of Rehabilitation Research 21 (1): 103­7. Barnett, S., and P. Franks. 1999. "Deafness and Mortality: Analyses of Gilbert, C., and A. Foster. 2001. "Childhood Blindness in the Context of Linked Data from the National Health Interview Survey and National VISION 2020--The Right to Sight." Bulletin of the World Health Death Index." Public Health Reports 114 (4): 330­36. Organization 79 (3): 227­32. Berglund, B., T. Lindvall, D. Schwela, and K.-T. Goh. 2000. Guidelines for Goelzer, B., C. H. Hansen, and G. A. Sehrndt, eds. 2001. Occupational Community Noise. Geneva: World Health Organization. http://www. Exposure to Noise: Evaluation, Prevention, and Control. Special Report who.int/docstore/peh/noise/guidelines2.html. S 64. Dortmund and Berlin: Federal Institute for Occupational Safety Berruecos, P. 2000. "Cochlear Implants: An International Perspective-- and Health. Latin American Countries and Spain." Audiology 39 (4): 221­25. Grote J. 2000. "Neonatal Screening for Hearing Impairment." Lancet Beynon, G. J., F. L. Thornton, and C. Poole. 1997. "A Randomized, 355 (9203): 513­14. Controlled Trial of the Efficacy of a Communication Course for First Joint Committee on Infant Hearing. 2000. "Year 2000 Position Statement: Time Hearing Aid Users." British Journal of Audiology 31 (5): 345­51. Principles and Guidelines for Early Hearing Detection and Bowman, R. J., A. Sillah, C. Van Dehn, V. M. Goode, M. Muquit, G. J. Intervention Programs." Pediatrics 106 (4) (October): 798­817. Johnson, and others. 2000. "Operational Comparison of Single-Dose http://www.jcih.org/jcih2000.pdf. Azithromycin and Topical Tetracycline for Trachoma." Investigative Kennedy, C., L. Kimm, R. Thornton, and A. Davis. 2000. "False Positives Ophthalmology and Visual Science 41 (13): 4074­79. in Universal Neonatal Screening for Permanent Childhood Hearing Brobby, G. W. 1989. "Personal View . . . Strategy for Prevention of Deafness Impairment." Lancet 356 (9245): 1903­4. in the Third World." Tropical Doctor 19 (4): 152­54. Keren, R., M. Helfand, C. Homer, H. McPhillips, and T. A. Lieu. 2002. Cheng, C. M., and B. McPherson. 2000. "Over-the-Counter Hearing Aids: "Projected Cost-Effectiveness of Statewide Universal Newborn Electroacoustic Characteristics and Possible Target Client Groups." Hearing Screening." Pediatrics 110 (5): 855­64. Audiology 39 (2): 110­16. Kumaresan, J. A., and J. W. Mecaskey. 2003. "The Global Elimination of Chidambaram, J. D., M. Melese, W. Alemayehu, E. Yi, T. Prabriputaloong, Blinding Trachoma: Progress and Promise." American Journal of D. C. Lee, and others. 2004. "Mass Antibiotic Treatment and Tropical Medicine and Hygiene 69 (Suppl. 5): 24­28. Community Protection in Trachoma Control Programs." Clinical Lee, D. J., O. Gomez-Marin, B. L. Lam, and D. D. Zheng. 2002. "Visual Infectious Diseases 39 (9): 95­97. Acuity Impairment and Mortality in U.S. Adults." Archives of Cohen, N. L., S. B. Waltzman, and S. G. Fisher. 1993. "A Prospective, Ophthalmology 120 (11): 1544­50. Randomized Study of Cochlear Implants." Department of Veterans Lutman, M. E., and F. Grandori. 1999. "Screening for Neonatal Hearing Affairs Cochlear Implant Study Group. New England Journal of Defects: European Consensus Statement." European Journal of Medicine 328 (4): 233­37. Pediatrics 158 (2): 95­96. Dandona, R., and L. Dandona. 2003. "Childhood Blindness in India: A Makhdoum, M. J., A. F. Snik, and P. van den Broek. 1997. "Cochlear Population Based Perspective." British Journal of Ophthalmology 87 (3): Implantation: A Review of the Literature and the Nijmegen Results." 263­65. Journal of Laryngology and Otology 111 (11): 1008­17. Loss of Vision and Hearing | 961 Mathers, C. D., K. Bernard, K. M. lburg, M. Inoue, D. Ma Fat, K, Shibuya, Taylor, R. S., S. Paisley, and A. Davis. 2001. "Systematic Review of the and others. 2003. "Global Burden of Disease in 2002: Data Sources, Clinical and Cost Effectiveness of Digital Hearing Aids." British Journal Methods and Results." GPP Discussion Paper No. 54. Geneva: World of Audiology 35 (5): 271­88. Health Organization. van Hasselt, P., and E. van Kregten. 2002. "Treatment of Chronic Mathers, C., A. Smith, and M. Concha. 2005. "Global Burden of Adult- Suppurative Otitis Media with Ofloxacin in Hydroxypropyl Onset Hearing Loss in the Year 2002." Paper in preparation, World Methylcellulose Ear Drops: A Clinical/Bacteriological Study in a Rural Health Organization, Geneva. http://www3.who.int/whosis/burden/ Area of Malawi." International Journal of Pediatric Otorhinolaryngoly gbd2000docs/Hearing%20loss.zip. 63 (1): 49­56. Mizoue, T., T. Miyamoto, and T. Shimizu. 2003. "Combined Effect of Smo- Vohr, B. R., J. E. Widen, B. Cone-Wesson, Y. S. Sininger, M. P. Gorga, R. C. king and Occupational Exposure to Noise on Hearing Loss in Steel Fac- Folsom, and S. J. Norton. 2000. "Identification of Neonatal Hearing tory Workers."Occupational and Environmental Medicine 60 (1): 56­59. Impairment: Characteristics of Infants in the Neonatal Intensive Care Morata, T. C. 1998. "Assessing Occupational Hearing Loss: Beyond Noise Unit and Well-Baby Nursery." Ear and Hearing 21 (5): 373­82. Exposures." Scandinavian Audiology 48 (Suppl.): 111­16. Waters, H. R., J. A. Rehwinkel, and G. Burnham. 2004. "Economic Mui, S., D. Reuben, M. Damesyn, G. Greendale, and A. Moore. 1998. Evaluation of Mectizan Distribution." Tropical Medicine and Inter- "Sensory Impairment as a Predictor of 10-Year Mortality and national Health 9 (4): 16­25. Functional Impairment." Journal of the American Geriatric Society 46 WHO (World Health Organization). 1986. Prevention of Deafness and (9): 19­20. Hearing Impairment: Report by the Director General. EB79/10. Geneva: Mulrow, C. D., C. Aguilar, J. E. Endicott, M. R. Tuley, R. Velez, W. S. Charlip, WHO. and others. 1990. "Quality-of-Life Changes and Hearing Impairment: ------. 1991. Report of the Informal Working Group on Prevention of A Randomized Trial." Annals of Internal Medicine 113 (3): 188­94. Deafness and Hearing Impairment Program Planning. WHO/ Naidoo, K. S., A. Raghunandan, K. P. Mashige, P. Govender, B. A. Holden, PDH/91.1. Geneva: WHO. G. P. Pokharel, and L. B. Ellwein. 2004. "Refractive Error and Visual ------. 1992. International Statistical Classification of Diseases and Related Impairment in African Children in South Africa." Investigative Health Problems. Rev. Geneva: WHO. Ophthalmology and Visual Sciences 44 (9): 3764­70. ------. 1993. International Classification of Disease, Injuries and Causes of Nirmalan, P. K., A. Padmavathi, and R. D. Thulasiraj. 2003. "Sex Death. 10th ed., rev. Geneva: WHO. Inequalities in Cataract Blindness Burden and Surgical Services in ------. 1997. Report of the First Informal Consultation on Future Program South India." British Journal of Ophthalmology 87 (7): 847­49. Developments for the Prevention of Deafness and Hearing Impairment. Paradise, J. L. 1999. "Universal Newborn Hearing Screening: Should We WHO/PDH/97.3. Geneva: WHO. Leap before We Look?" Pediatrics 103 (3): 670­72. ------. 1998. Prevention of Hearing Impairment from Chronic Otitis Parving, A. 2003. "The Hearing Aid Revolution: Fact or Fiction?" Acta Media. Report of a WHO/CIBA Foundation workshop, London, Otolaryngologica 123 (2): 245­48. November 12­21, 1996. WHO/PDH/98.4. Geneva: WHO. Pascolini, D., S. P. Mariotti, G. P. Pokharel, R. Pararajasegaram, D. Eya'ale, ------. 1999. Hearing Aids Services--Needs and Technology Assessment for A.-D. Négrel, and S. Resnikoff. 2004. "Available Data on Visual Impair- Developing Countries. Report of a WHO/CBM workshop, Bensheim, ment: 2002 Global Update." Ophthalmic Epidemiology 11 (2): 67­115. Germany, November 24­26, 1998. WHO/PDH/99.7. Geneva: WHO. Reacher, M. H., B. Munoz, A. Alghassany, A. S. Daar, M. Elbualy, and H. R. ------. 2000. Integrated Management of Childhood Illness: Handbook. Taylor. 1992."A Controlled Trial of Surgery for Trachomatous Trichiasis WHO/FCH/CAH/00.12. Geneva: WHO. http://www.who.int/child- of the Upper Lid." Archives of Ophthalmology 110 (5): 667­74. adolescent-health/publications/IMCI/WHO_FCH_CAH_00.12.htm. Resnikoff, S., D. Pascolini, D. Etya'Alé, I. Kocur, R. Pararajasegaram, G. P. ------. 2003. International Statistical Classification of Diseases and Related Pokharel, and S. P. Mariotti. 2004. "Global Data on Visual Impairment Health Problems, 10th rev. Geneva: WHO. http://www3.who. int/icd/ in the Year 2002." Bulletin of the World Health Organization 82: 844­51. vol1htm2003/fr-icd.htm. Richter, B., S. Eissele, R. Laszig, and E. Lohle. 2002. "Receptive and ------. 2004a. Changing History: The World Health Report. Geneva: Expressive Language Skills of 106 Children with a Minimum of 2 Years' WHO. Experience in Hearing with a Cochlear Implant." International Journal ------. 2004b. Chronic Suppurative Otitis Media: Burden of Illness and of Pediatric Otorhinolaryngoly 64 (2): 111­25. Management Options. Geneva: WHO. Ruben, R. J. 2000. "Redefining the Survival of the Fittest: Communication ------. 2004c. Guidelines for Hearing Aids and Services for Developing Disorders in the 21st Century." Laryngoscope 110 (2, part 1): 241­45. Countries. 2nd ed. Geneva: WHO. Shahin, H., T. Walsh, T. Sobe, E. Lynch, M. C. King, K. B. Avraham, and M. Yoshinaga-Itano, C., and J. S. Gravel. 2001. "The Evidence for Universal Kanaan. 2002. "Genetics of Congenital Deafness in the Palestinian Newborn Hearing Screening." American Journal of Audiology 10 (2): Population: Multiple Connexin 26 Alleles with Shared Origins in the 62­64. Middle East." Human Genetics 110 (3): 284­89. Yoshinaga-Itano, C., A. Sedey, D. K. Coulter, and A. L. Mehl. 1998. Solomon, A. W., M. J. Holland, N. D. Alexander, P. A. Massae, A. Aguirre, "Language of Early- and Later- Identified Children with Hearing Loss." A. Natividad-Sancho, and others. 2004. "Mass Treatment with Single- Pediatrics 102 (5): 1161­71. Dose Azithromycin for Trachoma." New England Journal of Medicine 351 (19): 1962­71. Zakzouk, S. 2002. "Consanguinity and Hearing Impairment in Developing Countries: A Custom to Be Discouraged." Journal of Laryngology and Sorri, M., M. Luotonen, and K. Laitakari. 1984. "Use and Non-Use of Otology 116 (10): 811­16. Hearing Aids." British Journal of Audiology 18 (3): 169­72. Zeng, F. G. 1995. "Cochlear Implants in China." Audiology. 34 (2): 61­75. Summerfield, A. Q., and D. H. Marshall. 1999. "Paediatric Cochlear Implantation and Health-Technology Assessment." International Journal of Pediatric Otorhinolaryngoly 47 (2): 141­51. Taylor, H. R., S. Katala, B. Munoz, and V. Turner. 1991. "Increase in Mortality Associated with Blindness in Rural Africa." Bulletin of the World Health Organization 69 (3): 335­38. 962 | Disease Control Priorities in Developing Countries | Joseph Cook, Kevin D. Frick, Rob Baltussen, and others Chapter 51 Cost-Effectiveness of Interventions for Musculoskeletal Conditions Luke B. Connelly, Anthony Woolf, and Peter Brooks BURDEN OF MUSCULOSKELETAL DISEASE over the age of 60 have OA. Table 51.1 provides an estimate of the contribution of musculoskeletal conditions to the global Musculoskeletal conditions are the most common cause of burden of disease, including a disaggregation by gender and chronic disability around the world. The importance of muscu- between the developed and developing world. The proportions loskeletal conditions as a cause of mortality and morbidity has presented in the second and third panels are the most notewor- been recognized by the designation of 2000­10 as the Bone and thy data in table 51.1. First, the second panel shows that mus- Joint Decade (Hazes and Woolf 2000) by the United Nations, culoskeletal conditions account for approximately 1.7 and World Health Organization (WHO),and more than 60 countries 2.4 percent of the burden of disease experienced by males and around the world.WHO (2003,2004) has highlighted the burden females, respectively, or, across both genders, approximately of musculoskeletal conditions. Estimates of the global burden of 2 percent of the global burden of disease. The disaggregation by these conditions have increased 25 percent over the past decade developing and developed regions, however, shows that while (WHO 2000). Conditions considered under this rubric include musculoskeletal conditions account for around 3.4 percent of osteoarthritis (OA); inflammatory arthritis (rheumatoid arthri- the total burden of disease in the developed world, they account tis and the seronegative spondyloarthropathies); back pain; mus- for 1.7 percent in the developing world. The data also show culoskeletal injuries, including sports injuries; crystal arthritis that, of the set of musculoskeletal conditions, OA accounts for (gout and calcium pyrophosphate disease), and metabolic bone the largest burden, approximately 52 percent of the total in disease, principally osteoporosis (OP). developing regions and 61 percent in developed regions. Back pain is extremely common in both industrial and Table 51.2 provides a further disaggregation of the estimated developing countries, with up to 50 percent of workers suffer- burden of musculoskeletal conditions by developing region ing an episode each year. Back pain causes 0.8 million disability- and mortality stratum. Note that the burden of disease caused adjusted life years (DALYs) each year and is a major cause of by musculoskeletal conditions varies considerably by region: absence from work and of correspondingly high economic in Africa, mortality stratum D, musculoskeletal conditions losses. Nearly 40 percent of back pain is due to occupational account for less than 1 percent of the burden from all causes, risk factors, and many of these factors can be prevented with while in the Western Pacific, mortality stratum B, they account the cooperation of labor, management, industrial engineers, for more than 3 percent of the total burden of disease. Similarly, ergonomists, and health workers. the relative importance of rheumatoid arthritis (RA) and OA OA is increasing among the world's aging populations and is varies considerably by region. In the African regions, where the the sixth leading cause of years lost because of disability glob- prevalence of RA is low, only 12 percent of the burden created ally. It accounts for nearly 3 percent of the total global years lost by musculoskeletal diseases is due to RA; in the Americas, how- to disability, and 10 percent of men and 18 percent of women ever, that proportion is approximately 24 to 27 percent. 963 Table 51.1 Estimated Burden of Musculoskeletal Diseases, by Gender and by Developed or Developing Regions, 2001 Developing Developed regions (both regions (both Total Males Females genders) genders) Numbers of DALYs (thousands) Rheumatoid arthritis 4,757 1,353 3,404 3,238 1,520 Osteoarthritis 16,372 6,621 9,750 11,049 5,323 Other musculoskeletal diseases 8,699 5,033 3,638 6,789 1,880 All musculoskeletal diseases 29,798 13,007 16,792 21,076 8,723 Percentage of total DALYs Rheumatoid arthritis 0.32 0.18 0.49 0.27 0.59 Osteoarthritis 1.12 0.86 1.39 0.91 2.05 Other musculoskeletal diseases 0.59 0.65 0.52 0.56 0.73 All musculoskeletal diseases 2.03 1.69 2.40 1.74 3.37 Percentage of musculoskeletal DALYs Rheumatoid arthritis 15.96 10.40 20.27 15.36 17.42 Osteoarthritis 54.94 50.91 58.07 52.43 61.02 Other musculoskeletal diseases 29.10 38.69 21.66 32.21 21.56 Source: Calculated from WHO (2004). Note: Totals may not sum due to rounding. Table 51.2 Estimated Burden of Musculoskeletal Conditions by Region and Mortality Stratum, Selected WHO Regions, 2001 Eastern Western Africa Americas Southeast Asia Mediterranean Pacific Condition D E B D B D B D B Numbers DALYs (thousands) Rheumatoid arthritis 127 141 532 83 117 855 99 218 1,065 Osteoarthritis 625 687 969 117 931 2,474 227 577 4,442 Other musculoskeletal diseases 285 316 677 107 516 1,756 159 408 2,590 All musculoskeletal diseases 1,037 1,144 2,178 307 1,564 5,085 485 1,203 8,097 Percentage of total DALYs Rheumatoid arthritis 0.09 0.07 0.66 0.47 0.19 0.24 0.43 0.19 0.44 Osteoarthritis 0.42 0.33 1.19 0.67 1.52 0.69 0.99 0.51 1.84 Other musculoskeletal diseases 0.19 0.15 0.83 0.60 0.84 0.49 0.69 0.36 1.06 All musculoskeletal diseases 0.70 0.55 2.68 1.74 2.55 1.42 2.11 1.06 3.34 Percentage of musculoskeletal DALYs Rheumatoid arthritis 12.28 12.29 24.45 27.23 7.50 16.82 20.38 18.11 13.19 Osteoarthritis 60.27 60.10 44.49 38.50 59.51 48.66 46.90 47.97 54.99 Other musculoskeletal diseases 27.44 27.61 31.06 34.27 32.99 34.53 32.71 33.92 31.82 Source: Calculated from WHO (2004). Notes: The letters in the column heads refer to mortality strata. B low child and low adult mortality, D high child and high adult mortality, E high child and very high adult mortality. RA has a prevalence of 0.7 to 0.1 percent worldwide and because of OP. Back pain, OA, trauma, and RA account for results in significant work disability and long-term treatment 32,948,765 DALYs, or 2.15 percent of the global total for 2000. costs. In addition, OP is increasing with the aging of popula- A recent review of the prevalence of rheumatic disorders in tions: one in three people over the age of 50 suffers a fracture Sub-Saharan Africa suggests that the frequency of RA is 964 | Disease Control Priorities in Developing Countries | Luke B. Connelly, Anthony Woolf, and Peter Brooks increasing in East, Central, and South Africa but is rare in West and mucocutaneous lesions. These conditions may follow gas- Africa (McGill and Oyoo 2002). Gout is also prevalent trointestinal or sexually acquired infections and can be associ- throughout the continent, and the HIV epidemic has spawned ated with HIV. Gout and other forms of crystal arthritis tend to a variety of associated spondyloarthropathies among the aging present as an inflammatory response to the presence of uric acid population. Countries such as Thailand are also recognizing an (gout) or various calcium crystals (chondrocalcinosis). increasing burden of disease caused by arthritis and trauma Much of the pain that produces complaints and reduced (Jitapunkul and others 2003). function does not emanate from a frank arthropathy, but from the soft tissues in or around a joint. When these pains are con- fined to a particular area of the body's surface, they can be NATURE, CAUSES, AND EPIDEMIOLOGY referred to as regional pain syndromes and may or may not be OF MUSCULOSKELETAL CONDITIONS related to injury or overuse. If these pains are more widespread and are associated with specific tender points, the condition is Osteoarthritis is the most common condition affecting human known as fibromyalgia. Fibromyalgia is well recognized in the joints and causes significant disability. The principal clinical industrial world and has also been noted in China and Malaysia features are pain, which varies in severity and character, and and among Tamil Indians. stiffness. Disability occurs as a result of pain, weakness, joint The major causes of infectious arthritis can be viral, instability, and reduced range of motion. bacterial, fungal, or helminthic. Each can present as either a The following are the major forms of inflammatory polyarticular presentation or a monarthritis. Many of the arthritis: helminthic infections present with more generalized aches and pains and involvement of muscle tissues as well as joints. All the · rheumatoid arthritis conditions have specific diagnostic features and treatments. · seronegative spondyloarthropathies OP is characterized by low bone mass and deterioration in -- ankylosing spondylitis the microarchitecture of the bone, which leads to fracture -- reactive arthritis after low or moderate trauma. The condition is defined by diag- -- enteropathic arthritis nostic criteria based on bone mineral density as follows: a bone -- psoriatic arthritis mineral density of more than 2.5 standard deviations below · juvenile chronic arthritis the average bone mineral density of young adult women. -- systemic The clinical features of OP are primarily due to its major -- pauciarticular outcome: fracture. The most important fractures occur in the -- polyarticular distal radius, vertebrae, or hip, often following minor trauma. · arthritis associated with systemic connective tissue diseases Vertebral fractures lead to loss of height, kyphosis, and back -- systemic lupus erythematosus pain. The incidence of fracture varies with country and with -- progressive systemic sclerosis type of fracture. Hip fractures are low in African countries but -- vasculitis high and increasingly reported in Australasia, Europe, and -- polydermatomyositis North America. Fracture risk increases with age and is begin- · crystal arthritis ning to have a significant impact on quality of life, mortality, -- gout and health care costs in many countries. -- calcium pyrophosphate deposition disease Rickets is caused by a mineralization defect of newly formed bone in the growing skeleton. This defect leads to an increase in RA has a prevalence of between 1 and 3 percent in most the amount of nonmineralized bone tissue (osteoid) and a countries for which figures are available, but it may be slightly thinning of the growth plates. This condition produces bone less common in tropical countries. The exact etiology of RA is pain, bone deformation, swelling of the joints, and growth unknown, but the evidence suggests an immune reaction, and retardation. Rickets is primarily caused by a lack of exposure to it presents as an inflammation affecting joints and other tissues. sunshine because of climate, pollution, or overuse of clothing Its clinical features can be divided into three groups: constitu- or sunscreens. Rickets is relatively rare in industrial countries, tional, articular, and extra-articular. Constitutional features but it does occur as a consequence of dietary deficiency or involve tiredness, fatigue, weight loss, and fever, and articular excess clothing. features involve principally the synovial joints, producing pain Osteomalacia is the adult equivalent of rickets. It is similarly and eventual deformity and disability. characterized by an increase in osteoid tissue and causes bone The seronegative spondyloarthropathies are primarily pain and fractures. It occurs primarily in the elderly in Europe inflammatory arthropathies and share several common features, and North America because of a lack of exposure to sunshine including familial aggregation, asymmetric joint involvement, that is not compensated for by adequate vitamin D intake. Cost-Effectiveness of Interventions for Musculoskeletal Conditions | 965 Osteomalacia may also occur in countries with abundant sun- PREVENTIVE STRATEGIES shine where clothing prevents sun exposure. Osteomalacia is commonly reported among migrants to Western Europe from Obesity brought about by increases in sedentary lifestyles and India and the Middle East. changes in eating patterns is becoming a major problem world- Back pain accounts for the majority of musculoskeletal disease wide. Weight reduction has been demonstrated to reduce pain presentations to health professionals, and its lifetime prevalence and disability from OA of the knee and other forms of lower exceeds 80 percent in most industrial countries. Spinal disorder limb arthropathy. In OA of the knee, weight reduction will not refers to a wide range of specific and nonspecific musculoskeletal only reduce pain and improve mobility, but it can put off the disorders affecting the spinal column. These conditions include time when surgical replacement of the weight-bearing joint is congenital lesions such as scoliosis, infective problems such as necessary. Obesity can also be associated with back pain, and osteomyelitis and neoplastic disorder (myeloma or secondary weight reduction is an important factor in reducing the recur- cancers), and trauma and referred back pain. rence of episodes of back pain and in reducing long-term dis- The majority of individuals with acute back pain will ability and chronic pain. improve significantly over a six-week period, although in many Smoking and excessive alcohol use are also associated with cases the pain may recur. Early diagnosis and treatment, partic- OP. Adequate calcium intake (1,000 to 1,500 milligrams per ularly of pain, by means of a modified exercise program will day) has been shown to maintain bone density and reduce the reduce long-term morbidity and disability. risk of axial (vertebral) fractures. Smoking also increases the Musculoskeletal injuries are extremely common, whether in risk of developing RA. the workplace or associated with sporting activities or with Weight reduction and diet are also important considera- daily living. Motor vehicle trauma, household accidents, and tions in the management of gout. Appropriate nutrition and occupational accidents occur frequently and are a major cause exercise underpin many of the preventive and treatment strate- of damage to the musculoskeletal system. gies for musculoskeletal disease. Table 51.3 shows the major genetic and environmental risk factors for musculoskeletal diseases. Lack of exercise and obesity are major contributors to soft tissue disorders, OA, and TREATMENTS back pain. Infectious forms of musculoskeletal disease depend on the environment and on the types of organisms that are A range of treatment approaches is available to address the prevalent. multiple aspects of musculoskeletal disorders. Symptomatic Treatments Table 51.3 Risk Factors for Musculoskeletal Disease Symptomatic treatments for musculoskeletal disease principally involve pain reduction. Nonpharmacological treat- Condition Genetic Environmental ments such as massage, heat, and ice, and physiotherapeutic Rheumatoid arthritis HLA DR techniques such as ultrasound may be useful in the short term. Seronegative HLA B27 Pure analgesic agents such as acetaminophen should be tried spondyloarthropathy initially; if no response occurs, compound analgesics or opioid Osteoarthritis Severe osteochondropathies Obesity derivatives, including codeine, may be useful. The side effects of Lack of exercise the latter compounds are significant, particularly in the elderly, Soft tissue Occupation with constipation and disorientation being the most common. Environment Table 51.4 shows the principal modalities of pain relief for Crystal arthritis Congenital (Lesch Nyhan) Obesity arthritis and soft tissue rheumatism. In many countries, com- Nutrition plementary medicines (traditional medicines) are also used Infectious Environment extensively, particularly for the management of pain. These Osteoporosis Lack of exercise compounds remain unproven, and clinical studies to explore Nutrition their worth should be encouraged. Metabolic bone disease Environment Recent years have seen the introduction of a number of Nutrition specific antiosteoarthritic agents, including glucosamine, chon- Back pain Obesity droitin sulfate, soybean extract, and injectable hyaluronic acid Occupation derivatives. Clinical trials have demonstrated that glucosamine Lack of exercise and chondroitin sulfate are beneficial in terms of pain reduc- Trauma Environment tion in patients with OA, but the effects are relatively small. 966 | Disease Control Priorities in Developing Countries | Luke B. Connelly, Anthony Woolf, and Peter Brooks Table 51.4 Treatment for Arthritis Specific Nonsteroidal Disease-modifying antiosteoarthritis anti-inflammatory antirheumatic Category Analgesics agents drugs drugs Pharmacological Acetaminophen Glucosamine Nonspecific NSAIDs Chloroquine treatments Compound analgesics Avocado extract Selective COX-1 sparing Hydroxychloroquine Codeine Soybean agents Sulfasalazine Opioid derivatives Chondroitin sulfate Gold Diacerein D-penicillamine Hyaluronic acid (injectable) Methotrexate Azathioprine Cyclophosphamide Leflunomide Cyclosporine A Corticosteroids Minocycline Biologics: Antitumor necrosis factor Anti-interleukin-1 receptor antagonist Nonpharmacological Massage n.a. n.a. n.a. treatments Heat Ice Ultrasound Source: Authors. n.a. not applicable Many cases of OA and soft-tissue rheumatism and most cases most common being methotrexate, hydroxychloroquine, and of the inflammatory forms of arthritis will require an anti- sulfasalazine. Corticosteroids are also used intermittently in inflammatory drug as well as or instead of a pure analgesic. The many cases. Patients with RA also need to receive information nonsteroidal anti-inflammatory drugs (NSAIDs) have been the about exercise programs and education on activities of daily mainstay for treating arthritic conditions for nearly a century. living so that they can make informed choices in relation to More recently, concern about the adverse gastrointestinal side their therapies. effects of NSAIDs led to the development of COX-1 (cyclo- oxygenase-1) sparing agents. These agents have similar effects on pain relief but a reduced incidence of gastrointestinal side Osteoporosis effects, although they may produce adverse events in the renal A number of therapies are available for OP, including and cardiovascular systems, such as hypertension, decreased calcitonin, calcium, bisphosphonates, hormone replacement renal function, and increased stroke and heart attacks. therapy (HRT), and selective estrogen receptor modifiers. Clinical trial data support the use of HRT, bisphosphonates, selective estrogen receptor modifiers, calcitonin, vitamin D and Rheumatoid Arthritis calcium supplementation, and calcitriol in reducing fracture The medical community now appreciates the importance of rates in high-risk patients. Calcium and vitamin D supplemen- early diagnosis and treatment of RA. All patients with RA tation are recommended to reduce hip fractures among the should be started on a specific antirheumatic drug on diagno- elderly living in assisted living accommodations and nursing sis. These drugs have been shown to be efficacious in random- homes. The recommended daily requirement for calcium ized controlled trials (Gabriel, Coyle, and Moreland 2001), but varies significantly between countries--for example, from each has a quite different spectrum of adverse side effects. Even 1,000 milligrams per day in the United States to less than with the new biologic agents, few patients with RA actually go 500 milligrams per day in India. Recommended levels of vita- into complete remission, and disease activity continues despite min D supplementation range from 500 to 1,000 international a reduction in endpoints, such as the number of painful and units per day, particularly for at-risk aging females. In addition swollen joints, function impairment, and pain. Most patients to these pharmacological interventions, attention to risk factors with RA now receive combinations of antirheumatic drugs, the for falling is also important. Cost-Effectiveness of Interventions for Musculoskeletal Conditions | 967 Surgical treatments vary, from the use of external splints for Another characteristic of economic evaluations in this field fractures, to interventions such as arthroscopy, internal fixation is that they have been performed almost entirely for developed for complicated fractures, and insertion of prosthetic devices, countries. In the sections that follow, we discuss the steps we most commonly total hip and knee replacements. Biomaterials have taken in an attempt to minimize the adverse consequences are increasingly being used to repair bone or cartilage defects in of reliance on the literature for developed countries. younger patients, particularly those with sporting or other Nevertheless, the pragmatic approach that we have adopted is traumatic injuries. subject to some important limitations and caveats. Rehabilitation treatments include a range of activities, from Cost-effectiveness is a relative concept, in the sense that single discipline interventions such as physiotherapy to multi- cost-effectiveness ratios (CERs) are useful only for comparing disciplinary programs, particularly for complex problems such alternative ways of achieving a desired outcome--for instance, as back pain. improving the quality and length of life. Assertions that an intervention is, in its own right, cost-effective are usually based ECONOMIC ISSUES on the notion that a particular CER represents a cutoff between those interventions that are efficient and those that are not. An economic discussion of health policies designed to prevent, Thresholds of this kind involve an assumption about the value treat, and manage musculoskeletal conditions in developing of life--for example, that a quality-adjusted life year (QALY) is countries is inherently difficult for a variety of reasons, but worth US$30,000. Nevertheless, the literature routinely uses primarily because of the lack of both epidemiological and cost-effectiveness rules that are based on thresholds without cost-effectiveness data for most developing countries. Some the theoretically necessary explicit consideration of implicit progress has been made by Symmons, Mathers, and Pfleger budget constraints. (2004a, 2004b), who provide incidence estimates for OA and We have tried to avoid using a threshold type of approach in RA from epidemiological data on prevalence and relative mor- relation to the discussion of cost-effectiveness. Instead, we tality risks, although data from many areas are scant. critically reviewed the cost-effectiveness literature in rheuma- Perhaps a more important constraint on economic evalua- tology to provide an indication of the relative costs and tions in this field is the surprising number of interventions for consequences of available interventions. In some cases, an which trial data on efficacy are inadequate. Another issue, cur- intervention appears to be inefficient because it costs more and rently the target of a concerted effort to improve practice in the produces fewer benefits than a competing alternative or field, is the lack of cross-study comparability of the results of because two interventions produce identical effects but one economic evaluations of interventions for OA, RA, and OP. costs less than the other. Nevertheless, we have provided a sum- One of the most important variables is the choice of compara- mary of our views--for ease of reference--as table 51.5. This tor used to assess the cost-effectiveness of interventions. table summarizes our thoughts on the weight of the current The Outcome Measures in Rheumatology Clinical Trials effectiveness and cost-effectiveness evidence and the likelihood Economics Working Group, which was established in 1996, has that developing countries might realistically consider each made some progress toward redressing this problem. In princi- intervention. For the reasons given above, though, we have ple, the relevant comparator is generally the next-best alterna- articulated the evidence in more detail in the text. tive or alternatives to the intervention of interest. The choice of Cross-country differences in the epidemiology of conditions comparator is especially important for cost-effectiveness analy- of interest, the age structure of populations, and the access to sis, because cost-effectiveness is a relative, not an absolute, con- health care, along with differences in relative prices, are liable to cept; whether a particular intervention is considered efficient affect the cost-effectiveness of any given intervention. Some of depends on the efficiency of other interventions and on budget the substantive gaps between the developed and developing constraints. This issue is a fundamental one, because a great worlds may compound the problem. For example, if the price many health sector innovations involve new ways of producing of labor relative to that of capital is consistently lower in the desirable effects with existing technology. The relevant developing countries, capital-intensive interventions may be consideration in such cases is the additional benefits that the relatively less attractive than they are in the developed coun- innovation is expected to confer and the relative cost of tries, especially if labor-intensive alternatives exist. achieving those benefits. In such circumstances, the computa- To improve comparability across the literature, we adjusted tion of incremental cost-effectiveness ratios (ICERs) on the reported CERs by converting them to 2001 U.S. dollar prices basis of a no-treatment alternative is of limited use, unless that (therefore, those we report generally differ from those the orig- scenario is genuinely under consideration. Unfortunately, the inal authors cite). Generally, we adjusted outcomes to U.S. dol- no-treatment (or, more accurately, the placebo treatment) lars for studies outside the United States that reported CERs option is precisely the comparator that much of the literature in local currencies using the official exchange rate in effect at has used. that time, but wherever such studies reported only U.S. dollar 968 | Disease Control Priorities in Developing Countries | Luke B. Connelly, Anthony Woolf, and Peter Brooks Table 51.5 Summary of the Economic Evaluation of Interventions for Musculoskeletal Conditions Considered Generally cost-effective recommended Conditions and in developed for developing treatment options countries?a countries?b References Additional comments Osteoporosis Primary prevention Physical activity No Yes (for low-cost Katzmarzyk Gledhill, and Shephard Based on consensus interventions) 2000; Geelhoed, Harris, and Prince 1994; Patrick and others 2001 Calcium plus vitamin D Yes Yes Willis 2002 HRT Yes -- Geelhoed, Harris, and Prince 1994; Armstrong and others 2001; Kanis and others 2002 Raloxifene No No Armstrong and others 2001; Based on evidence Kanis and others 2002 Secondary prevention Screening No No Norlund 1996 Calcium and calcium plus vitamin D Yes Yes Kanis and others 2002 HRT Yes Yes Fleurence, Torgerson, and Reid Differences in life expectancy 2002; Kanis and others 2002 and incidence of OP will affect age at which recommended Raloxifene No No Kanis and others 2002 Calcitonin, alendronate, No No Coyle and others 2001; and biphosphonates Kanis and others 2002 Fluoride No No Kanis and others 2002 Alfacalcidol No No Kanis and others 2002 More randomized clinic trials needed Osteoarthritis Primary prevention No evidence Yes Based on consensus Secondary prevention Education program No Further research Lord and others 1999 needed Exercise program No evidence Low-cost programs Patrick and others 2001 may be useful Nonselective NSAIDs No for nabumetone McCabe and others 1998 Gastroprotective agents Yes, but several Van Dieten and others 2000; qualifiers Gabriel, Campion, and O'Fallon 1994 Synovial fluid replacement Yes No Torrance and others 2002 Different comparators, relative price Tertiary interventions Total hip arthroplasty Yes -- Chang, Pellissier, and Hazen 1996 Knee replacement Yes -- Segal and others 2004 Rheumatoid arthritis Inpatient or outpatient No evidence -- Telephone help line Yes Yes Nordstrom and others 1996; With good communications and Hughes and others 2002 low levels of access to medical care (Continues on the following page.) Cost-Effectiveness of Interventions for Musculoskeletal Conditions | 969 Table 51.5 Continued Considered Generally cost-effective recommended Conditions and in developed for developing treatment options countries?a countries?b References Additional comments Disease-modifying antirheumatic drugs Auranofin Not effective No Thompson and others 1988 Cyclosporine, azathioprine, Equal efficacy; No Anis and others 1996 Cost, monitoring, and adverse D-penicillamine cyclosporine should events be used after cheaper, more effective drugs Combination therapy Yes, in some Possibly Verhoeven and others 1998 studies Biologics No data No -- Need trials in developing countries, but current costs are prohibitive Corticosteroids Yes Possibly Bae and others 2003 Side effects Low back pain Back schools No evidence No Van Tulder 2003 Massage Little evidence Yes, if low cost Furlan and others 2002 Early interventions Yes Yes Gatchel and others 2003 Depends on labor market conditions Ankylosing spondylitis Spa exercise Yes, but ICERs No Van Tubergen and others 2002 Does not provide compelling were sensitive to evidence indirect costs Biologics No No -- Unattractive because of their high price Source: Authors. n.a. not available. a. Based on a cost-effectiveness threshold value of approximately US$30,000 or other favorable quantitative data on costs and benefits. b. Based on authors' judgment of generally favorable/unfavorable cost-effectiveness evidence. See text for precise cost-effectiveness (for example, cost per QALY) data. equivalents, we took these as given. Finally, we used the U.S. perimenopausal and postmenopausal women, with no estab- Bureau of Labor Statistics consumer price index data for 2004 lished history of OP. to inflate (deflate) the U.S. dollar CERs to 2001 prices. Thus, unless otherwise stated, all price data are expressed in 2001 U.S. Physical Activity. The prophylactic effects of physical activity dollars. are generally well appreciated, and a large proportion of pre- For clarity, we have classified cost-effectiveness results by ventable disease is sometimes attributed to sedentary lifestyles. condition and also according to whether the intervention Katzmarzyk, Gledhill, and Shephard (2000) estimate the rela- constitutes a primary, secondary, or tertiary intervention. The tive risks for those who are inactive compared with those who exception is RA, for which the management protocols are less are physically active for a range of conditions, including OP. amenable to this type of abstraction. For RA, we found that cat- Their results for Canada suggest mean OP relative risk factors egorizing the evidence according to a taxonomy that is problem of 1.56 to 1.90 for sedentary versus active women, depending or intervention based was more useful. on race, and indicate that the population-attributable fraction of OP caused by inactivity was approximately 27 percent and accounted for more than 16 percent of the direct economic Primary Interventions costs of physical inactivity. This section reviews the evidence on the cost-effectiveness of The effectiveness and cost-effectiveness of programs interventions designed to prevent the onset of OP. The works intended to encourage lifestyle changes are generally not well surveyed analyzed interventions in healthy people, primarily established. Geelhoed, Harris, and Prince (1994) consider the 970 | Disease Control Priorities in Developing Countries | Luke B. Connelly, Anthony Woolf, and Peter Brooks effect of an intervention in Australia involving exercise and is a dominant long-term therapy for U.S. women at average risk calcium supplements for healthy postmenopausal women to (in this case, a 10 percent lifetime risk) of breast cancer: the 5- prevent osteoporotic fractures. They find that the cost of the and 10-year period ICERs were US$37,620 and US$33, 472 per intervention was US$96,119 per QALY; however, note that QALY, respectively. For women at a 30 percent or higher risk of the authors assumed no toxic effect of the lifestyle regimen on breast cancer, the ICER for raloxifene versus HRT was less than diseases other than OP. US$4,160, and decreased with risk. Kanis and others (2002) argue that existing evidence on Calcium Plus Vitamin D. Willis (2002) analyzes the cost- raloxifene suggests that it has no significant effect on either effectiveness of administering calcium plus vitamin D3to healthy appendicular fractures or CHD. On the basis of the existing postmenopausal women in Sweden and demonstrates that this cost-effectiveness evidence, the use of raloxifene as a prophy- intervention is a cost-saving one for 50-, 60-, and 70-year-old lactic intervention for OP in the developing regions has little to women with a maternal family history of hip fracture and for 60- recommend it. and 70-year-old women with either a history of fragility fractures or a smoking habit. In developing regions, calcium plus vitamin D therapy may be a cost-effective or cost-saving intervention if Secondary Interventions targeted at older, asymptomatic women with maternal histories The following studies were concerned with interventions in of hip and other fragility fractures--especially those who smoke. people with some indication of OP, either from a bone mineral A targeted strategy of this kind is likely to be the most cost- density assessment or a fracture. Some of the general studies effective in regions where environmental uptake of these ele- include Jönsson and others' (1999) study based on Swedish epi- ments is limited for dietary or other reasons. demiological data. The authors consider two different levels of intervention costs, those associated with HRT and those associ- Hormone Replacement Therapy. Geelhoed, Harris, and ated with HRT plus bisphosphonates, and find that the higher- Prince's (1994) cost-effectiveness analysis of interventions in cost intervention (HRT plus bisphosphonates therapy) was a hypothetical cohort of 100,000 healthy postmenopausal dominant for the 80-year-old group modeled. In the context of women includes several HRT strategies: (a) estrogen from age developing countries, note both the relatively higher incidence 50 for life, (b) estrogen from age 50 for 15 years, and (c) estro- of osteoporotic fractures among 80-year-olds and the relatively gen from age 65 for life. Compared with a no-therapy alterna- larger size of this demographic group in Sweden. tive, the cost per QALY was US$8,609 for strategy c, US$13,268 for strategy a, and US$30,183 for strategy b. Screening. Norlund (1996) conducted a cost-benefit analysis Armstrong and others (2001) compare HRT with a no- of fracture prevention in osteoporotic women age 50 to 54 in therapy scenario in healthy postmenopausal women and exam- Sweden, assuming 70 percent participation in the screening ine how the risks of breast cancer and coronary heart disease program and an offer of HRT with 30 percent acceptance. The (CHD) might influence the cost-effectiveness of the interven- study provides evidence of a negative net benefit, indicating tions over 5- and 10-year periods, as well as a lifetime interven- that the costs of a population screening program of this kind tion of approximately 31 years. They report a relatively low cost exceed its benefits. Thus, population-based bone mineral den- per QALY of US$2,238 to US$2,850 for women at a 10 to sity screening programs aimed at perimenopausal or post- 15 percent risk of breast cancer. The cost-effectiveness of HRT menopausal women are likely to be a poor use of health fell as the risk of breast cancer increased. resources in the developing world. Both the base cases in these studies assume that HRT reduces hip fracture rates, and Armstrong and others (2001) Calcium and Calcium Plus Vitamin D. Citing trial evidence, also assume reductions in CHD. These constitute important Kanis and others (2002) assume that calcium supplements assumptions because, as Kanis and others (2002) point out, alone reduce only vertebral fracture risks in women with estab- data from randomized clinical trials (RCTs) support the lished OP. Assuming a compliance rate of 70 percent, the hypothesis of no effect of HRT on either appendicular fractures authors find that the intervention's cost per QALY for 50-, 60-, or CHD. Thus, the ICERs reported by both the studies may 70-, and 80-year-old cohorts were approximately US$64,995, either understate or overstate the true cost per QALY produced US$31,548, US$10,271, and US$10,527, respectively. They also by using HRT as a primary prevention. examine the cost-effectiveness of calcium plus vitamin D on the basis of trial evidence that this combination also reduces Raloxifene. Armstrong and others' (2001) study also includes appendicular fractures. Assuming a 70 percent compliance a cost-effectiveness analysis of raloxifene use (compared with rate, they find that calcium plus vitamin D was cost saving in HRT and no intervention) in healthy postmenopausal women. 80-year-olds and either cost saving or a low-cost intervention Their results indicate that, by comparison with raloxifene, HRT (mean cost per QALY of US$584) in 70-year-olds. For 50- and Cost-Effectiveness of Interventions for Musculoskeletal Conditions | 971 60-year-olds, the mean costs per QALY were US$29,357 and statistically insignificant. When they assume that fluoride has US$13,730, respectively. Thus, in developing regions, calcium a neutral effect on hip fractures, the authors find that the cost plus vitamin D therapy may be an attractive investment for per QALY was in the acceptable range for interventions in elderly women with established OP. the United Kingdom--that is, less than US$46,684. Fluoride is unlikely to be a desirable intervention for preventing OP in Hormone Replacement Therapy. Fleurence, Torgerson, and developing countries. Reid (2002) demonstrate an ICER of US$12,800 to US$19,700 for HRT for their Scottish sample. Kanis and others (2002) show Alfacalcidol. Kanis and others (2002) report wide confidence that while HRT was generally a dominant therapy for 80-year- intervals on the cost per QALY of an alfacalcidol intervention. olds, with a cost per QALY of US$4,527, it was an expensive This result is largely due to substantial variation in the appar- therapy for 50-year-olds at a cost per QALY of US$42,940. These ent vertebral, hip, and humeral fracture risk available from results suggest that HRT is likely to be an attractive intervention RCTs. Thus, alfacalcidol does not appear to be a good invest- for established OP for some age groups in the developing ment for developing economies; however, additional RCTs regions. Differences in life expectancy and the underlying inci- are required to reduce the uncertainty regarding the cost- dence of OP will, however, have a considerable bearing on the effectiveness of this intervention. age at which HRT interventions may be considered desirable in each region. COST-EFFECTIVENESS OF INTERVENTIONS Raloxifene. Kanis and others (2002) find that the cost per FOR OA QALY associated with raloxifene was approximately US$835,622 in 50-year-olds, and although this cost generally fell Primary Interventions with age, it remains an expensive intervention. Raloxifene ther- Despite clear evidence of an association of OA with obesity and apy is not an attractive investment for the developing regions. of a reduction in symptoms and progression of the disease with weight reduction, no formal studies of the cost-effectiveness of Calcitonin and Bisphosphonates. The cost-effectiveness evi- this intervention are available. dence on nasal calcitonin is unambiguous. It is a particularly expensive intervention and represents an unattractive invest- ment of health care resources even in wealthy developed coun- Secondary Interventions tries. The most favorable cost-effectiveness results for nasal cal- Patient education programs, exercise programs, medications, citonin come from a study by Coyle and others (2001), who and synovial fluid replacement have demonstrated varying find that both calcitonin and alendronate reduced wrist, hip, levels of cost-effectiveness. and vertebral fractures in postmenopausal women but that etidronate had no such effect on hip and wrist fractures. The Education Programs. Lord and others (1999) evaluated the ICERs for nasal calcitonin for 65-year-old women for five years cost-effectiveness of a nurse-led education program for of therapy were US$34,166 per QALY compared with no ther- patients with OA of the knee in the United Kingdom, using apy and US$23,952 per QALY compared with etidronate. The usual care as the comparator. They found that the costs for the results of this study were sensitive to the underlying fracture intervention group were greater than for the control group, but rate. that the outcomes for the two groups were not statistically Kanis and others (2002, iv) also conclude that calcitonin is different. "not cost-effective at any age largely because of its costs." The cost-effectiveness of education programs for OA Indeed, their estimates of costs per QALY for 70- and 80-year- patients in the developing countries is unknown. Education old women, the groups for which the intervention is most programs will be subject to diminishing returns, and their cost-effective, equate to approximately US$245,373 and marginal effectiveness may depend directly on the basic level US$181,109, respectively. By contrast, both alendronate and of education of those targeted. Though the scant evidence etidronate were dominating interventions for 80-year-olds. At presented here suggests that education programs may not be current prices, calcitonin therapy is not an attractive invest- cost-effective, further research on their effectiveness and cost- ment for the developing regions. effectiveness in developing countries is required. Fluoride. Kanis and others (2002) find that fluoride was gen- Exercise Programs. Patrick and others (2001) analyzed the erally a dominant intervention in women with established OP, cost-effectiveness of an aquatic exercise program for the man- because it appears to decrease the risk of vertebral fracture but agement of OA and compare it with usual care. The study to increase the risk of hip fracture, although the latter result is involved a 20-week randomized trial of aquatic classes for 249 972 | Disease Control Priorities in Developing Countries | Luke B. Connelly, Anthony Woolf, and Peter Brooks adults age 55 to 75 with a confirmed diagnosis of OA. The Tertiary Interventions results were generally unfavorable. In many cases (24 percent of Total joint replacement for arthritis is one of the most com- the bootstrapped estimates), the exercise program was domi- monly performed and cost-effective operations in developed nated by usual care, and the 95 percent confidence interval countries. In developing countries, however, the availability of ranged from dominated to US$498,700 per QALY gained. this intervention is constrained by the availability of surgeons Evidence of the cost-effectiveness of exercise programs for able to perform the operation. If the surgical expertise is avail- established OA is currently meager. Nevertheless, as part of a able, the cost-effectiveness of total joint replacement is likely to diversified portfolio, low-cost exercise programs may still play be as good as in Australia, Europe, and North America. a useful role in the aging populations of developing regions and confer some benefit on those with established OA, particularly if they are associated with weight reduction. Total Hip Arthroplasty. Chang, Pellissier, and Hazen (1996) assess the cost-effectiveness of total hip arthroplasty in vari- Nonselective NSAIDs. In a U.K. study, McCabe and others ous age groups compared with nonsurgical management. Their (1998) consider the cost-effectiveness of the use of five different analyses suggest that, in 60-year-old white women, total hip NSAIDs (nabumetone, diclofenac, ibuprofen, piroxicam, and arthroplasty is dominant compared with nonsurgical manage- naproxen) in RA and OA. Taking the least and most expensive of ment. For 85-year-old men, the cost per QALY is US$6,893. the five NSAIDs--namely, ibuprofen and nabumetone, which Generally, their results suggest that, when total hip arthroplasty were also at the high- and low-risk ends of the spectrum in terms is used as a treatment for OA of the hip with significant of adverse gastrointestinal events--the authors conclude that functional limitation, it is cost-effective. nabumetone is not a cost-saving prescription. Knee Replacement. Segal and others (2004) review a number Gastroprotective Agents. The most common side effects of of interventions for OA and suggest a cost per QALY of NSAIDs are gastrointestinal; therefore, evaluating therapies to US$5,407 for knee replacement in Australia. reduce these events is important. Van Dieten and others (2000) review the literature on the cost-effectiveness of misoprostol in reducing adverse gastrointestinal events in OA and RA patients who take NSAIDs. Unfortunately, the reviewed studies evi- COST-EFFECTIVENESS OF INTERVENTIONS dently reported CERs based on such nongeneralizable mea- FOR RA sures as cost per patient ratios. Nevertheless, van Dieten and others (2000) argue that strong evidence exists that gastropro- The result of a decade of vigorous debate about the appropri- tection is cost-effective for OA and RA patients taking NSAID ate treatment strategies for RA appears to be a consensus that therapy. This finding appears to be true in relation to several of patients with moderate or severe RA should be treated early the reviewed studies, which produced estimates of cost savings and aggressively, if possible, by combining several disease- derived from prophylaxis. However, van Dieten and others' modifying antirheumatic drugs (DMARDs) (Maetzel and (2000) study is at variance with that of Gabriel, Campion, and others 2002). O'Fallon (1994), who conclude that misoprostol was generally The complex medical management of RA can involve the dominant in that it provided no greater quality-of-life use of a large number of agents, including NSAIDs, low-dose improvement and cost more. corticosteroids, and a long list of DMARDs. The economic literature for interventions in RA has, for good reason, tended Synovial Fluid Replacement. In a Canadian study, Torrance to focus on the cost-effectiveness of the alternatives that arise and others (2002) analyzed the cost-effectiveness of synovial when a particular management strategy fails. In that sense, the fluid replacement in a randomized, one-year, multicenter trial intervention-based taxonomy used in earlier sections of this of 255 patients with OA of the knee. Patients were randomized chapter is a less helpful way to characterize some of the contri- to appropriate care with hylan G-F 20 or to appropriate care butions to this field. Thus, the subsections used in this part of without hylan G-F 20. The mean QALY gain in the intervention the chapter reflect at times a problem-based taxonomy and at group was 0.071, and the resulting ICER was US$5,233 per other times an intervention-based classification. QALY (with similar results from sensitivity analyses). However, the relevant incremental comparators in developing regions are likely to be quite different from those used by the foregoing Treatment Modalities study. Also, the relative price of this product is likely to be One of the challenges common to the developing world is that higher. Thus, we cannot find strong grounds for recommend- specialized medical expertise is often scarce. Thus, the consider- ing that developing regions adopt this intervention. ation of a variety of treatment modalities is worthwhile, Cost-Effectiveness of Interventions for Musculoskeletal Conditions | 973 especially those that involve labor substitution between special- from a six-month RCT of 311 patients with RA. The authors ist and nonspecialist categories. Unfortunately, relatively few report that the cost of auranofin was approximately US$692 studies of this kind, let alone large randomized studies, are greater than for the placebo treatment, but the lack of efficacy available for RA. of auranofin means that it is now rarely used in RA treatment. In a nonrandomized study of 26 patients in Finland, Nordstrom and others (1996) compare the costs of treating RA Cyclosporine, Azathioprine, and D-penicillamine. In a patients either as inpatients or as outpatients. Even though the Canadian study, Anis and others (1996) conduct a cost- authors find that the cost of treating patients as outpatients was effectiveness analysis of cyclosporine use in patients with RA approximately one-sixth the cost of inpatient treatment, the based on the results of a meta-analysis of five RCTs. Their small sample size and possible bias associated with the non- comparators included a placebo control, azathioprine, and D- randomized design mean that the study's results cannot be penicillamine and analyses based on societal costs or third- generalized. party payer costs, but the ICERs were expressed as the cost per An interesting and possibly cost-effective strategy for man- patient per year improved, so the results are difficult to inter- aging RA involves the use of a telephone help line staffed by pret in the context of a general priority-setting exercise for specialist nurses. Hughes and others (2002) examine the health expenditures. For the purposes of this chapter, perhaps costs and benefits of such an intervention in the United the study's most useful result was that it found no statistically Kingdom and conclude that it was cost saving. Their work was significant differences between cyclosporine, azathioprine, and based on a sample of 87 RA patients who used the telephone D-penicillamine. help line in a given month. A large proportion of respondents Given that the existing evidence on cyclosporine, azathio- indicated that they used the help line in place of a visit to a gen- prine, and D-penicillamine indicates similar levels of efficacy, eral practitioner, and on this basis, the authors computed that cyclosporine should be used only after less expensive and more the service produced a net saving. effective therapies for the management of RA, including The existing evidence on the effectiveness and cost- azathioprine and D-penicillamine. effectiveness of a telephone help line for RA patients is based on a relatively small sample. Nevertheless, this type of intervention Combination Therapy. Verhoeven and others (1998) analyze may be useful in developing regions with good communications the use of combination therapy using data from the but low levels of access to medical care. This type of intervention Combinatietherapie Bij Reumatoide Artritis, or COBRA study, may be particularly worthy of consideration when direct access conducted in Europe between 1993 and 1995. The study was a to a medical practitioner is associated with large travel costs. 56-week trial that involved treating an intervention group with sulfasalazine, methotrexate, and prednisolone versus sul- fasalazine alone as a control. Even though the authors conclude Disease-Modifying Antirheumatic Drugs that combined therapy is cost-effective, they qualify the results DMARDs include cyclosporine, azathioprine, D-penicillamine, by stating that the study was probably underpowered. Despite sulfasalazine, etanercept, hydroxychloroquine, methotrexate, the lack of good cost-effective data, the standard approach to leflunomide, and gold compounds. Gabriel, Coyle, and RA treatment is to use combination therapy with DMARDs Moreland (2001) provide a comprehensive review of the effec- and to maintain corticosteroids at 7.5 milligrams per day or less tiveness and cost-effectiveness of DMARDs, including a com- if possible. prehensive literature search in which they retrieved 30 articles from 500 identified for possible relevance. Only six of those Biologics. A number of trials have shown that the biologic papers included economic evaluations, and of those six, only agents (tumor necrosis factor inhibitors and others) are the three included measures of both benefits and costs. Only one most effective agents available for reducing inflammation in of the articles used a nonclinical outcome measure (QALYs). RA. Their cost (US$10,000 to US$15,000 per patient per year); Thus, the cost-effectiveness evidence for the use of DMARDs mode of administration (intramuscular, subcutaneous, or by to treat patients with RA is generally scant. intravenous infusion); and potential side effects (particularly The three full economic evaluations of DMARDs that the reactivation of tuberculosis) preclude their use in develop- Gabriel, Coyle, and Moreland identified were studies of auran- ing countries. Until trials are carried out in developing envi- ofin (oral gold) (Thompson and others 1988), cyclosporine ronments and are combined with robust cost-effectiveness (Anis and others 1996), and combined therapy (Verhoeven and data, we cannot recommend their use. others 1998). Corticosteroids. Bae and others (2003) analyze the cost- Auranofin. Thompson and others (1988) compare the cost- effectiveness of low-dose corticosteroids for the long-term effectiveness of auranofin with that of a placebo using data treatment of RA. They compare the results of corticosteroid 974 | Disease Control Priorities in Developing Countries | Luke B. Connelly, Anthony Woolf, and Peter Brooks treatment with treatment using any DMARDs plus a cortico- nonmeasurement of indirect costs may introduce substantial steroid and with treatment using DMARDs and NSAIDs. Their bias into estimates of the cost-effectiveness of interventions. modeling includes a consideration of the rates of relevant side This problem is potentially serious because, in some cases, effects of the treatments. They also look at NSAID-use scenarios investigators have estimated the indirect costs of low back pain that included Proton Pump Inhibitor prophylaxis and the use of at more than twice those of the direct medical costs (Bolten, COX-2­specific inhibitors rather than nonspecific NSAIDs. Kempel-Waibel, and Pforringer 1998). The results generally showed that corticosteroids dominate nonselective NSAIDs in terms of their cost-effectiveness. The Back Schools. Van Tulder and others (2003) review 15 RCTs of exceptions were when the adverse events rate for cortico- back schools for patients with recurrent and chronic low back steroids was assumed to be 1.5 times that of the base case and pain, but they consider only three of these to be of high quality. when the comparators' adverse events rates were assumed to They conclude that the evidence is only moderate that back be 0.5 to 1.0 times the base case rate. In the latter case, the cost schools have better short-term effects than other treatments for per QALY was US$114,168, and in the former, NSAIDs were chronic back pain. They also find some evidence that back dominant. The comparison of COX-2­specific inhibitors with schools are more effective than placebos or waiting list controls corticosteroids produced a higher-cost and higher-utility in occupational settings. However, the authors note that little is outcome: the resulting incremental cost per QALY was currently known about the cost-effectiveness of back schools. US$132,880. Thus, evidence is insufficient to provide a recommendation on The authors also produced useful age-specific estimates the probable cost-effectiveness of back schools for low back of the cost-effectiveness of the two alternative treatment pain in developing countries; however, early intervention, approaches. Their ICERs show that corticosteroids dominate education, and exercise programs should be encouraged. for the management of 50- and 60-year-olds with RA. For 40-, 30-, and 20-year-olds, the ICERs were US$11,258, US$30,938, Massage. Furlan and others (2002) examine the effectiveness and US$46,981, respectively. and cost-effectiveness of a variety of massage techniques for The evidence suggests that a management strategy of nonspecific low back pain by comparing them with (a) sham or DMARDs plus low-dose corticosteroids is a less costly and placebo massage, (b) other medical treatments, or (c) no treat- more effective strategy than DMARDs plus NSAIDs in older ment. The authors conclude that massage might be beneficial age groups, largely because of the higher risks of adverse gas- for patients with subacute and chronic nonspecific low back trointestinal events in those groups. For developing countries, pain, especially when combined with exercises and education. a relevant question to consider is the extent to which life The evidence suggests that acupuncture massage is more expectancy and risk factors for adverse gastrointestinal events effective than classic massage, but this finding needs to be will differ from the age groups Bae and others (2003) studied. confirmed. Corticosteroid-induced OP is also a long-term risk, but for Little is known about the cost-effectiveness of massage for women with RA starting corticosteroid therapy, watchful wait- low back pain. On the basis of the existing evidence, in coun- ing is recommended as preferable to screening, as long as the tries or regions in which massage--especially acupunctural steroid dose remains below 7 milligrams per day. massage--is routinely available at low cost, the intervention may be cost-effective. Where acupunctural interventions are considered, the costs of bloodborne disease transmission must COST-EFFECTIVENESS OF INTERVENTIONS FOR also be weighed against the expected benefits of the interven- OTHER MUSCULOSKELETAL CONDITIONS tion. This consideration may be important in countries where the prevalence of bloodborne viruses is high, particularly if Low Back Pain strict infection control measures are not routinely followed. Low back pain is as common in developing countries as it is in the developed world. Health professionals now generally agree Early Interventions. The few studies of early intervention that conservative care for acute lower back pain is the initial programs to reduce the progression of acute low back pain to treatment of choice, unless there is structural evidence of chronic pain have tended to report considerable cost savings. pathology that is amenable to surgical intervention (Gatchel Gatchel and others (2003) published a prospective trial of early and others 2003). Evidence also indicates that programs that interventions in individuals with acute low back pain and a incorporate some physical activity may reduce the costs of both high risk of the pain progressing to chronicity. The authors acute and chronic low back pain compared with those that do screened approximately 700 patients and designated them as not involve activity. being at either low or high risk. The patients were then assigned For economic evaluations, one of the important complicat- to early or nonintervention groups and followed for 12 ing factors associated with low back pain is that the months. The early intervention was generally conducted over a Cost-Effectiveness of Interventions for Musculoskeletal Conditions | 975 three-week period and involved an intensive, multidisciplinary IMPLEMENTATION OF CONTROL STRATEGIES: approach that included exercise classes, biofeedback and pain LESSONS OF EXPERIENCE management classes, group education sessions, case manager and occupational therapist sessions, and interdisciplinary team Given the increasing burden of musculoskeletal conditions conferences. worldwide, addressing ways of preventing musculoskeletal The early intervention resulted in statistically significant dif- conditions is an important step. Few examples are available of ferences in return-to-work outcomes, number of health care the implementation of strategies aimed specifically at prevent- visits, and number of disability days caused by back pain. It also ing musculoskeletal conditions, but many recommendations resulted in a variety of pain surrogates. Furthermore, the mean are aimed at modifying determinants that affect other aspects cost savings were approximately US$9,000 per patient. The of health in addition to musculoskeletal health. These recom- direct costs for the intervention group were approximately mendations include ensuring adequate physical activity to US$2,500 higher than those for the intervention group, but this maintain physical fitness; maintaining an ideal body weight; finding was largely the result of the up-front costs of the inter- ensuring a balanced diet that meets the recommended daily vention program itself. The direct costs of health care visits and allowances for calcium and vitamin D; avoiding smoking; pharmaceuticals were considerably lower for the intervention balancing the use of alcohol and avoiding alcohol abuse; and group. putting in place accident prevention programs to reduce mus- The evidence suggests that an intensive, multidisciplinary, culoskeletal injuries related to road traffic accidents, leisure early intervention program is cost saving for individuals with activities, and workplaces. Various programs involve changes in acute low back pain who are at high risk of having the pain the behavior of individuals and control of environmental haz- progress to chronicity; however, the cost savings associated with ards (these programs are considered elsewhere in this volume). this intervention are attributable to improved labor market out- Personal behavior changes can be achieved by education but comes and earnings for injured individuals. The intervention may require resources such as sports facilities. A safe environ- itself may increase costs in the short term, but it appears to be ment will involve all sectors, and successful implementation associated with medium and long-term net benefits. Labor mar- may require legislation. The benefits of these interventions on ket conditions, including wages, along with the age of low back musculoskeletal health are not quantified, but in any case they pain sufferers, may have an important bearing on the viability of are probably small. Physical activity and an ideal weight will this type of intervention in developing countries. benefit the broadest range of musculoskeletal conditions. Ankylosing Spondylitis Osteoarthritis Although the direct costs associated with ankylosing spondyli- The strategy for managing OA is pain management with sim- tis are relatively low, its impact on indirect costs, including pain ple analgesics or NSAIDs, along with education to facilitate and suffering, are substantial. To date, little work has been done self-management and rehabilitative programs to improve func- on the economics of interventions for ankylosing spondylitis. tion, activities, and participation. These strategies include gen- Pharmaceutical interventions are currently typically limited to eral and specific exercise programs, devices such as walking NSAIDs and DMARDs such as methotrexate and sulfasalazine. sticks, and environmental modifications. Joint replacement Van Tubergen and others (2002), however, analyze the cost- surgery should be considered for end-stage joint damage. effectiveness of a spa exercise intervention. The intervention Simple analgesics can be accessible over the counter or through period was three weeks, and although the authors argue that health clinics. Education and rehabilitative programs can also the cost-effectiveness of the intervention was favorable, they be delivered through health clinics. Joint replacement surgery also note that the ICERs were sensitive to variations in assump- requires resources in terms of physical facilities, financial tions about indirect costs. resources, and expertise. The cost-effectiveness of arthroplasty Although a spa exercise program is apparently beneficial is greatly affected by complications such as infection or failure and may even be considered cost-effective for ankylosing of the prosthesis, both of which are related to inadequate spondylitis sufferers in developed countries, the current cost- resources. effectiveness evidence does not provide a compelling case for Pain management should be available to those who have widespread adoption of the intervention in developing regions. disabling symptoms and is dependent on access to drugs and Patients, however, should be encouraged to exercise-- education with respect to the benefits and risks. Access to especially to swim. The cost-effectiveness of tumor necrosis arthroplasty will be the greatest challenge, given the increasing factor­inhibiting drugs is not yet evident for ankylosing needs in the developing world brought about by aging popula- spondylitis, but the drugs are currently unattractive invest- tions and increases in other risk factors such as reduced physi- ments for developing countries because of their high price. cal activity and increasing obesity. 976 | Disease Control Priorities in Developing Countries | Luke B. Connelly, Anthony Woolf, and Peter Brooks Rheumatoid Arthritis Various local programs aim at identifying and treating those The greatest successes in recent years include advances in man- with OP. Their costs relate not only to diagnostic tests but also aging RA and the ability to control disease activity; to prevent to treatment. The cost of bisphosphonates is high compared tissue damage; and to improve function, activities, and with the income levels in those countries likely to experience participation. Methotrexate is a readily accessible, effective the greatest increase in the burden of OP. The speed of benefits antirheumatic drug. Monitoring full blood count and liver of bisphosphonates is good, with clinical trials demonstrating function is recommended, but the rarity of serious adverse fracture risk reduction within 12 to 18 months. The role of events may necessitate a review of this recommendation if the HRT is not clear at present, but because of the likely increased costs and difficulties of monitoring would deny access to the risk of cardiovascular disease and strokes outweighing the ben- drug. Symptomatic therapy with NSAIDs and multidiscipli- efit of fracture prevention, it is not currently recommended for nary rehabilitation are also key components of the manage- preventing OP in unselected women. However, the benefit-risk ment of RA. Central to this approach is ensuring early and ratio will be favorable in those at increased risk of fracture and accurate diagnosis with long-term expert review, which low risk of cardiovascular disease, and the costs of HRT are requires public awareness about arthritis and adequate compe- more feasible. Adherence to treatments for OP to prevent frac- tency and facilities in the community and in health clinics for ture is poor because of the silent nature of their effect, and diagnosis and management. Education and rehabilitation can patient education to modify expectations is important. The also be delivered in these settings. Paramedical workers can be effectiveness of any fracture prevention program depends on trained to undertake much of this work. The problem is to adherence for as long as possible, given the long-term charac- ensure adequate training and experience of health care workers ter of the condition. for a condition that affects 0.3 to 0.5 percent of adults in devel- oping countries. Without treatment, the effects of RA can be great, so effective management can yield significant gains. The RESEARCH AND DEVELOPMENT costs are relatively low, because paramedical workers can deliver much of the care and because the drugs are not new and Another important issue with musculoskeletal disease is the are widely available. development of a research agenda. Several established market economies have set up early arthritis clinics, but running such clinics effectively may be Size and Nature of the Burden of Disease more difficult without a system to encourage and enable early access to specialist care. The management of established RA is Uncertainty about the epidemiology of some musculoskeletal usually undertaken by specialists in partnership with primary conditions is still considerable, especially in developing coun- care. Programs for managing RA are usually self-sustaining and tries (WHO 2003). Incidence estimates for OA and RA have expandable because of the chronic incurable nature of the con- recently been generated from available epidemiological data on dition and the general public's gradual recognition of what can prevalence, relative mortality risks, and so on, though for some be achieved. regions even these basic epidemiological data are scant (Africa, Asia, and South America for RA and Africa, Eastern Europe, and South America for OA). Additional primary measurement Osteoporosis is required to produce a more accurate picture of interregional The prevention of fractures related to OP is based on a "bone- and intraregional epidemiological variations. This deficiency healthy" lifestyle of individuals who have adequate dietary limits estimates of the overall burdens of the various muscu- calcium, vitamin D, and weight-bearing exercise and who avoid loskeletal conditions and the extent to which they might be smoking and excessive alcohol consumption. Implementing reduced. such measures requires raising public awareness and educating The studies of the Community-Oriented Program for primary care personnel. In addition, those at high risk of frac- Control of Rheumatic Disease (COPCORD) (Darmawan and ture who would benefit from a specific intervention need to be others 1995) are, in part, meeting this need. The Bone and Joint identified by the presence of risk factors, including low bone Monitor Project is also undertaking initiatives to standardize density. The limitation to this approach is a lack of access to data collection and reporting in epidemiological studies to bone densitometry, in which case the decision to treat may have improve the collation and comparison of needed epidemiolog- to be made on clinical grounds alone. The occurrence of low- ical data. Global burden-of-disease data concentrate on specific trauma fracture is a good indicator of OP and, in the presence diagnoses, but a far greater burden that has yet to be estimated of other clinical risk factors, may be reason enough to treat. In relates to regional and generalized musculoskeletal problems particular, multiple vertebral fractures are virtually diagnostic characterized by pain with disability. More research is neces- of OP. sary in this area. Cost-Effectiveness of Interventions for Musculoskeletal Conditions | 977 Evaluation of Existing Interventions being achieved because of a lack of prioritization and Few reliable economic evaluations of available interventions resources. Strategies for preventing musculoskeletal conditions have been done. One reason is that a surprising number of have a wide range of other health benefits; they need to be interventions have inadequate trial data or efficacy measures, jointly promoted, and the additional benefits need to be better and many reasonably well-established interventions need to be recognized. The various determinants of ill health--such as tested in trials against appropriate comparators. The hetero- lack of physical activity or obesity--that pose a risk to muscu- geneity of comparators used in different economic evaluations loskeletal health need to be quantified along with their other and other methodological differences pose material risks to detrimental effects. The benefits and cost-effectiveness of mod- comparisons of the costs and consequences of different ifying these determinants of health, with regard to preventing interventions for musculoskeletal conditions, and some stan- or modifying the outcome of musculoskeletal conditions, need dardization is needed. to be quantified and compared with strategies that focus on Many studies have used only clinical outcomes and not life personal interventions. years saved or QALYs as outcome measures. Positive steps are being taken to correct the situation. The Outcome Measures Implementation in Rheumatology Clinical Trials Economics Working Group (Gabriel, Tugwell, and Drummond 2002) has made some Improving musculoskeletal health requires implementing progress toward redressing this problem in the rheumatology strategies for preventing and controlling certain diseases and economics literature. Although work is not yet complete, this environmental risk factors. Selected strategies must be in line concerted approach to standardization holds promise for the with local needs, priorities, and resources. Informed decision literature. making at the policy level requires data on the burden and the avertable burden of musculoskeletal conditions, plus the costs Avertable Proportion of the Burden for particular populations by strategy. Changes in local demo- graphics that are likely to increase the effect of musculoskeletal New data are needed to estimate the burden that could be conditions also need to be considered before developing plans averted by implementing the proposed strategies. Available for implementation. data are currently limited and relate predominantly to individ- ual interventions in short-term clinical trials with outcomes that do not enable reliable estimates of the avertable burden. CONCLUSIONS: PROMISES AND PITFALLS Interventions need to be evaluated, often in combination with Musculoskeletal diseases are the most common cause of outcome measures that enable the burden on individuals and chronic disability worldwide and will become increasingly society to be measured meaningfully and in more naturalistic important as aging populations require relief from chronic pain circumstances. Concordance cannot be assumed, in particular and disability. One of the characteristics of musculoskeletal for these largely chronic conditions, and it is influenced by diseases is that they are not fatal and do not have the high many personal and environmental factors. The impact of profile of other conditions, such as cancer and heart disease. musculoskeletal conditions is pervasive, underrecognized, and However, they are preventable in many cases, and simple inter- underestimated. Therefore, data are needed not only on what is ventions, such as maintaining ideal body weight and partici- theoretically avertable, but also on what is being averted and pating in an exercise program, may have a significant effect on the reasons for any disparities. long-term morbidity. However, the field of musculoskeletal disease is thwarted by a significant lack of epidemiological and Resource Requirements outcome data across a broad spectrum of geography, condi- The implementation of strategies for preventing and control- tion, and treatments. ling musculoskeletal conditions is multisectoral, and the The designation of 2000­10 as the Bone and Joint Decade resource consequences of this need to be established. Different by the United Nations, WHO, and 60 countries will certainly models for the delivery of a strategy may have different help raise the profile of these diseases in local communities. All resource implications. These variations need to be explored to nations have a significant opportunity to embrace the decade, ensure appropriate investment and provision of resources. to ensure that their populations understand the importance of these diseases, and to encourage the training of a range of Likely Effectiveness of Interventions on Both Health and health professionals to deal with this burgeoning epidemic. Nonhealth Benefits Improving knowledge among health workers at all levels Musculoskeletal conditions are common and have a major about musculoskeletal conditions is important for early diag- impact on individuals and society; however, they are inade- nosis and intervention, as is the provision of access to special- quately treated, and the success of simple interventions is not ist services, such as orthopedic surgery. Simple programs that 978 | Disease Control Priorities in Developing Countries | Luke B. Connelly, Anthony Woolf, and Peter Brooks emphasize the importance of obesity and lack of exercise as Fleurence, R., D. J. Torgerson, and D. M. Reid. 2002. "Cost-Effectiveness predictors of poor musculoskeletal outcomes are low cost, of Hormone Replacement Therapy for Fracture Prevention in Young Postmenopausal Women: An Economic Analysis Based on a but their implementation and their influence on health out- Prospective Cohort Study." Osteoporosis International 13 (8): 637­43. comes need to be assessed in properly conducted studies. Furlan, A. D., L. Brosseau, M. Imamura, and E. Irvan. 2002. "Massage Medications--particularly analgesic and anti-inflammatory for Low-Back Pain: A Systematic Review within the Framework of drugs for arthritis and pain and vitamin D and calcium sup- the Cochrane Collaboration Back Review Group." Spine 27 (17): 1896­910. plementation to prevent OP--need to be widely available. Gabriel, S. E., M. E. Campion, and W. M. O'Fallon. 1994. "A Cost-Utility Exciting advances in the treatment of inflammatory forms Analysis of Misoprostol Prophylaxis for Rheumatoid Arthritis Patients of arthritis with biologics need to be evaluated from an eco- Receiving Nonsteroidal Antiinflammatory Drugs." Arthritis and nomic perspective, particularly in developing nations, where Rheumatology 37 (3): 333­41. the risk of exacerbating underlining infections such as tuber- Gabriel, S. E., D. Coyle, and L. W. Moreland. 2001 "A Clinical and Economic Review of Disease-Modifying Antirheumatic Drugs." culosis is much higher than in developed countries. Currently, Pharmacoeconomics 19 (7): 715­28. biologic agents are not cost-effective in developing countries, Gabriel, S. E., P. Tugwell, and M. Drummond. 2002. "Progress towards but they may be in the future. Access to hip and knee replace- an OMERACT-ILAR Guideline for Economic Evaluations in ments, probably the most cost-effective surgical intervention Rheumatology." Annals of the Rheumatic Diseases 61 (4): 370­73. available, is important but depends on the availability of a qual- Gatchel, R. J., P. B. Polatin, C. Noe, M. Gardea, C. Pulliam, and ified staff. J. Thompson. 2003. "Treatment- and Cost-Effectiveness of Early Intervention for Acute Low-Back Pain Patients: A One-Year Musculoskeletal diseases will continue to present a challenge Prospective Study." Journal of Occupational Rehabilitation 13 (1): to the health systems of both developing and developed 1­9. countries, but as we solve some of the issues related to commu- Geelhoed, E., A. Harris, and R. Prince. 1994. "Cost-Effectiveness Analysis nicable diseases, the hope is that more resources will become of Hormone Replacement Therapy and Lifestyle Intervention for Hip Fracture." Australian Journal of Public Health 18 (2): 153­60. available for tackling the burgeoning epidemic of noncommu- Hazes, M., and A. D. Woolf. 2000. The Bone and Joint Decade 2000­2010. nicable disease, including musculoskeletal conditions. Journal of Rheumatology 27:1­3. Hughes, R. A., M. E. Carr, A. Huggett, and C. E. Thwaites. 2002. "Review of the Function of a Telephone Helpline in the Treatment of Outpatients with Rheumatoid Arthritis." Annals of the Rheumatic Diseases 61 (4): ACKNOWLEDGMENTS 341­45. Jitapunkul, S., C. Kunanusont, W. Phoolcharoen, P. Suriyawongpaisal, and We gratefully acknowledge the capable research assistance pro- S. Ebrahim. 2003. "Determining Public Health Priorities for an Ageing vided by Richard Supangan. Population: The Value of a Disability Survey." Southeast Asian Journal of Tropical Medicine and Public Health 34 (4): 929­36. Jönsson, B., J. Kanis, A. Dawson, A. Oden, and O. Johnell. 1999. "Effect and Offset of Effect of Treatments for Hip Fracture on Health Outcomes." REFERENCES Osteoporosis International 10 (3): 193­99. Anis, A. H., P. X. Tugwell, G. A. Wells, and D. G. Stewart. 1996. "A Cost Kanis, J. A., J. E. Brazier, M. Stevenson, N. W. Calvert, and M. Lloyd Jones. Effectiveness Analysis of Cyclosporine in Rheumatoid Arthritis." 2002. "Treatment of Established Osteoporosis: A Systematic Review Journal of Rheumatology 23 (4): 609­16. and Cost-Utility Analysis." Health Technology Assessment 6 (29): 1­146. Armstrong, K., T.-M. Chen, D. Albert, T. C. Randall, and J. S. Schwartz. Katzmarzyk, P. T., N. Gledhill, and R. J. Shephard. 2000. "The Economic 2001. "Cost-Effectiveness of Raloxifene and Hormone Replacement Burden of Physical Inactivity in Canada." Canadian Medical Associa- Therapy in Postmenopausal Women: Impact of Breast Cancer Risk." tion Journal 163 (11): 1435­40. Obstetrics and Gynaecology 98 (6): 996­1003. Lord, J., C. Victor, P. Littlejohns, F. M. Ross, and J. S. Axford. 1999. Bae, S.-C., M. Corzillius, K. M. Kuntz, and M. H. Liang. 2003. "Cost- "Economic Evaluation of a Primary Care-Based Education Effectiveness of Low Dose Corticosteroids versus Non-steroidal Anti- Programme for Patients with Osteoarthritis of the Knee." Health inflammatory Drugs and COX-2 Specific Inhibitors in the Long-Term Technology Assessment 3 (23): 1­55. Treatment of Rheumatoid Arthritis." Rheumatology 42 (1): 46­53. Maetzel, A., V. Strand, P. Tugwell, G. Wells, and C. Bombardier. 2002. Bolten, W., A. Kempel-Waibel, and W. Pforringer. 1998. "Analysis of the "Cost-Effectiveness of Adding Leflunomide to a Five-Year Strategy of Cost of Illness in Backache." Medizinische Klinik 93 (6): 388­93. Conventional Disease-Modifying Antirheumatic Drugs in Patients with Rheumatoid Arthritis." Arthritis and Rheumatism 47 (6): 655­61. Chang, R. W., J. M. Pellissier, and G. B. Hazen. 1996. "A Cost-Effectiveness Analysis of Total Hip Arthroplasty for Osteoarthritis of the Hip." McCabe, C. J., R. L. Akehurst, J. Kirsch, M. Whitfield, M. Backhouse, A. D. Journal of the American Medical Association 275 (11): 858­65. Woolf, and others. 1998. "Choice of NSAID and Management Strategy in Rheumatoid Arthritis and Osteoarthritis: The Impact on Costs and Coyle, D., A. Cranney, K. M. Lee, V. Welch, and P. Tugwell. 2001. "Cost Outcomes in the U.K." Pharmacoeconomics 14 (2): 191­99. Effectiveness of Nasal Calcitonin in Postmenopausal Women: Use of Cochrane Collaboration Methods for Meta-Analysis within Economic McGill, P. E., and G. O. Oyoo. 2002. "Rheumatic Disorders in Sub-Saharan Evaluation." Pharmacoeconomics 19 (5, part 2): 565­75. Africa." East African Medical Journal 79 (4): 214­16. Darmawan, J., H. A. Valkenburg, K. D. Muirden, and R. D. Wigley. 1995. Nordstrom, D. C. E., Y. T. Kontinnen, S. Solovieva, C. Friman, and S. "The Prevalence of Soft Tissue Rheumatism in Indonesia: A WHO- Santavirta. 1996. "In- and Out-Patient Rehabilitation in Rheumatoid ILAR COPCORD Study." Rheumatology International 15 (3): 121­24. Arthritis." Scandinavian Journal of Rheumatology 25 (4): 200­6. Cost-Effectiveness of Interventions for Musculoskeletal Conditions | 979 Norlund, A. 1996. "Prevention of Osteoporosis: A Cost-Effectiveness van Tubergen, A., A. Boonen, R. Landewe, M. Rutten­van Molken, D. van Analysis Regarding Fractures." Scandinavian Journal of Rheumatology der Heijde, A. Hidding, and S. van der Linden. 2002. "Cost 25 (Suppl. 103): 42­45. Effectiveness of Combined Spa-Exercise Therapy in Ankylosing Patrick, D. L., S. D. Ramsey, A. C. Spencer, S. Kinne, B. Belza, and T. D. Spondylitis: A Randomized Controlled Trial." Arthritis and Topolski. 2001. "Economic Evaluation of Aquatic Exercise for Persons Rheumatism 47 (5): 459­67. with Osteoarthritis." Medical Care 39 (5): 413­24. van Tulder, M. W., R. Esmail, C. Bombardier, and B. W. Koes. 2003. "Back Segal, L., S. E. Day, A. B. Chapman, and R. H. Osborne. 2004. "Can We Schools for Non-specific Low Back Pain." Cochrane Library (2), Reduce Disease Burden from Osteoarthritis." Medical Journal of Update Software, Oxford, U.K. Australia 180 (Suppl. 5): S11­17. Verhoeven, A. C., J. C. Bibo, M. Boers, G. L. Engel, and S. van der Linden. Symmons D., C. Mathers, and B. Pfleger. 2004a. Global Burden of 1998. "Cost-Effectiveness and Cost-Utility of Combination Therapy in Osteoarthritis in the Year 2000. Geneva: World Health Organization. Early Rheumatoid Arthritis: Randomized Comparison of Combined Step-Down Prednisolone, Methotrexate, and Sulphasalazine with ------. 2004b. The Global Burden of Rheumatoid Arthritis in the Year 2000. Sulphasalazine Alone." British Journal of Rheumatology 37 (10): Geneva: World Health Organization. 1102­9. Thompson M. S., J. L. Read, H. C. Hutchings, M. Paterson, and E. D. J. WHO (World Health Organization). 2000. "Global Burden of Disease." Harris. 1988. "The Cost Effectiveness of Auranofin: Results of a Global Programme on Evidence for Health Policy Discussion Paper 50, Randomized Clinical Trial." Journal of Rheumatology 15 (1): 35­42. WHO, Geneva. Torrance, G. W., J. P. Raynauld, V. Walker, C. H. Goldsmith, N. Bellamy, and ------. 2003. "The Burden of Musculoskeletal Conditions at the Start of P. A. Band, and others. 2002. "A Prospective, Randomized, Pragmatic, the New Millennium." WHO Technical Report Series 919, WHO, Health Outcomes Trial Evaluating the Incorporation of Hylan G-F 20 Geneva. into the Treatment Paradigm for Patients with Knee Osteoarthritis (Part 2 of 2): Economic Results." Osteoarthritis and Cartilage 10 (7): ------. 2004. Annex 3: Burden of Disease in Disability-Adjusted Life-Years 518­27. (DALYs), by Cause, Sex, and Mortality Stratum, in WHO Regions, van Dieten, H. E. M., I. B. C. Korthals-De Bros, M. W. van Tulder, W. F. Estimates for 2001. Geneva: WHO. http://www.who.int/whr/2002/ Lems, B. A. C. Dijkmans, and M. Boers. 2000."Systematic Review of the annex/en/print.html. Cost Effectiveness of Prophylactic Treatments in the Prevention of Willis, M. S. 2002. "The Health Economics of Calcium and Vitamin D3 Gastropathy in Patients with Rheumatoid Arthritis or Osteoarthritis for the Prevention of Osteoporotic Hip Fractures in Sweden." Taking Non-steroidal Anti-inflammatory Drugs." Annals of the International Journal of Technology Assessment in Health Care 18 (4): Rheumatic Diseases 59 (10): 753­59. 791­807. 980 | Disease Control Priorities in Developing Countries | Luke B. Connelly, Anthony Woolf, and Peter Brooks Chapter 52 Pain Control for People with Cancer and AIDS Kathleen M. Foley, Judith L. Wagner, David E. Joranson, and Hellen Gelband The undertreatment of chronic pain is a global problem, espe- or excruciating. With numerical scales, patients rate their pain cially for people in the final stages of cancer and, increasingly, by choosing a number--for example, from 0 (no pain) to 10 AIDS. The pain of dying is often severe, but it can be controlled (worst pain). Visual analog scales often take the form of a ruled for most people by a simple and inexpensive intervention: oral line, anchored on the left by the words no pain and on the right analgesic drugs, including morphine and other opioids. by worst possible pain. Although it was long known that opioid drugs were essential Pain measurement instruments have been validated in clin- for the relief of moderate to severe pain, even in the 1980s the ical trials of analgesic therapies and subsequently used in amounts being used globally were so low that only a minority national and international surveys, repeatedly demonstrating of those dying could have had adequate pain relief. Since then sensitivity and reliability for both cancer and AIDS patients. progress has been made, mainly in resource-rich countries, These instruments include the Brief Pain Inventory (Bernabei widening the gap between rich and poor. The absence of opi- and others 1998; Cleeland and others 1996; Daut, Cleeland, and oids in developing countries is not merely a problem of supply Flanery 1983), the Memorial Pain Assessment Card (Fishman or costs, however. This chapter lays out the institutional and and others 1987), the Memorial Symptom Assessment Scale political barriers that restrict their availability in most low- and (Portenoy and others 1994), and the Edmonton Symptom middle-income countries. Assessment Scale (Chang, Hwang, and Feuerman 2000). BURDEN OF PAIN FROM CANCER AND AIDS Effects of Pain Pain is"an unpleasant sensory and emotional experience associ- Pain dramatically affects quality of life. Patients with persistent ated with either actual or potential tissue damage or described in serious pain cease participating in social activities and may be terms of such damage"(Task Force on Taxonomy 2004). Pain, in unable to work or care for their families (Daut, Cleeland, and its various manifestations, is the most common symptom lead- Flanery 1983). Psychological effects, including depression and ing patients to seek medical assistance (box 52.1). anxiety,increase with pain intensity (Rosenfeld and others 1996). The suffering of an individual radiates throughout house- Measurement of Pain holds, neighborhoods, and villages. Caregivers suffer distress, Pain is a subjective experience, but it can be described by anxiety, and depression. They may have to give up their own patients and assessed using validated questionnaires and scales employment to care for a dying relative. The loss of income of (Cleeland 1990). In categorical scales, the patient describes the the patient and the caregiver may dramatically lower the pain using specific words, for example, mild, moderate, severe, family's social status (Murray and others 2003). 981 Box 52.1 Classification of Pain Pain is classified according to two main characteristics: º results in significant changes in psychological, func- temporal and physiologic. Temporal categories are tional, and social status. · acute pain Physiologic pain categories are º characterized by a well-defined onset and self- · somatic pain limited end º originates in ligaments, tendons, bones, blood ves- º allows clear description of location, character, and sels, and nerves timing º sharp or dull, but typically well localized and º shows signs of autonomic nervous system intermittent hyperactivity--for example, tachycardia, hyperten- · visceral pain sion, profuse sweating (diaphoresis), dilated pupils º originates in body organs and results from activa- (mydriasis), or pallor tion of nociceptive receptors and efferent nerves · chronic or persistent pain º characterized by deep aching and cramping, often º long lasting, usually defined as at least three referred to cutaneous sites months · neuropathic pain º characterized by a localization, character, and timing º results from direct injury to peripheral receptors, that is often more vague than with acute pain nerves, or the central nervous system º characterized by adaptation of the autonomic nerv- º typically burning and dysesthetic (abnormal and ous system, so signs of hyperactivity disappear unpleasant), often in area of sensory loss. Source: Authors. Pain in Patients with Cancer and AIDS day of moderate or severe pain requiring an opioid drug for Several well-defined acute and chronic pain syndromes are asso- relief. The elements that determine the number of cancer and ciated with cancer, with HIV/AIDS, and with their treatment AIDS pain days in a population are the numbers dying from (Breitbart 2003; Foley 1979). In low-income countries, where each condition and the average prevalence and duration of patients usually present late in the course of illness, pain from severe pain associated with dying. the disease itself is more common than treatment-related pain. Patterns of pain from specific cancers and from AIDS at Researchers consistently report that 60 to 90 percent of given stages appear to be similar everywhere. However, because patients with advanced cancer experience moderate to severe different cancers produce different symptoms, the mix of can- pain, regardless of age and gender and whether ambulatory or cers in a country will influence the overall pattern and burden hospitalized (Cleeland and others 1988; Cleeland and others of pain reflected in the total number of pain days. 1996; Daut and Cleeland 1982; Foley 1979, 1999; Stjernsward About 2.1 million deaths from cancer and about 3 million and Clark 2003). The intensity, degree of pain relief, and effect from AIDS occur annually in low- and middle-income countries of pain vary according to the type of cancer, treatment, and (LMICs) worldwide, and these numbers are increasing. Using personal characteristics, but prevalence and severity of pain expert opinion,we estimate that about 80 percent of people dying usually increase with disease progression. No population-based from cancer and 50 percent of those dying from AIDS experience studies of AIDS-related pain have been published, but several moderate or severe pain, lasting for an average of 90 days. researchers report that up to 80 percent of patients in the last phase of illness experience significant pain requiring analgesics INTERVENTIONS FOR PAIN RELIEF (Larue, Fontaine, and Colleau 1997; Schofferman and Brody 1990; Singer and others 1993). The goal of pain control is not to cure disease, but to allow patients to function as effectively as possible and to minimize Pain Days pain. Interventions for pain relief include drugs, radiotherapy, No standard metric has been developed to describe the pain and anesthetic, neurosurgical, psychological, and behavioral burden for people at the end of life. We have adopted a approaches (see table 52.1). However, analgesic drugs are the transparent and direct measure--the pain day--defined as a mainstay of treatment and the focus of this chapter. According 982 | Disease Control Priorities in Developing Countries | Kathleen M. Foley, Judith L. Wagner, David E. Joranson, and others Table 52.1 Procedures Used to Control Specific Types of Cancer Pain Type of procedure Most common indications Anesthetic Inhalation therapy with nitrous oxide Breakthrough pain, incidental pain in patients with diffuse, poorly controlled pain Intravenous barbiturates (for example, sodium pentobarbital) Diffuse body pain and suffering inadequately controlled by systemic opioids Local anesthetic by intravenous, subcutaneous, Neuropathic pain in any site with local application to the area of hyperesthesia or allodynia or transdermal application Trigger point injections Focal muscle pain Nerve block Peripheral Pain in discrete dermatomes in chest and abdomen or in distal extremities Epidural Unilateral lumbar or sacral pain Midline perineal pain Bilateral lumbosacral pain Intrathecal Midline perineal pain Bilateral lumbosacral pain Autonomic Stellate ganglion Reflex sympathetic dystrophy Lumbar sympathetic Reflex sympathetic dystrophy of the lower extremities Lumbosacral plexopathy Vascular insufficiency of lower extremity Celiac plexus Midabdominal pain from tumor infiltration Intermittent or continuous epidural infusion with local anesthetics Unilateral and bilateral lumbosacral pain Midline perineal pain Neuropathic pain from the midthoracic region down Intermittent or continuous epidural or intrathecal with Unilateral and bilateral pain below the midthoracic region; often local opioid analgesics combined with local anesthetics Intermittent or continuous intraventrical infusions Head and neck pain and upper chest with opioid analgesics Chemical hypophysectomy Diffuse bone pain Neuroablative Nerve root: rhizotomy Somatic and neuropathic pain from tumor infiltration of the cranial and intercostal nerves Spinal cord: dorsal root entry zone lesion Unilateral neuropathic pain from brachial, intercostal, and lumbosacral plexopathy and postherpetic neuralgia Spinal cord: cordotomy Unilateral pain below the waist; often combined with local neurolytic blocks in perineal and bilateral lumbosacral plexopathy; may be performed bilaterally Spinal cord: myelotomy Midline pain below the waist, but rarely used because it involves extensive surgery Brain stem: mesencephalic tractomy Pain in the nasopharynx and trigeminal region Thalamus: thalamotomy Unilateral neuropathic pain in the chest and lower extremity Cortex: cingulotomy Useful through a stereotactic approach for diffuse pain Pituitary: transsphenoidal hypophysectomy Bone metastases in endocrine-dependent tumors, breast, and prostate Neurostimulatory Peripheral nerve: transcutaneous and percutaneous electrical Dysesthesias from tumor infiltration of nerve or trauma nerve stimulation Spinal cord: dorsal column stimulation Of limited use in neuropathic pain in the chest, midline, and lower extremities Thalamus: thalamic stimulation Of rare use in neuropathic pain in the chest, midline, or lower extremity Radiotherapy External beam Bone and brain metastases Nerve and spinal cord compression (Continues on the following page.) Pain Control for People with Cancer and AIDS | 983 Table 52.1 Continued Type of procedure Most common indications Physical Cutaneous stimulation (superficial heat, cold, massage) Dysesthesias from tumor infiltration of nerve or trauma Transcutaneous electrical nerve stimulation Cutaneous nerve injury pain Acupuncture For focal or diffuse pain syndrome Bed rest Reduced movement­related pain syndrome Psychological Hypnosis Provides distraction and cognitive approach to reduce pain Relaxation, imagery, biofeedback, distraction, reframing Provides distraction and cognitive approach to reduce pain Patient education Source: Breitbart 2003; Authors. Pain persists · Step 3 is for pain requiring a strong opioid. No specific Opioid for moderate or increases to severe pain dosages are recommended for opioid drugs because the Opioid for mild to Morphine concept of a standard dose does not apply: effective doses of moderate pain oral morphine range from as little as 5 mg to more than Codeine 1,000 mg every four hours. Adjuvant drugs are also essential Pain persists to treat side effects of analgesics or to provide additive or increases analgesia (table 52.2). Nonopioid + Nonopioid + Nonopioid Aspirin or Aspirin or Aspirin or acetaminophen acetaminophen acetaminophen Controlled field testing and clinical experience has demon- ± adjuvant drug ± adjuvant drug ± adjuvant drug strated that 70 to 90 percent of cancer patients can achieve pain Step 1 Step 2 Step 3 control using the ladder (Goudas, Carr, and Bloch 2001). Although the ladder has not been validated in formal studies Source: WHO 1990. for patients with AIDS, recent clinical reports describe its suc- cessful application (Anand, Carmosino, and Glatt 1994; Figure 52.1 The Three-Step Analgesic Ladder Kimball and McCormick 1996; McCormack and others 1993; Newshan and Lefkowitz 2001; Newshan and Wainapel 1993; Schofferman and Brody 1990). to the World Health Organization (WHO), "A palliative care In an ideal world, a trained professional would prescribe pain programme cannot exist unless it is based on a rational nation- medication throughout the course of illness, in accordance with al drug policy," and this includes "regulations that allow ready the ladder. However, most patients self-medicate pain with anal- access of suffering patients to opioids" (WHO 2002, 87). gesics and traditional medicines that they buy over the counter until they have late-stage disease and severe pain that can be WHO Three-Step Analgesic Ladder and Its Effectiveness treated only with a strong opioid. That is when they are most WHO has developed a "three-step analgesic ladder" (fig- likely to seek formal medical care, which would start on step 3 of ure 52.1) for cancer pain and its treatment (WHO 1986), the ladder. Unfortunately, the opioid they need is unlikely to be which includes a strong opioid (morphine) (table 52.2). The unavailable in LMICs, even from health professionals. ladder is equally appropriate for patients with HIV/AIDS Adequacy of and Barriers to Pain Control and Palliative Care (O'Neill, Selwyn, and Schietinger 2003). in Developing Countries The steps in the ladder represent increasing pain severity and the drugs that should be used in each case: The adequacy of pain control in populations is not easily measured. A useful and available surrogate is the per capita · Step 1 is limited to nonopioids, including drugs that are consumption of morphine (Joranson 1993), a figure based on widely available (for example, acetaminophen, aspirin, or mandatory annual reports by national governments to the nonsteroidal anti-inflammatory drugs, or NSAIDs). International Narcotics Control Board (INCB). Of the 27 mil- · Step 2 describes moderate pain that requires a combination lion grams of morphine used legally in 2002, 78 percent went of a nonopioid and opioid for relief. to six countries--Australia, Canada, France, Germany, 984 | Disease Control Priorities in Developing Countries | Kathleen M. Foley, Judith L. Wagner, David E. Joranson, and others Table 52.2 Basic Drug List for Cancer and AIDS Pain Relief: Analgesics and Adjuvant Drugs Category Analgesics Basic drugs Alternatives Nonopioids Acetylsalicylic acid (aspirin) Choline magnesium trisalicylate Acetaminophen Diflunisal Ibuprofen Naproxen Indomethacin Diclofenac Opioids for mild to moderate pain Codeine Dihydrocodeine Hydrocodone Tramadol Opioids for moderate to severe pain Morphine Methadone Hydromorphone Oxycodone Pethidine Buprenorphine Fentanyl Opioid antagonists Naloxone Nalorphine Adjuvant drugs for analgesia and symptom control Antiemetics Prochlorperazine Metoclopramide Ondansetron Laxatives Senna Cisacodyl Sodium docusate Bran Mineral oil Dantron Lactulose Sorbitol Magnesium hydroxide Antidiarrheal agents Loperamide Paregoric Diphenoxylate hydrochloride and atropine sulfate Antidepressants (adjuvant analgesics) Amitriptyline Imipramine Paroxetine Antipsychotic Haloperidol Thorazine Anticonvulsants (adjuvant analgesics) Gabapentin Valproic acid Carbamazepine Corticosteroids Prednisone Prednisolone Dexamethasone Anxiolytics Diazepam Clonazepam Lorazepam Midazolam Psychostimulants Methylphenidate Pemoline Source: Foley, Aulino, and Stjernsward 2003. the United Kingdom, and the United States. The rest went to use of opioids will lead to drug abuse and addiction, and legal the other 142 countries that reported. Morphine is largely restrictions on opioids. Medical, religious, gender, social, and unavailable in Africa, the eastern Mediterranean, and Southeast cultural factors also present barriers (see box 52.2). With AIDS, Asia (figure 52.2). social and self-stigmatization work against adequate care of any The major barriers to palliative care in LMICs are scarce kind. In addition, most of the emphasis in poor countries has resources, lack of national policies or low priority for pain been on prevention and, more recently, on antiretroviral drugs. relief, lack of awareness by health professionals and the public In all cases, even less care is in place for children than for adults that cancer and AIDS pain can be relieved, concern that medical (Joranson, Rajagopal, and Gilson 2002). Pain Control for People with Cancer and AIDS | 985 Box 52.2 Living with and Dying from Cancer in Scotland and Kenya Physical suffering dominates the lives of people with their families. They are comforted and inspired by reli- advanced cancer in rural Kenya. In contrast, the concerns gious beliefs and by the support of their communities. of cancer patients in Scotland, whose physical needs are They accept their fate. met, focus on the prospect of death. A recent study com- In Scotland, health care is free and of high quality. pared these two groups. Patients are able to get primary treatment for the cancer The suffering in Kenya stems largely from poverty and and, when needed, palliative care. They are likely to be the high cost of basic health care. Hospital care is limited, angry about their illness rather than accepting, and many and patients feel happier at home. Families care for feel isolated from family and friends. Although patients' patients without drugs or supplies or even the knowledge physical needs are met routinely, psychosocial needs are of what to expect and how to help the patient. Patients are met only for some. concerned about the physical and financial burden on Source: Murray and others 2003. mg morphine 80 70 60 USA 50 40 30 UK 20 10 Romania Chile Uganda 0 Source: INCB 2003 and authors' calculations. Figure 52.2 Global Morphine Consumption, by Country (Per Capita, 2002) Legal Controls on Opioid Drugs in 1968 by the Single Convention, is the independent, quasi- The Single Convention on Narcotic Drugs of 1961, amended by judicial organization that implements the Single Convention. the 1972 Protocol (United Nations 1961), is an international The Single Convention requires that all countries (even treaty that aims both to prevent the illicit production of, traf- nonsignatories) intending to make opioids available for ficking in, and use of narcotic drugs and to ensure their avail- medical use estimate national opioid needs and report annually ability for medical and scientific needs. The INCB, established on imports, exports, and distribution to the retail level. It also 986 | Disease Control Priorities in Developing Countries | Kathleen M. Foley, Judith L. Wagner, David E. Joranson, and others Box 52.3 Pain Control in Romania and Chile Palliative care has developed in Romania since the early Palliative care has developed in Chile over the past 1990s, largely through the Romanian Association for 15 years, largely through nongovernmental organizations. Palliative Care. Support has come predominantly from The Ministry of Health's cancer program has also played U.K. charities and from the Open Society Institute. a role by including palliative care in its National Cancer Services are provided throughout the country by paid staff Control Program initiative and by moving to reform and volunteers in 10 hospital-based inpatient services, drug laws. As a result, morphine consumption increased 9 hospice home care teams, 2 day care services, and 1 pal- from less than 5 kg in 1990 to 55 kg in 2000. Despite liative care training center that provides services. Coverage these efforts, only a minority of patients have access to is still low: only 15 percent of cancer patients are treated oral morphine for chronic pain related to cancer or with opioid analgesics. Morphine is available (paid for by AIDS. The remaining barriers include inadequately the government) only for terminally ill cancer patients. trained clinicians, a lack of national standards and guide- Prescription restrictions and extra authorization needed lines, cure-oriented cancer treatment policies, fear of for releasing morphine to each patient are still so burden- addiction on the part of both professionals and the pub- some that patients may die before the paperwork is com- lic, and a lack of resources to improve the health care pleted. The situation is improving gradually, however. infrastructure. Sources: International Observatory on End of Life Care 2005. sets out the following principles on which countries can Costs Included in the Estimates develop their own policies and regulations: The quantitative analysis presented here is restricted to the costs, before such drugs reach the patient, of oral morphine · Individuals must be authorized to dispense opioids by and the adjuvant drugs needed to treat its side effects. We men- virtue of their professional license or be specially licensed to tion other costs associated with delivering oral morphine to do so. terminal AIDS and cancer patients later in this section, but for · Opioids may be transferred only between authorized parties. reasons we discuss, we have not assigned dollar values to them. · Opioids may be dispensed only with a medical prescription. Costs of Oral Morphine. The appropriate measure of drug Many governments have imposed even tighter restrictions, cost is the sum of costs to all payers--governments, insurers, such as limiting the number of days for which an opioid pre- charities, and patients--for the drug itself, but that sum does scription can be written. not include the costs of personnel to administer the drug or otherwise care for the patient. COSTS AND COST-EFFECTIVENESS Oral morphine can be purchased in bulk powder or finished form and administered as a tablet or liquid (De Lima and This section describes the costs and benefits of providing oral others 2004; Rajagopal and Venkateswaran 2003). The cost to morphine and essential adjuvant drugs to terminally ill cancer the final payer is influenced by import taxes, if any; require- and AIDS patients who require it. It assumes the drugs are used ments to document the chain of custody of the product; costs according to the WHO analgesic ladder. We recognize that to local manufacturers of excipients, salts, diluents, and other other analgesics can also contribute significantly to patients' materials required to produce finished forms; and price costs and pain relief, but at least some such drugs (acetamino- markups. The actual cost of oral morphine in LMICs is diffi- phen, for example) are available relatively cheaply in most cult to document because it is unavailable in so many places or places. Although not everyone has access to such drugs, we are is manufactured for finished use at different points in the dis- unaware of any data that could be used to estimate that pro- tribution chain. The price of a 30-day supply of immediate- portion. Costs are estimated for three countries at differing release oral morphine in 2003 ranged from US$10 in India to income levels and with different patterns of cancer and AIDS US$254 in Argentina, among the few countries for which prices deaths: Chile, Romania, and Uganda (see box 52.3). were reported (De Lima and others 2004). Pain Control for People with Cancer and AIDS | 987 Morphine is likely to cost less where it is produced locally average of three months. (Merriman 2002; personal commu- and used in easy-to-reach, urban locations. Liquid prepara- nication, L. De Lima, International Association for Hospice tions made by mixing morphine powder will cost less than and Palliative Care, June 2004; personal communication, tablets. Even with these variations, if barriers to access to oral M. R. Rajagopal, Amrita Institute of Medical Sciences, Kochi, morphine are removed, a total drug cost of 1 cent per mil- Kerala, India, June 2004). ligram or less for immediate-release oral morphine should be Using the inputs above, we estimate the cost of oral mor- achievable for most countries. A realistic and conservative esti- phine for a cancer or AIDS patient with severe pain near mate of the cost of oral morphine is 0.5 cent to 1.0 cent per mil- the end of life at about 30 to 75 cents per day, or US$9.00 to ligram in the countries in our analysis. US$22.50 per month, which is needed for an average of three The cost of morphine per patient depends on the number months. of days that opioids are required and the average daily dose, recognizing that the required dosage typically increases with Costs of Other Necessary Drugs. Morphine's most common increasing pain nearer to death. An average daily dose in side effects are constipation, nausea, and (less frequently) palliative care programs in developing countries is roughly psychosis. Representative drugs to treat these conditions are 60 to 75 mg per day, and patients require this dose for an senna, a laxative, available to some government purchasers for Table 52.3 Background Data, Assumptions, and Results of Cost Analysis Item Uganda Chile Romania Economic, demographic, and health characteristics Population, 2001 22,800,000 15,400,000 22,400,000 Gross national income per capita, 2001 (US$) 260 4,590 1,720 World Bank income designation Low Upper middle Lower middle Percentage of the population living in rural areas 85 14 45 Number of cancer deaths, 2000 10,504 18,315 38,360 Number of AIDS deaths, 2001 84,000 220 350 Prevalence of cancer and AIDS deaths (per million population) 4,145 1,204 1,728 Morphine use per capita, 2001 (mg) 0.1 2.1 2.2 Total morphine use, 2001 (mg millions) 2.191 31.770 48.809 Assumptions and estimates used to determine the costs of oral morphine Percentage of cancer patients requiring end-of-life care with oral morphine 80 80 80 Average number of days of oral morphine required for cancer patients 90 90 90 Average daily dose of oral morphine for cancer patients (mg) 60­75 60­75 60­75 Effectiveness of intervention, cancer (percentage of pain days averted per day of therapy) 80 80 80 Percentage of patients with cancer already receiving adequate end-of-life pain care 0.5 20.0 15.0 Percentage of AIDS patients requiring end-of-life care with oral morphine 50 50 50 Average number of days of oral morphine required for AIDS patients 90 90 90 Average daily dose of oral morphine for AIDS patients (mg) 60­75 60­75 60­75 Effectiveness of intervention, AIDS (percentage of pain days averted per day of therapy) 80 80 80 Percentage of patients with AIDS already receiving adequate end-of-life pain care 0.5 20.0 15.0 Average daily cost of related drugs for cancer and AIDS patients (US$) 0.18­0.33 0.18­0.33 0.18­0.33 Cost analysis results (all costs in 2002 US$) Total incremental annual cost of oral morphine (US$ millions) 2.2­4.9 0.6­1.2 1.1­2.6 Incremental annual cost per capita (US$) 0.10­0.21 0.03­0.07 0.05­0.11 Incremental number of pain days per year avoided with use of oral morphine (millions) 3.6 0.9 1.9 Incremental cost per person per day of pain avoided (US$) 0.60­1.35 0.60­1.35 0.60­1.35 Incremental cost per year of pain-free life added (US$) 216­420 216­420 216­420 Sources: Income and demographic data, World Bank 2003; cancer deaths, Ferlay and others 2001; AIDS deaths, UNAIDS and WHO 2002a, 2002b, 2002c; per capita morphine use, INCB 2003; authors' calculations. 988 | Disease Control Priorities in Developing Countries | Kathleen M. Foley, Judith L. Wagner, David E. Joranson, and others about 3 cents per day; prochlorperazine, an antiemetic, about Other Costs. Security and recordkeeping related to stocking 8 cents per day; and haloperidol, an antipsychotic, about 15 and distributing opioids, required by the Single Convention, cents per day (Management Sciences for Health 2003). Retail entail additional fixed and ongoing costs. Because most hospi- prices after markups would add 20 to 30 percent. tals handle injectable opioids (for example, pethidine), these Under the assumptions of this analysis, oral morphine for all costs would be less for hospital-based programs than for dying cancer and AIDS patients would cost between 3 cents and community-based programs. 21 cents per capita per year (table 52.3) in Chile, Romania, and Professional training and education is required for all per- Uganda. The cost per pain day avoided by oral morphine is the sonnel involved in the use of opioids for dying patients, in part same in all three countries, assuming that each country can to overcome fears and in part to ensure proper use. These costs acquire and dispense morphine equally efficiently. are likely to be highest where the health care system is most deficient. The costs of changing national policy toward opioids is sub- Cost-Effectiveness stantial in terms of cost, time, expertise, and leadership (see, for The analysis indicates that the drug costs of oral morphine example,Pain and Policy Studies Group 2003 and 2004 and other come to about US$216 to US$420 per year of pain-free life annual reports at http://www.medsch.wisc.edu/painpolicy/ gained in the three sample countries. The next question is publicat/annrepts.htm). The time expended is an opportunity whether the pain relief that could be achieved would be worth cost, but it may be amortized over a long time if the effort the cost. We know that patients value pain-free days highly. A succeeds. day lived with the certainty of experiencing severe pain is of very low value, perhaps even lower than death itself (Furlong Potential Cost Savings. In some circumstances, making oral and others 2001; Le Gales and others 2002). Bryce and others morphine available through a palliative care system could (2004) find that people are willing to give up several months of actually save money--for example, if it enabled some termi- healthy life for access to good end-of-life care. Patients in low- nally ill patients who would otherwise be admitted to the hos- income countries place as great or even greater value on pain pital for pain control to die at home, or if it shortened their relief as patients in high-income countries (Cleeland and oth- period of hospitalization. This outcome is more likely in places ers 1988; Murray and others 2003). with good medical infrastructure, but even in low-income countries, patients in unbearable pain are often brought to hospitals by distressed relatives who are willing go into debt Costs Not Included in the Analysis to ease the suffering. The analysis presented includes only the most basic costs--the costs of oral opioids and associated drugs--that would be incurred in a pain control program. Clearly, there are many IMPLEMENTATION OF STRATEGIES TO IMPROVE other costs, ranging from the costs of services at the individual PAIN CONTROL patient level to the costs of changing drug laws and policies at the national level. The most significant additional costs are Providing adequate oral morphine involves medical, political, discussed below. legal, and societal change. Model programs, such as Hospice Uganda and efforts in India, have demonstrated the feasibility Incremental Costs of Care Delivery. In addition to requiring of providing good palliative care, including oral morphine, the drugs themselves, implementation of the three-step ladder even for poor rural dwellers. WHO and the INCB have sup- requires trained individuals to assess and monitor patients. ported these efforts. Where health care systems are well developed, the incremental cost of adding oral morphine will be low. If it involved one additional health center visit, the cost per patient would WHO Guidance on Palliative Care and Pain Relief increase by about US$8 in Chile, US$6 in Romania, and US$4 WHO has affirmed the need for palliative care and has defined in Uganda, amounting to less than 1 cent per capita in all three the elements of model programs in several reports. In 2002, countries. WHO's executive board called for the integration of pain and Where primary health care is weak, widespread access to palliative care into national cancer control programs (box 52.4; oral opioids depends on the development of new systems, such WHO 2002). The Joint United Nations Programme on as community- or hospital-based palliative care networks. HIV/AIDS and the WHO AIDS Program consider pain and Clearly, allocating the full cost of upgrading the health care palliative care to be essential and pain management to be inte- system, or even the development of new palliative care pro- gral to AIDS care (Foley, Aulino, and Stjernsward 2003). grams, to oral morphine alone, would be inappropriate. WHO, in collaboration with the INCB and the WHO Pain Control for People with Cancer and AIDS | 989 Box 52.4 WHO Pain Relief and Palliative Care Recommendations Based on Resource Level WHO's recommendations are as follows: · Countries with medium levels of resources should · Countries with low levels of resources should ensure ensure that minimum standards for pain relief and pal- that minimum standards for pain relief and pallia- liative care are progressively adopted at all levels of care tive care are progressively adopted at all levels of care and that, nationwide, coverage of patients is increasing throughout the country. Countries also should through services provided by health care workers and ensure that patient coverage is high through services home-based care. provided mainly by home-based care. Home-based · Countries with high levels of resources should ensure care is generally the best way to achieve good-quality that national pain relief and palliative care guidelines care and coverage in countries with strong family sup- are adopted by all levels of care and that, nationwide, port and poor health care infrastructure. However, patient coverage is high through a variety of options, many patients are cared for in large cancer hospitals, including home-based care. and these institutions should have pain relief and palliative care available to all patients. Source: WHO 2002. Collaborating Center (WHOCC), has developed guidelines for The cost of treating a patient in Kampala and Mbarara is national authorities to self-diagnose their regulatory systems about US$7 per week, including one home visit. For patients for problems that might lead to a lack of access to needed drugs who come to the hospice, the cost is about US$4 per week (WHO 2000). These organizations also sponsor national and (personal communication, A. Merriman, Hospice Uganda, international workshops to help national authorities evaluate February 2003). Even at those prices, most patients cannot their policies, develop action plans, implement policy change, afford even the medicines, which are subsidized by and evaluate outcomes. In addition to WHO and the INCB, a contributions. number of programs and organizations are making resources and expertise available to assist countries in various ways India. For decades, the only morphine available in India was (box 52.5). injectable and used for postoperative pain. The enactment of a strict, national narcotics law caused morphine use to decline Hospice Uganda. Hospice Uganda began in 1993 with an old even further, from a high of 573 kilograms in 1985 to 18 kilo- Land Rover, a grant to last three months, and a mandate to grams in 1997, among the lowest per capita in the world. During become Africa's model home-based hospice for dying cancer the period of declining use, international efforts to promote and AIDS patients (Ramsay 2001). By July 2004, the original pain control and palliative care programs began to reach India. Kampala location had served about 4,500 nearby patients. In 1992, pain relief and the availability of morphine were desig- Two additional sites--Mobile Hospice Mbarara and Little nated priorities in the National Cancer Control Programme Hospice Hoima--had served about 2,500 patients (Merriman (Joranson, Rajagopal, and Gilson 2002; Rajagopal, Joranson, 2004). Hospice Uganda's influence has spread across Africa and Gilson 2001; Rajagopal and Venkateswaran 2003). through its reputation and the training programs it runs The Ministry of Health convened national workshops from (Merriman 2004). 1992 to 1994 to ascertain why morphine use continued to At the beginning, morphine was largely unavailable, and the decline. The following experience from a referral hospital, law required that a physician prescribe it. Hospice Uganda's recounted by a former narcotics commissioner of India, is founder, Anne Merriman, convinced the government to amend instructive: the law to allow specialist palliative care nurses and clinical offi- cers to prescribe morphine (Merriman 2003). Now, morphine, the Institute has not been able to procure a single tablet [to] paid for by the government, is available for dying patients in date, primarily due to the stringent state laws and multiplicity about 15 of Uganda's 56 districts (Merriman 2003). In 1998, of licenses. After a lot of effort, the Institute had been able Uganda became the first nation in Africa to list palliative care to obtain the licenses in 1994 and had approached [a manufac- as an essential clinical service. turer] for a supply of tablets . . . [but] by the time the tablets 990 | Disease Control Priorities in Developing Countries | Kathleen M. Foley, Judith L. Wagner, David E. Joranson, and others Box 52.5 Selected Resources for Developing National Palliative Care Programs Resources available to countries include the following: opioid availability. The Open Society Institute also sup- · The WHOCC for Policy and Communications in ports the development of implementation strategies in Cancer Care at the University of Wisconsin serves as a 12 of these countries. That effort involves experts in critical resource for palliative care education and coun- pain and palliative care, cancer, and AIDS and repre- try policy makers interested in assessing their opioid sentatives from ministries of health and financing and drug regulations and developing strategies for change. health insurance programs. See http://www.soros.org/ Its Web site links to WHO guidelines in several lan- initiatives/health/focus_areas/international. guages and provides articles and reports on efforts to · The Journal of Pain and Symptom Management has improve national policy and opioid availability in published three special supplements over the past nine Africa, Asia, Eastern Europe, and Latin America. See years in association with the International Associa- http://www.medsch.wisc.edu/painpolicy/. tion for the Study of Pain. The supplements describe · The WHOCC publishes Cancer Pain Release every countries' efforts to advance opioid availability and quarter. The journal includes topical analysis of cur- palliative care. See http://www.elsevier.com/wps/find/ rent issues in cancer pain management and palliative journaldescription.cws_home/505775/description# care and reviews recent international research and edu- description. cational resources. See http://www.whocancerpain. · The International Observatory on End of Life Care is a wisc.edu/. clearinghouse on palliative care in resource-poor coun- · The Open Society Institute sponsored workshops in tries that is aimed particularly at policy makers. The cooperation with WHO's Essential Drug and Cancer initial focus has been on Eastern and Central Europe, units to bring together pain and palliative care experts but the intent is to cover all resource-poor countries. and drug policy makers from Central and Eastern See http://www.eolc-observatory.net/global_analysis/ Europe and the former Soviet Union to develop strate- index.htm. gies for implementing regulatory change to improve Source: Authors. could be arranged, the licenses had expired. The doctors at the controller exempted palliative care programs from needing a Institute and the associated pain clinic have stopped prescrib- drug license, thereby eliminating the need for a pharmacist; ing morphine tablets. (Joranson, Rajagopal, and Gilson 2002, and (c) a palliative care network was established, which consists 153). of about 50 small programs. Statewide, coverage has increased In 1999, the INCB called on the government of India to take to about 20 percent of those needing palliative care. measures to make morphine available for medical uses. In 1994, an initiative begun by the WHOCC, the Indian RECOMMENDATIONS FOR RESEARCH Association of Palliative Care, and the Pain and Palliative Care AND DEVELOPMENT Society systematically studied the reasons for the lack of mor- phine. In 1997, the WHOCC developed a proposal to reduce Policy makers and program implementers need practical tools the number of licenses and extend their period of validity, and to improve pain control and palliative care. They need survey the following year all state governments were instructed to instruments, guidance on how to effect policy and legal adopt a model rule based on the proposal. Gradually, rules changes, and palliative care models for resource-poor settings. have begun to change. By 2002, 7 of 28 states or territories had Many tools exist, but those could be made more accessible adopted the model rule, but it has been implemented through the use of toolkits, distance learning, Web sites, and so successfully only in Kerala. on. Each country also should gather information to assess its The success in Kerala can be attributed to three things: own capabilities and needs, such as the following: (a) the state government simplified the licensing process and stipulated that for oral morphine to be available from a center, · In relation to the national level: it must have at least one doctor with at least one month of º Study the incidence and prevalence of pain related to practical experience in palliative care; (b) the national drugs major causes of illness and death using methodologies Pain Control for People with Cancer and AIDS | 991 adapted from developed countries (Breivik, Collett, and REFERENCES Ventafridda forthcoming). Anand, A., L. Carmosino, and A. E. Glatt. 1994. "Evaluation of Recalcitrant º Survey existing pain and palliative care programs to Pain in HIV-Infected Hospitalized Patients." Journal of Acquired identify national and local leaders in pain control Immune Deficiency Syndromes 7 (1): 52­56. and palliative care and to catalog national guidelines and Bernabei, R., G. Gambassi, K. Lapane, F. Landi, C. Gatsonis, R. Dunlop, standards for acute and chronic pain. For hospice and and others. 1998. "Management of Pain in Elderly Patients with Cancer." Journal of the American Medical Association 279 (23): 1877­82. palliative care services, assess the extent of available care, Breitbart, W. 2003. "Pain." In A Clinical Guide to Supportive and Palliative service delivery models, national and local policies, and Care for HIV/AIDS, ed. J. F. O'Neill, P. Selwyn, and H. Schietinger, professional and public knowledge about pain control 85­122. Washington, DC: Health Resources and Services and palliative care. Administration. º Assess national and local regulatory barriers to opioid Breivik, H., B. Collett, and V. Ventafridda. Forthcoming. "The Pain in availability using WHO (2000) guidelines and needs Europe Survey: Detailed Results and Analysis." European Journal of Pain. assessment protocols (Higginson 1997) to help countries Bryce, C. L., G. Arnold, R. M. Schooler, J. Loewenstein, R. S. Wax, and identify the patient-related, physician-related, and insti- D. C. Angus. 2004. "Quality of Death: Assessing the Importance Placed tutional issues that impede drug distribution. on End-of-Life Treatment in the Intensive-Care Unit." Medical Care 42 º Study costs that affect opioid availability in several coun- (5): 423­31. tries to document the costs of licensing, obtaining, stor- Chang, V. T., S. S. Hwang, and M. Feuerman. 2000. "Validation of the Edmonton Symptom Assessment Scale." Cancer 88 (9): 2164­71. ing, keeping records for, and dispensing opioid drugs. Cleeland, C. S. 1990. "Assessment of Pain in Cancer: Measurement Issues." º Study the costs of alternative delivery models for pain In Advances in Pain Research and Therapy, ed. K. M. Foley, J. J. Bonica, control medications in LMICs. and V. Ventafridda, 47­55. New York: Raven Press. º Assess the offsetting savings achievable by reducing hos- Cleeland, C. S., J. L. Ladinsky, R. C. Serlin, and N. C. Thuy. 1988. pital days by means of better outpatient access to oral "Multidimensional Measurement of Cancer Pain: Comparisons of U.S. and Vietnamese Patients." Journal of Pain and Symptom Management morphine, to document potential savings in representa- 3 (1): 23­27. tive countries, which might help reduce barriers to Cleeland, C. S., Y. Nakamura, T. R. Mendoza, K. R. Edwards, J. Douglas, access. and R. C. Serlin. 1996. "Dimensions of the Impact of Cancer Pain in a · In relation to model programs: Four Country Sample: New Information from Multidimensional º Construct an inventory of model programs for pain con- Scaling." Pain 67 (2­3): 267­73. trol and palliative care. Include their infrastructure and Daut, R. L., and C. S. Cleeland. 1982. "The Prevalence and Severity of Pain in Cancer." Cancer 50 (9): 1913­18. personnel needs, their operating costs, and so on, in Daut, R. L., C. S. Cleeland, and R. C. Flanery. 1983. "Development of the easy-to-use formats such as toolkits and education and Wisconsin Brief Pain Questionnaire to Assess Pain in Cancer and training programs for policy makers and implementers. Other Diseases." Pain 17 (2): 197­210. Regularly add new information from ongoing and new De Lima, L., C. Sweeney, J. L. Palmer, and E. Bruera. 2004. "Potent initiatives. Analgesics Are More Expensive for Patients in Developing Countries: A Comparative Study." Journal of Pain and Palliative Care º Devise additional models, particularly for poor rural Pharmacotherapy 18 (1): 59­70. communities, for providing palliative care and pain con- Ferlay, J., F. Bray, P. Pisani, and D. M. Parkin. 2001. Globocan 2000: Cancer trol practically, efficiently, and sustainably. Incidence, Mortality, and Prevalence Worldwide, Version 1.0, IARC CancerBase 5. Lyon, France: IARC Press. Fishman, B., S. Pasternak, S. L. Wallenstein, R. W. Houde, J. C. Holland, and K. M. Foley. 1987. "The Memorial Pain Assessment Card. A Valid CONCLUSIONS Instrument for the Evaluation of Cancer Pain." Cancer 60 (5): 1151­58. Foley, K. M. 1979. "Pain Syndromes in Patients with Cancer." Advances Unrelieved acute and chronic pain is a serious public health in Pain Research and Therapy, ed. K. M. Foley, J. J. Bonica, and problem worldwide, and 80 percent of cancer patients and V. Ventafridda, 59­75. New York: Raven Press. 50 percent of AIDS patients experience severe pain during the ------. 1999. "Pain Assessment and Cancer Pain Syndromes." In Oxford last months of life. Relief for these patients is possible only with Textbook of Palliative Medicine, 2nd ed., ed. D. Doyle, G. Hank, and N. MacDonald, 310­31. New York: Oxford University Press. oral morphine or another opioid, but developing countries face Foley, K. M., F. Aulino, and J. Stjernsward. 2003. "Palliative Care in many barriers in this respect. Nevertheless, model pain and Resource-Poor Settings." In A Clinical Guide to Supportive and palliative care programs have demonstrated the feasibility of Palliative Care for HIV/AIDS, ed. J. F. O'Neill, P. Selwyn, and H. providing opioid treatment safely, effectively, and inexpen- Schietinger, 387­407. Washington, DC: Health Resources and Services Administration. sively in resource-poor settings. To this end, national govern- Furlong, W. J., D. H. Feeny, G. W. Torrance, and R. D. Barr. 2001. ments must resolve the legal and regulatory barriers to opioid "The Health Utilities Index (HUI) System for Assessing Health- availability, but they need the expertise and support of the Related Quality of Life in Clinical Studies." Annals of Medicine 33 (5): global community to make pain relief a reality. 375­84. 992 | Disease Control Priorities in Developing Countries | Kathleen M. Foley, Judith L. Wagner, David E. Joranson, and others Goudas, L., D. B. Carr, and R. Bloch. 2001. Management of Cancer Pain. Collaborating Center for Policy and Communications in Cancer Evidence Report/Technology Assessment 35, Publication 02-E002. Care. Rockville, MD: Agency for Healthcare Research and Quality. ------. 2004. Improving Cancer Pain in the World, Report for 2003. Higginson, I. 1997. Palliative and Terminal Care. Abingdon, England: Madison, WI: World Health Organization Collaborating Center for Radcliffe Medical Press. Policy and Communications in Cancer Care. INCB (International Narcotics Control Board). 2003. Report of the Portenoy, R. K., H. T. Thaler, A. B. Kornblith, J. M. Lepore, H. Friedlander- International Narcotics Control Board for 2003. Geneva: INCB. Klar, N. Coyle, and others. 1994. "Symptom Prevalence, International Observatory on End of Life Care. 2005. "Hungary." Characteristics, and Distress in a Cancer Population." Quality of Life http://www.eolc-observatory.net/global_analysis/hungary.htm. Research 3 (3): 183­89. Joranson, D. E. 1993. "Availability of Opioids for Cancer Pain: Recent Rajagopal, M. R., D. E. Joranson, and A. M. Gilson. 2001. "Medical Use, Trends, Assessment of System Barriers: New World Health Misuse, and Diversion of Opioids in India." Lancet 358 (9276): 139­43. Organization Guidelines and the Risk of Diversion." Journal of Pain Rajagopal, M. R., and C. Venkateswaran. 2003. "Palliative Care in India: and Symptom Management 8 (6): 353­60. Successes and Limitations." Journal of Pain and Palliative Care Pharmacotherapy 17 (3­4): 121­28. Joranson, D. E., M. R. Rajagopal, and A. M. Gilson. 2002. "Improving Access to Opioid Analgesics for Palliative Care in India." Journal of Pain Ramsay, S. 2001. "Raising the Profile of Palliative Care for Africa." Lancet and Symptom Management 24 (2): 152­59. 358 (9283): 734. Kimball, L. R., and W. C. McCormick. 1996. "The Pharmacologic Rosenfeld, B., W. Breitbart, M. V. McDonald, S. D. Passik, H. Thaler, and Management of Pain and Discomfort in Persons with AIDS Near the R. K. Portenoy. 1996. "Pain in Ambulatory AIDS Patients. II: Impact of End of Life: Use of Opioid Analgesia in the Hospice Setting." Journal of Pain on Psychological Functioning and Quality of Life." Pain 68 (2­3): Pain and Symptom Management 11 (2): 88­94. 323­28. Larue, F., A. Fontaine, and S. M. Colleau. 1997. Underestimation and Schofferman, J., and R. Brody. 1990. "Pain in Far Advanced AIDS." Undertreatment of Pain in HIV Disease: Multicentre Study." British Advances in Pain Research and Therapy, ed. K. M. Foley, J. J. Bonica, Medical Journal 314 (7073): 23­28. and V. Ventafridda, 379­86. New York: Raven Press. Le Gales, C., C. Buron, N. Costet, S. Rosman, and P. R. Slama. 2002. Singer, E. J., C. Zorilla, B. Fahy-Chandon, S. Chi, K. Syndulko, and W. W. "Development of a Preference-Weighted Health Status Classification Tourtellotte. 1993. "Painful Symptoms Reported by Ambulatory HIV- System in France: The Health Utilities Index 3." Health Care Infected Men in a Longitudinal Study." Pain 54 (1): 15­19. Management Science 5 (1): 41­51. Stjernsward, J., and D. Clark. 2003. "Palliative Medicine: A Global Management Sciences for Health. 2003. International Drug Price Indicator Perspective." In Oxford Textbook of Palliative Medicine, 3rd ed., ed. Guide, 2003. Boston: Management Sciences for Health. D. Doyle, G. W. C. Hanks, N. Cherny, and K. Calman, 1199­222. New York: Oxford University Press. McCormack, J. P., R. Li, D. Zarowny, and J. Singer. 1993. "Inadequate Treatment of Pain in Ambulatory HIV Patients." Clinical Journal of Task Force on Taxonomy. 2004. Classification of Chronic Pain, 2nd ed., ed. Pain 9 (4): 279­83. H. Merskey and N. Bogduk. Seattle: International Association for the Study of Pain Press. Merriman, A. 2002. Palliative Medicine: Pain and Symptom Control in the Cancer and/or AIDS Patient in Uganda and Other African Countries. UNAIDS (Joint United Nations Programme on HIV/AIDS) and WHO 3rd ed. Kampala: Hospice Africa Uganda. (World Health Organization). 2002a. Epidemiological Fact Sheets on HIV/AIDS and Sexually Transmitted Infections, 2002 Update, Chile. ------. 2003. "Model Programmes in Africa: Uganda 1993­2003." Geneva: UNAIDS and WHO. Background for presentation at White House Conference Center, February 25, 2003. [typescript]. Hospice Africa Uganda, Kampala. ------. 2002b. Epidemiological Fact Sheets on HIV/AIDS and Sexually Transmitted Infections, 2002 Update, Romania. Geneva: UNAIDS and ------. 2004. "Some Facts about Hospice Uganda, July 2004." [type- WHO. script]. Hospice Africa Uganda, Kampala. ------. 2002c. Epidemiological Fact Sheets on HIV/AIDS and Sexually Murray, S. A., E. Grant, A. Grant, and M. Kendall. 2003. "Dying from Transmitted Infections, 2002 Update, Uganda. Geneva: UNAIDS and Cancer in Developed and Developing Countries: Lessons from Two WHO. Qualitative Interview Studies of Patients and Their Carers." British Medical Journal 326 (7385): 368­71. United Nations. 1961. "Single Convention on Narcotic Drugs." United Nations. http://www.incb.org/e/conv/1961/. Newshan, G., and M. Lefkowitz. 2001. "Transdermal Fentanyl for Chronic Pain in AIDS: A Pilot Study." Journal of Pain and Symptom World Bank. 2003. 2003 World Development Indicators. Washington, DC: Management 21 (1): 69­77. World Bank. Newshan, G. T., and S. F. Wainapel. 1993. "Pain Characteristics and Their WHO (World Health Organization). 1986. Cancer Pain Relief. Geneva: Management in Persons with AIDS." Journal of the Association of WHO. Nurses in AIDS Care 4 (2): 53­59. ------. 1990. Cancer Pain Relief and Palliative Care. Technical Report O'Neill, J. F., P. A. Selwyn, and H. Schietinger. 2003. A Clinical Guide to Series 804. Geneva: WHO. Supportive and Palliative Care for HIV/AIDS. Washington, DC: Health ------. 2000. Achieving Balance in National Opioids Control Policy: Resources and Services Administration. Guidelines for Assessment. Geneva: WHO. Pain and Policy Studies Group. 2003. Improving Cancer Pain in the ------. 2002. National Cancer Control Programmes: Policies and World, Report for 2002. Madison, WI: World Health Organization Managerial Guidelines. 2nd ed. Geneva: WHO. Pain Control for People with Cancer and AIDS | 993 Part Three Strengthening Health Systems · Strengthening Public Health Services · Strengthening Personal Health Services · Capacity Strengthening and Management Reform Chapter 53 Public Health Surveillance: A Tool for Targeting and Monitoring Interventions Peter Nsubuga, Mark E. White, Stephen B. Thacker, Mark A. Anderson, Stephen B. Blount, Claire V. Broome, Tom M. Chiller, Victoria Espitia, Rubina Imtiaz, Dan Sosin, Donna F. Stroup, Robert V. Tauxe, Maya Vijayaraghavan, and Murray Trostle What gets measured gets done. --Anonymous Public health surveillance is the ongoing systematic collection, focus on the use of data to improve public health interventions analysis, and interpretation of data, closely integrated with the (USAID 2005). Additionally, the guidelines for implementing timely dissemination of these data to those responsible for the 2004 draft revised International Health Regulations require preventing and controlling disease and injury (Thacker and World Health Organization (WHO) member states to have Berkelman 1988). Public health surveillance is a tool to esti- key persons and core capacities in surveillance (http:// mate the health status and behavior of the populations served www.who.int/csr/ihr/howtheywork/faq/en/#draft). by ministries of health, ministries of finance, and donors. Just as decision makers require competent, motivated econ- Because surveillance can directly measure what is going on in omists to provide quality technical analyses, they also need the population, it is useful both for measuring the need for competent staff members to provide scientifically valid surveil- interventions and for directly measuring the effects of inter- lance information and communicate the results as information ventions. The purpose of surveillance is to empower decision for action. Competent epidemiologists and surveillance staff makers to lead and manage more effectively by providing members are not a luxury in developing countries; they are a timely, useful evidence. necessity for rational planning, implementation, and interven- Increasingly, top managers in ministries of health and tion (Narasimhan and others 2004). finance in developing countries and donor agencies are recog- nizing that data from effective surveillance systems are useful for targeting resources and evaluating programs. The HIV and DEFINITIONS AND BASIC CONCEPTS severe acute respiratory syndrome (SARS) epidemics under- scored the critical role of surveillance in protecting individual In this chapter, we use the following definitions: nations and the global community. For example, in 2005, China rapidly began to expand its surveillance and response capacity · Indicator: a measurable factor that allows decision makers through its Field Epidemiology Training Program (FETP); to estimate objectively the size of a health problem and Brazil and Argentina chose to use World Bank loans to develop monitor the processes, the products, or the effects of an surveillance capacity; and the U.S. Agency for International intervention on the population (for example, the number Development (USAID) redesigned its surveillance strategy to of new cases of diarrhea, the proportion of children fully 997 immunized in a district, or the percentage of high school of cholera, or "rash illness" rather than measles). Because students who report that they smoke at least one cigarette a syndromic surveillance is inexpensive and is faster than sys- day). tems that require laboratory confirmation, it is often the · Active surveillance: a system employing staff members to first kind of surveillance begun in a developing country. regularly contact heath care providers or the population to However, because of the lack of specificity (for example, a seek information about health conditions. Active surveil- "rash illness" could be anything from the relatively minor lance provides the most accurate and timely information, rubella to devastating hemorrhagic fevers), reports require but it is also expensive. more investigation from higher levels. Also an increase in · Passive surveillance: a system by which a health jurisdiction one disease causing a syndrome may mask an epidemic of receives reports submitted from hospitals, clinics, public another (for example, rotavirus diarrhea decreases at the health units, or other sources. Passive surveillance is a rela- same time cholera increases). tively inexpensive strategy to cover large areas, and it pro- In the specialized area of surveillance for biologic terror- vides critical information for monitoring a community's ism, syndromic surveillance refers to active surveillance of health. However, because passive surveillance depends on syndromes that may be caused by potential agents used by people in different institutions to provide data, data quality biologic terrorists and sometimes refers to alternative meas- and timeliness are difficult to control. ures such as increases in the use of over-the-counter drugs · Routine health information system: a passive system in which or increases in calls to emergency departments. regular reports about diseases and programs are completed · Behavioral risk factor surveillance system (BRFSS): an active by public health staff members, hospitals, and clinics. system of repeated surveys that measure behaviors that are · Health information and management system: a passive sys- known to cause disease or injury (for example, tobacco or tem by which routine reports about financial, logistic, and alcohol use, unprotected sex, or lack of physical exercise). other processes involved in the administration of the public Because the aim of many intervention program strategies is health and clinical systems can be used for surveillance. to prevent disease by preventing unhealthy behavior, these · Categorical surveillance: an active or passive system that surveys provide a direct measure of their effect in the popu- focuses on one or more diseases or behaviors of interest to lation, often long before the anticipated health effects are an intervention program. These systems are useful for pro- expected. These surveys are useful for providing timely gram managers. However, they may be inefficient at the dis- measures of program effectiveness for both communicable trict or local level, at which staff may need to fill out multi- and noncommunicable disease interventions. ple forms on the same patient (that is, the HIV program, the tuberculosis program, the sexually transmitted infections program, and the Routine Health Information System). At OBJECTIVES OF SURVEILLANCE SYSTEMS higher levels, allocating the few competent surveillance experts to one program may leave other programs under- Public health surveillance provides the scientific and factual served, and reconciling the results of different systems to database essential to informed decision making and appropriate establish the nation's official estimates may be difficult. public health action. The key objective of surveillance is to pro- · Integrated surveillance: a combination of active and passive vide information to guide interventions. The public health systems using a single infrastructure that gathers informa- objectives and actions needed to make successful interventions tion about multiple diseases or behaviors of interest to determine the design and implementation of surveillance sys- several intervention programs (for example, a health tems. For example, if the objective is to prevent the spread of facility­based system may gather information on multiple epidemics of acute infectious diseases, such as SARS, managers infectious diseases and injuries). Managers of disease- need to intervene quickly to stop the spread of disease. specific programs may be evaluated on the results of the Therefore, they need a surveillance system that provides rapid integrated system and should be stakeholders. Even when an early warning information from clinics and laboratories. In integrated system is functioning well, program managers contrast, chronic diseases and health-related behaviors change may continue to maintain categorical systems to collect slowly. Managers typically monitor the effect of programs to additional disease-specific data and control the quality of change risky behaviors such as tobacco smoking or chronic dis- the information on which they are evaluated. This practice eases once a year or even less often. A surveillance system to may lead to duplication and inefficiency. measure the population effects of a tuberculosis control pro- · Syndromic surveillance: an active or passive system that uses gram might provide information only every one to five years-- case definitions that are based entirely on clinical features for example, through a series of demographic and health sur- without any clinical or laboratory diagnosis (for example, veys. The principle is that different public health objectives and collecting the number of cases of diarrhea rather than cases the actions required to reach them require different information 998 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others Box 53.1 Different Objectives, Different Data, Different Systems Objective Action Data System Detect epidemics Epidemic response Early warning information Active surveillance Monitor intervention programs Program monitoring Program indicators Health information Monitor impact of policy change Health policy Health indicators Health information Monitor health system Resource allocation Administrative data Health information and management Source: Nsubuga, Eseko, and others 2002. systems. The type of action that can be taken, when or how often Table 53.1 Utility of Surveillance Data that action needs to be taken, what information is needed to Immediate detection of Epidemics take or monitor the action, and when or how frequently the information is needed should determine the type of surveil- Newly emerging health problems lance or health information system (box 53.1). Changes in health practices Changes in antibiotic resistance Changes in distribution of population at risk for PRINCIPLES AND USES OF SURVEILLANCE disease Periodic dissemination for Estimating magnitude of a health problem, Foege, Hogan, and Newton (1976) state that "the reason for including cost collecting, analyzing, and disseminating information on a dis- Assessing control activities ease is to control that disease. Collection and analysis should Setting research priorities not be allowed to consume resources if action does not follow." Determining risk factors for disease The fundamental principle of public health surveillance is that Facilitating planning the surveillance should be designed and implemented to pro- Monitoring risk factors vide valid (true) information to decision makers in a timely Monitoring changes in health practices manner at the lowest possible cost. Because managers are unlikely to need to make interventions to address small differ- Documenting distribution and spread of disease and injury ences between areas, sacrificing precision makes sense to Stored information for Describing natural history of diseases improve timeliness and save resources that can be used for public health interventions. The utility of surveillance data can Facilitating epidemiologic and laboratory research be viewed as immediate, annual, and archival, on the basis of Validating use of preliminary data the public health actions that can be taken (table 53.1; Thacker and Stroup 1998b). Setting research priorities Documenting distribution and spread of disease and injury ESTABLISHING AND MAINTAINING Source: Adapted from Thacker and Stroup 1998b, 65. A SURVEILLANCE SYSTEM and social environment, these steps are linked continuously What is worth doing is worth doing right. Managers who (figure 53.1; Thacker and Stroup 1998a). decide to use public health surveillance as a management tool must recognize that they will need to commit political support and human and financial resources. As with every health sys- ANALYSIS AND DISSEMINATION tem, competent, motivated health workers need to be found or OF SURVEILLANCE DATA trained and provided with career paths and supervision. After a manager decides to create a surveillance system, there are six Surveillance information is analyzed by time, place, and steps to establishing the system. Because the system must adapt person. Knowledgeable technical personnel should review data constantly to changes in the population and the physical regularly to ensure their validity and to identify information Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 999 maintenance and evolution of the standard. Public health data Establish goals needs extend into multiple areas beyond clinical medicine (for example, environmental toxins, unintentional injury, and food safety). Develop case definitions One international standard computer program used in many countries' information systems is Epi Info, an epidemiol- ogy surveillance and biostatistics program widely used around Select appropriate personnel the world for the analysis of surveillance data (http://www.cdc. gov/epiinfo). CDC created, maintains, and distributes Epi Info at no cost to users. Acquire tools and clearances for collection, analysis, and dissemination SURVEILLANCE AS A COMPONENT OF NATIONAL Implement surveillance system PUBLIC HEALTH SYSTEMS WHO and the World Bank cite public health surveillance as an Evaluate surveillance activities essential function of a public health system (World Bank 2001). When linked to policy and program units, surveillance infor- mation improves the efficiency and effectiveness of health Source: Adapted from Thacker and Stroup 1998a, 119. services by targeting interventions and documenting their effect on the population. Figure 53.1 Elements in Establishing and A critical challenge in the health sector in developing coun- Maintaining a Surveillance System tries is to ensure quality and effectiveness of surveillance and public health response in an environment of decentralization. National-level program and surveillance system managers may lose control of the quality and timeliness of locally col- of use to top managers. Simple tables and graphs are most lected data. This situation can be avoided by training local useful for summarizing and presenting data. Timely dissemi- decision makers in how to use information to meet their nation of data to those who make policy and implement needs and negotiating with them over the core information intervention programs is critical to the usefulness of surveil- collected by each district local unit. National-level managers lance data. or donors can also improve information quality by sponsor- The rapidly evolving field of public health informatics, ing national surveillance scientific and quality assurance net- which deals with collection, classification, storage, retrieval, works, linking funding to provision of adequate data, and analysis, and presentation of large amounts of health data, performing periodic surveys to confirm the results of local offers the potential for truly integrated public health surveil- reporting. If the responsibility for implementing programs is lance based on data standardization, a communications infra- devolved to local managers, then national-level managers need structure, and policies on data access and sharing. Surveillance only a few summary indicators, rather than the detailed infor- will benefit by incorporating a systematic approach to stan- mation they may be used to. District or local managers tend dards for data content. For example, the U.S. Centers for to prefer integrated systems to minimize filling out redundant Disease Control and Prevention (CDC) has used standards- forms. based systems to support automatic electronic reporting of Donors need surveillance data to target and evaluate their diagnostic laboratory results of notifiable diseases, thereby investments. If they perceive weakness in the national system, both increasing the number of cases reported and receiving they may create parallel nongovernmental surveillance systems results more rapidly (Effler and others 1999). Use of data stan- to gather data directly to meet their needs. These systems dards facilitates comparability of surveillance information invariably pay workers more than government jobs do, so the over time (for example, measurement of effect of program most competent people in the government system may leave to interventions), across different surveillance approaches (for work for the parallel system. Although this system meets example, facility-based reporting compared with sample sur- donors' short-term needs, it invariably weakens government veys), and across countries and regions. To be credible, a stan- systems. Parallel systems may weaken the very ministries that dard should be developed through an open, participatory they are meant to help and may not be sustainable after exter- process, by an internationally recognized accredited standards- nal funding ends. Therefore, parallel systems are inherently development organization that is also capable of long-term inequitable and should be used only as a last resort. 1000 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others Public health Data collection Response Data Data generation use hemisphere hemisphere Interpretation Analysis Source: Authors. Figure 53.2 Surveillance and Response Conceptual Framework SURVEILLANCE AS A TOOL TO IMPROVE capacity has limited the ability of developing countries to build PUBLIC HEALTH national surveillance systems that respond to both international public health threats and local health concerns. This capability Managers need focused, timely, scientifically sound information is essential to the sustained development of countries. that provides evidence to make decisions on interventions for improving the health of the population in their jurisdiction. Simply collecting data and presenting them are not enough. Role of Field Epidemiologists in Providing Evidence Developed countries have constructed their public health and Using Surveillance Information for disease control strategies by using the principles of field epi- Evidence-Based Decisions demiology. Developing countries need to build and sustain human capacity in field epidemiology. Strengthened field A major gap in promoting effective surveillance lies between epidemiologic capacity can serve a country in the following the production of data and the ability to convert those data into specific areas: usable information and then initiate the appropriate public health action. Surveillance and response can be described in · providing a response to acute problems terms of a data generation hemisphere and a data use hemisphere · providing the scientific basis for program and policy (USAID 2005). The data generation hemisphere is the tradi- decisions tional view of surveillance, whereas the data use hemisphere is · implementing disease surveillance systems the public health response that begins with interpretation of · supporting national health planning the data from the surveillance system (figure 53.2). · making resource allocation decisions Substantial attention and the accompanying resources in · allocating the human capacity base for national health surveillance have been devoted to the prompt and complete priorities. production of surveillance data. Although developing countries Specific competencies that should be developed include, but experience weaknesses in both hemispheres, more attention is are not limited to, the ability to accomplish the following: needed to creating and strengthening the local capacity within developing countries to identify and manage effective responses · design, implement, and evaluate surveillance for a health to disease outbreaks and public health conditions of national event and international concern. In some disease-specific programs, · identify and assess an actual public health problem this capacity has to be imported through short-term expatriate · design and conduct a scientific investigation assistance. Even when local capacity is developed, it is often · analyze and interpret data from an investigation specific to the disease program, making transfer of skills to · recommend logical and practical public health actions after other areas problematic. The failure to develop this indigenous the analysis and interpretation of data Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1001 · be proficient in all aspects of diseases of public health evidence gathered efficiently and presented effectively. Because importance (for example, HIV and AIDS, sexually transmit- these needs differ, depending on management's needs, a num- ted diseases, malaria, tuberculosis, and zoonoses). ber of different strategies have been developed. Here are some of the most useful. These competencies need to be tailored to the various levels of the health care system. Since 1975, CDC and WHO have collaborated with more Sentinel Surveillance than 30 countries to strengthen health systems and address In a sentinel surveillance system,a prearranged sample of report- training needs for disease detection and response in a country- ing sources agrees to report all cases of defined conditions,which specific, flexible, and sustainable manner. More than half of the might indicate trends in the entire target population (Birkhead world's population now lives in a country where surveillance, and Maylahn 2000).When properly implemented, these systems investigation, and response are carried out by staff members offer an effective method of using limited resources and enable and trainees of FETPs and allied programs, which include prompt and flexible monitoring and investigation of suspected the Epidemic Intelligence Service in the United States, the public health problems. Examples of sentinel surveillance are European Program for Intervention Epidemiology Training, networks of private practitioners reporting cases of influenza or and Public Health Schools without Walls (PHSWOWs). These a laboratory-based sentinel system reporting cases of certain programs generally function within central ministries of health bacterial infections among children. Sentinel surveillance is and may not be visible outside the public health system. It can excellent for detecting large public health problems,but it may be be argued that these programs provide most of the surveillance insensitive to rare events, such as the early emergence of a new of and response to emerging infections in the world, in addi- disease, because these infections may emerge anywhere in the tion to training most of the public health workers who manage population. surveillance systems at the top level. FETPs are two-year courses designed to provide a ministry with a motivated, professional Periodic Population-Based Surveys group of field epidemiologists with the expertise to respond to managers' needs for evidence, perform surveillance, respond to Population-based surveys can be used for surveillance if they outbreaks, and train and supervise technical personnel at other are repeated on a regular basis (Thacker and Berkelman 1988). levels (White and others 2001). Examples of population-based surveys in surveillance include Other models have evolved. Guatemala's marriage of its the BRFSS in the United States, HIV-prevalence surveys, FETP (part of a larger, Central American FETP) with the Data household surveys, and the demographic and health sur- for Decision Making program (Pappaioanou and others 2003) veys that many developing countries conduct every five years exemplifies this successful local adaptation. Data for Decision (http://www.orcmacro.com). Population-based surveys require Making recruits health workers from the community and sub- careful attention to the methodology, particularly the use of district levels to receive training in surveillance and outbreak standard protocols, supervision of interviewers, comparable investigation in the context of their daily work. This training is sampling strategy, and standard questionnaires. These surveys delivered as a series of linked workshops with practical field- require a clear definition of the target population to which based projects, providing short-term service at the local levels. the results can be generalized, and they need careful attention The most promising graduates of the course are selected for to the sample size, based on efficiency and the epidemiologic further training in an FETP. India, with its decentralized sys- characteristics of the health condition under surveillance (for tem, complex cultural and population dynamics, and wide example, rare conditions require substantial samples). variance in the sophistication of public health institutions, Supervising interviewers and maintaining high response rates provides another model for strengthening national surveil- are critical to avoid bias. Because the surveys are repetitive, pop- lance. The World Bank initiated the Integrated Disease ulation changes (caused, for example, by mortality or mobility) Surveillance Project, which develops the capacity of local and might bias results. midlevel surveillance workers in India. Additionally, FETP graduates are recruited as the project's surveillance officers at Laboratory-Based Surveillance the state level to coordinate the surveillance activities of the The methods used for infectious disease surveillance form a hundreds of local health workers throughout the states. spectrum that evolves with the economic development of a country. Foodborne disease (FBD) surveillance, for example, SELECTED SURVEILLANCE STRATEGIES is divided into four distinct levels of surveillance. Each level is more complex and has greater capacity for controlling and Surveillance systems need to be designed and implemented to detecting disease, but it also depends on more resources and meet top management's needs for focused, reliable, timely infrastructure (figure 53.3). 1002 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others and Tauxe 2002). A collaborative WHO program called Improving capacity and infrastructure Global Salm-Surv promotes the use of Salmonella serotyping internationally among countries that wish to upgrade their No formal Syndromic Lab-based Integrated surveillancea surveillanceb surveillancec national capacity for FBD surveillance (http://www.who.int/ food-chain surveillanced salmsurv/en). Molecular subtyping is now expanding the power of laboratory-based surveillance to detect outbreaks in the back- Outbreak detection can occur at all levels, but ability increases ground of sporadic cases by distinguishing the molecular along the spectrum "fingerprint" of an outbreak strain. CDC maintains PulseNet, an Internet-based network of all U.S. public health laboratories that uses a standardized genotyping method called pulsed- Source: Authors. field-gel-electrophoresis (PFGE) as the basis for a national a. Without a formal surveillance, only large or unusual outbreaks can be detected. database of PFGE subtypes (Swaminathan and others 2001). b. Syndromic surveillance is based on groups of signs or symptoms indicative of a common diagnosis, such as acute gastroenteritis. Standardized subtyping protocols have now been developed for c. Laboratory-based surveillance relies on laboratory-confirmed pathogens, such as seven pathogens, and next-generation, gene-based technologies Salmonella or Shigella. d. Integrated food-chain surveillance uses data from across the food chain. are under development for the future. Similar networks are developing around the world, with PulseNet Europe and Figure 53.3 Spectrum of Case-Based Foodborne Disease PulseNet Canada already active and discussions rapidly Surveillance advancing for PulseNet Asia Pacific and PulseNet Latin America. As with Salmonella serotyping itself, the global use of standard genotyping will facilitate the detection of multiconti- nent clusters. For FBD, surveillance for clinical syndromes is the most common method of surveillance in the developing world. Surveillance of FBD outbreaks that are investigated by public Integrated Disease Surveillance and Response health authorities is often a useful means of monitoring both The Integrated Disease Surveillance and Response (IDSR) the safety of the food supply and the activities of the public strategy, first developed in Africa, links epidemiologic and lab- health system. Although both surveillance for clinical syn- oratory data in communicable disease surveillance systems at dromes and outbreak surveillance will remain important, all levels of the health system, with emphasis on integrating the future in FBD is in laboratory-based surveillance. If surveillance with response (WHO 1993, 1998). Districts microbiologic diagnosis is sought routinely for a sample of were identified as a focus for strengthening efforts in collecting patients with acute gastroenteritis, then surveillance based on timely data, analyzing the collected data, and using the gener- those diagnoses is possible. For enteric bacterial pathogens ated information for public health responses. The IDSR strategy such as Salmonella or Shigella, determining the serotype of the is based on core activities, including case-patient detection, strains in central reference laboratories allows more rapid and registration, and confirmation; reporting, analysis, use, and complete identification of epidemics, which may otherwise feedback of data; and epidemic preparedness and response (for lead to preventable death and disability. example, outbreak investigations, contact tracing, and public Laboratory-based surveillance systems require resources, health interventions). Support functions include coordination, facilities, and training. A central public health reference labora- supervision or performance evaluation, training, and resource tory is essential for quality assurance and quality control and provision for infrastructure, including communication support. Such a laboratory-based system might begin with sys- (Nsubuga, Eseko, and others 2002). tematic referral of a sample of strains isolated at a sample of Key steps in implementing the IDSR strategy include sensi- sentinel clinics, plus those strains that are part of outbreaks. A tizing key health authorities and stakeholders; conducting situ- systematic sampling scheme provides better data than a more ational analysis; preparing a strategic IDSR plan; identifying haphazard attempt at universal reporting. Regular sharing of and training a motivated, competent workforce; developing information between the public health microbiology laborato- national IDSR technical guidelines; implementing the plan; ry and epidemiologists is critical for the information to be used and monitoring and evaluating implementation to improve successfully. performance (WHO 2000b). Assessment of the existing The utility of serotyping as an international language national surveillance and response activities provides baseline for Salmonella subtypes has led to its widespread adoption. In a data to measure progress; to identify and build consensus on recent survey, 61 countries reported that they used Salmonella the national priority communicable diseases; to identify serotyping for public health surveillance (Herikstad, Motarjemi, surveillance gaps of the selected priority diseases; to document Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1003 Table 53.2 Steps in the Development of the Philippine National Epidemic Sentinel Surveillance System Steps Data side Human capacity side 1. Identify the health problems thought to cause burden Consult top managers, donors, international agencies, and experts. disease. 2. Determine who will make interventions. Involve users in design. 3. Determine information users' need to make interventions. Involve users in design. 4. Decide how often decision makers need reports. Involve users in design. 5. Identify who collects, tabulates, and analyzes reports and Identify manager and staff to analyze, report, and enter data. who disseminates information. 6. Design report. Involve users and staff in design. 7. Make table shells. Involve staff in design. 8. Design questionnaire. Involve staff in design. 9. Pilot questionnaire. Involve staff in implementation and evaluation. 10. Pilot data flow and analysis. Involve staff in implementation and evaluation. 11. Pilot system. Train staff in system and involve them in evaluation. 12. Run system. Involve staff in ongoing training and quality assurance monitoring. 13. Evaluate system: Was information used? Are data and Involve staff and users in design of external evaluation and in review of analysis of good quality? evaluator's report. Source: Adapted from White and McDonnell 2000, 311. the strengths, weaknesses, and opportunities of the existing on the skills of the workforce, PDOH was able to avoid the systems; and to make appropriate recommendations. The duplications, inefficiencies, and sustainability problems of mul- WHO Regional Office for Africa, collaborating with its part- tiple vertical systems (White and McDonnell 2000). ners, has prepared tools and guidelines for implementation of IDSR at the country level. Indicators to monitor the perform- ance of the surveillance and response systems have been pre- Informal Networks as Critical Elements pared and field tested and are now in use in Africa. of Surveillance Systems WHO and other agencies frequently receive telephone calls or informal reports about urgent health events.WHO publishes an Example: The Philippine National Epidemic informal list of these"rumors,"which allows public health work- Surveillance System ers to respond to health risks promptly rather than waiting for In the late 1980s, the Philippine Department of Health formal reports (http://www.who.int/csr/don/en/). The gradu- (PDOH), relying on its integrated management information ates of FETPs, PHSWOWs, and similar programs that provide system, detected less than one outbreak per year in a population competency-based on-the-job training in ministries of health of more than 60 million people. In 1989, the PDOH designed make up one of the most important informal networks. the National Epidemic Sentinel Surveillance System, a hospital- FETPs and allied programs both train epidemiologists and based sentinel surveillance system that encompasses both the provide service to their ministries of health. For example, a stu- flow of data and the personnel requirements needed to make dent in the Brazilian FETP was assigned to review routine data the surveillance system work effectively (table 53.2). After the on patients with leishmaniasis. She noted that some patients pilot study demonstrated promising results, the PDOH created had symptoms of heavy metal poisoning, and further study personnel positions and a supervisory structure for sentinel indicated that a drug being used to treat leishmaniasis was con- physicians, nurses, and clerks in regional epidemiology and taminated with heavy metals. The drug was reformulated, and surveillance units (RESUs) integrated into the public health the problem was resolved. When this study was presented at a system. In 1995 alone, the system detected and formally inves- regional meeting of the Training Programs in Public Health tigated about 80 outbreaks, including 25 bacteriologically Interventions Network (a network of FETPs and allied training confirmed outbreaks of typhoid and 5 of cholera. As the programs), other countries banned the drug until it was refor- Philippines developed HIV serological and behavioral risk mulated (CDC 2004a). surveillance, the RESU staff conducted surveys in their com- Large categorical surveillance systems are expensive, and munities. By integrating surveillance functions that were based staff members might become complacent, especially if the 1004 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others disease under surveillance is rare. For example, the polio sur- epidemiologic studies (http://www.phac-aspc.gc.ca/cfep-pcet/ veillance system for acute flaccid paralysis in the Western outbreaks_e.html). Hemisphere detected no cases in July 2000. A trainee from the The success of this global effort to control the first new epi- FETP of the Dominican Republic, while investigating a case demic disease of the 21st century depended on a combination of suspected poisoning in a child, documented the first out- of open collaboration among scientists and politicians of many break of circulating vaccine-type poliovirus in the Western countries and the rapid and accurate communication of sur- Hemisphere since 1991. There were 13 confirmed cases in the veillance data within and among countries. Rapid global spread Dominican Republic and 8 cases in Haiti. Her investigation was recognized, and a global surveillance network was estab- led to national immunization days in both countries, which lished on the basis of an agreed-upon case definition that was raised immunization levels and stopped the outbreak (Kew sufficiently specific to ensure effective reporting. and others 2002). Public health surveillance is critical to recognizing new cases of SARS and differentiating this disease from other causes of severe respiratory illness, especially influenza (Heymann and THE ROLE OF SURVEILLANCE IN Rodier 2004). Ongoing research into sources in the environ- MAJOR OUTBREAKS ment as well as clinical, laboratory, and epidemiologic concerns will improve surveillance for this critical public health prob- It seems incredible that a disease as devastating as AIDS could lem. Notably, this highly contagious disease--for which there is have spread silently to many countries over many years before neither a vaccine nor a cure--was controlled by competent, it was detected and before effective control measures were dedicated health workers with access to excellent communica- implemented in the 1980s. In recent years, surveillance and tions. SARS presented a greater challenge than smallpox, for response systems at all levels have been more effective at iden- which long incubation periods and vaccine facilitate control tifying and preventing spread of infectious diseases. (Mack 2005). Although it is reassuring that national, regional, and global systems were effective in controlling SARS, there is no reason to rest on our laurels. The only certainty is that there Example: Surveillance and Global Response to SARS will be more new challenges, very possibly including further An epidemic of severe pneumonia of unknown etiology was outbreaks of SARS. detected in Guangdong province, China, in November 2002, and control measures were instituted on the basis of the way the disease spread from person to person. In February and Example: Avian Influenza in Thailand March 2003, the disease spread to Hong Kong (China) and The disastrous pandemic (worldwide epidemic) of influenza in then to Vietnam, Singapore, Canada, and elsewhere (WHO 1918 is thought to have originated from epidemics in birds, as 2003b). This new disease was named severe acute respiratory were the influenza pandemics of 1957 and 1968 (Ungchusak syndrome, and a preliminary case definition was established on and others 2005). In early 2004, large epidemics of avian the basis of initial epidemiologic investigations. A novel coro- influenza were recognized in birds in eight Asian countries; by navirus (SARS-CoV) was identified as the causative agent in November, the disease had spread from birds to 44 humans, March, and mapping of the full genome was completed in 73 percent of whom died (Ungchusak and others 2005). This April. This global pandemic ended in July 2003, as transmission contagion sparked fears that the highly lethal avian virus might was interrupted in Taiwan (China), after more than 8,000 be adapting to spread from person to person, which could patients in 26 countries and five continents were affected and cause extensive health and economic damage around the 774 deaths were confirmed (Peiris and others 2003). world. In Thailand, avian influenza was investigated by FETP WHO spearheaded the global effort to control this pan- graduates and others in the Thai Ministry of Health in part- demic, working with national and subnational health workers. nership with CDC. By applying field epidemiologic techniques In China, the FETP, which was initiated in October 2001 in the supported by laboratory studies, they detected that the virus China Center for Disease Control, mobilized all 20 of its was being spread from human to human in a family. It is likely trainees, and they contributed substantially to the surveillance, that person-to-person transmission may have occurred in investigation, and control of the SARS outbreak, working with other countries, where field epidemiology was not used. local health officials (CDC 2003a). In Canada, which had the The Thai example is important for achieving the following: most cases of SARS outside Asia, 8 of the 10 FETP residents (a) raising global awareness of the potential of a global catas- were involved in the SARS outbreak. They instituted surveil- trophe early enough that plans can be made to avert or decrease lance, conducted epidemiologic investigations, designed pre- harm and (b) demonstrating that, as with SARS, the disease vention and control guidelines, responded to inquiries from could be controlled with proven field epidemiologic methods the media and the public, and planned and implemented supplemented by good communications, without vaccines, Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1005 drugs, or a high-technology laboratory or surveillance system The Ugandan Ministry of Health invested in developing (Mack 2005). competent, motivated health workers through the PHSWOW, an active partnership with Makerere University, the Rockefeller Example: Ebola in Uganda, the Role of the PHSWOW Foundation, CDC, and WHO. Both students and graduates On October 8, 2000, a second-year student in the Ugandan contributed to the ministry's ability to rapidly identify and PHSWOW returned to Gulu district in northern Uganda for control this dangerous epidemic. Because the minister had his field project. He found a hospital jammed with patients timely evidence, he was able to notify other countries quickly with high fevers, diarrhea, and bleeding. He diagnosed viral and to bring in international teams before the disease spread hemorrhagic fever. He called the Ministry of Health in further. Partially because of the lessons learned from this epi- Kampala, where that weekend a graduate of the PHSWOW was demic, Uganda has become one of the leading countries in in charge of taking calls about epidemics. She agreed with implementing the IDSR program. his diagnosis and arranged for samples to be rushed to the National Institute for Virology in South Africa, the nearest WHO reference center for viral hemorrhagic fevers. When the SURVEILLANCE FOR SPECIFIC CONDITIONS minister of health arrived at his office the next day, the gradu- ate briefed him. Recognizing the gravity of the situation, the Surveillance systems are important tools for targeting, moni- minister sent the graduate to head the public health team sur- toring, and evaluating many health risks and interventions. veillance and control team in Gulu, and the student headed the Because managers need a wide variety of information for spe- clinical team that established infection control in hospitals and cific interventions, systems have been developed and tested to treated patients. meet those needs. Laboratory tests quickly confirmed that the illness was Ebola hemorrhagic fever, which usually kills more than 50 per- cent of those infected (Heymann 2004). Public health surveil- Environmental Public Health Surveillance lance was difficult for several reasons. Because the disease was Surveillance for environmental public health practice requires severe and rapidly fatal, rural villagers feared that they might be the collection, analysis, and dissemination of data on hazards, stigmatized if the government knew about cases in their area. exposures, and health outcomes (figure 53.4; Thacker and Some sought out traditional healers; others fled as soon as they others 1996). realized they had been exposed, which prompted outbreaks in Health outcomes of relevance include death, disease, two other districts. Gulu was a politically unstable area, and injury, and disability. However, relating those outcomes to some villages were difficult to reach because of rebel or bandit specific environmental hazards and exposures is critical to activity. The Ugandan government mobilized its military to help with case finding and invited WHO, CDC, and other international teams to assist. Patients with Ebola infection Agent is a hazard require intense nursing and medical attention to control bleed- ing, diarrhea, and fevers. Some patients bleed easily, and all Agent is present in environment their secretions can be highly infectious. Hospitals in Gulu were Hazard desperately short of supplies to control the spread of infection surveillance from so many patients simultaneously. In spite of this situation, Route of exposure exists Ugandan health workers selflessly cared for the sick. By January 23, 2001, a total of 425 cases had occurred, the largest Host is exposed to agent Ebola outbreak recorded. Only 53 percent of the patients had died, a proportion far less than the 88 percent reported in the Exposure 1976 Ebola outbreak in the Democratic Republic of Congo (for- Agent reaches target tissue surveillance merly Zaire) and other previous epidemics (WHO Report of an International Commission 1978). Sadly, 22 health care workers Agent produces adverse effect were infected. Because the team from the Ugandan Ministry of Health set up active surveillance nationwide, the other two Outcome outbreaks, started when infected Gulu residents fled to distant Adverse effect becomes clinically apparent surveillance villages, were quickly detected and controlled. International Source: Thacker and others 1996. observers commented, "National notification and surveillance efforts led to the rapid identification of these foci and to effec- Figure 53.4 The Process of Adverse Effects and the tive containment" (CDC 2001). Corresponding Surveillance 1006 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others environmental public health surveillance. Hazards include 5 million persons each year and causing high rates of disability. toxic chemical agents, physical agents, biomechanical stressors, People from all economic groups are at risk for injuries, but and biologic agents that are located in air, water, soil, food, and death rates caused by injury tend to be higher in developing other environmental media. Exposure surveillance is the mon- countries (Peden, McGee, and Sharma 2002). Injury surveil- itoring of members of the population for the presence of an lance includes monitoring the incidence, causes, and circum- environmental agent, its metabolites, or its clinically inappar- stances of fatal and nonfatal injuries. Injuries are classified by the ent (for example, subclinical or preclinical) effects. intention of the act into two groups: unintentional injuries and Four challenges complicate environmental public health violence-related injuries.WHO (Holder, Peden, and Krug 2001) surveillance. First, the ability to link specific environmental and the Pan American Health Organization (Concha-Eastman causes to adverse outcomes is limited by our poor understand- and Villveces 2001) have developed guidelines for establishing ing of disease processes, long lead times, inadequate measures injury surveillance systems in developing countries. of exposure, and multiple potential causes of disease. Second, If the range of fatal and nonfatal injuries, as well as the risk data collected for other purposes rarely include sufficient infor- factors that can lead to injury, are to be fully captured, surveil- mation to meet a case definition for a condition caused by lance systems need to be established in multiple settings. Fatal an environmental agent. Third, public alarm is often out of injuries can be captured by using forensic or death certificate proportion to the hazard of concern, and sentiment will often data. A far greater number of injuries are nonfatal and can influence public policy disproportionately to scientific infor- be tracked through hospital- or primary care­based systems. mation. Fourth, biologic markers will become increasingly Systematic information on nonfatal injuries, including preva- critical elements of environmental exposure surveillance. lence, incidence, and related risk behaviors can also be obtained Obtaining data on exposure, which can include estimates through ongoing population-based surveys. derived from hazard data through sophisticated modeling Critical points should be addressed when planning an injury or direct measurements of individual exposure obtained from surveillance system in a developing country. First, data sources use of personal monitors (for example, passive air samplers), need to be clarified. In some developing countries, routine data is generally impractical in developing countries. Childhood on injuries are not always captured in health information blood lead levels are the only biomonitoring data that are col- systems. It is therefore necessary to consider other sources of lected routinely in several countries, either in national surveys data--for example, law enforcement agencies, coroners, or or from screening programs for children at high risk. medical examiners. Next, the events and variables in an injury Health outcome surveillance as applied to environmental surveillance system should be defined according to the objec- public health is similar to traditional surveillance efforts. In the tives of the system. Criteria such as the intentionality (violence- United States, the focus is on surveillance for birth defects; related injuries versus unintentional injuries); the outcome developmental disabilities (for example, cerebral palsy, autism, (fatal injuries versus nonfatal injuries); and the nature of and mental retardation); asthma and other chronic respiratory violence-related injuries (physical, sexual, psychological, depri- diseases (for example, bronchitis and emphysema); cancer; and vation, or neglect) should be considered when establishing the neurological diseases (for example, Parkinson's disease, multi- system. Finally, case definitions and coding procedures should ple sclerosis, and Alzheimer's disease) (McGeehin, Qualters, and be defined before implementing the system. Niskar 2004). Other nations have different sets of priority con- For example, the Nicaraguan Ministry of Health, in collab- ditions for surveillance. Disease registries, vital statistics data, oration with CDC and the Pan American Health Organization, annual health surveys, and administrative data systems (for began developing and implementing an injury surveillance example, hospital discharge data) are sources that have been system in 2001 (Clavel-Arcas, Chacon, and Concha-Eastman used for monitoring health conditions. The challenges men- 2004). The system, based on the medical facility emergency tioned previously have constrained our ability in all nations, department (ED), collects data on injuries in keeping with the regardless of level of development, to establish and maintain Injury Surveillance Guidelines established by WHO (Holder, effective and comprehensive environmental public health sur- Peden, and Krug 2001). Under the system, a reportable case is veillance systems. As we invest in understanding the enlarging defined as a patient who died from or was treated for an injury threats in the global environment, we must overcome these in the ED. Cases include patients with unintentional and vio- challenges and establish improved surveillance systems. The lence-related injuries. health of the global community depends on this investment. ED staff members identify cases and collect data in five hospitals in Nicaragua. Information used to complete the instrument is collected directly from the patients or their Injury Surveillance representatives. An ED admission clerk collects basic demo- Injuries are a major public health problem and are among the graphic data on the patient's arrival. ED medical staff members 10 leading causes of death worldwide, killing an estimated (physicians and nurses) collect the remaining information Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1007 (for example, location, mechanism of injury, nature, severity, discussed elsewhere in this chapter. These core surveillance and circumstances surrounding the injury) during triage and tools should be robust before new data types can be considered assessment. for supplementing public health surveillance. The hospital epidemiologist collects data collection forms Syndromic surveillance is an investigational approach by daily from the ED, reviews the quality of data, and requests data which health department staff members, assisted by automated from the ED staff if the forms are incomplete. The statistician data acquisition and generation of statistical signals (comput- reviews data daily. The country project coordinator also moni- erized algorithms), monitor disease indicators continually to tors the quality of the data periodically. Using Epi Info 2002 detect outbreaks of disease earlier and more completely than programs developed specifically for this project, the project might otherwise be possible with traditional reportable disease coordinators analyze trends and identify potential risk factors methods (Buehler and others 2004). (Noe and others 2004). The information is used to produce CDC's list of biologic terrorism agents and diseases can be monthly reports for dissemination. Information is reported at found at http://www.bt.cdc.gov and an updated list of refer- both the regional and the country levels. ences dealing with syndromic surveillance is at http://www.cdc. Injury prevention programs in Nicaragua use surveillance gov/epo/dphsi/syndromic/. data to assess the need for new policies or programs and to evaluate the effectiveness of existing policies and programs. For example, the municipality of León is using the information Complex Emergency Surveillance from the hospital to monitor the increase in suicide attempts The key elements in planning a disaster surveillance system are among youths abusing pesticides and to evaluate an intersec- establishing objectives, developing case definitions, determin- toral campaign to promote life that includes primary through ing data sources, developing simple data collection instru- tertiary prevention strategies. ments, field testing the methods, developing and testing the analysis strategy, developing a dissemination plan for the report or results, and assessing the usefulness of the system. Surveillance for Biologic Terrorism The surveillance needs are different in the preimpact, impact, Surveillance for biologic terrorism is conducted primarily for and postimpact phases (Binder and Sanderson 1987). outbreak detection and management. Surveillance must sup- The role of surveillance in disaster situations has included port early detection of an incident of biologic terrorism and its the following broad framework of activities: characterization in the same manner as for the detection and control of naturally occurring outbreaks of infectious diseases. · predisaster activities (for example, hazard mapping, provi- Early detection of outbreaks can be achieved by the following sion of guidelines, and training for medical and rescue (Buehler and others 2004): teams) · continuous monitoring and surveillance for priority health · timely and complete receipt, review, and investigation of problems in affected populations (for example, in the post- disease case reports, including the prompt recognition and tsunami surveillance in Tamil Nadu, India, a one-page reporting to or consultation with health departments by instrument was used for 10 priority health conditions for physicians, health care facilities, and laboratories daily active surveillance in displaced populations at camps) · improvement of the ability to recognize patterns indicative · prospective surveillance of affected populations focusing on of a possible outbreak early in its course (for example, by the natural history of exposure and health effects and long- using analytic tools that improve the predictive value of data term effects of stress disorders among survivors. at an early stage of an outbreak or by lowering the threshold for investigating possible outbreaks) · receipt of new types of data (such as purchases of health Surveillance in Refugee Populations care products, absences from work or school, symptoms Support of relief efforts following national and global disasters presented to a health care provider, or orders for laboratory has been a relatively new application of epidemiologic practice tests) that can signify an outbreak earlier in its course. for the public health professionals. Nevertheless, since the ini- tial CDC involvement with the United Nations in a large-scale Environmental detection systems for microbial pathogens relief effort concerning approximately 20 million displaced and toxins of concern for biologic terrorism might also be people affected by the 1967­70 civil war in Nigeria, CDC staff categorized as new types of data early in the course of an out- members have participated in several assessments of the health break, before infection (Meehan and others 2004). The primary needs, damage, and nutrition in refugee populations resulting surveillance tools for event detection and management are from man-made and natural disasters. The more notable and the traditional disease-reporting systems for notifiable diseases extended actions were conducted in the 1979­82 Khmer 1008 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others Thailand-Cambodia refugee-relief action, followed by long- cost effectiveness of interventions and strategies. In all these term public health surveillance of Somalian refugees areas, consideration of noncommunicable (mostly chronic) (1980­83), periodic but comprehensive health and nutritional conditions becomes critical. In 1999, noncommunicable assessments of Afghan refugees in Pakistan (1980­2002), and diseases were estimated to cause approximately 60 percent of growth and nutritional assessments of internally displaced the deaths in the world and 43 percent of the global burden of populations--especially children--in the Democratic People's disease (WHO 2000a). WHO forecasts that by 2020 the burden Republic of Korea (1990s) and southern Sudan. Although these of disease from noncommunicable diseases for developing and relief efforts occurred many years and many thousands of miles newly industrialized countries will have increased more than apart, they shared several important characteristics: 60 percent (Murray and Lopez 1996). Some developing countries have found it difficult to acquire · Large numbers of people were in fixed camps or on the and analyze accurate mortality statistics regularly, let alone move searching for food and shelter. These needs were usu- morbidity and quality-of-life information. Ensuring develop- ally addressed by external aid agencies and many times ment, implementation, and widespread use of noncommuni- caused local environmental degradation (fuel, temporary cable disease data for better decisions on resource allocation is housing, water pollution, and so on). critical to improving the quality of lives and promoting a more · Refugees, after the initial phase, competed with indigenous equitable future for health within and between countries. populations for scarce jobs, leading to social strife and Hypertension, elevated blood cholesterol, tobacco use, stress. Refugees were also exploited and suffered violence-- excessive alcohol consumption, obesity, and the multiple dis- additional factors leading to stress and social maladjust- eases linked to these risk factors are a global public health prob- ment. lem. In one study, smoking, high blood pressure, and high · No administrative structure to provide and coordinate cholesterol alone explained approximately two-thirds to three- assistance of the necessary magnitude existed before the fourths of heart attacks and strokes (Vartiainen and others crisis, and thus, it had to be created after the fact. 1995). Until recently, surveillance for risk factors was an activ- · Assistance was complicated by the uncertainty associated ity commonly associated with developed countries (Holtzman with military activity, crime, and hostile governments. 2003). However, recently WHO has increased attention to non- · Data that were relatively simple to gather and analyze communicable disease surveillance by developing tools and provided health workers and administrators information working to achieve data comparability between countries needed to plan and monitor assistance and its impact. (WHO 2003c). Data on key health behaviors, obesity, hyper- · Close collaboration with other local and international tension, lipids, and diabetes are collected inconsistently in relief organizations (such as the United Nations High developing countries, especially in Africa. Data on tobacco use Commissioner for Refugees, the International Red Cross, are available through the Global Youth Tobacco Survey the United Nations Children's Fund, WHO, and USAID) (http://www.cdc.gov/tobacco/global). was essential to instituting and sustaining a meaningful sur- Incidence data (the number and proportion of new cases in veillance system for refugees that led to interventions. a population) are limited in developing countries. However, India's National Cancer Registry program may serve as a notable The major goal of these activities is to identify and eliminate exception (http://icmr.nic.in/ncrp/cancer_regoverview.htm). preventable causes of morbidity and mortality. Planning In 1981, the Indian Council of Medical Research, recognizing requires effective use of existing knowledge about characteris- that there was a lack of information on follow-up of cancer tic or predictable demographic patterns, easily applied health patients to assess quality of care, instituted a cancer registry net- indicators, and avoidable errors of omission or commission. As work.The network provides data on the magnitude and patterns in disasters, the principles of surveillance (data collection, data of cancer in eight areas of India to enable studies of the histo- analysis, response to data, and assessment of response) and logic features correlating with prognosis and association studies other public health techniques should be an integral part of (for example, whether a history of vasectomy is associated with relief efforts. Retrospective evaluation of these efforts has also cancer of the prostate). Another important example relates to proved useful (CDC 1983). the widespread use of folic acid in China and the resultant reduction in incidence of birth defects (Kelly and others 1996; Wald 2004). Chronic Disease Surveillance Systems Surveillance data have been critical in establishing the Development and evaluation of policies for health improve- importance of obesity as a public health priority in the United ment require a reliable assessment of the burden of disease and States. Data for individual states provided by CDC's BRFSS injury, an inventory of the disposition of resources for health, have enabled individual health departments to document their assessment of the policy environment, and information on the obesity epidemic (Sturm 2003). These data provide a measure Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1009 of the effectiveness of interventions to meet the control Table 53.3 Health Expenditures by National Income Level of objectives. The BRFSS is a practical tool for developing and Countries, 2001 middle-income countries, as Jordan demonstrated when it Government expenditures Total expenditures implemented a BRFSS in 2002; the first survey documented on health as a on health as a substantial levels of obesity, especially among women, com- percentage of gross percentage of gross bined with low levels of physical activity (CDC 2003b). Income groupa domestic product domestic product High income 6.30 10.74 Upper-middle income 3.68 6.41 ECONOMICS OF PUBLIC HEALTH Lower-middle income 2.58 5.63 SURVEILLANCE SYSTEMS Low income 1.22 4.78 The outbreak of SARS in 2003 demonstrates the far-reaching Source: World Bank 2004. economic impact of not having an effective global public a. All World Bank member economies with populations of more than 30,000 are classified into income groups, divided according to 2003 gross national income per capita, calculated using the health surveillance system in place, with an estimated reduc- World Bank Atlas method. The groups are low income, US$765 or less; lower-middle income, tion in real gross domestic product of more than US$1.0 billion US$766 to US$3,035; upper-middle income, US$3,036 to US$9,385; and high income, US$9,386 or more. in Canada (Darby 2003) and estimated income losses in the range of US$12.3 billion to US$28.4 billion for East and Southeast Asia as a whole (Fan 2003). Public health surveillance is considered a global public good are reportedly the weak link in the global surveillance frame- (Zacher 1999), particularly when it is used for eradication of work, although they bear the greatest burden of disease, emerg- such diseases as poliomyelitis. As eradication campaigns ing and reemerging old pathogens, and drug-resistant decrease the number of cases, maintaining systems to find the pathogens (U.S. GAO 2001). The greatest need for surveillance last few cases becomes more expensive. Often, the majority of systems is in these countries, but most lack both the resources the costs for these systems fall on hard-pressed developing and the political will to build human capacity and finance the countries. This factor raises questions of fairness and equity. systems (table 53.3). Resource constraints and intense pressure For example, as poliomyelitis becomes rare, it ceases to be a sig- to provide care and treatment services lead public health nificant risk to national populations, whereas other diseases, authorities in the poorest countries to spend resources on such as malaria and diarrhea, typically are major causes of surveillance (U.S. GAO 2001). Because the costs and benefits morbidity and mortality. In such countries, it seems most fair derive from surveillance systems spilling across national bor- and efficient for the global community to finance eradication ders, donors should assist with capacity building in countries campaigns, leaving national systems free to address the diseases that have been unable to invest the human and material that most affect their populations. The negative impact of resources required. globally mandated eradication surveillance systems can be An interesting and unresolved feature of these global public mediated or reversed by leveraging on the eradication pro- goods--the solution to their adequate provision and supply gram's infrastructure to gather surveillance data for diseases of rests at local, national, and sometimes regional levels--has concern to local governments (Nsubuga, McDonnell, and prompted the international health community to advocate for others 2002). A similar case can be made for influenza early capacity building in developing countries rather than for con- warning systems in countries that gather information that will solidation of the fragmented systems at the global level (WHO be used to create vaccines that will benefit other populations 2002). but not their own. Equity demands that the countries that Standard tools of economic evaluation (Meltzer 2001) have benefit from such systems finance them. been used to compare the benefits and costs of several public Public health surveillance systems serve an essential func- health interventions. The public good characteristics of surveil- tion in preventing and controlling disease spread within and lance systems, with benefits that are not easy to quantify, make across national borders. Although the private sector benefits, the use of such tools difficult to implement in practice. However, it lacks the incentive to invest in public health surveillance economic evaluation of laboratory surveillance systems to systems, and sovereign states depend on the contribution of detect specific disease-causing organisms have been undertaken others (WHO 2002); this situation has important implications in the developed world by comparing benefits and costs now for the financing of public health surveillance systems. Even and in the future (Elbasha, Fitzsimmons, and Meltzer 2000). within national borders, the difficulty of quantifying the bene- These evaluations have not been done in developing countries fits of surveillance systems for individual communities leads to and are needed. At best, an analysis of the benefits and costs of neglect by local authorities, providing the economic rationale existing or proposed surveillance systems is feasible. This analy- for funding by the national government. Developing countries sis requires an estimate of the cost of illness and answers the 1010 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others question of how many cases of a particular disease need to be Analysis of surveillance data can also be transformed by prevented by the surveillance system to be exactly equal to the using available technology. Software that is Web-enabled, expenditure on the system. together with the advances in geographic information system Given expenditures on specific health interventions or pro- software and global positioning devices, means that anyone grams, one can, by using traditional econometric tools, apply with Internet access can potentially apply the latest version of the data on health outcomes from the surveillance systems as software running on a distant server in the national capital to inputs to economic analysis. Surveillance also clearly leads to a local data to generate up-to-date maps and graphs describing cost saving if it prevents the need for expenditure on treating health status in that jurisdiction. patients. Use of surveillance data can also be transformed. Sophisticated algorithms can be applied to data as it is collected to determine when (and how) an alert should be sent to local, FUTURE OF SURVEILLANCE national, or even international health officials to indicate a need for immediate investigation. Increasingly sophisticated Public health agencies, ministries of finance, and international visual display techniques and creation of custom channels with donors and organizations need to transform surveillance from data of particular relevance to groups of data users are just dusty archives of laboriously collected after-the-fact statistics some of the tools already being used to put public health con- to meaningful measures that provide accountability for local tent on the desktop of anyone with broadband, secure Internet health status or that deliver real-time early warnings for devas- access. tating outbreaks. This future depends in part on developing Realization of this future vision does not require technology consensus on critical surveillance content and developing beyond what is already feasible, but the following factors are commitment on the part of countries, funding partners, and needed: multilateral organizations to invest in surveillance system infra- structure and to use surveillance data as the basis for decision · the organizational and political will to develop and coordi- making. This vision of the future assumes a coherent, integrated nate the needed systems and standards that will enable those approach to surveillance systems that is based on matching the systems surveillance objective with the right data source and modality · appropriate attention to individual privacy and system and on paying attention to country-specific circumstances security while maintaining global attention to data content needs. · removal of the financial and logistical barriers to broadband Information technology and informatics can help in attain- Internet access ing this vision. Specifically, technology can facilitate the collec- · a strategic multisectoral approach to accelerating national tion, analysis, and use of surveillance data, if data standards are infrastructure among the poorest developing nations. developed and compatible systems are established. Data collec- tion for surveillance would be an automatic by-product of any electronic systems used to support clinical care. Under this sce- GLOBAL SURVEILLANCE NETWORKS nario, an automatic electronic message would be sent to the responsible public health jurisdiction with information about Globally, infectious disease surveillance is implemented a health event (for example, death, disease, or injury), includ- through a loose network that links parts of national health care ing all relevant information from the electronic health record systems with the media, health organizations, laboratories, and about the patient, provider's name, patient's home address, risk institutions focusing on particular disease conditions. WHO factors, previous immunizations, and treatments. Even before has described a "network of networks" (U.S. GAO 2001) that this ideal capacity becomes widespread, technology such as cell links existing regional, national, and international networks of phone­based systems could accelerate collection of key data laboratories and medical centers into a surveillance network (for example, occurrence of a viral hemorrhagic fever out- (figure 53.5). break). The rapid penetration of cell phones in developing Government centers of excellence (for example, CDC, the countries might obviate the need for prohibitively expensive French Pasteur Institutes, and FETPs) along with WHO coun- landline-based systems. An accelerated system of wireless try and regional offices also contribute to disease and health Internet access might also transform the capacities to which a condition reporting. Military networks, such as the U.S. local health post or a district health official might have access. Department of Defense's Global Emerging Infectious Disease These systems should also be considered as means for collect- System, and Internet discussion sites, such as ProMed ing information beyond traditional data. For example, (http://www.promedmail.org) and Epi-X (http://www.cdc.gov/ telemedicine access can permit views of a rash illness to be epix), also supplement the reporting networks. In 1997, WHO shared with national or international medical specialists. started the Global Outbreak Alert and Response Network, and Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1011 WHO collaborating centers and UNHCR laboratories Epidemiology and UNICEF training country networks offices National Military public laboratory health networks authorities WHO Global regional and Public Health country Intelligence offices Network Internet Non- discussion governmental sites organizations Media press Formal Informal Source: U.S. GAO 2001. a. UNHCR represents the United Nations High Commissioner for Refugees. b. UNICEF represents the United Nations Children's Fund. Figure 53.5 Global Infectious Disease Surveillance Frameworks it was formally adopted by WHO member states in 2000. The capacity of both epidemiologists and laboratorians to collect, network has more than 120 partners around the world and use, and interpret surveillance and outbreak data (for example, identifies and responds to more than 50 outbreaks in develop- the collaborative WHO program in foodborne diseases called ing countries each year (Heymann and Rodier 2004). the WHO Global Salm-Surv) are also important components The International Health Regulations are the only binding in developing global surveillance networks. international agreements on disease control. The regulations provide a framework for preventing the international spread of RESEARCH AGENDA IN PUBLIC HEALTH disease through effective national surveillance coupled with the SURVEILLANCE international coordination of response to public health emer- gencies of global concern by using the guiding principle of Developing nations share surveillance needs with the rest of the maximum protection, minimum restriction (WHO 2003a). world, yet they are challenged by economic limitations, weak The current regulations apply only to cholera, plague, and yel- public health infrastructure, and the overwhelming challenges low fever; they require WHO member states to notify WHO of of poverty and disease. As a result, countries in the developing any cases of these diseases that occur in humans within their world often depend on the research efforts of others, or they territories and then give further notification when the territory collaborate with others to conduct the research necessary for is free of infection. The regulations are being revised to include their surveillance needs. Within individual countries, surveil- the development of national core capacities and national focal lance systems are essential in measuring disease and injury bur- persons who have the competencies of graduates of FETPs and den as a first step in establishing public health priorities that allied training programs. Programs established to improve the lead to policies and programs. 1012 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others The major research question for surveillance is how to · conduct surveillance for multiple competing risk factors develop and maintain a cadre of competent, motivated surveil- that lead to a single condition (for example, smoking, cho- lance and response workers in developing countries. Other lesterol, hypertension, and overweight for heart disease) questions include how to design and maintain surveillance sys- · conduct surveillance for the adverse effects of drugs tems for these problems, especially morbidity systems for · interpret ecologic data relative to data collected on individ- chronic diseases. Standard methods can be used to evaluate uals (Greenland 2004) existing surveillance systems, which, in turn, will help define · measure cost-effectiveness of alternative approaches to sur- surveillance needs (Romaguera, German, and Klaucke 2000). veillance (for example, integrated compared with categori- Developing countries have used the IDSR strategy, which pro- cal approaches) vides an efficient approach to data collection and analysis. · link data sources effectively (for example, hazard, exposure, Unfortunately, the majority of developing countries have lim- and outcome data for environmental diseases) ited surveillance systems for noninfectious diseases; instead, · build and sustain human infrastructure in developing existing data systems (for example, vital records, motor vehicle countries crash records, or insurance claims data) are potential sources of · strengthen evidence-based decision-making cultures in min- surveillance data. In other settings, even these data sources are istries of health and finance. scarce, and approaches such as verbal autopsies and recurrent surveys might be alternatives (White and McDonnell 2000). Surveillance for risk factors is another challenge, and CONCLUSION BRFSSs need to be validated and applied more widely in devel- oping countries. Surveillance for injuries, environmental haz- Public health surveillance is an essential tool for ministries of ards (such as traffic intersections that are associated with high finance, ministries of health, and donors to effectively and rates of injuries), and exposures to chemical or biological efficiently allocate resources and manage public health inter- agents is a key public health concern with few examples of ventions. To be useful, public health surveillance must be effective application anywhere in the developed or less devel- approached as a scientific enterprise, applying rigorous meth- oped parts of the world. Rigorous research is required in this ods to address critical concerns in this public health practice field (Thacker and others 1996). (Thacker, Berkelman, and Stroup 1989). Although the surveil- The burgeoning use of electronic data systems and the lance needs in the developing world appear to differ from those almost universal availability of the Internet provide a tremen- in the developed world, the basic problems are similar. In a time dous opportunity for more timely and comprehensive surveil- when we are confronted with SARS and avian influenza, the lance in all parts of the world. Yet in this rapidly emerging field, need to integrate global networks is undeniable, and research in critical needs exist, including the following: how these concerns are addressed is essential. Collaboration among practitioners, researchers, nations, and international · competent, motivated health workers organizations is necessary to address the global needs of public · data standards (Lober, Trigg, and Karras 2004) health surveillance. · global policies and practices for international surveillance · useful software (Dean 2000) · evaluation of the effectiveness of all these applications. ACKNOWLEDGMENTS New approaches that must be evaluated by using standard The authors would like to acknowledge ORISE Fellow Danielle methods (Romaguera, German, and Klaucke 2000) include the Backes of the Division of International Health, Coordinating following: Office for Global Health, U.S. Centers for Disease Control and · IDSR for infectious diseases Prevention, for her invaluable help in preparing this chapter. · syndromic surveillance (CDC 2004b) for terrorism and emergency response · laboratory-based surveillance methods to enhance diagnos- REFERENCES tic accuracy and increase timeliness of recognition of out- Binder, S., and L. M. Sanderson. 1987. "The Role of the Epidemiologist in breaks and interventions (Swaminathan and others 2001). Natural Disasters." Annals of Emergency Medicine 16 (9): 1081­84. Many research questions remain about surveillance Birkhead, G. S., and C. M. Maylahn. 2000. "State and Local Public Health Surveillance." In Principles and Practices of Public Health Surveillance, methodology, including how to do the following: ed. S. M. Teutsch and R. E. Churchill, 270. New York: Oxford University Press. · use data for forecasting or temporal and spatial analysis for Buehler, J. W., R. S. Hopkins, J. M. Overhage, D. M. Sosin, and V. Tong. aberration detection 2004. "Framework for Evaluating Public Health Surveillance Systems Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1013 for Early Detection of Outbreaks: Recommendations from CDC Holtzman, D. 2003. "Analysis and Interpretation of Data from the U.S. Working Group." MMWR Recommendations and Reports 53 (RR-5): Behavioral Risk Factor Surveillance System (BRFSS)." In Global 1­11. Behavioral Risk Factor Surveillance, ed. D. V. McQueen and P. Puska. CDC (U.S. Centers for Disease Control and Prevention). 1983. New York: Kluwer Academic Press. "Surveillance of Health Status of Kampuchean Refugees--Khao Kelly, A. E., A. C. Haddix, K. S. Scanlon, C. G. Helmick, and J. Mulinare. I-Dang Holding Center, Thailand, December 1981­June 1983." 1996. "Cost Effectiveness of Strategies to Prevent Neural Tube Morbidity and Mortality Weekly Report 32 (31): 412­15. Defects." In Cost-Effectiveness in Health and Medicine, ed. M. R. Gold, ------. 2001. "Outbreak of Ebola Hemorrhagic Fever, Uganda, August J. E. Siegel, L. B. Russell, and M. C. Weinstein, 313­48. New York: 2000­January 2001." Morbidity and Mortality Weekly Report 50 (5): Oxford University Press. 73­77. Kew, O., V. Morris-Glasgow, M. Landaverde, C. Burns, J. Shaw, Z. Garib, ------. 2003a. "Efficiency of Quarantine during an Epidemic of Severe and others. 2002. "Outbreak of Poliomyelitis in Hispaniola Associated Acute Respiratory Syndrome--Beijing, China, 2003." Morbidity and with Circulating Type 1 Vaccine-Derived Poliovirus." Science 296 Mortality Weekly Report 52 (43): 1037­40. (5566): 356­59. ------. 2003b."Prevalence of Selected Risk Factors for Chronic Disease-- Lober, W. B., L. Trigg, and B. Karras. 2004. "Information System Jordan, 2002." Morbidity and Mortality Weekly Report 52 (43): 1042­44. Architectures for Syndromic Surveillance." Morbidity and Mortality Weekly Report 53 (Suppl.): 203­8. ------. 2004a. Partnerships in Excellence: Charting the Future in Global Health. Atlanta: CDC, Epidemiology Program Office, Division of Mack, T. M. 2005. "The Ghost of Pandemics Past." Lancet 365 (9468): International Health. 1370­72. ------. 2004b. "Syndromic Surveillance: Reports from a National McGeehin, M. A., J. R. Qualters, and A. S. Niskar. 2004. "National Conference, 2003." Morbidity and Mortality Weekly Report 53 (Suppl.): Environmental Public Health Tracking Program: Bridging the 18­22. Information Gap." Environmental Health Perspectives 112 (14): 1409­13. Clavel-Arcas, C., R. Chacon, and A. Concha-Eastman. 2004. "Hospital Based Injury Surveillance Systems in Nicaragua and El Salvador, Meehan, P. J., N. E. Rosenstein, M. Gillen, R. F. Meyer, M. J. Kiefer, S. 2001­2002." Paper presented at the Seventh World Conference on Deitchman, and others. 2004."Responding to Detection of Aerosolized Injury Prevention and Safety Promotion, Vienna, June 6­9. Bacillus Anthracis by Autonomous Detection Systems in the Workplace." MMWR Recommendations and Reports 53 (RR-7): 1­12. Concha-Eastman, A., and A. Villveces. 2001. Guidelines for the Design, Implementation and Evaluation of Epidemiological Surveillance Systems Meltzer, M. I. 2001. "Introduction to Health Economics for Physicians." on Violence and Injuries. Washington, DC: Pan American Health Lancet 358 (9286): 993­98. Organization. Murray C. J. L., and A. D. Lopez, eds. 1996. Global Burden of Disease and Darby, Paul. 2003. "The Economic Impact of SARS." Publication 434-05, Injury Series, Volume I: The Global Burden of Disease--A Conference Board of Canada, Ottawa. Comprehensive Assessment of Mortality and Disability from Diseases, Injuries, and Risk Factors in 1990 and Projected to 2020. Cambridge, Dean, A. D. 2000. "Computerizing Public Health Surveillance Systems." In MA: Harvard University Press. Principles and Practices of Public Health Surveillance, ed. S. M. Teutsch and R. E. Churchill, 229­52. New York: Oxford University Press. Narasimhan, V., H. Brown, A. Pablos-Mendez, O. Adams, G. Dussault, G. Elzinga, and others. 2004. "Responding to the Global Human Effler, P., M. Ching-Lee, A. Bogard, M. C. Ieong, T. Nekomoto, and D. Resources Crisis." Lancet 363 (9419): 1469­72. Jernigan. 1999. "Statewide System of Electronic Notifiable Disease Reporting from Clinical Laboratories: Comparing Automated Noe, R., J. Rocha, C. Clavel-Arcas, C. Aleman, M. E. Gonzales, and C. Reporting with Conventional Methods." Journal of the American Mock. 2004. "Occupational Injuries Identified by an Emergency Medical Association 282: 1845­50. Department Based Injury Surveillance System in Nicaragua." Injury Prevention 10 (4): 227­32. Elbasha, E. H., T. D. Fitzsimmons, and M. I. Meltzer. 2000. "Costs and Benefits of a Subtype-Specific Surveillance System for Identifying Nsubuga, P., N. Eseko, W. Tadesse, N. Ndayimirije, C. Stella, and S. Escherichia coli O157:H7 Outbreaks." Emerging Infectious Diseases McNabb. 2002. "Structure and Performance of Infectious Disease 6 (3): 293­97. Surveillance and Response, United Republic of Tanzania, 1998." Bulletin of the World Health Organization 80 (3): 196­203. Fan, E. X. 2003. SARS: Economic Impacts and Implications. ERD Policy Brief Series 15. Manila: Asian Development Bank. Nsubuga, P., S. M. McDonnell, B. Perkins, R. Sutter, L. Quick, M. E. White, and others. 2002. "Polio Eradication Initiative in Africa: Influence on Foege, W. H., R. C. Hogan, and L. H. Newton. 1976. "Surveillance Projects Other Infectious Disease Surveillance Development." BMC Public for Selected Diseases." International Journal of Epidemiology 5 (1): Health 2 (1): 27. 29­37. Pappaioanou, M., M. Malison, K. Wilkins, B. Otto, R. A. Goodman, R. E. Greenland, S. 2004. "Ecologic Inference Problems in the Analysis of Churchill, and others. 2003. "Strengthening Capacity in Developing Surveillance Data." In Monitoring the Health of Populations, ed. R. Countries for Evidence-Based Public Health: The Data for Decision- Brookmeyer and D. F. Stroup, 315­40. New York: Oxford University Making Project." Social Science and Medicine 57 (10): 1925­37. Press. Peden, M., K. McGee, and G. Sharma. 2002. The Injury Chartbook: A Herikstad, H., Y. Motarjemi, and R. V. Tauxe. 2002. "Salmonella Graphical Overview of the Global Burden of Injuries. Geneva: World Surveillance: A Global Survey of Public Health Serotyping." Health Organization. Epidemiology and Infection 129 (1): 1­8. Peiris, J. S., K. Y. Yuen, A. D. Osterhaus, and K. Stohr. 2003. "The Severe Heymann, D. L. 2004. Control of Communicable Diseases Manual. Acute Respiratory Syndrome." New England Journal of Medicine Washington, DC: American Public Health Association. 349 (25): 2431­41. Heymann, D. L., and G. Rodier. 2004. "Global Surveillance, National Romaguera, R., R. R. German, and D. N. Klaucke. 2000. "Evaluating Public Surveillance, and SARS." Emerging Infectious Diseases 10 (2): 173­75. Health Surveillance." In Principles and Practices of Public Health Holder, Y., M. Peden, and E. Krug. 2001. Injury Surveillance Guidelines. Surveillance, ed. S. M. Teutsch and R. E. Churchill, 176­93. New York: Geneva: World Health Organization. Oxford University Press. 1014 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others Sturm, R. 2003."Increases in Clinically Severe Obesity in the United States, White, M. E., S. M. McDonnell, D. H. Werker, V. M. Cardenas, and S. B. 1986­2000." Archives of Internal Medicine 163 (18): 2146­48. Thacker. 2001. "Partnerships in International Applied Epidemiology Swaminathan, B., T. J. Barrett, S. B. Hunter, and R. V. Tauxe. 2001. Training and Service, 1975­2001." American Journal of Epidemiology "PulseNet: The Molecular Subtyping Network for Foodborne Bacterial 154 (11): 993­99. Disease Surveillance, United States." Emerging Infectious Diseases 7 (3): WHO (World Health Organization). 1993. "Epidemiological Surveillance 382­89. of Communicable Disease at the District Level." WHO Regional Thacker, S. B., and R. L. Berkelman. 1988. "Public Health Surveillance in Committee for Africa, 43rd session, AFR/RC43/18. WHO, Geneva. the United States." Epidemiologic Reviews 10: 164­90. ------. 1998. "Integrated Disease Surveillance: Regional Strategy for Thacker,S.B.,R.L.Berkelman,and D.F.Stroup.1989."The Science of Public Communicable Diseases." WHO Regional Committee for Africa, 48th Health Surveillance." Journal of Public Health Policy 10 (2): 187­203. session. AFR/RC48/8. WHO, Geneva. Thacker, S. B., and D. F. Stroup. 1998a. "Public Health Surveillance." In ------. 2000a. Global Strategy for the Prevention and Control of Applied Epidemiology: Theory to Practice, ed. R. C. Brownson and D. B. Noncommunicable Diseases. Report by the Director-General, 53rd Petitti, 105­35. New York: Oxford University Press. World Health Assembly. Provisional Agenda Items 2.11: A53/14. Geneva: WHO. ------. 1998b. "Public Health Surveillance and Health Services Research." In Epidemiology and Health Services, ed. H. K. Armenian and S. ------. 2000b. "An Integrated Approach to Communicable Disease Shapiro, 61­82. New York: Oxford University Press. Surveillance." Weekly Epidemiological Record 75 (1): 1­7. Thacker, S. B., D. F. Stroup, R. G. Parrish, and H. A. Anderson. 1996. ------. 2002. Global Public Goods for Health: The Report of Working Group "Surveillance in Environmental Public Health: Issues, Systems, and 2 of the Commission on Macroeconomics and Health. Geneva: WHO. Sources." American Journal of Public Health 86 (5): 633­38. ------. 2003a. Revision of the International Health Regulations World Ungchusak, K., P. Auewarakul, S. F. Dowell, R. Kitphati, W. Auwanit, Health Assembly. Resolution WHA 56.28. Geneva: WHO. P. Puthavathana, and others. 2005. "Probable Person-to-Person ------. 2003b. "Severe Acute Respiratory Syndrome (SARS)." Weekly Transmission of Avian Influenza A (H5N1)." New England Journal of Epidemiological Record 78 (12): 81­83. Medicine 352 (4): 333­40. ------. 2003c. The SURF Report 1. Surveillance of Risk Factors Related USAID (U.S. Agency for International Development). 2005. "Infectious to Noncommunicable Diseases: Current State of Global Data. Geneva: Disease and Response Strategy 2005." USAID, Washington, DC. WHO. U.S. GAO (U.S. General Accounting Office). 2001. Challenges in Improving WHO Report of an International Commission. 1978. "Ebola Infectious Disease Surveillance Systems. GAO-01-722. Washington, DC: Haemorrhagic Fever in Zaire, 1976." Bulletin of the World Health U.S. General Accounting Office. Organization 56 (2): 271­93. Vartiainen, E., C. Sarti, J. Tuomilehto, and K. Kuulasmaa. 1995. "Do World Bank. 2001. World Development Report 2000­2001: Attacking Changes in Cardiovascular Risk Factors Explain Changes in Mortality Poverty. New York: Oxford University Press. from Stroke in Finland?" British Medical Journal 310 (6984): 901­4. ------. 2004. World Development Indicators 2004. Washington, DC: Wald, N. J. 2004. "Folic Acid and the Prevention of Neural-Tube Defects." World Bank. http://devdata.worldbank.org/hnpstats/query/default. New England Journal of Medicine 350 (2): 101­3. html. White, M. E., and S. M. McDonnell. 2000. "Public Health Surveillance in Zacher, M. 1999. Global Epidemiological Surveillance: Global Public Goods. Low and Middle Income Countries." In Principles and Practices of New York: Oxford University Press. Public Health Surveillance, ed. S. M. Teutsch and R. E. Churchill, 287­315. New York: Oxford University Press. Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1015 Chapter 54 Information to Improve Decision Making for Health Sally K. Stansfield, Julia Walsh, Ndola Prata, and Timothy Evans The new source of power is not money in the hands of a few, but information in the hands of many. --John Naisbitt and Patricia Aburdene, Megatrends 2000 This chapter focuses on the collection and management of INTRODUCTION public health information, in contrast to clinical information, which concerns individual patient care encounters. Even From infancy on, we receive information that gives form to our when aggregated, clinical data are necessary, but not sufficient, thinking and problem solving. The method by which a phe- to inform efforts to improve the health of populations. While nomenon is measured shapes societal perceptions of it and the substantial attention has been focused on these facility-based collective efforts to affect it. Likewise, the choices we make in clinical consultations and the health management information the collection and use of information for health will determine system (HMIS) used to track the relevant data, we focus here our effectiveness in detecting problems, defining priorities, on the broader health information system (HIS) needed to identifying innovative solutions, and allocating resources for inform decisions at individual, facility, district, and national improved health outcomes. Despite those fundamental reali- levels. Considered here are the routine data collection systems ties, there has been little awareness to date of the ramifications upon which program management, planning, monitoring, that greater information use can have for advancing health, and and evaluation depend. Information needs for specific tasks, even less attention has been given to systems needed to provide such as for research or for program evaluation, are discussed timely, accurate, and relevant information. in the chapters on research (chapters 4 and 7). Other chapters An example of the formative power of information for pol- in this volume refer to information needs to enable disease icy change lies in the history of the United Nations' Standard control or to evaluate programs and improve the delivery of System of National Accounts, created by Richard Stone more interventions. Those interested in these issues should also pay than 50 years ago. The annual reporting of these accounts by special attention to chapter 53 and chapters 70­73. This most countries shapes our impressions of the relative position chapter bridges the global and the local issues; it makes the of nations, defines our views of the differential opportunities case for strengthening the evidence base for action through offered to their citizens, and drives the content of national and comprehensive health information systems that include cen- global political discourse (Jolly 2002). Another example is the sus, vital events, monitoring, public health surveillance, measurement of disability-adjusted life years (DALYs), which resource tracking, facility-based service statistics, and house- has shaped priorities for investment in global health over the hold surveys. past decade. 1017 However, data or information alone will not transform out- donor-driven and disease-specific initiatives have actually comes. Data, which are simple measures of characteristics of undermined efforts to develop a comprehensive HIS by creat- people and things, have little inherent meaning or value. ing separate, parallel, and often duplicative systems to meet the Analysis of the data enables the identification of patterns, need for each funding source. thereby creating information. Finally, the use of information to Health information and the systems for its supply are a pub- generate recommendations, rules for action, and behavior lic good, meeting the defining criteria of being nonexcludable change signifies the creation of knowledge that is used to make (in that, once the information is in the public domain, it is dif- decisions and change human behavior. ficult to withhold from users) and nonrival (in that consump- Good decisions on effective policies, services, and behaviors tion of the information does not lessen its availability for use by require timely, accurate, and relevant information. Health in- others).As a public good, the supply of health information is the formation is required for strategic planning and the setting of primary responsibility of governments: national governments priorities; clinical diagnosis and management of illness or for information within these jurisdictions, and international injury; quality assurance and quality improvement for health agencies and national governments together for international services; detection and control of emerging and endemic dis- comparative information and global summary data. ease; human resource management; procurement and manage- Harmonizing the data collection, standards, best practices, ment of health commodities (including drugs, vaccines, and and other elements of a national and global HIS has several diagnostics); regulation of toxic exposures; program evaluation; advantages. Standardization enables economies of scale for research; and other types of policies and programs (Walsh and training, hardware and software, and processes. Routine health Simonet 1995). Citizens require such information to choose information is a summative good in that the collation of each healthy behaviors, to demand effective policies and services, and contribution produces a cumulative increase in the value of the to hold their governments accountable for the allocation and public good, strengthening the credibility and importance of use of resources for health. Internationally, information is that information. Furthermore, standardization of systems required to meet transnational needs, such as for the detection improves the reliability and comparability of information, both and control of consequences of epidemics and infectious dis- within nations and across national and regional boundaries eases, results-based management of development assistance (Cibulskis and Hiawalyer 2002). programs, and advocacy for increased financing for health. Several recent trends further enhance the pressures to SYSTEMS AND SUPPLY OF HEALTH INFORMATION deliver better health information. Global epidemics, such as of severe acute respiratory syndrome (SARS) and "bird flu," have To create an effective HIS, governments must finance the sys- amply demonstrated the need and potential benefits of sensi- tem, create the necessary policy environment (for example, tive and transparent systems for tracking health events. through legislation and regulation), and develop systems and Donors, including the Global Fund and the Global Alliance for services for the collection, collation, dissemination, and use of Vaccines and Immunization (GAVI), increasingly demand health information. A substantial portion of the national health performance measures and detailed evidence to justify new information is fully within the control of government health requests for support. "Basket" funding and sectorwide ap- officials. However, information from the private health sector proaches place further responsibilities on countries to define and other parts of the government is also required. Table 54.1 their own priorities. Decentralization and devolution of budg- lists some of the data required and their sources. A principal etary controls have shifted much of this growing burden to challenge is the integration of these intra- and extrasectoral the periphery, requiring districts to provide local evidence as a functions into a single, comprehensive HIS. basis for decisions. Tracking progress toward the Millennium Development Goals for health requires empowering countries Direct Expenditures for Health Information to measure key indicators and produce evidence-based strate- As for most public goods, the production of health information gic plans to achieve and document that progress. Furthermore, is mostly financed by government appropriation. Budget sup- nearly every chapter in this volume cites the need for better port for the HIS comes through both the ministry of health information, including through research dependent on a and a national statistics office (NSO) in most countries. The health information system, to accelerate improvements in NSO is usually responsible for collecting information through health.1 the national census and most household surveys. For the least Yet there is a striking disconnect between the need for infor- developed countries especially, bilateral and multilateral mation and the ability to respond to that need. To collect, donors are essential sources of finance, particularly for HIS collate, analyze, and communicate the necessary information planning, infrastructure development, and training. In Africa, in a timely and understandable fashion requires organized it has been estimated that grants or loans from donors account processes and procedures and a comprehensive HIS. However, for between 20 and 70 percent of the financing for statistical 1018 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others Table 54.1 Health Information from Sources Outside in the management of health information. There should be pro- the Health Sector tection from political interference and full empowerment of the health statistics office to make public statements in response to Health information Responsible agency criticisms of reports and the underlying methods. Ethical prac- Census and national surveys: National statistical office tices for protecting privacy and confidentiality must be well Income and poverty distribution understood, and procedures should be in place to deal with Household expenditure for health breaches in these standards. Accuracy and reliability should be Coverage with health interventions stated as expectations and ensured through periodic review of National expenditures for health, economic Ministry of finance data collection methods and through benchmarking with inter- development indicators, and industrial nationally credible definitions of indicators. A client orientation production and distribution data should be instilled and users of data regularly consulted in Employment data: Ministry of labor defining outputs and formats for the presentation of data. Human resources for health Occupational health information Systems for Collection, Management, and Analysis Import data: Ministry of trade Most developing countries have no comprehensive strategy for Pharmaceuticals and vaccines information management, reflecting the fractal nature of Capital equipment and health commodities donor and national investments in these systems. Interventions Food production and security information and Ministry of agriculture to improve the HIS in the least developed countries, often nutritional status data donor driven, have often focused only on a specific subsystem, Military health service statistics Ministry of defense primarily for health service statistics, and have neglected other Patterns of transportation injury (including Ministry of transportation components of the HIS. motor vehicle accidents) An effective HIS requires an overarching architecture that Literacy rates and school health program Ministry of education defines the data elements, processes, and procedures for collec- information tion, collation, presentation, and use of information for deci- sion making throughout the health sector (see box 54.1). This Source: Authors. information architecture promotes comparison and integra- tion of data elements from a variety of subsystems. As O'Carroll systems overall. Revenue generated by selling statistical prod- (2003) points out, such a comprehensive design enables phased ucts and services accounts for 10 to 20 percent of the financing system development, reduces redundancy, increases efficiency, for national statistical systems (Economic Commission for and improves interoperability. Interoperability is critical to Africa 2003). User fees or taxes for use of information products ensuring, for example, that census data, vital statistics, and and services can partially offset the costs of developing and health facility data can be integrated to generate rates, ratios, maintaining the information system. In many countries, taxes cost-effectiveness estimates, and other information required to and tariffs on computer equipment and government regulation compare options for health investment. of communications and Internet use remain barriers to public The Pan American Health Organization (PAHO) has led the access to health information. Cost must not be a barrier to use Regional Core Health Data Initiative "to facilitate speedy access of health information for the public good. to basic information on the health situation in the countries of the Region." This initiative has involved an international con- Information Policy sultation and agreement on the priority data, collection meth- Sound information systems require a legislative and regulatory ods, and indicators. The initiative has shown that it is possible environment that encourages and supports effective HIS to create a regional database of essential, consistent, valid, stan- development. At the global level, many efforts have been made dardized, timely, and regular information. PAHO has used the to establish international standards and policy frameworks for information to set its priorities, whereas countries have applied statistical data (United Nations Statistical Commission 1994). the results to design health programs and to allocate resources These policy frameworks are used to establish mandates for to upgrade their information systems. In the future, the plan is collection of basic health data (such as a decennial census or to expand the systems to subnational districts (PAHO 2004). surveillance for reportable infectious diseases), to ensure the Other WHO regions, including the Asia Pacific, are institu- independence of official statistical agencies, to reinforce pro- ting similar initiatives with Web-based publication of core fessional ethics, and to create norms for data quality and health indicators. dissemination. Another key policy intervention, less tangible though equally Data Collection. No single mechanism for data collection is critical, is the creation of a culture of quality and transparency adequate to meet the needs for public health decision making. Information to Improve Decision Making for Health | 1019 Box 54.1 The Health Metrics Network: Harmonizing Investment in HIS Development Developing countries, multilateral agencies, bilateral architecture and a plan for development of national health donors, and technical resource agencies have recently information systems. This HMN Framework includes a come together to form a global Health Metrics Network blueprint for iterative improvements in the HIS; descrip- (HMN) that is designed to provide guidance for the tions of core data collection subsystems (census, surveys, development of the HIS, both in meeting national infor- vital events monitoring, service statistics, and resource mation needs and in producing the required indicators for tracking); and procedures for management and dissemi- tracking progress toward global goals. The HMN will pro- nation of information. vide the first consensus technical framework for HIS Source: Authors. These needs can be met using a combination of the six key health fact, generally undertaken to compensate for the lack of infor- information subsystems: census, household surveys, public mation available through routine systems (AbonZahr and health surveillance, vital events monitoring, health service sta- Boerma 2005). The investment in surveys has thereby enabled tistics, and resource tracking. Surveys are conducted on a sam- donors and developing countries to sustain their neglect of the ple in order to limit costs,whereas the other subsystems are more development of comprehensive and sustainable national health often designed to cover the entire population. In most develop- information systems. The United Nations Population and ing countries, public health surveillance--except for certain dis- Statistics Divisions and the European Statistical Office (EURO- ease-specific efforts--is conducted through passive reporting STAT) also support household survey work. Differences in from health facilities. Especially where utilization rates are low, methodologies among these surveys are currently a barrier to this facility-based surveillance may be considered a sample or the comparison of results. The World Bank's Managing for "sentinel"surveillance strategy.Vital events monitoring is, ideal- Development Results Roundtable, held in Marrakech, ly, universal; however, many countries use a phased introduction Morocco, in 2004, recommended harmonization of these sur- of vital events monitoring that makes it functionally a sentinel or veys to eliminate duplication. sample-based data collection effort during the transition to uni- Nonetheless, surveys offer an important source of informa- versal coverage. tion that transcends most of the selection bias that is inherent A national census every 10 years is an irreplaceable compo- in service statistics. Especially in the least developed countries, nent of a national information system because it provides where vital events registration systems and census taking are denominator data for so many indicators and sampling frames embryonic or nonexistent, surveys represent the only source for subsequent sample surveys. The major costs of a census of unbiased information about demography, socioeconomic come from activities to establish the census maps, enumerate status, coverage, morbidity, mortality, health expenditures, and populations, enter data, and analyze the results. The carto- other characteristics of the population. Where substantial pro- graphic costs can often be shared with other government portions of the population use private health services, house- departments, because the resulting updated maps can be hold surveys are particularly important. Even industrial coun- instrumental in carrying out other critical public functions. tries rely on periodic community-based sample surveys for Sample surveys of households are a mainstay of health infor- immunization coverage, for health service utilization rates, and mation collection in the developing world. They provide data for information on household health expenditures (Perrin, on service utilization; coverage of health interventions Kalsbeek, and Scanlan 2004). (for example, immunization); morbidity (self-reported illness The World Health Organization (WHO) recommends using or disability); pregnancy outcomes; mortality levels, differen- periodic surveys to monitor coverage, such as for immunization tials, and trends; and causes of death (through associated verbal programs, especially in view of the shortcomings of service sta- autopsy; that is, expanded interviews in the case of death to tistics for obtaining these measures (Murray and others 2003). determine cause on the basis of signs and symptoms before Some household surveys collect biological and clinical speci- death). mens, such as blood, saliva, urine, and self-collected vaginal Surveys are, almost without exception, funded externally in swabs, or they check swabs for anemia, HIV, disease antibodies, the least developed countries and are not seen within the coun- vitamin A, and other conditions. However, the performance try as being part of a health information "system." They are, in characteristics of most diagnostic technologies (for example, 1020 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others cost, ease of field use, sensitivity, and specificity) are designed monitoring systems function poorly and may be found only as for clinical use and do not lend themselves readily to use in pop- remnants of past colonial administrations. In 2003, 115 of 192 ulation surveys, especially in remote areas of developing coun- WHO member states reported mortality data with causes of tries. Moreover, the collection of diagnostic information along death, capturing about one-third of global deaths, or 18.6 mil- with individual identifiers introduces complex ethical issues in lion deaths per year. In South Asia, only 60 percent of births are the notification of people with treatable conditions, the financ- registered (22.5 million), and in Africa, only 30 percent (17 mil- ing of any required treatments, and the use of the specimens for lion). Alternatives to universal registration include the sample other studies (Ties Boerma, Holt, and Black 2001). registration systems used in China and India and the demo- Public health surveillance has been defined as the "ongoing graphic surveillance sites in Tanzania. The International systematic collection, analysis, and interpretation of data on Network of Field Sites with Continuous Demographic specific health events affecting a population, closely integrated Evaluation of Populations and Their Health in Developing with the timely dissemination of these data to those responsi- Countries (INDEPTH), an association of longitudinal vital and ble for prevention and control" (Thacker and others 1996). In health statistics surveillance sites in 17 countries, can provide developing countries, surveillance usually focuses primarily on technical support and training for development and manage- a set of notifiable diseases, mainly infectious, which health care ment of these demographic surveillance sites (http://www. providers and laboratories are often required by law to report. indepth-network.org/). The UN Statistics Division has devel- Some nations also track risk factors for important diseases, oped principles and recommendations for vital statistics injury events, adverse drug reactions, cancers, and pregnancy systems to guide countries in their development (http:// outcomes. Surveillance may be intensified over a period of unstats.un.org/unsd/demographic/sources/civilreg/civilreg years to enable targeting of special interventions for the control methods.htm.) or elimination of diseases such as polio, tetanus, or measles. Vital events monitoring systems may also be enhanced to Active surveillance or screening of populations for target determine causes of death, whether those deaths occur within diseases may also be used in specific circumstances, such as health facilities or in the community. When deaths occur out- during peak seasons for the disease or during natural disasters, side the health care system, a verbal autopsy, or structured inter- when the potential for epidemics may be high. view of the relatives of the deceased, can assist in determining Most passive surveillance data, however, are incomplete. the cause of death. Verbal autopsies can, however, be used reli- Reliance on surveillance for reportable diseases diagnosed in ably to diagnose only those few conditions that have character- health facilities omits diseases diagnosed among those who go istic clinical signs or patterns of signs that can be recognized by to private providers or who are too poor to go to any health family members or by the health workers who review the inter- facilities. Even in health facilities, reportable diseases are often view data. WHO is now developing standardized tools for underrecognized or cannot be confirmed in laboratories that verbal autopsy that will enhance the sensitivity and specificity have inadequate resources. Sentinel surveillance methods and of these instruments and permit comparisons over time and registries maintained in a few selected sites may be more repre- across geographics. sentative of the entire population and more cost-effective in Health service statistics are critical management tools for identifying and reporting the target diseases or health condi- both preventive and curative services. The statistics are col- tions; however, an outbreak may go undetected in a geographic lected at each level: community outreach service points, pri- area without a sentinel site. Special regional surveillance may mary care facilities, and district and regional referral hospitals. also be used where populations are vulnerable to special health Information from clients and providers documents the quantity risks. The Vigisus project in Brazil, for example, has developed and quality of services and enables managers to detect and a system of epidemiologic and environmental surveillance for solve problems in order to improve health outcomes and effi- the prevention and control of disease among indigenous ciencies. This health information subsystem must be "flexible populations in the Amazon region (http://www.br.undp.org/ and capable of adapting to local needs, while at the same time propoor/BRA97028a.htm). Despite the methodological haz- allowing for standardization of health care quality assurance ards, public health surveillance is essential for both national indicators, and subsequent ability to measure and compare the and global planning and preparedness, especially in view of the quality performance of health facilities nationwide" (Duran- risks of regional expansion (for example, of meningitis and Arenas and others 1998). A principal barrier to improving serv- polio) or global spread of recent epidemics (for example, of ice quality in many health care facilities is the lack of reliable SARS and bird flu). systems for managing and retrieving individual patient Vital events monitoring is the continuous, compulsory, and records. (in most cases) universal civil registration of key vital events, Service statistics are especially powerful when they can be such as births, deaths (sometimes including fetal deaths), mar- compared with population-based measures from censuses and riages, divorces, and migrations. In many countries, vital events surveys to estimate rates and ratios, such as disease incidence or Information to Improve Decision Making for Health | 1021 service coverage rates. Most service statistics subsystems track areas where no telephone or cable access exists, satellite tech- data only from public sector providers and facilities. Future nology can provide access to e-mail. Several countries, such as improvements must implement systems and incentives to en- Bolivia and Peru, have successfully used satellite telephone sure reporting of service data from the private sector. technologies to enable continuous Web-based updating of The resource-tracking subsystem must enable measurement health databases. Because the effectiveness of epidemic control and management of human resources; facilities; commodities often depends on timely detection and reporting of outbreaks, (pharmaceuticals, vaccines, and other consumables); and e-mail and telephone technologies have shown particular finances. Human resource tracking provides a mechanism for promise for use in disease surveillance. In Peru, for example, licensing health service providers and accrediting health facili- 100 percent reporting was achieved and sustained within six ties. Licensure and accreditation can be paired with incentives months of rollout of a pilot surveillance system using cellular to ensure service quality and private sector contributions to telephones (Lescano and others 2003). The system is to be achieving public health goals. expanded to national coverage this year. The national health account (NHA) framework provides Although individual citizens will not soon have equal access methods for measuring total national expenditures for health to ICT, these technologies can immediately be better used to from household, public, private, and donor sources. NHA data improve public health. Automation of data entry and analysis document the sources of health financing, the amount spent can ease data capture, validation, analysis, and transmittal of for services, the distribution of funds across services and inter- health information. District managers can generate reports ventions, and the distribution of health benefits from those with tables and charts and transmit them to central levels, services and interventions. An NHA framework tracks the flow which can then apply this knowledge to improve local manage- of funds, for example, from the ministry of health to health ment. Special prompts and "exception reports" can alert man- providers and government service programs or from house- agers to unexpected findings that require double-checking or holds to pharmacies and private providers. These internation- immediate interventions (for example, outbreaks of infectious ally comparable data enable benchmarking of performance disease, low immunization coverage, or other management among countries (Peters and others 2000). problems). Of the 68 countries that have implemented NHAs, only Use of free software, such as the U.S. Centers for Disease one-third have used the framework more than once. However, Control and Prevention's Epi Info, can lower costs, but often 19 of 21 countries studied can report at least one instance in these software packages require substantial adaptation to local which the NHA system has informed and shaped policies (De needs, along with additional training and technical support. and others 2003). For example, South Africa's NHA analysis Acquisition of computer equipment should be viewed not as a documented a higher per capita health expenditure in the rich- one-time capital expenditure but as a long-term commitment est districts, leading to intensified efforts to mitigate these to buy periodic upgrades, maintenance, and technical support. inequities (Abt Associates 2003). Experience shows that purchase of inexpensive software and computers, such as in the Eastern Cape Province of South Information and Communications Technologies. The rapid Africa, may actually increase overall costs when they require evolution of information and communications technologies early replacement with more adequate alternatives. (ICT) over the past 30 years has immense implications for the Geographic information system (GIS) technologies have potential speed, cost, and effectiveness of an HIS. But a "digital also been successfully used in districts in several countries to divide" persists, with poor countries failing to benefit fully from enable mapping and visual representation of the geographic these ICT advances. Lack of access to reliable power sources, distribution of risk factors, disease, and services. A desktop GIS absence of Internet connectivity, inability to procure computer viewer and mapping software are available in several shareware equipment and appropriate software, and inadequate technical versions, including the WHO's "Health Mapper," so that maps support are some of the barriers. African users account for only can be produced at little cost. Other potentially promising 1 percent of the world's Internet traffic, 80 percent of which is technologies include electronic scanners and personal in South Africa (http://www3.sn.apc.org/africa). Although less digital assistants for data capture (http://www.healthnet.org) than 0.001 percent of the Internet use in Africa is among health and global positioning systems to facilitate the mapping professionals, this usage is growing rapidly. process. Internet access in health facilities can make the HIS more The principal barriers to improved information systems, effective and efficient by enabling instantaneous transmittal of however, are human, not technological. Substantial investment data to central locations. Internet access in facilities can also in training and technical support must accompany the intro- speed data transmission and improve clinical outcomes by pro- duction of any new technology. If the HIS is not functionally viding access to evidence-based decision support for clinical solid, introducing ICT will likely only worsen existing care (Godlee and others 2004; McLellan 2001). Even in remote problems. 1022 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others Dissemination and Use of Health Information Information will "allow the public, their elected representa- Information is a means to the end of improving health, but the tives, or donors to determine whether they are obtaining value availability of reliable information does not guarantee its use or for money" (Cibulskis and Hiawalyer 2002; see also Mackay improved decision making. Because decisions are often driven 1998). Providing full access to the media will help to accelerate as much by politics as by evidence, it is critical to design infor- expectations of evidence-based decision making and account- mation systems to meet the needs of decision makers and to ability. Civil society, including nongovernmental organizations, create a culture of evidence that provides incentives and ac- should be the principal users of information to create and sustain countability for evidence-based decision making. Extensive citizen demand for quality services. The Healthy Communities dissemination promotes widespread use and accountability. Foundation's "dashboard" of lead indicators of health system The many users of information include the following: performance exemplifies one promising example of the visual display of data (http://whatcom.healthycities.org/demo/ aboutus.htm). Such dissemination and use of health informa- · health ministries at national, regional, and district levels tion has enhanced government accountability for improved · researchers and evaluators health in Papua New Guinea, where reports of local government · legislative and policy analysts performance in improving health systems transformed election · nongovernmental organizations and consumer organizations results (G. Hiawalyer, personal communication). · advocacy groups · private sector health providers and insurers · communities, including groups of patients BENEFITS, COSTS, AND COST-EFFECTIVENESS · journalists OF IMPROVED INFORMATION · donors and international agencies concerned with health · individuals and families. There is broad agreement that information--plus the knowl- edge it enables--creates value. Yet it is challenging, indeed, to The literature on health information systems is replete with quantify the added value of information. Information, after all, complaints of the neglect of existing information, yet remark- is necessary but never sufficient to achieve improved outcomes. ably little is known regarding the effectiveness of interventions Other resources--human, material, and financial--are to improve the use of information. The NHA experience (De required for change. Nonetheless, it is possible to define the and others 2003) suggests that policy makers are most likely to interventions necessary to improve health information and to use information when it contributes to and informs a preferred draw on a few studies to estimate the cost and cost-effectiveness government direction, especially if that information is not of these investments. available to stakeholders outside the health ministry. But sys- tems and dissemination patterns for information can be engi- Strengthening of Systems neered to ensure that clients, providers, and managers will seek The steps involved in strengthening HIS include securing fund- and use information to inform decisions. Standard procedures ing for a review of the current HIS and planning reforms and can be developed to ensure analysis and use of data at the level then using that plan to secure funding for implementing the at which it is collected. Training of health workers can be reforms. The reforms depend on legislation and regulations designed to include both basic and refresher training in the that delineate the requirements, incentives, and disincentives analysis and interpretation of data that are relevant to each job. for collecting the needed information. Finally, the review of the Expectations of information use can be built into routine job current HIS includes a situational analysis and outline of a plan requirements, including use of evidence for planning, data that involves a comprehensive information architecture that is requirements for periodic reporting to supervisors, and use of linked to both national and international needs. information during performance reviews. Groups of managers The HMN Framework includes assessment and planning can be convened across districts or regions for benchmarking, tools and HIS standards that will guide strengthening of sys- in which each manager presents and compares performance tems. Full implementation will likely take at least 36 months, data and is rewarded for transparency and learning. These and the effects on decision making and health outcomes will be practices will result only from intense training in analysis and detectable only after approximately five years. use. For example, Loevinsohn (1994) demonstrated that fewer than half of midlevel managers were able to use the informa- tion system even to identify best- and worst-performing dis- Benefits and Effectiveness of Improved Information tricts. Nonetheless, if managers use the information, and if The value of health information can be characterized in terms improved efficiency and coverage with interventions is the of cost savings; system efficiencies (for example, increased result, the HIS becomes exceedingly cost-effective. coverage or quality of services); or improved health outcomes Information to Improve Decision Making for Health | 1023 (for example, DALYs saved or improved health equity). Table 54.2 Cost of Essential HIS Subsystems Information can also be used to increase overall resources for Total cost Per capita health. Publications such as this volume, World Development (US$ million) cost (US$) Report 1993: Investing in Health (World Bank 1993), and the report of the Commission on Macroeconomics and Health Low High Low High HIS subsystem income income income income (2000), are important examples of evidence that has been used to change health policies and increase resources for health. Health service statistics 4.8 25.9 0.16 1.66 The industrial world holds examples of the use of informa- Public health surveillance 0 0 0 0 tion to make service provision more effective and efficient. A (included with health quality improvement and evidence-based decision assistance service statistics) program for diabetes patients in the United States created a net Census 7.5 30.0 0.25 1.0 savings of US$510,133,2 primarily by averting hospitalizations Household surveys 0.6 1.0 0.02 0.03 (Petrakos 1998). The U.S. Institute of Medicine estimates that a Vital events surveillance 1.5 6.0 0.05 0.20 computerized system for managing physician orders for med- Resource tracking 1.5 3.0 0.05 0.10 ications costing US$1 million to US$2 million could "pay for Total 15.9 65.9 0.53 2.99 itself in three to five years" and prevent injury to hundreds of Source: Authors. patients per year (Kohn, Corrigan, and Donaldson 2001). Note: Table is based on a population of 30 million. Household survey costs are based on the There are also promising examples of the benefits and effec- experience of the demographic and health surveys during 2001­2003 (Macro International, personal communication). Costs vary by sample size and by length of the survey instrument; tiveness of improved information from developing countries. Macro International estimates, an average cost of US$100 per survey participant. A sample of Quality improvements driven by better information in Bolivia 6,000 is assumed for the low-income setting, and a sample size of 10,000 is assumed for the resulted in a 300 percent increase in hospital utilization rates high-income setting. Cost estimates for vital events monitoring are based on demographic surveillance sites. In the high-income setting, the annual costs are assumed to quadruple. (Pappaioanou and others 2003). In rural Mali, populations Resource-tracking costs are based on the experience of national health accounts (Abt Associates, enrolled in a community-based information system calculated personal communication), and the Egyptian Budget Tracking system. Similar costs are estimated for human resources and commodities. delivery costs for childhood immunization to be US$1.47 per child, compared with US$2.79 per child among populations not registered (Zayan, Berggren, and Doumbia 1992). Better information can also improve efficiencies in the man- brought for treatment. The additional costs of program-specific agement of pharmaceutical resources. For example, imple- surveillance (for example, in support of polio eradication or menting a subnational information system in the Eastern Cape tetanus elimination programs) could be assumed with a province of South Africa led to improved access to pharmaceu- minor marginal investment in addition to facility-based and ticals, with a 39 percent reduction in stockouts of essential community-based information systems, including for vital drugs. Such improvements undoubtedly lead to better health events surveillance. outcomes, which may result in increased productivity and con- The calculated range for per capita annual costs of a com- sequently an increase in the growth rate of the gross domestic prehensive HIS--US$0.53 to US$2.99--compares closely to the product (Jamison, Sachs, and Wang 2001; Nordhaus 2002). estimates from a country setting in which those data have been obtained, including a low-resource country (Tanzania) with a per capita cost of approximately US$0.50 (Rommelmann and Costs of Improved Health Information others 2004) and a high-resource country (Mexico) with a per Few studies have documented the costs of an HIS. Kleinau capita cost of approximately US$1.00. The Health Metrics (2000) estimated the resource requirements for health service Network (HMN) Technical Task Force South Africa has also statistics, the most expensive of the six subsystems. Using sim- estimated costs of the HIS at approximately US$26 million ilar assumptions, we have calculated updated costs. (165 million rand) for a population of 43 million, yielding a per This estimate includes only the public sector facilities, not capita cost of US$0.60. The highest range of the estimate would private sector reporting systems. Reporting from private apply in countries with higher salaries and a more comprehen- providers would likely include a more limited set of reported sive HIS. data: diseases, vaccinations, possibly staffing, and minimal uti- lization. Table 54.2 summarizes the total annual costs and per capita costs of the six health information subsystems. Estimations of the Cost-Effectiveness of Interventions The costs of a facility-based services statistics subsystem of to Improve Health Information the HIS (table 54.3) can be assumed in most developing coun- The Tanzania Essential Health Interventions Program (TEHIP) tries to include routine public health surveillance, because is perhaps the best source of evidence for the cost-effectiveness these data are obtained at health facilities when ill patients are of improved health information. The project was designed to 1024 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others Table 54.3 Annual Costs of the Facility-Based Services Statistics Subsystem of an HIS HIS cost Low-resource setting High-resource setting Personnel Primary care facility One person (salary US$4,514/year) spends 10 percent Two people (salary US$10,351/year each) spend of time at each of 6,000 facilities (US$2,708,400) 20 percent of time at each of 6,000 facilities (US$24,842,400) First referral level One person (salary US$4,514/year) spends 25 percent Two people (salary US$10,351/year each) spend of time at each of 1,000 facilities (US$1,128,500) 75 percent of time at each of 1,000 facilities (US$15,526,500) District hospital Two people (salary US$4,514/year each) spend Two people (salary US$10,351/year each) spend 20 percent of time at each of 300 facilities 100 percent of time at each of 300 facilities (US$541,680) (US$6,210,600) Regional level Three people (salary US$10,962/year each) spend Three people (salary US$25,134/year each) spend 50 percent of time at each of 15 facilities 100 percent of time at each of 15 facilities (US$246,645) (US$1,131,030) National level Six people (salary US$10,962/year each) spend Ten people (salary US$25,134/year each) spend 50 percent of time (US$32,886) 100 percent of time (US$251,340) Subtotal (personnel) US$4,658,111 US$47,961,870 Data collection instruments and supplies Primary care facility US$100/year US$400/year First referral level US$250/year US$1,000/year District hospital US$500/year US$2,000/year Regional US$1,500/year US$5,000/year National US$5,000/year US$30,000/year Subtotal (supplies) US$7,350 US$38,400 Information technology: computers and software Primary care facility 0 0 First referral level 0 0 District hospital 0 20 percent use of each of two computers with software at US$1,100 at each of 300 facilities (US$132,000) Regional level 0 Two dedicated computers with software at US$1,100 at each of 15 facilities (US$33,000) National level 50 percent use of each of four computers with 10 dedicated computers with software at US$1,100 software at US$1,100 (US$2,200) (US$11,100) Subtotal (information technology) US$2,200 US$176,100 Training cost US$180,000 US$1,730,000 Total cost US$4,847,661 US$49,906,370 Per capita cost US$0.16 US$1.66 Source: Authors. Note: Based on a model country with a total population of 30 million. test how evidence can be used to decentralize health sector cost-effective interventions. The information systems included planning at the district level and to what extent evidence-based a district burden-of-disease intervention priority profile, dis- priority setting would result in improved health outcomes. The trict health accounts, a district cost information system, and project budgeted for a marginal investment of US$2.00 per district health service mapping. Management and technical capita for the information and for health interventions, support strengthened the district and regional health sector although only US$0.80 per capita was actually spent. The use of the information for management and administration. slightly increased investment covered training in the use of the Communities participated in the ownership and management information to set priorities and to better manage the most of health facilities. The cost-effectiveness estimates in this Information to Improve Decision Making for Health | 1025 Box 54.2 The Tanzania Essential Health Interventions Program TEHIP is a partnership between Tanzania's Ministry of money from national investments in health. Interventions Health and the International Development Research included door-to-door collection of data and training or Centre. The project was established to determine the fea- technical support for managers in the analysis and use of sibility of an evidence-based approach to health planning the data for decision making. TEHIP districts allocated at the district level. Testing the premise of the World services to high-burden diseases, resulting in a tripling of Bank's (1993) World Development Report 1993: Investing in clinic utilization rates and increased treatment effective- Health, TEHIP enabled district health planners in two of ness. With a per capita increase in spending of only Tanzania's 117 districts to collect and use burden-of- US$0.80, district health managers achieved a 47 percent disease and cost-effectiveness data to get the best value for reduction in child mortality rates. Source: Authors. Table 54.4 The Effectiveness of Evidence-Based Resource Allocation in Improving Health Probability Number of of dying Mortality DALYs DALYs children (birth to rate Total Deaths gained/ Total discounted Year 5 years 5 years) ( 5 years) deaths averted death DALYs at 3 percent 1999 31,000 135.5 34 1,054 -- -- -- -- 2000 31,500 119.0 25 791 263 41 10,850 11,511 2001 32,000 110.0 25 803 251 41 10,332 10,643 2002 32,661 114.0 26 853 202 41 8,303 8,304 Total DALYs gained 29,487 30,458 Source: Authors. Note: 1999 is baseline year; therefore, no deaths were averted. section are based solely on the declines in mortality of children Table 54.5 Costs of Evidence-Based Resource Allocation for under five years of age, even though adult mortality also Improving Health decreased. To ensure a conservative estimate of the costs of the Total cost at US$2 Discounted HIS, we used a per capita cost of US$2.00--higher than the Year Population per capita (US$) cost (US$) actual investment for TEHIP and at the high end of the range 1998 186,809 373,618 420,510 of costs for a comprehensive HIS estimated in table 54.2-- US$0.53 to US$2.99. All costs were ascribed to the information 1999 191,012 382,024 417,448 system, because there were no improvements in the interven- 2000 196,515 393,030 416,966 tions themselves. Expenditures and deaths before 2002 were 2001 202,176 404,352 416,482 discounted by 3 percent annually (see box 54.2). 2002 208,000 416,000 416,000 The demographic and epidemiologic data were taken from Total costs 1,969,024 2,087,406 the Rufiji district, where the most complete data were available. Source: Authors. The estimate of the number of children under age five (32,661) is based on the 2002 census results. The Ministry of Health, fewer children than the preceding year. DALYs saved from each census, and Rufiji Demographic Surveillance System estimates child death averted is estimated at 41.2. The resulting calcula- range from 31,000 to 36,000 children for 2003. Because of this tions of effectiveness are summarized in table 54.4. discrepancy, the decline in the total fertility rate, from 5 to 4.7 The estimates of cost are based on population size projected (5 percent), is taken into account in estimating the number of back from the 2002 census results, assuming an average annual children less than five years of age for 1999 to 2001. The prob- growth rate of 2.8 percent. Costs incurred in 1998 are included ability of dying before five years of age declined by 15.6 percent, because we assume that it takes at least two years (1998 to 2000) and because of declining fertility, each year has 1.5 percent of improving the HIS before health benefits accrue. 1026 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others Cost per DALY saved (US$) capita less than US$997; World Bank 2002). Figure 54.1 shows 800 that, for the high-prevalence countries (Miller and McCann 2000), the investment in a comprehensive HIS is highly cost- 700 effective (US$159 to US$126 per DALY saved for low-cost set- 600 tings and US$757 to US$597 per DALY saved for high-cost settings), even if the investment results in only minor increases 500 in immunization coverage. A similar analysis for countries with 400 a lower prevalence rate of hepatitis B demonstrates that the cost per DALY saved is higher, but the investment in an HIS still 300 yields a savings of DALYs at a cost that is well below the GNP 200 per capita for the majority of the low-income countries. These calculations of the cost-effectiveness of investments 100 in an HIS are highly conservative, because they consider health 0 benefits within a single population group (children, in the case 5 10 20 30 40 of TEHIP) or a single disease problem (hepatitis B). They Percentage increase in hepatitis B immunization coverage therefore underestimate the true cost-effectiveness of invest- attributable to investment in HIS ment in an HIS, which can drive improvements in program High-cost setting (US$2.00 per capita) efficiency and effectiveness across a broader range of health Low-cost setting (US$0.80 per capita) interventions. Source: Authors. Figure 54.1 Cost-Effectiveness of Health Information Systems: Cost FINANCING OF IMPROVED HEALTH per DALY Saved Because of Increases in Coverage Attributable to HIS INFORMATION The annual per capita cost, estimated earlier, of US$0.53 to Using these figures for effectiveness (table 54.4) and cost US$2.99 for a comprehensive HIS, represents a substantial (table 54.5), we find that the cost-effectiveness of the HIS that portion of the current per capita health expenditure for many results in improved evidence-based resource allocation and developing countries. These figures include capital and recur- child health may be conservatively estimated at US$68.50 per rent costs, although they do not include the costs of any external DALY gained (US$2,087,406 to gain 30,457 DALYs). Even in technical assistance. Because most countries have already made the poorest countries, this is well below the gross national a substantial investment in a HIS, the actual incremental costs to product (GNP) per capita benchmark for what is considered improve the existing HIS likely are much less. Salaries, which worthwhile for government investment in health. account for more than 90 percent of HIS costs, are expenditures This analysis for the TEHIP project is based solely on child that are already being made in most settings, so the marginal deaths averted. But the improvement in health information cost of HIS improvements would be primarily the initial devel- would also yield substantial benefits for adult populations. For opment costs of planning, training, technical assistance, and example, HIS-driven increases in coverage with hepatitis B vac- information technology upgrades. Furthermore, the costs of cine have varied between 5 and 33 percent (Miller and McCann HIS improvements may be fully offset or even exceeded by the 2000). These increases in coverage with hepatitis B immuniza- savings from the resulting improvements in efficiencies in the tion will result in incremental reductions in death and disability health care system. among adults attributable to hepatitis B­induced cirrhosis and Existing funding is adequate to strengthen systems substan- liver cancer, thereby averting the loss of substantial numbers of tially in all low-income and lower-middle-income countries DALYs in low-income countries (World Bank 2002). Hepatitis B primarily through the major international initiatives (Global vaccine is a cost-effective addition to an existing immunization Fund to Fight AIDS, Tuberculosis, and Malaria; President's program, with a cost per death averted of US$11 to US$15 Emergency Plan for AIDS Relief; and Multi-country AIDS (US$193 to US$262 per DALY saved). But efficiency and cover- Program of the World Bank). All these funders recommend age can be substantially improved with an additional invest- that 3 to 7 percent of grants and loans be allocated to monitor- ment in the HIS. The cost per DALY saved by incremental ing and evaluation. Several bilateral development agencies and investment in the HIS can be calculated using estimates of costs the multilateral development banks will provide financing for of the HIS from table 54.2, plus the estimates of cost and deaths HIS reform, including the U.S. Agency for International averted because of immunization from Miller and McCann Development (USAID) through the MEASURE (Monitoring (2000) for populations in all low-income countries (GNP per and Evaluation to Assess and Use Results) Project, which is Information to Improve Decision Making for Health | 1027 designed to improve and institutionalize the collection and use opment of the district-level HIS. The HMN will create an of data for health policy development and program monitor- alliance of countries committed to a parsimonious consensus ing. The HMN offers some financial assistance to countries that technical framework and encourage donors to cooperate with are preparing for and planning HIS reform and will assist coun- and strengthen the HMN-sanctioned HIS architecture in tries in negotiating financing packages that blend loan funding participating countries. with grants from bilateral donors to implement those reforms. The predictors of success in developing and maintaining an Several international agencies support strengthening sys- HIS are as follows: tems for national statistics that extend beyond the health sec- tor. STATCAP (Statistical Capacity Building), which is a new · high-level commitment to HIS development and the linked lending program offered by the Partnership in Statistics for changes in management Development in the 21st Century (PARIS21) through the · a champion of HIS reform who engages the stakeholders World Bank, supports the development of national statistical and can work across sectors systems. The separate Trust Fund for Statistical Capacity · an information architecture that is simple, is structured to Building offers smaller amounts of grant funding to prepare drive decision making at the level that data are collected, the statistical master plan that is required for obtaining a provides incentives and accountability for performance, and STATCAP loan. Although short-term project funding can often links health information subsystems be secured for system development, the resulting system and its · investment in training and increased status for the people recurrent costs must be within the country's capacity to sustain who manage the HIS. it, both technically and financially. RESEARCH AND DEVELOPMENT IMPLEMENTATION OF CHANGE: An effective HIS delivers routine information that enables LESSONS OF EXPERIENCE informed policy making and management but also promotes health research. Routine information systems may serve as a Underinvestment is the root cause of the nearly universal research platform, but the HIS itself should also be a subject weaknesses in the HIS in developing countries. This failure is of research. Research should drive the continual refinement of reflected in the poorly paid and undervalued HIS staff; in the HIS methods and tools, thereby ensuring expanding and well- irregular and unreliable transmittal of data from the periphery; documented returns on our investments in health. in the underreporting of events, including births, deaths, and The instruments and methods of the HIS must be continu- morbidity; and in the failures to base planning and decision ally refined to improve its effectiveness and reduce its costs. For making--at both the district and the central levels--on credi- the phased introduction of vital events monitoring, for exam- ble evidence (Azubuike and Ehiri 1999). ple, there is a pressing need for the development and validation When the need for HIS improvement is identified, min- of methods for projecting subnational results to national rates istries of health should explicitly state the characteristics they of birth and death. More research is needed to develop and test need in a reformed system and quantify the expected benefits. new methods for rapid assessment in order to obtain timely A common mistake made in implementing HIS change is fail- and affordable information to solve management problems. As ing to recognize the associated need for change in management field-appropriate and cost-effective diagnostic technologies are processes and organizational culture. In contrast, recent HIS developed, research should be performed to document the util- reforms in Niger (Mock and others 1993) and Uganda ity of obtaining biomarkers in household surveys. (Gladwin, Dixon, and Wilson 2003) have had unprecedented Documenting improved outcomes and lower costs will pro- success because they have been aligned with broader manage- vide evidence for policy makers on the effectiveness of HIS ment reforms and changes in organizational culture. Failure to investments. To better decide how to improve the HIS, decision adjust management roles with HIS changes can constrain makers will need documentation of the costs and effects of effectiveness, such as when HIS managers are not given the introducing ICT in support of the HIS. Existing and emerging necessary increased status and authority to demand reports technologies should be tested for their cost and effectiveness in and trigger corrective actions (Gladwin, Dixon, and Wilson assisting field-based data capture, instantaneous data transmis- 2003). Failure to invest adequately in training, especially in sion, GIS-based mapping of indicators, and compelling presen- skills for presentation and communication of results, may also tation for decision making by policy makers, managers, and inhibit the use of health information. The demand from inter- other stakeholders. Research and development efforts are national organizations and global programs, such as the needed to devise software--or preferably shareware--that is Expanded Program on Immunization and Stop TB, for reports specifically tailored to support the consensus technical frame- on vast numbers of indicators has retarded the smooth devel- work developed by the HMN. 1028 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others Bailey and Pang (2004) point out the need for more research improvements in health if they are engineered to reflect, rein- in the developing world to better understand users' informa- force, and even drive health sector reforms. Even more com- tion needs. In fact, research is needed to better document the pellingly, investments in the HIS can make health the "thin entire information value chain, with special attention to edge of the wedge," giving governments and politicians a posi- improving the identification of information needs, to over- tive experience with information sharing and overcoming the coming the natural disincentives to information sharing, and natural disincentives to transparency and accountability. HIS to enabling better use of information for constructive change. investments hold the promise, therefore, not only of trans- At present, there is still a need to improve the access to infor- forming public health, but also of accelerating progress toward mation and knowledge in the developing world. However, the good governance in every sector. future will bring the larger challenge of improving the man- agement and use of information and the knowledge such infor- mation can bring. Research in the HIS will be instrumental NOTES in both accelerating equitable access to information and 1. Sauerborn and Lippeveld (2000) have defined a health information improving the management and use of knowledge for system as the "set of components and procedures organized with the objec- improved health. tive of generating information that will improve health management decisions at all levels of the health system." 2. The dollar amounts given are quoted from the references and are not CONCLUSIONS adjusted for current dollar value. More than ever before, it is in the mutual interest of the devel- oping and industrial worlds to invest in strengthening systems REFERENCES for collection and management of health information AbonZahr, C., and T. Boerma. 2005. "Health Information Systems: The (Stansfield 2005). Foundations of Public Health." Bulletin of the World Health The trend toward "basket" funding for health and sector- Organization 83 (8): 578­83. wide approaches makes the need for priority setting all the Abt Associates. 2003. Primer for Policymakers--Understanding National Health Accounts: The Methodology and Implementation Process. more acute. Priority setting depends on accurate information. Bethesda, MD: Partners for Health Reformplus Project, Abt Associates. The success of efforts to reduce poverty and health inequity Azubuike, M. C., and J. E. Ehiri. 1999. "Health Information Systems in will depend on the existence of information systems to detect Developing Countries: Benefits, Problems, and Prospects." Journal of those problems, facilitate the design of solutions, and track the Royal Society for the Promotion of Health 119 (3): 180­84. progress toward eliminating the problems. Countries and Bailey, C., and T. Pang. 2004. "Health Information for All by 2015?" Lancet 364 (9430): 223­24. donors must, therefore, accelerate and harmonize their invest- Boerma, Ties, J., E. Holt, and R. Black. 2001. "Measurement of Biomarkers ments in information systems. in Surveys in Developing Countries: Opportunities and Problems." Within countries, the trend toward decentralization of Population and Development Review 27 (2): 303­14. authority for management of health resources has led to fur- Cibulskis, R. E., and G. Hiawalyer. 2002. "Information Systems for Health ther challenges for the HIS, as well as to greater reliance on the Sector Monitoring in Papua New Guinea." Bulletin of the World Health Organization 80 (9): 752­58. information it provides to inform decision making. It is clear Commission on Macroeconomics and Health. 2000. from the instructive failures of underresourced systems that the De, S., T. Dmytraczenko, D. Brinkerhoff, and M. Tien. 2003. Has Improved accuracy and value of information reported to the national Availability of Health Expenditure Data Contributed to Evidence-Based level will depend on that information's perceived value in the Policymaking? Country Experiences with National Health Accounts. periphery. Information is relevant only if it is used to solve a Technical Report 022. Bethesda, MD: Partners for Health Reformplus local problem or if it helps to generate innovation that solves a Project, Abt Associates. local problem (Bailey and Pang 2004). Therefore, the decen- Duran-Arenas, L., C. C. Rivero, S. F. Canton, R. S. Rodriquez, F. Franco, R. W. Luna, and J. Catino. 1998. "The Development of a Quality tralization of authority will be successful only with better infor- Information System: A Case Study of Mexico." Health Policy and mation systems to support decisions at the periphery, and Planning 13 (4): 446­58. evidence-based decision making will be possible only if author- Economic Commission for Africa. 2003. "Workshop on Organization ity can be devolved to the periphery. This decentralization, and Management of Statistical Systems." Report on workshop in Addis Ababa, December 8­12. http://www.unstats.un.org/unscl/ along with increasing cooperation and collaboration across methods/statorg/workshops/AddisAbaba/presentation_session8b_ sectors to improve health outcomes, makes it all the more financing.pdf. critical to present data in simpler ways that are understandable Gladwin, J., R. A. Dixon, and T. D. Wilson. 2003. "Implementing a New and compelling to a broader and nontechnical audience. Health Management Information System in Uganda." Health Policy and Planning 18 (2): 214­24. Although historically neglected, investments in comprehen- Godlee, F., N. Pakenham-Walsh, D. Ncayiyana, B. Cohen, and A. Packer. sive development of the HIS will clearly deliver good value for 2004. "Can We Achieve Health Information for All by 2015?" Lancet money. Improvements in the HIS can accelerate broad 364 (9430): 295­300. Information to Improve Decision Making for Health | 1029 Jamison, D. T., J. Sachs, and J. Wang. 2001. "Mortality Changes and Pappaioanou, M., M. Malison, K. Wilkens, B. Otto, R. A. Goodman, R. E. Economic Welfare in Sub-Saharan Africa, 1960­2000." Commission on Churchill, and others. 2003. "Strengthening Capacity in Developing Macroeconomics and Health, Background Paper for Working Group 1, Countries for Evidence-Based Public Health: The Data for Decision World Health Organization, Geneva. Making Project." Social Science and Medicine 57: 1925­37. Jolly, R. 2002. "Statisticians of the World Unite: The Human Development PARIS21 (Partnership in Statistics for Development in the 21st Century). Challenge Awaits." Journal of Human Development 3 (2): 263­72. 2004. "Meeting the Data Challenge: A Funding Proposal for PARIS21 Kleinau, E. 2000. "Management of Health Information Systems." In and the Trust Fund for Statistical Capacity Building for 2004 to 2006." Design and Implementation of Health Information Systems, ed. PARIS21, Paris. T. Lippeveld, R. Sauerborn, and C. Bodart. Geneva: World Health Perrin, E. B., W. D. Kalsbeek, and T. M. Scanlan, eds. 2004. Toward a Health Organization. Statistics System for the 21st Century. Summary of a workshop, Division Kohn, L., J. Corrigan, and M. Donaldson, eds. 2001. To Err Is Human: of Behavioral and Social Sciences and Education, National Research Building a Safer Health System. Washington, DC: National Academy Council and Committee on National Statistics, National Academy of Press. Sciences. Washington, DC: National Academy Press. Lescano, A. G., M. Ortiz, R. Elgegren, E. Gozzer, E. Saldarriaga, I. Soriano, Peters, D. H., A. E. Elmendorf, K. Kandola, and G. Chelleraj. 2000. and others. 2003. "Alerta DISAMAR: Innovative Disease Surveillance "Benchmarks for Health Expenditures, Services, and Outcomes in in Peru." Paper presented at the American Society of Tropical Medicine Africa during the 1990s." Bulletin of the World Health Organization and Hygiene, Philadelphia, December 5. 78 (6): 761­68. Loevinsohn, B. P. 1994."Data Utilization and Analytical Skills among Mid- Petrakos, C. 1998. "Finding a Cure: Disease Management Aids the Search Level Programme Managers in a Developing Country." International for Better Outcomes." Modern Physician, September 1, 2004. Journal of Epidemiology 23 (1): 194­200. http://www.modernphysician.com. Mackay, K. 1998. "Public Sector Performance: The Critical Role of Rommelmann, V., P. Setel, Y. Hemed, H. Mponezya, G. Angeles, and Evaluation." In Public Sector Performance--The Critical Role of T. Boerma. 2004. "Costs and Results of Information Systems for Evaluation: Selected Proceedings of a World Bank Seminar, ed. K. Mackay, Poverty Monitoring, Health Sector Reform, and Local Government ix­xvi. Washington, DC: World Bank. Reform in Tanzania." MEASURE Evaluation, Adult Morbidity and Mortality Project, University of Newcastle upon Tyne, U.K., and McLellan, F. 2001. "Information Technology Can Benefit Developing University of North Carolina, Chapel Hill. Countries." Lancet 358 (9278): 308. Sauerborn, R., and T. Lippeveld. 2000. "Why Health Information Miller, M. A., and L. McCann. 2000."Policy Analysis of the Use of Hepatitis Systems?" In Design and Implementation of Health Information Systems, B, Haemophilus influenzae type B, Streptococcus Pneumoniae- ed. T. Lippeveld, R. Sauerborn, and C. Bodart. Geneva: World Health Conjugate and Rotavirus Vaccines in National Immunization Organization. Schedules." Health Economics 9: 19­35. Stansfield, S. 2005. "Structuring Information Systems to Improve Health." Mock, N., J. Setzer, I. Sliney, G. Hadizatou, and W. Bertrand. 1993. Bulletin of the World Health Organization 83 (8): 562. "Development of Information-Based Planning in Niger." International Journal of Technology Assessment in Health Care 9 (3): 360­68. Thacker, S. B., D. F. Stroup, R. G. Parrish, and H. A. Anderson. 1996. "Surveillance in Environmental Public Health: Issues, Systems, and Murray, C. J. L., B. Shengelia, N. Gupta, S. Moussavi, A. Tanjon, and Sources." American Journal of Public Health 86 (5): 633­38. M. Thieren. 2003. "Validity of Reported Vaccination Coverage in 45 Countries." Lancet 362 (9389): 1022­27. Walsh, J. A., and M. Simonet. 1995. "Data and Data Needs for Health Sector Reform." Health Policy 32 (1­3): 295­306. Nordhaus, W. 2002. "The Health of Nations: The Contribution of Improved Health to Living Standards." Cowles Foundation Discussion World Bank. 1993. World Development Report 1993: Investing in Health. Paper 1355, Yale University, New Haven. Washington, DC: World Bank. O'Carroll, P. W. 2003. "The Context for Public Health Informatics." In ------. 2002. World Development Indicators. Washington, DC: World Public Health Informatics and Information Systems, ed. P. W. O'Carroll, Bank. W. Yasnoff, M. E. Ward, L. H. Ripp, and E. L. Martin. New York: Zayan, A., W. Berggren, and F. Doumbia. 1992. The Price of Immunization Springer. and the Value of Information. Westport, CT: Save the Children. PAHO (Pan American Health Organization). 2004. "Ten Year Evaluation of the Regional Core Health Data Initiative." Epidemiological Bulletin 25 (3): 1­7. 1030 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others Chapter 55 Drug Resistance Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, Philip Jenkins, Thomas O'Brien, Iruka N. Okeke, Ariel Pablo-Mendez, and Keith P. Klugman The control of infectious diseases is seriously threatened by the insufficient controls on drug prescribing; inadequate steady increase in the number of micro-organisms that are resist- compliance with treatment regimens; poor dosing; lack of ant to antimicrobial agents--often to a wide range of these infection control; increasing frequency and speed of travel, agents. Resistant infections lead to increased morbidity and pro- which lead to the rapid spread of resistant organisms; and longed hospital stays, as well as to prolonged periods during insufficient incentives for patients, physicians, or even govern- which individuals are infectious and can spread their infections ments to care about increasing resistance. It is important to dis- to other individuals (Holmberg, Solomon, and Blake 1987; tinguish between risk factors for the emergence of resistance Rubin and others 1999). The problem is particularly severe in (de novo resistance) and those for the spread of resistance (pri- developing countries, where the burden of infectious diseases is mary resistance). relatively greater and where patients with a resistant infection are The molecular basis of resistance may give a clue to the less likely to have access to or be able to afford expensive second- likelihood of resistance emerging. If a single DNA base pair line treatments, which typically have more complex regimens mutation leads to the development of resistance, then its selec- than first-line drugs. Furthermore, the presence of exacerbating tion is likely to be widespread, especially if the biological fitness factors,such as poor hygiene,unreliable water supplies,civil con- cost of the mutation is low. De novo or acquired resistance flicts, and increased numbers of immunocompromised patients results in the appearance of a resistant strain in a single patient. attributable to the ongoing HIV epidemic, can further increase Subsequent transmission of such resistant strains from an infec- the burden of antimicrobial resistance by facilitating the spread tious case to other persons leads to disease that is drug resistant of resistant pathogens. In this chapter, we discuss the causes and from the outset, a phenomenon known as primary resistance burden of drug resistance and evaluate interventions that address (IUATLD 1998). Independent, cumulative events result in mul- the resistance problem in developing countries.Although a num- tidrug-resistant bacteria or tuberculosis (MDR-TB). Both the ber of the interventions we discuss are relevant to drug resistance creation and the transmission of drug resistance contribute to in HIV/AIDS and other forms of antiviral resistance, chapter 18 its prevalence in a given population. This mechanism also includes a more in-depth discussion of this subject. holds true in the case of antimalarials; that is, resistance devel- ops when malaria parasites encounter drug concentrations that are strong enough to eradicate the susceptible parasite popula- RISK FACTORS tion, but they fail to inhibit the multiplication of naturally occurring resistant strains. Commonly used antimalarial drugs Drug Use in Humans are not mutagenic. The evolution of drug resistance is facilitated by a number of fac- In the case of tuberculosis, spontaneous mutations leading tors, including increasing use of antibiotics and antimalarials; to drug resistance occur rarely in Mycobacterium tuberculosis, 1031 and multidrug regimens can prevent the emergence of clinical countries, without controls on over-the-counter use, has led to drug resistance (Cohn, Middlebrook, and Russell 1959). some of the highest rates of resistance in the world, as was seen Resistance is thus an avertable phenomenon resulting from with penicillin resistance in Vietnam. Relatively wealthy coun- inadequate treatment, which, in turn, is often the result of an tries such as the Republic of Korea and Japan also have lax con- irregular drug supply, prescription of inappropriate regimens, trol and even greater access to funds to purchase antibiotics or poor adherence resulting from a lack of supervision. In the (Song and others 1999). Patterns of resistance differ by antimi- case of malaria, the widespread misuse of chloroquine as pro- crobial class, and resistance to several classes has been linked to phylaxis is believed to be an important factor in the emergence particular patterns of use in developing countries. Macrolide and spread of resistance to this drug. use in children in China may be preferred to the use of beta- Despite conventional wisdom, the highest rates of antibiotic lactams, which are known to be associated in rare instances resistance in the pneumococcus bacterium globally are not for with serious anaphylactic reactions, and in Beijing and penicillins or macrolides, which usually require multiple DNA Shanghai, the highest global rates of macrolide resistance are mutations or the import of foreign genes, respectively, but for encountered in nasopharyngeal isolates from children (Wang sulfamethoxazole-trimethoprim, which can be selected from and others 1998; Yang, Zhang, and McGee 2001). Tetracycline among a population of susceptible pneumococci by a single use remains widespread in developing countries, and poor base change in the dihydrofolate reductase gene (Adrian and African countries, such as the Central African Republic, may Klugman 1997). The direct selection of resistance following have higher rates of resistance to tetracycline than to beta- exposure of children carrying pneumococci has been shown in lactams or macrolides (Rowe and others 2000). a prospective study in Malawi to occur in 42 percent of children The relationship between compliance and resistance emer- exposed to sulfadoxine-pyremethamine for a week and in 38 gence in the treatment of acute and largely self-limiting infec- percent of children a month after exposure to drug treatment tions is less robust than in the case of chronic infections such as for malaria (Feikin and others 2000). tuberculosis (TB). It is likely that resistance selection occurs Evolutionary biology suggests that drug selection pressure is more readily in the commensal flora (for example, the pneu- an important factor in the emergence and spread of drug resist- mococcal flora of the nasopharynx) than among the organisms ance. Although the relationship between antimicrobial use and causing the acute infection. Thus, shorter courses (and reduced drug resistance (in the pneumococcus, for example) is well compliance) may reduce the selection of resistance in com- established in developed countries (Bronzwaer and others mensal flora. In contrast, in TB, selection takes place in the 2002), direct evidence to support this hypothesis is less forth- infecting pathogen, and poor compliance is associated with the coming in developing countries because of a lack of data on selection of resistant strains. antibiotic use. Resistance to antimicrobials is less likely to arise in the poorest developing countries simply because of the lower levels of antibiotic use associated with poorer socioeconomic Antibiotic Use in Animals status. For instance, India--a large country with scant control Many developed countries use antibiotics for veterinary uses, over antibiotic prescribing--has very low rates of resistance both for improving feed efficiency and rate of weight gain among systemic isolates of pneumococci, at least in rural areas (subtherapeutic use) and for disease prevention and treatment (INCLEN 1999). These low rates exist despite wide antibiotic (therapeutic use) (Levy 1992). Although the extent of anti- availability, probably because extreme poverty limits the dura- biotic use in animals in developing countries is unknown, one tion of antibiotic exposure for the treatment of acute pneumo- study from Kenya reported that tetracyclines, sulfonamides, coccal infections. Rising incomes and increased affordability of and aminoglycosides were the most commonly used antimi- antibiotics will likely change this low incidence of resistance; crobials for veterinary purposes (Mitema and others 2001). the same may be true of quinolones, which are widely available Over 90 percent of the antibiotics used were for therapeutic at relatively affordable prices, even in semirural and rural pop- purposes, and there was no evidence of use for growth ulations. This trend may be responsible for the emergence of promotion. nalidixic acid resistance to Shigella in Bangladesh and fluoro- There is strong evidence that the use of antibiotics in farm quinolone resistance to Salmonella typhi in India. animals promotes the development of drug-resistant bacteria Recent evidence suggests that shorter courses of antibiotics in animals (Aarestrup and others 2001). Because routes for the may select for less resistance in the pneumococcus compared movement of these resistant bacteria to humans are available, with longer courses (when patients comply with those courses) there is sufficient circumstantial evidence that drug resistance (Schrag and others 2001). Very low levels of resistance have also in bacteria associated with food animals can influence the level been found in isolated rural African communities (Mthwalo of resistance in bacteria that cause human diseases (Wegener and others 1998). This observation, however, should not lead and others 1999). Furthermore, mathematical models indicate to complacency. Increased access to antibiotics in developing that the effect of subtherapeutic use on resistance in humans 1032 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, and others is greatest when resistance levels are undetectable (Smith and Disease Burden others 2002). The appearance of drug-resistant strains of Although no estimates of disease burden are currently available Enterococcus faecium in broiler meat products at retail outlets that are specific to drug resistance, the contribution of drug declined after the ban of antimicrobial growth promoters in resistance to the burden of infectious diseases is believed to be Denmark (Emborg and others 2003). Salmonella has been large. Resistance has emerged in malaria, HIV, TB, and other recovered from chicken (35 percent), turkey (24 percent), and bacterial infections that together constitute a significant pro- pork (16 percent) samples obtained from area supermarkets in portion of the burden of disease in developing countries. Washington, D.C. (White and others 2001). There is evidence An indication of the extent of the problem is provided by that dissemination of tetracycline-resistance-encoding plas- the burden of diseases for which drug resistance is a problem mids between aquaculture and humans has already occurred in (table 55.1), as well as by the levels of drug resistance among Europe (Rhodes and others 2000). The global nature of this these pathogens (table 55.2 and figures 55.1 and 55.2). problem became apparent in 2001, when authorities in some European countries found residues of chloramphenicol in tiger Pneumococci. Surveillance of drug resistance in pneumococci shrimp imported from China, Indonesia, and Vietnam shows several general trends. The numbers of strains that are (Holmstrom and others 2003). fully susceptible to penicillin-G, once nearly universal in most of the world, have declined by 30 to 50 percent in many countries and by 75 percent in some, as resistant clones have spread widely Transmission of Resistant Pathogens but irregularly throughout the world (Sa-Leao and others 2002). Once resistance has emerged in a population, it can spread At the same time, percentages resistant to macrolides and to both geographically and between age groups. Unsafe drinking sulfamethoxazole-trimethoprim have increased, especially water, unsanitary conditions, and poor infection control in where those drugs have been widely used, and resistance to hospitals are risk factors for the transmission of all infections, tetracycline or chloramphenicol has fluctuated widely. Linked including resistant ones. The transmission of resistant strains resistance to these drugs results in a growing percentage of from children to adults has been suggested by anecdotal reports strains resistant to many or all of them. Resistance to fluoro- as far back as the 1980s (Klugman and others 1986). That asso- quinolones is still rare but is beginning to be observed in many ciation is strongly supported by the role of conjugate pneumo- places (Ho and others 2001; Quale and others 2002). coccal vaccine in reducing antimicrobial resistance among Certain Streptococcus pneumoniae clones have been widely adult pneumococcal bacteremic isolates in the United States disseminated. A penicillin-, chloramphenicol-, and tetracy- (Whitney and others 2003). The association of HIV infection cline-resistant clone (and sometimes erythromycin) of Spanish with pediatric serotypes and antimicrobial resistance in pneu- origin (Spain23F-1) has, since its original description, been iso- mococci suggests the potential utility of this approach in lated in other parts of Europe, the United States, South and reducing the burden of antimicrobial resistance in pneumo- Central America, South Africa, and East Asia (McGee and cocci in developing countries where the burden of disease is others 2001). It is likely that this clone is even more widespread overwhelmingly associated with HIV infection in both children and that the absence of reports from other areas reflects the (Madhi and others 2000) and adults (Jones and others 1998). absence of molecular testing techniques needed to delineate Table 55.1 Estimated Burden of Disease in Disability-Adjusted Life Years, by Cause and Gender, 2001 Both sexes Males Females DALYs Percentage DALYs Percentage DALYs Percentage Condition (Thousands) of total (Thousands) of total (Thousands) of total Infectious and 359,377 24.5 184,997 24.1 174,380 24.9 parasitic diseases Respiratory infections 94,037 6.4 49,591 6.5 44,446 6.4 Diarrheal diseases 62,451 4.3 31,633 4.1 30,818 4.4 Gonorrhea 3,320 0.2 1,437 0.2 1,883 0.3 Tuberculosis 36,040 2.5 22,629 2.9 13,411 1.9 Malaria 42,280 2.9 20,024 2.6 22,256 3.2 Source: WHO 2002b, annex table 3, 194. Drug Resistance | 1033 Table 55.2 Prevalence of S. pneumoniae Not Susceptible to Three or More Drug Classes, Alexander Project 1998­2000 Percentage multiresistant defined as Any three drug classes Any three drug classes Any five or more Region and country N excluding penicillin including penicillin Any four drug classes drug classes Africa 540 14.3 24.8 13.5 3.3 Kenya 277 3.6 16.6 2.2 0.0 South Africa 263 25.5 33.5 25.5 6.8 Eastern Europe 1,109 10.1 11.7 6.0 1.0 Czech Rep. 275 0.7 1.1 0.4 0.0 Poland 453 13.0 15.2 6.4 1.1 Russian Fed. 161 10.6 12.4 3.7 1.2 Slovak Rep. 220 15.5 17.3 14.1 1.8 Western Europe 3,328 14.7 18.4 11.9 4.1 Austria 149 2.7 4.7 2.0 0.0 Belgium 230 13.9 15.7 7.0 2.6 France 444 35.6 49.1 34.9 11.7 Germany 321 4.7 5.9 1.6 0.0 Greece 431 18.6 19.5 13.9 2.1 Italy 304 19.7 22.4 9.9 1.0 Netherlands 185 0.0 1.1 0.0 0.0 Portugal 328 6.1 9.5 5.5 1.8 Ireland 54 9.3 14.8 9.3 1.9 Spain 295 27.8 32.9 25.4 15.3 Switzerland 349 5.7 7.7 4.9 2.3 United Kingdom 238 5.9 6.3 5.0 2.1 Far East 730 53.2 63.2 40.6 23.0 Hong Kong, China 193 76.2 79.3 70.5 60.1 Japan 404 48.3 63.1 29.2 6.4 Singapore 133 34.6 39.9 31.6 19.6 Middle East 314 11.2 18.2 10.5 4.1 Israel 148 8.8 12.2 8.8 2.0 Saudi Arabia 166 13.3 23.5 12.1 6.0 Latin America 2,861 13.3 20.1 12.1 1.9 Brazil 181 2.8 5.0 1.1 0.0 Mexico 248 21.0 31.1 20.2 3.2 United States 2,432 16.2 25.8 15.5 7.0 All isolates 8,882 17.5 23.7 14.6 5.9 Source: Jacobs and others 2003. Note: Drug classes were defined as follows: -lactams (penicillin MIC 0.12 mg/L), macrolides (erythromycin MIC 0.5 mg/L), tetracyclines (doxycycline MIC 0.5 mg/L), phenicols (chloramphenicol MIC 8 mg/L), folate pathway inhibitors (co-trimoxazole MICs 1 mg/L based on trimethoprim component), and quinolones (ofloxacin MIC 8 mg/L). clones, rather than an absence of the organisms themselves. these strains to spread to areas where resistance is uncommon Other globally disseminated S. pneumoniae include specific can no longer be considered remote. clones of serotypes 19F, 14, 19A, 9N, 9V, 3, and 6 (McGee and others 2001). Spread of these pandemic clones has continued, Shigella. In many regions where Shigella, especially Shigella even in areas where successful interventions have reduced dysenteriae, is prevalent and an important cause of infant selective pressure from antimicrobial use (Arason and others mortality, resistance first to sulfamethoxazole-trimethoprim, 2002). With increasing international travel, the potential of then to ampicillin, and commonly to tetracycline and 1034 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, and others Percentage failure 100 80 60 40 20 0 CAR Togo Faso EthiopiaBurundiEritreaRwanda Kenyaanzania Gabon CongoChad CongoGhanaLiberia NigerGuinea Benind`lvoire T ZambiaUganda Senegal Mali LeoneGambiaNigeria Botswana Zimbabwe Cameroon DR Mauritania Mozambique Côte Burkina Sierra Eastern, southern, Great Lakes block Central block Western block Source: WHO Regional Office for Africa, 1997­2002. WHO has established 126 sentinel surveillance sites in 36 African countries that monitor the efficacy of locally used antimalarial drugs by following up patients in clinics. According to standard protocol (13, 14), results are expressed as I) early treatment failure (ETF); II) late clinical failure (LCF): in the future, late parasitological failure (LPF) will be considered as well. Treatment failure for policy change as shown here consists of the sum of ETF + LCF. Note: The box indicates the 25th/75th percentile, the line limits lower/upper values, and where the lines cross, the median. Figure 55.1 Chloroquine Treatment Failure in Africa Prevalence 0.9% 1.0%­2.9% 3.0%­6.4% 6.5% Source: WHO-IUATLD 2004. Figure 55.2 Prevalence of MDR-TB among New TB Cases, 1994­2002 Drug Resistance | 1035 chloramphenicol has emerged and, over recent decades, spread MDR-TB in previously treated cases ranged from 0 to 58.3 per- to half or more of the strains sampled. In the 1990s, resistance cent, with a median of 7.0 percent (WHO 2004). has begun to emerge and spread to fluoroquinolones and third- An estimated 273,000 (at a 95 percent confidence interval generation cephalosporins, which in many places are the last [CI]; 185,000 to 414,000) new cases of MDR-TB occurred effective oral drugs available (Ding and others 1999). In the worldwide in 2000. By simple extrapolation, 70 million people past two decades, emergence and spread of Shigella dysenteriae could be latently infected with MDR-TB, and more than 1 mil- type 1 resistant to sulfamethoxazole-trimethoprim, ampicillin, lion active MDR-TB cases could remain among previously tetracycline, chloramphenicol, and--increasingly--nalidixic treated patients. Despite its threatening potential, MDR-TB acid has reduced the effectiveness of these inexpensive and is--and will probably remain--generally rare. Decades after widely available antimicrobials in the empiric management of the introduction of TB drugs, the global prevalence of MDR- epidemic dysentery (Cunin and others 1999; Hoge and others TB in new patients remains less than 2 percent (Dye and 1995). The alternatives, ciprofloxacin and ceftriaxone, are rela- Espinal 2001). Old animal studies (Cohn and others 1954) and tively expensive and not always available. As a consequence, recent analyses using molecular epidemiology (Garcia-Garcia high fatality rates have been observed in a number of recent and others 2000) suggest that MDR-TB strains might be, on dysentery outbreaks (Legros and others 1998). The emergence average, less infectious. And unlike most other bacteria, of fluoroquinolone-resistant strains has quickly followed. The M. tuberculosis replicates rather slowly (low mutation rate) and unchecked spread of these pathogens could pose a major pub- shares little if any genetic material. Thus, even in the absence of lic health challenge (Sarkar and others 1979). widespread treatment of MDR-TB, its prevalence may not necessarily explode (Kam and Yip 2001). Gonorrhea. Newly drug-resistant strains of gonococci tend to spread rapidly because of their peculiar epidemiology and the Malaria. Chloroquine-resistant strains of Plasmodium falci- lack of control programs. Therefore, it is important to detect parum malaria appeared a half-century ago in Southeast Asia microepidemics of such strains, but this need is rarely met. The and South America and spread across Africa, especially East past half-century has witnessed the successive emergence and Africa, in the past quarter-century (Wellems and Plowe 2001). spread of gonococcal strains resistant to each new drug that The use of molecular markers testing indicates the wide becomes widely used to treat gonorrhea, including sulfon- geographical reach of pfcrt polymorphism for chloroquine amides, penicillin, tetracycline, and sulfamethoxazole- resistance, and dhfr and dhps polymorphisms for sulfadoxine- trimethoprim (Tapsall 2002). Within less than a decade, such pyrimethamine. Current levels of treatment failure of strains have commonly come to account for half or more of the chloroquine are in figure 55.1. There is evidence that malaria isolates in many regions. The recent emergence of resistance to mortality, especially in children under the age of five, is rising as a fluoroquinolones leaves only less available parenteral drugs, consequence of increasing resistance to chloroquine (Greenberg such as spectinomycin or ceftriaxone, as the reliable therapy and others 1989; Trape 2001). In response to increasing treat- (Ison and others 1998; Palmer, Leeming, and Turner 2001). ment failure, many countries, including Malawi, South Africa, and Tanzania, adopted sulfadoxine-pyrimethamine as first-line Tuberculosis. The emergence and spread of multidrug-resistant treatment; however, resistance to this drug too is growing in tuberculosis, which is defined as combined resistance to isoni- many parts of Africa. In Southeast Asia, the emergence of mul- azid and rifampicin, threaten the control of TB globally (Kochi, tidrug resistance to sulfadoxine-pyrimethamine and mefloquine Vareldzis, and Styblo 1993). Patients infected with strains over the past decade and a half has prompted the use of combi- resistant to multiple drugs are very difficult to cure (Espinal nation treatments that include artemisinin (Wongsrichanalai and others 2000; Goble and others 1993), particularly if they and others 2001). are HIV-infected or malnourished (Fischl and others 1992), and alternative treatment is much more toxic and expensive (Drobniewski and Balabanova 2002). A patient with MDR-TB ECONOMIC BURDEN may remain infectious much longer than a patient with drug- susceptible organisms. Among new cases, prevalence of resist- Although few estimates have been made of the economic ance to at least one TB drug ranges from 0 percent in some impact of drug resistance in developing countries, there is some Western European countries to 57.1 percent in Kazakhstan, indication that this burden is substantial. Estimates for costs with a median of 10.2 percent. Multidrug resistance among associated with the loss of antibiotic effectiveness in outpatient untreated patients ranged from 0 percent in eight countries to prescriptions in the United States range from US$378 million 10.0 to 14.2 percent in six others. In previously treated cases, to as high as US$18.6 billion (Elbasha 1999). A report by the resistance to at least one drug ranged in different settings from Office of Technology Assessment to the U.S. Congress estimated 0 to 82.1 percent, with a median of 18.4 percent. Prevalence of the annual cost associated with antibiotic resistance in 1036 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, and others hospitals (attributable to five classes of hospital-acquired compliance with dosing, reduce the likelihood that a resistant infections from six antibiotic-resistant bacteria) to be at least pathogen will survive and proliferate. Prolonging the effective US$1.3 billion in 1992 dollars (Office of Technology therapeutic life of existing drugs is not sufficient, however. Assessment 1995). The U.S. Centers for Disease Control and Increasing incentives for pharmaceutical firms to bring new Prevention (CDC) estimated that the cost of all hospital- drugs to markets may also be called for. acquired infections, including both antibiotic-resistant and antibiotic-susceptible strains, was US$4.5 billion. Patients infected with resistant strains are more likely to be Drug Treatment Strategies sicker, to be hospitalized for longer periods of time, and to die The appropriate choice of drug treatment is an important step of the infection (Carmeli and others 2002). Both the duration in delaying the evolution of drug resistance. Drug combina- of hospitalization and the attributable cost of treating tions that include drugs with different targets were first used methicillin-resistant Staphylococcus aureus were found to be in the treatment of tuberculosis and have now become rou- nearly three time as large as those for a susceptible infection tine practice in the treatment of cancer and HIV/AIDS. (Abramson and Sexton 1999). One problem with estimating Combinations of artemisinin and its derivatives with other the attributable morbidity and mortality that is caused by antimalarials, notably mefloquine, have accelerated recoveries, resistant pathogens is that patients who are infected with resist- increased cure rates, and reduced transmissibility. In the ant strains are more likely to have been sicker in the first place. refugee camps on the western border of Thailand, where most Therefore, the ability to appropriately control for the underly- of the recent studies with artemisinin combinations have been ing severity of the illness that causes hospitalization is a conducted, the use of combinations delayed the development concern. of resistance and reduced the incidence of disease (Nosten and Another important cost of resistance comes from the need others 2000). The rationale behind drug combinations is that, to move to second-line treatments, which are often much more if resistance results from spontaneous genetic mutations, the expensive than the first-line treatment that is no longer effec- chance that a parasite will emerge that is simultaneously resist- tive. For instance, treating the roughly 300 million cases of ant to two drugs with unrelated modes of action (that is, drug malaria with artemisinin-based combinations would involve targets) is the mathematical product of the individual parasite an excess burden of roughly US$200 million each year in drug mutation frequencies multiplied by the total number of para- costs. Periodically changing first-line treatment may also involve sites exposed to the drugs (White 1998, 1999). Combinations, costs of assessing alternate treatment regimens, retraining therefore, reduce enormously the probability that a resistant health care providers, and restocking health care facilities. mutant will arise. Sequential deployment of the drugs is much Though all these impose a significant economic burden, espe- less effective, because it does not exploit the multiplicative cially in poorer countries, they may be an inevitable conse- reduction in selection risk. quence of past drug use. A focus on the cost of resistance alone In the context of antibiotics, combinations have typically may be misleading, because it is potentially possible to elimi- been used to broaden the spectrum of antimicrobial coverage nate drug resistance by not using any drugs. To appropriately rather than to reduce the likelihood of the emergence of assess the net benefits of drug use, one must include the cost of resistance. With the development of new penicillins and increased resistance and the benefits of antibiotic or antimalar- cephalosporins with broader spectra of activity a decade ago, ial use in treating infections and preventing their spread to most serious infections have been treated with monotherapy. uninfected individuals. The use of combination therapy to preserve new classes of antibiotics from the emergence of resistance at a societal level may be rational, but it has not been implemented because of INTERVENTIONS cost concerns and the potential for enhanced toxicity associ- ated with the use of more agents than necessary to effect a cure In this section, we discuss interventions to address the chal- in an individual patient. lenge of drug resistance (table 55.3). Many interventions to Other strategies include periodic withdrawal of a drug or address the problem of resistance are the same as those that rotation between different drugs. These strategies depend on reduce the burden of disease (these are discussed in detail in the extent of the fitness cost of resistance1 and the extent of the relevant disease-specific chapters in this volume). Reducing multidrug resistance, which may vary with the specific combi- disease diminishes the need for drug treatment and, therefore, nation of pathogen and drug. Withdrawal or drastic decline in lowers the likelihood that resistant strains will emerge. Some antimicrobial use is occasionally but not always accompanied interventions, such as the use of drug combinations, reduce the by the replacement of resistant strains with sensitive ones. The likelihood that resistance will emerge, whereas other interven- effects of antimicrobial removal have best been assessed tions, such as improvements in drug prescribing and patient for antibacterial drugs in northern Europe, where drug use is Drug Resistance | 1037 Table 55.3 Potential Nonclinical Interventions: Evidence from Developing Countries Strategy Intervention Reference Description Study location Treatment Combination therapy Nosten and others 2000 Use of an artesunate-mefloquine combination Thailand strategies for malaria was found to reduce incidence of mefloquine resistance in Plasmodium falciparum malaria. Cycling strategy Kublin and others 2003 Replacement of chloroquine with sulfadoxine- Malawi pyrimethamine resulted in a decline in chloroquine-resistant strains over an eight-year period to levels that permit reintroduction of the drug. Drug heterogeneity Bonhoeffer, Lipsitch, and Modeling studies demonstrated the superiority n.a. Levin 1997; Laxminarayan and cost-effectiveness of policies involving and Weitzman 2002 use of different antibiotics on different patients compared with those using the same antibiotics on all patients. Directly observed Balasubramanian, Oommen, Directly observed therapy reduced the Kerala, India therapy short course and Samuel 2000 probability of treatment failure. Dye and others 2002 Directly observed therapy for TB was 2.8 times South Africa cheaper to deliver and between 2.4 and 4.2 times more effective than conventional treatment. Reducing selection Training providers Bexell and others 1996 Continuing education seminars for paramedical Lusaka, Zambia pressure prescribers resulted in patients being prescribed antibiotics less frequently at intervention centers (34 percent) compared with control centers (42 per- cent). Drug choice and dosing were also improved. Santoso, Suryawati, and One-on-one educational interventions and seminars Yogyakarta and Prawaitasari 1996 for medical and paramedical prescribers reduced Central Java antimicrobial prescription by 17 and 10 percent, provinces, respectively (p 0.001). Indonesia Training drug sellers Agyepong and others 2002 Training drug dispensers on patient communication Dangme West resulted in modest improvements in the proportion District, Ghana of patients showing strict, full adherence to antimalarial regimen. Treatment guidelines Qingjun and others 1998 Blister packages increased compliance with Hunan province, with education chloroquine therapy to 97 percent, from 83 percent China in the control group. Direct education of Helitzer-Allen and Introduction of a nonbitter antimalarial tablet and Malawi patients others 1993 a new educational message were effective in improving antimalarial prophylaxis compliance among pregnant women by 57 to 91 percent. Paredes and others 1996 Video, radio, and printed bulletins were used to Lima, Peru educate women in an intervention community on the management of watery infantile diarrhea. The overuse of nonindicated medicines (antibiotics and antidiarrheals) dropped 11 percent in the intervention group and only 7 percent in the control group. Reducing spread of Hand washing Kurlat and others 1998 Training of nursing staff in hand washing, handling of Argentina resistance pathogens infants, and care of intravenous lines resulted in 40 percent reduction in bacterimia rates. Bednets (malaria) Maxwell and others 2002 Use of netting resulted in a 55 to 75 percent Tanzania reduction in malaria morbidity and consequent conservation of antimalarial drug use. Vaccination Klugman 2001 Pneumococcal vaccines target the serotypes most South Africa commonly encountered clinically, which are more likely to be resistant to antimicrobials. Source: Authors. 1038 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, and others tightly regulated and susceptibility patterns are closely moni- Prescribing Patterns. Studies in developing countries have tored (Seppala and others 1997). A recent study demonstrated shown that as much as a third of drug prescriptions, account- that after chloroquine was replaced by sulfadoxine- ing for 20 to 50 percent of drug costs, are irrational and that pyrimethamine in Malawi because of a loss of effectiveness antimicrobials are among the most frequently prescribed med- attributable to resistance, chloroquine-susceptible Plasmodium ications (Bosu and Ofori-Adjei 2000). Although altering pre- falciparum strains appear to have returned (Kublin and others scribing behavior is an important intervention to control drug 2003). Though the results of this study offer promise for stop- resistance, the widespread availability of drugs without a pre- ping or reversing resistance trends, if antimicrobials are more scription limits its effectiveness. The prescribing problem may sparingly and less indiscriminately applied, little is known be worse among private practitioners than among public prac- about the rate at which the resistance to chloroquine may titioners (Siddiqi and others 2002). Continuing education for reemerge with widespread use of this drug. practicing health workers is one type of intervention that has Rotating between two or more antibiotics has been proposed been tested in several countries. In the United States, a decline to address the problem of nosocomial drug resistance in the in antimicrobial prescribing in pediatric ambulatory care has United States (Bergstrom, Lipsitch, and McGowan 2000; been attributed to educational programs directed at physicians McGowan 1986), even though there is not much supporting as well as the public (McCaig, Besser, and Hughes 2002). In empirical evidence to date. In one hospital-based study, switch- developing countries, successful educational programs for pre- ing from gentamicin to other aminoglycosides reduced resist- scribers have improved diagnostic quality, dispelled percep- ance to gentamicin. However, when gentamicin was reintro- tions of patient pressure, reduced unjustified antimicrobial duced, resistance developed rapidly (Gerding 2000). Modeling prescription (Chuc and others 2002; Hadiyono and others studies have indicated that a superior strategy may be to increase 1996), and reduced polypharmacy (Hadiyono and others 1996) antimicrobial heterogeneity so that different patients are treated among private as well as public providers, including nonphysi- with drugs to which mechanisms of resistance are independent cians (Chakraborty, D'Souza, and Northrup 2000; Chuc and (Bonhoeffer, Lipsitch, and Levin 1997; Laxminarayan and others 2002). In general, these measurable outcomes were Weitzman 2002).Although this may be difficult to implement in improved by 5 to 20 percent by a single intervention--a modest many developing countries, the approach incorporates an evo- but significant change that is best combined with parallel inter- lutionary perspective that may help deal with drug resistance. ventions. Although cost-effectiveness was not a focus of the studies, the resultant reduction in drug use would ultimately Reducing Selection Pressure result in cost savings. Important components of educational interventions are long-term commitment and refresher Inappropriate antimicrobial use constitutes selective pressure courses, and complementary interventions are also desirable without a corresponding benefit to individual or public health. (le Grand, Hogerzeil, and Haaijer-Ruskamp 1999). (Eliminating all antibiotic use could, of course, eliminate the Prescription guidelines, essential drug lists, and formularies problem of drug resistance, but this strategy is clearly undesir- are essential for defining policy and provide a useful framework able.) This multifaceted problem arises from behaviors of on which educational interventions can be based (Laing, prescribers (not always physicians),dispensers (not always phar- Hogerzeil, and Ross-Degnan 2001). The World Health macists), and consumers (not always infected). An important Organization (WHO) recommends standard treatment guide- factor in overprescription is the issue of externalities; physicians, lines as one of several approaches for promoting rational drug patients, and pharmacists have few incentives to consider the use. Also, guidelines proposed by pharmacy and therapeutics effects of their prescriptions or drug use on overall levels of committees or external advisers have been applied in developing resistance and the burden imposed on the rest of society. countries, with mixed results. Although standard treatment Physicians, both in private practice and in hospital settings, may guidelines reduced antibiotic use for respiratory infections by also derive income from drugs sold and may,therefore,prescribe 50 percent in Fiji (le Grand, Hogerzeil, and Haaijer-Ruskamp antibiotics more frequently than is desirable. In China, for 1999), they did not alter prescription patterns in a Ugandan instance, many hospitals rely on selling drugs for the bulk of study and produced a detectable but insignificant effect in Sri their revenue (Hu and others 2001).Patient pressure demanding Lanka (Angunawela, Diwan, and Tomson 1991). In general, a prescription is known to influence prescribing in developed follow-up was essential for success, and the use of standard countries but could be less important in developing countries. treatment guidelines was more effective with nonphysician prescribers (le Grand, Hogerzeil, and Haaijer-Ruskamp 1999). Interventions at the Provider Level Education must form part of any treatment guideline interven- In this section, we discuss interventions directed at health care tion, and evidence suggests that, if anything, educational providers and local retail pharmacies, such as education and programs are more effective than simply formulating guidelines professional accountability. (Laing, Hogerzeil, and Ross-Degnan 2001). Rigid guidelines Drug Resistance | 1039 such as preprinted order forms or prepackaged drug kits for the educational interventions in developing countries vary but can management of community-acquired infections have been per- be broadly classified into focus group discussions (Hadiyono ceived as excessively prescriptive and have not been successful and others 1996) and large seminars (Bexell and others 1996). intervention tools (le Grand, Hogerzeil, and Haaijer-Ruskamp Both models have been found to be of comparable effective- 1999). Devising incentives for compliance could potentially ness; however, applicability and cost are situation-specific lower the higher prescription rates among private providers, (Santoso, Suryawati, and Prawaitasari 1996). where treatment guidelines by themselves are less likely to work. Peer and supervisory monitoring increases professional Drug Quality. Ensuring drug quality is important, both to accountability, thereby promoting the application of knowl- benefit the individual patient and to ensure that a patient is not edge to practice. The requirement that antibiotic prescriptions subjected to suboptimal doses that would promote drug resist- for inpatients be countersigned by an infectious disease con- ance. The few studies that have been conducted indicate that sultant was successful in reducing prescriptions by 50 percent, more than half of the antimicrobials marketed in developing with a resultant cost savings of about US$350,000 over two countries do not match their labels. Hence, even when pre- years in a Panama hospital (Saez-Llorens and others 2000). scribers and consumers are using the drugs responsibly, thera- Such a program would have limited applicability in other peutic failure and subinhibitory levels of antimicrobials may countries and settings where the number of trained medical occur. Substandard drugs are those that are degraded as a result professionals is small. A supervisory program in Vietnam, with of expiration or improper storage or that are counterfeit medical equipment incentives, reduced the number of patients (Okeke and Lamikanra 2001; Prazuck and others 2002; Taylor for whom antibiotics were prescribed and increased the num- and others 2001). ber who received a correct dose regimen (Chalker 2001). Diagnostic Tests. Bacterial culture and susceptibility testing, a Interventions at the Patient Level necessary component of rational antimicrobial prescribing, is Improving communication between patients and providers uncommon in many developing countries (Okeke, Lamikanra, could improve adherence to prescribed antimalarial regimens and Edelman 1999). Furthermore, diagnostic tests to confirm or (Agyepong and others 2002). Compliance can also be positively refute infections are also commonly unavailable or unreliable,so affected by packaging. Blister packages, when combined with diagnoses are made largely on the strength of clinical signs and proper instruction about drug use, have been shown to pro- symptoms (Berkley and others 2001). Laboratory tests are duce modest increases in antimalarial compliance, particularly expensive and routinely cost more than an empiric drug that for long-term regimens, such as with primaquine (Qingjun and could be effective. In contrast, malaria dipsticks can be an inex- others 1998). Blister packages are also time-savers for primary pensive tool for case detection and may be cost-effective in low health care workers, potentially allowing them more time to transmission settings (Rimon and others 2003). Clinicians have advise patients on drug use. However, the introduction of blis- been known to use chemotherapy as a diagnostic tool: a cure ter packages must be accompanied by clear directions to avoid would confirm a diagnosis. Susceptibility testing of at least some injury following ingestion of blister packages. specimens will provide much-needed surveillance data to sup- Reducing patient self-medication may also desirable, port empiric prescribing, although efforts should be made to although the effect on drug resistance remains to be precisely take into account spatial heterogeneity in resistance patterns. quantified. Enforcement of prescription-only regulations for most antimicrobials reduces self-medication in developed Retail Pharmacies and Outlets. Drug distributors, including countries (Carey and Cryan 2003; Goff, Koff, and Geiling not only pharmacists but also pharmacy attendants, patent 2002; Pechere 2001) and may be desirable in developing coun- medicine stallkeepers, and itinerant drug sellers, often sell tries as well. However, such a strategy may be difficult to drugs without prescription and are an important source of pri- implement in developing countries. There have been only rare mary care for people in many developing countries (Igun 1994; reports of reduction in antibiotic use following blanket Indalo 1997). Patients in search of convenient and accessible enforcement of prescription supply legislation in areas where health care often seek treatment at drug retail premises. Many antibiotics are freely available (Bavestrello, Cabello, and drug sellers have not been formally trained in diagnosis and Casanova 2002). Opponents to enforcement demand a heavy prescription but often have financial incentive to perform those financial and political investment, implying that a black mar- tasks, with varying degrees of competence. Despite the poten- ket for medicines could emerge, particularly if the demand for tial loss of business to storekeepers, educational interventions these drugs is not lowered (Bhutta and Balchin 1996; le Grand, have been successful in increasing prescription requirement Hogerzeil, and Haaijer-Ruskamp 1999). The sale of antimicro- demands and promoting referral advice, all steps in the right bials is a lucrative business, even when illegal, because of high direction (Chuc and others 2002). Models for delivering demand. 1040 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, and others Educational interventions directed at consumers have been storekeepers and former TB patients (Barker, Millard, and proposed to reduce self-medication and increase compliance, Nthangeni 2002; Pungrassami and Chongsuvivatwong 2002). but evidence for the effectiveness of this strategy remains Although DOTS was not designed to cure patients with inconclusive, largely because so few studies have been con- MDR-TB, it succeeds in 50 percent of cases (Espinal and others ducted. Gonzalez Ochoa and others (1996) demonstrated that, 2000). Controversy has emerged about the best approach although refresher training for health personnel managing to MDR-TB in resource-constrained settings. Although some acute respiratory infections reduced antibiotic prescribing by experts assert that standard TB control prevents the emergence 9 to 19 percent in two intervention areas in Havana, no benefit of MDR-TB in a cost-effective way (Chaulet, Raviglione, and was seen from community education programs when used Bustreo 1996) and that expensive treatment of MDR-TB would alone or in conjunction with prescriber training. In contrast, divert scarce resources from struggling DOTS programs, others Denis (1998) was able to show that a poster had 5 percent effec- argue that it is unethical to abandon MDR-TB patients and tiveness alone and 20 percent effectiveness when used with a maintain that, if untreated, MDR-TB strains will become dom- video to promote appropriate use of quinine and tetracycline inant, undermining tuberculosis control in future generations regimens for malaria among Cambodian villagers. The best (Farmer, Becerra, and Kim 1999). These arguments are of par- method to deliver information about the consequences of ticular consequence for programs in poor countries. antibiotic resistance remains to be identified and will need to be modified to suit different cultures (Marin and others 1995). Application of Antibiotics for Nonhumans In the case of tuberculosis, first-line directly observed thera- The use of antimicrobial agents in agriculture and aquaculture py (DOT) remains one of the most cost-effective of all public can contribute to the spread of antimicrobial resistance in health strategies (WHO 1994). The discovery that DOT can be humans, although the extent of this contribution has not been administered successfully as a short course (DOTS) has been precisely quantified. It is believed that agricultural antibiotic pivotal to successful implementation. Relatively simple, the use hastens the emergence of resistant pathogens in humans, DOTS approach can improve patient compliance, cure the and antibiotic use in humans contributes to the spread of vast majority of new TB patients, and prevent transmission of resistance once it has emerged (Smith and others 2002). There the disease and the emergence of MDR-TB (Balasubramanian, are no reliable estimates of the extent to which antibiotics are Oommen, and Samuel 2000; Dye and others 2002). Un- being used for such nonhuman purposes even in developed fortunately, many countries have been slow to adopt and imple- countries. In developing countries, one would expect use to ment DOTS programs correctly, and only a minority of TB increase with rising incomes and greater industrialization of patients worldwide are managed according to this protocol (Dye agriculture and food production. and others 2002; Pungrassami and others 2002). Because many WHO has recommended that antimicrobials normally pre- patients either are treated outside the DOTS regimen or do not scribed for humans should no longer be used to promote adhere to the long-term chemotherapy necessary to eradicate growth in animals (WHO 2000). Some countries in Europe, the causative organism, MDR-TB is likely to emerge and treat- including Denmark and Sweden, have already phased out such ment costs are likely to escalate to as high as 1,000 times the cost use, and under the current plan, the European Union will ban of conventional treatment of drug-sensitive infection. the use of antibiotics for growth promotion by 2006. Evidence A DOTS program requires good laboratory support for case of the effect of a ban on antibiotic use in swine production is identification and is highly employee intense, requiring health available from Denmark. Although the ban significantly low- workers to observe the ingestion of every dose of antimicro- ered the use of antibiotics in growth promotion and raised the bials over several months. However, the potential economic cost of swine production by less than 1 percent, the resulting gains are immense because of the high costs of allowing the higher incidence of disease among swine increased the use of continued spread of the disease and of managing resistant antibiotics used for veterinary therapeutic purposes (Hayes patients and because of the benefit of reduced hospital admis- and Jensen 2003). A similar increase was noted in Sweden, sions (Dye and others 2003). Several studies have investigated where antibiotic use in animals was banned in 1986; however, means for reducing the cost of conventional DOTS programs this problem was temporary, and in the longer term, livestock without compromising effectiveness. Lwilla and others (2003) producers were able to move to an effective production system demonstrated that both institutional and community-based with lower antibiotic use (Wierup 2001). DOTS programs are effective, permitting cost-effective imple- mentation in remote areas. Furthermore, although DOTS involves supervised drug dosing, highly trained health workers Containing the Spread of Resistant Micro-organisms are not essential. Studies in South Africa, Haiti, and Thailand Although selection is a necessary component of the resistance have found that DOTS was effective when drug administration archetype, the dissemination of resistant organisms may have a was supervised by appropriately trained volunteers, including far greater effect on the current situation (Zaidi and others 2003). Drug Resistance | 1041 Interventions that block this dissemination have the added ben- also across the whole range of sectors involved. International efit of improving health by interrupting disease transmission travel and trade, particularly of food products, have facilitated and reducing the need for antibiotics in the first place. A case in the rapid globalization of resistance, and actions undertaken by point is the reduction in drug-resistant Streptococcus pneumoni- any single nation have consequences for other countries. Drug ae infections in the United States following the introduction of a resistance may threaten health gains made in other spheres of multivalent vaccine that protected against the serovars with public health. Thus, coinfection with antimicrobial-resistant which resistance is commonly associated (Whitney and others pathogens and HIV can lead to more rapid disease progression 2003). By contrast, a drop in antibiotic use had no detectable and enhanced dissemination of resistant pathogens. The emer- effect on resistance within a short period (Arason and others gence of antimicrobial resistance is considered a major threat 2002). Because these types of interventions are easier to evaluate, to the future security and political stability of some regions and in many cases are cheaper to implement (Coast and others (CIA 1999). Figure 55.3 shows the likely geography of the 2002),they are likely to be of great value to public health in devel- emergence, spread, and evolution of chloroquine resistance. oping countries. Concerted international action is needed to contain antimi- Interventions that interfere with the spread of many infec- crobial resistance. Failure of individual countries to act to tious diseases will have a parallel effect on the dissemination of contain resistance could lead to both national and interna- resistant micro-organisms. Because the potential health bene- tional consequences. For instance, the use of artemisinin as fits are obvious, these types of interventions are likely to be monotherapy in any single country could potentially lead to sustained (Wilson and Chandler 1993). A simple, cost-effective the rapid evolution of resistant strains, which could threaten example is hand washing, which could reduce diarrhea by the use of this valuable drug in any other part of the world. 47 percent while also having beneficial effects on acute There is a considerable need for greater international col- respiratory tract and other community-acquired infections laboration on surveillance of antimicrobial resistance, both for (Curtis and Cairncross 2003) (see chapters 19 and 35). routine surveillance and as an early warning system for unusual Similarly, insecticide-impregnated bednets are another impor- resistance events. Although existing laws at the international tant intervention for the control of malaria (see chapter 21). level require reporting of some infectious diseases, they do not The emergence, persistence, and intra- and interhospital include any systematic reporting of antimicrobial resistance. spread of multidrug-resistant organisms have all been facilitated Certain multiresistant pathogens, such as methicillin-resistant by inadequate infection control practices. Furthermore, the S. aureus, are now notifiable at the national level in some coun- emergence and spread of drug-resistant nosocomial pathogens tries, but the global nature of the resistance problem means from hospitals to the community are also a concern, and a that national legal measures alone are not enough. At the same history of hospitalization has been identified as a signifi- time, the creation of new international duties would be under- cant risk factor for the acquisition of a resistant infection in mined if they were not incorporated into national law (Fidler family members (OR 4.5, p 0.007) (Zaidi and others 2003). 1998). Unfortunately, we lack good clinical trials that compare the To support such surveillance, the World Health Organi- different approaches to infection control programs and their zation makes software (WHONET) available to enable clinical ability to control antimicrobial resistance in hospitals and other laboratories to enter their drug susceptibility test results into health care facilities (Duse and Smego 1999). It seems reason- databases that can be analyzed for local management of resist- able to assume, however, that if the overall frequency of noso- ance. Those results can also be merged, creating, thus far, comial infections is decreased in a health care facility, then the approximately 80 national surveillance databases. Statutory need for antimicrobial agents may be reduced. Furthermore, notification about pathogens with new resistance phenotypes is well-structured and rational infection control strategies that under active discussion in several countries. Interpretation of balance resources with the magnitude of the local problem existing surveillance data is hampered by the multiplicity must surely play an important role in decreasing morbidity, of methods used to measure resistance and by the difficulties of mortality, and costs (direct and indirect) to the patient, his or assessing the quality of the data. Enhanced laboratory capacity her family, the hospital, and the health care sector in general. is needed in many countries to provide effective diagnostic Hospital-acquired infections rank among the most important services and resistance surveillance. Multicountry external causes of death, either directly or indirectly, in the developing quality assurance schemes already exist but need to be extended world (Duse and Smego 1999; Ponce-de-Leon and Rangel- to cover more resource-poor nations. WHO has begun estab- Frausto 1993). lishing international surveillance standards by issuing antimi- crobial resistance surveillance standards (WHO 2002a), guide- Global Coordination lines for the management of drug-resistant tuberculosis (WHO Antimicrobial resistance is a global challenge that requires 1997), and protocols for detection of antimicrobial drug resist- urgent global action, not just across national boundaries but ance (WHO 1996). Monitoring the use of antimicrobial agents 1042 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, and others 1980s 1960s 1970s 1957 1959 1980s 1960s 1978 1960s 1970s 1980s Source: Cell 1997. Figure 55.3 Global Spread of Chloroquine-Resistant Strains of P. falciparum is an important aspect of surveillance and needs to be strength- than monotherapies. The underlying rationale is that this ened internationally. strategy would both save lives in the short term and delay the Coordinated surveillance for drug resistance to TB may be emergence of resistance in the long term, thereby benefiting all a useful prototype to follow. By 2003, a global network of 20 malarious countries (Arrow, Panosian, and Gelband 2004). supranational reference laboratories and five regional subnet- works, known as the World Health Organization­International Encouraging the Development of New Drugs Union against Tuberculosis and Lung Disease (WHO­ The development of new products has not kept pace with the IUATLD) Global Project, was reporting data on representative problem of addressing microbial resistance to drugs that are cases of TB from 62 countries. The network has monitored par- used to treat infections in human and veterinary medicine. In ticipants' laboratory testing methods and developed uniform the context of drug resistance, it is not sufficient to design poli- definitions, such as establishing clinical categories for new cases cies that will encourage only the development of new drugs. It and previously treated cases. The WHO­IUATLD Global is also important that such policies give pharmaceutical firms Project, which covers approximately 33 percent of the world's (a) an incentive to invest in new classes of drugs, rather than population and 28 percent of its reported cases of TB, and other products that belong to existing drug classes and that are, surveillance efforts reveal great variations from place to place in therefore, more likely to be rendered obsolete by resistance, the prevalence of resistance to antituberculosis drugs. and (b) an incentive to care about drug resistance when mak- The WHO Global Strategy for Containment of Anti- ing decisions about how to market and sell their product. (See microbial Resistance (WHO 2001) recommends more than 60 chapter 6 for a discussion of product development not specific interventions that can slow the emergence and reduce the spread to drug resistance.) of resistance in diverse settings, and it provides a blueprint for Current public and private investment in drugs, vaccines, global coordinated action. A recent report by the Institute of and other products to control major infectious diseases in Medicine calls for such coordinated action at a global level to developing countries has recently been less than 2 percent of subsidize artemisinin-based combination treatments for total health research expenditures throughout the world. Of the malaria for deployment in developing countries to ensure that 1,393 new chemical entities registered by Western health combination treatments are available at the same or lower price authorities between 1975 and 1999, only 13 (less than 1 percent) Drug Resistance | 1043 were intended for the treatment of tropical diseases, and half of IMPLEMENTATION OF STRATEGIES: those came from veterinary research. TWO LESSONS FROM EXPERIENCE Evidence suggests that in recent years, a steadily decreasing proportion of pharmaceutical company profits has been Integrated Management of Childhood Illness invested in antimicrobial research and development (R&D). The program on the integrated management of infant and One reason given for this decline is increased development childhood illness (IMCI) was initiated by WHO, the United costs, which result from more complex clinical trials, longer Nations Children's Fund, and other technical partners. The development times, and the relative attractiveness of returns program, which provides a framework for stepwise assessment from investing in drugs for chronic diseases, which must be and management of sick children, has been successful in avert- taken continuously rather than just for the duration of an infec- ing unnecessary antibiotic and antimalarial use and in reduc- tion. Efforts to reduce inappropriate antibiotic use may para- ing the cost of medications (Gove 1997). An underlying objec- doxically also reduce incentives for pharmaceutical firms to tive of IMCI is to detect, where present, conditions other than invest in developing new products (Monnet and Sorensen those responsible for the primary complaint. The result is that 2001; Philipson and Mechoulan 2003). The related issue, of rational prescribing may actually increase, but because IMCI using shorter treatment courses for many infections, may also guidelines reduce irrational prescribing, the net effect is a affect R&D incentives (Pichichero and Cohen 1997). reduction in drug use, including antimicrobials (Oluwole, Encouraging research into antimicrobial agents that will be Mason, and Costello 2000). Drug costs associated with treating used primarily in low-resource countries poses particular chal- children in Kenya were found to be reduced from an average of lenges, given the need for drug companies to make a profit. 44 cents per patient (U.S. currency) to between 16 and 39 cents Various incentives to the industry, including both push and per patient when IMCI guidelines were followed (Boulanger, pull mechanisms, have been discussed (Kettler 2000). Push Lee, and Odhacha 1999). A study that compared prescriptions mechanisms consist of incentives to offset R&D costs, such as arising from standard and IMCI-guided consultation with a research grants, tax credits, public investment in applied health worker in Kaduna state, Nigeria, found a reduction in research, cost-sharing between companies, and establishment polypharmacy when IMCI guidelines were used, from a medi- of local development facilities. Pull mechanisms are designed to an of five drugs down to two drugs, and an 80 percent reduc- create a market, thereby improving the likelihood of a return tion in the cost of all medicines (Wammanda, Ejembi, and on investments. They could include an international purchase Iorliam 2003). This reduction represented savings of 93.4 and fund for a new antimicrobial that meets specific criteria 68.6 percent for antibacterials and antimalarials, respectively, (United Kingdom Government 2001), tax credits on sales, and although part of the savings could be attributed to the substi- favorable patent policies, such as extension of patent terms or tution of tablets for more expensive syrup formulations. market exclusivity on new products. Time-limited exclusivities The contribution of IMCI to promoting judicious drug use is on new, clinically useful formulations might stimulate the worth noting. A study that compared four districts of Kenya-- additional pharmaceutical and clinical studies that are needed two districts that had IMCI programs (Morogoro and Rufiji) to support licensure. Better patent protection for new anti- and two that did not (Ulanga and Kilombero)--found that in microbial agents in areas of the world where patent laws are not 73 percent (95 percent CI; 65­80) of consultations studied in the enforced today would also help. IMCI intervention districts, a child needing an oral antibiotic or Patent policies to encourage investment in developing new antimalarial was prescribed correctly, compared with 35 percent antibiotics should also take into consideration the effect of those (95 percent CI; 25­45) in the control districts (Armstrong policies on incentives that would encourage pharmaceutical Schellenberg and others 2004). Also, in the IMCI districts, firms to conserve the effectiveness of their products. If effective 86 percent (95 percent CI; 80­92) of children who did not need patent lengths are too short, pharmaceutical companies will be antibiotics left the facility without an antibiotic, compared with less likely to care about growing drug resistance to their product 57 percent (95 percent CI; 48­66) in the control district. and will be more interested in maximizing sales of their prod- ucts during the period of patent production. Extending patent lengths may not solve the problem, however. If different firms Directly Observed Therapy make closely related antibiotics that have linked mechanisms of Strict compliance and cure verification make it possible to con- action, no single firm may have an incentive to care about drug duct directly observed therapy using shorter-term regimens resistance (Laxminarayan 2002). Increasing patent breadth, one (the DOTS program). Outpatient treatment of a single TB case solution to this problem, would have the added advantage in Pakistan, with complete compliance, was recently estimated of creating an incentive for firms to invest in developing new at US$164 and increased to US$310 when an institution-based classes of drugs rather than introducing drugs that are closely DOTS program was applied (Khan and others 2002). In related to existing products in modes of action. Beijing, the cost of saving one disability-adjusted life year 1044 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, and others (DALY) is 10 times higher when a DOTS TB program is not parts of the developing world could further increase drug use. used (Xu, Jin, and Zhang 2000). Recent expansion of DOTS Because developing countries will be less able to bear the costs care in India has resulted in savings of about 0.2 million lives of increasing resistance, it is important that the patterns of and US$400 million in indirect costs (Khatri and Frieden overuse and misuse observed in high-income countries not be 2002). Progress with DOTS in India was initially slow, but repeated. The right financial incentives must be in place for following better implementation, recent findings suggest that both patients and physicians to face the full cost of using huge successes have been made in the past five years. The suc- antibiotics and antimalarials and to ensure that these drugs are cesses can largely be attributed to improved management of not overused. the program, area-specific appraisals, infrastructure, personnel, and technical support as well as continuous supervision (Khatri and Frieden 2002). This model is one that may apply in Recommended Interventions other developing countries. Sets of interventions specific to the diseases discussed in this In New York city, individualized chemotherapy based on chapter are described below. drug susceptibility testing was nearly as effective in new patients with MDR-TB as in those with drug-susceptible TB Pneumococci. A number of affordable interventions can be (Telzak and others 1995), and the number of MDR-TB cases considered. Data suggest that hand washing can interrupt not decreased by more than 90 percent within the decade (Fujiwara only the spread of pathogens causing diarrhea but also the and others 1997). The relatively large number of MDR-TB transmission of nosocomial pathogens, and it can reduce respi- cases in high-income countries made second-line drugs more ratory infection­related mortality and morbidity. The promo- available and affordable, and experience in the management of tion of a culture of hand washing requires access to clean, MDR-TB grew considerably over the past decade (Mukherjee sufficient water--an urgent goal. Patient education to reduce and others 2004). Yet questions remained about which inter- demand for antibiotics for viral infections, and alteration of ventions had led to the city's success. In 1999, WHO created a perverse incentives for physicians to prescribe antibiotics exces- working group on DOTS-plus to assess the feasibility and cost- sively are recommended. Pneumococcal conjugate vaccine has effectiveness of treating MDR-TB in low- and middle-income been shown to reduce the burden of antibiotic-resistant invasive countries (Espinal and others 1999). DOTS-plus has already disease, reduce transmission of resistant strains, and reduce negotiated a 90 percent price reduction from the pharmaceuti- antibiotic use in vaccine recipients and their siblings. Strategies cal industry (Gupta and others 2001), and experience from for introducing this type of vaccine into the high-risk popula- Peru suggests feasibility and a mean cost of US$211 per DALY tions of developing countries are urgently needed. Better gained (Suarez and others 2002). diagnostics are a key to more appropriate and focused prescrib- ing, but their application is likely to be delayed in poor countries. The long-term effects of blanket recommendations AGENDA FOR ACTION for antimicrobial prophylaxis among HIV-infected populations must be closely monitored. Although the evolution of resistance is a biological phenome- non, it is influenced strongly by the behavior of physicians, Shigella. Interventions must include improved strategies for patients, and hospital administrators. In the language of eco- case definition, clinical recognition, and appropriate therapy of nomics, drug resistance is an externality associated with the use dysentery. These strategies require a focus on educating physi- of antibiotics--a consequence not taken into consideration cians and caregivers as well as restrictions on over-the-counter by those who use antibiotics or antimalarials. From a public availability of antibiotics in developing countries. Recent policy perspective, there may be an economic case for societal strategies for reducing antimicrobial prescribing for diarrhea at intervention, because patients, physicians, and nations acting the community level also include coadministration of zinc with on self-interest alone may produce a higher degree of drug oral rehydration therapy. This strategy has been shown to lead resistance than is ideal for society as a whole. From a global to significant reduction in antimicrobial prescribing for diar- perspective, there is a case for international coordination to rhea in Bangladesh and may be a useful adjunctive strategy ensure that the actions of any single country or region do not (Baqui and others 2002). increase the likelihood of emergence of resistance, which could then spread to other parts of the world. Tuberculosis. The top priority must be the expansion of DOTS Rising incomes in the developing world are likely to encour- (Iseman, Cohn, and Sbarbaro 1993), which itself can prevent age greater use of Western medicine and, consequently, greater the emergence of MDR-TB (Dye and others 2002). In individ- use of antibiotics. Moreover, the adoption of government- ual countries or parts of countries, however, additional strate- sponsored or employer-sponsored insurance plans in many gies may be appropriate. Pablos-Méndez, Gowda, and Frieden Drug Resistance | 1045 (2002) have grouped countries according to the proportion of chapter and has been accomplished by other groups. Here, we TB patients completing treatment successfully and the level of restrict our focus to the following five priorities: MDR-TB among previously untreated patients. The resulting matrix provides a reasonable framework for deciding whether · accounting for attributable morbidity and mortality and the to use second-line drugs in a national program. Countries with economic burden of drug resistance in the developing world treatment success of less than 70 percent should introduce or · measuring the cost-effectiveness of interventions to im- improve the DOTS strategy as the top priority. In settings with prove prescribing and patient compliance primary MDR-TB of less than 1.5 percent, treating MDR-TB is · researching incentives for firms to invest both in developing not a public health priority, although individual cases could be new drugs and in maintaining the effectiveness of existing referred to clinical experts. The hotspots--those areas with pri- drugs mary MDR levels above 5 percent--are international public · identifying socioeconomic, demographic, and cultural fac- health emergencies. Infection control practices must be empha- tors that determine antibiotic use and misuse and projecting sized in such settings. Intermediate situations are ideal for addi- how antibiotic use will change in future years tional research comparing DOTS with various individualized · designing international coordinating mechanisms for sur- regimens against MDR-TB. veillance to report resistance outbreaks and coordinate strategies for appropriate drug use, recognizing the global nature of drug resistance. Malaria. With declining effectiveness of chloroquine and rapidly emerging resistance to its replacement, sulfadoxine- pyrimethamine, it is imperative that significant attention be CONCLUSION paid not just to the choice of an appropriate first-line treatment Modern medicine rests on the bedrock of effective anti- for malaria, but also to strategies to prolong the effective ther- infective drugs. Unfortunately, the use of drugs creates selective apeutic life of the new treatment. WHO has recommended that pressure for resistance to arise, and thus, the growth of resist- new artemisinin derivatives be used only in combinations with ance may be an unavoidable consequence of our actions in other drugs. There is evidence that artemisinin has already treating disease. It is, however, important for governments to found its way into shops as monotherapy; discouraging the use intervene to ensure that the effectiveness of our current arsenal of this valuable drug in monotherapy through public subsidies of anti-infectives is not depleted at an excessively rapid rate. for combinations, by mandate, or through a combination of Given the potential for international spillovers of resistant measures is a necessary first step. Discouraging the use of pathogens and the ability of actions taken in one region or artemisinin as monotherapy is not sufficient, however. It is also nation to affect other parts of the world where a disease is important to ensure that the partner drug in the combination prevalent, a strong case can be made for coordinated interna- used with the artemisinin derivative is effective and, hence, able tional action--similar to another urgent global situation, the to protect the artemisinin. Therefore, discouraging the use of depletion of the ozone layer and the subsequent Montreal the partner drug as monotherapy, except in cases in which no Protocol to phase out the use of chlorofluorocarbons--to safe alternative exists, such as for sulfadoxine-pyrimethamine manage the evolution of resistance. in intermittent preventive treatment of malaria in pregnant Some interventions that we recommend, such as more women, is an important step in ensuring the long-term sus- restrictive prescribing policies and the use of combinations, tainability of malaria treatment. Training shopkeepers and could, in the absence of subsidies from the state, place a burden other purveyors of antimalarial treatments to recognize symp- on the poorest patients. For instance, an overly restrictive policy toms of malaria and to use diagnostics to detect malaria would on drug sales at the retail level could harm those who have less help reduce malaria treatment to instances for which it is access to formal medical care and prescriptions. There may be appropriate and would reduce the likelihood of the emergence similar effects from mandating that antimalarial drugs be sold in of resistance. Finally, steps to reduce the burden of malaria combinations that the poor cannot afford. Efforts to manage through the use of insecticide-treated bednets and, in some resistance should not be balanced on the backs of the poor, how- areas, residual household spraying would help reduce the need ever, because the rationale for these efforts is that society as a for antimalarial treatment and thereby reduce treatment selec- whole gains from them. It is important that state subsidies be tion pressure. used to ensure that interventions to manage for resistance do not reduce patients' access to effective and affordable drugs. Huge gains in life expectancy have come from the introduc- Research Agenda tion of effective drugs to treat infectious diseases. Our history A description of a research agenda to explain the full range of of treating infections successfully is brief, however, and dates issues related to drug resistance is outside the scope of this back only 50 or 60 years. Sustaining this ability in the long term 1046 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, and others requires a willingness to invest in interventions both to extend Diarrhoea on Morbidity and Mortality in Bangladeshi Children: the therapeutic life of existing drugs and to discover and Community Randomised Trial." British Medical Journal 325 (7372): 1059. develop new ones. Some of these interventions, such as better Barker, R. D., F. J. Millard, and M. E. Nthangeni. 2002. "Unpaid infection control, introduction of affordable vaccines, and Community Volunteers--Effective Providers of Directly Observed proper dosing, would benefit patients immediately; others, Therapy (DOT) in Rural South Africa." South African Medical Journal such as using combination treatments for malaria and invest- 92 (4): 291­94. ing in new drugs, may not bear fruit in the near term. Without Bavestrello, L., A. Cabello, and D. Casanova. 2002. "Impact of Regulatory Measures in the Trends of Community Consumption of Antibiotics in a sustainable, long-term vision of coexistence with harmful Chile" (in Spanish). Revista Medica de Chile 130 (11): 1265­72. microbes and imaginative solutions to the problem of resist- Bergstrom, C. T., M. Lipsitch, and J. E. McGowan Jr. 2000. "Nomenclature ance, our ability to control infectious diseases stands in peril. and Methods for Studies of Antimicrobial Switching (Cycling)." Paper prepared for the Conference on Antibiotic Resistance: Global Policies and Options, Harvard University, Cambridge, MA. NOTE Berkley, J. A., I. Mwangi, C. J. Ngetsa, S. Mwarumba, B. S. Lowe, K. Marsh, and C. R. Newton. 2001. "Diagnosis of Acute Bacterial Meningitis in 1. The fitness cost of resistance is an evolutionary disadvantage placed Children at a District Hospital in Sub-Saharan Africa." Lancet 357 on resistant pathogens. However, some argue that although most (9270): 1753­57. resistance-determining mutations engender some fitness cost, these costs Bexell, A., E. Lwando, B. von Hofsten, S. Tembo, B. Eriksson, and V. K. are likely to be ameliorated by subsequent compensatory mutations. Diwan. 1996. "Improving Drug Use through Continuing Education: A Randomized Controlled Trial in Zambia." Journal of Clinical Epidemiology 49 (3): 355­57. REFERENCES Bhutta, T. I., and C. Balchin. 1996. "Assessing the Impact of a Regulatory Intervention in Pakistan." Social Science and Medicine 42 (8): Aarestrup, F. M., A. M. Seyfarth, H. D. Emborg, K. Pedersen, R. S. 1195­202. Hendriksen, and F. Bager. 2001. "Effect of Abolishment of the Use Bonhoeffer, S., M. Lipsitch, and B. R. Levin. 1997. "Evaluating Treatment of Antimicrobial Agents for Growth Promotion on Occurrence of Protocols to Prevent Antibiotic Resistance." Proceedings of the National Antimicrobial Resistance in Fecal Enterococci from Food Animals in Academy of Sciences, U.S.A. 94: 12106­11. Denmark." Antimicrobial Agents and Chemotherapy 45 (7): 2054­59. Bosu, W. K., and D. Ofori-Adjei. 2000. "An Audit of Prescribing Practices Abramson, M. A., and D. J. Sexton. 1999. "Nosocomial Methicillin- in Health Care Facilities of the Wassa West District of Ghana." West Resistant and Methicillin-Susceptible Staphylococcus Aureus Primary African Journal of Medicine 19 (4): 298­303. Bacteremia: At What Costs?" Infection Control and Hospital Epidemiology 20 (6): 408­11. Boulanger, L. L., L. A. Lee, and A. Odhacha. 1999. "Treatment in Kenyan Rural Health Facilities: Projected Drug Costs Using the WHO- Adrian, P. V., and K. P. Klugman. 1997. "Mutations in the Dihydrofolate UNICEF Integrated Management of Childhood Illness (IMCI) Reductase Gene of Trimethoprim-Resistant Isolates of Streptococcus Guidelines." Bulletin of the World Health Organization 77 (10): 852­58. Pneumoniae." Antimicrobial Agents and Chemotherapy 41 (11): 2406­13. Bronzwaer, S., O. Cars, U. Buchholz, S. Molstad, W. Goettsch, I. K. Veldhuijzen, and others. 2002. "A European Study on the Relationship Agyepong, I. A., E. Ansah, M. Gyapong, S. Adjei, G. Barnish, and D. between Antimicrobial Use and Antimicrobial Resistance." Emerging Evans. 2002. "Strategies to Improve Adherence to Recommended Infectious Diseases 8 (3): 278­82. Chloroquine Treatment Regimes: A Quasi-Experiment in the Context of Integrated Primary Health Care Delivery in Ghana." Social Science Carey, B., and B. Cryan. 2003. "Antibiotic Misuse in the Community--A and Medicine 55 (12): 2215­26. Contributor to Resistance?" Irish Medical Journal 96 (2): 43­44, 46. Angunawela, I. I., V. K. Diwan, and G. Tomson. 1991. "Experimental Carmeli, Y., G. Eliopoulos, E. Mozaffari, and M. Samore. 2002."Health and Evaluation of the Effects of Drug Information on Antibiotic Economic Outcomes of Vancomycin-Resistant Enterococci." Archives Prescribing: A Study in Outpatient Care in an Area of Sri Lanka." of Internal Medicine 162 (19): 2223­28. International Journal of Epidemiology 20 (2): 558­64. Cell. 1997. Global spread of chloroquine-resistant strains of P. falciparum. Arason, V. A., A. Gunnlaugsson, J. A. Sigurdsson, H. Erlendsdottir, S. [cover graph.] 91 (5). Gudmundsson, and K. G. Kristinsson. 2002. "Clonal Spread of Chakraborty, S., S. A. D'Souza, and R. S. Northrup. 2000. "Improving Resistant Pneumococci Despite Diminished Antimicrobial Use." Private Practitioner Care of Sick Children: Testing New Approaches in Microbial Drug Resistance 8 (3): 187­92. Rural Bihar." Health Policy and Planning 15 (4): 400­7. Armstrong Schellenberg, J., J. Bryce, D. de Savigny, T. Lambrechts, C. Chalker, J. 2001. "Improving Antibiotic Prescribing in Hai Phong Mbuya, L. Mgalula, and others. 2004. "The Effect of Integrated Province, Viet Nam: The `Antibiotic-Dose' Indicator." Bulletin of the Management of Childhood Illness on Observed Quality of Care of World Health Organization 79 (4): 313­20. Under-Fives in Rural Tanzania."Health Policy and Planning 19 (1): 1­10. Chan, F. K., J. J. Sung, P. Y. Tan, K. H. Khong, and J. W. Lau. 1997. "`Blister Arrow, K., C. Panosian, and H. Gelband, eds. 2004. "Saving Lives, Buying Pack'­Induced Gastrointestinal Hemorrhage." American Journal of Time: Economics of Malaria Drugs in an Age of Resistance." Gastroenterology 92 (1): 172­73. Washington, DC: Institute of Medicine. Chaulet,P.,M.Raviglione,and F.Bustreo.1996."Epidemiology,Control and Balasubramanian, V. N., K. Oommen, and R. Samuel. 2000. "DOT or Not? Treatment of Multidrug-Resistant Tuberculosis." Drugs 52 (Suppl. 2): Direct Observation of Anti-Tuberculosis Treatment and Patient 103­8. Outcomes, Kerala State, India." International Journal of Tuberculosis Chuc, N. T., M. Larsson, N. T. Do, V. K. Diwan, G. B. Tomson, and T. and Lung Disease 4 (5): 409­13. Falkenberg. 2002. "Improving Private Pharmacy Practice: A Multi- Baqui, A. H., R. E. Black, S. El Arifeen, M. Yunus, J. Chakraborty, S. Ahmed, Intervention Experiment in Hanoi, Vietnam." Journal of Clinical and P. Vaughan. 2002. "Effect of Zinc Supplementation Started during Epidemiology 55 (11): 1148­55. Drug Resistance | 1047 CIA (U.S. Central Intelligence Agency). 1999. The Global Infectious Disease Feikin, D. R., S. F. Dowell, O. C. Nwanyanwu, K. P. Klugman, P. N. Threat and Its Implications for the United States. Washington, DC: Kazembe, L. M. Barat, and others. 2000. "Increased Carriage of Central Intelligence Agency. Trimethoprim/Sulfamethoxazole-Resistant Streptococcus Pneumoniae Coast, J., R. Smith, A. M. Karcher, P. Wilton, and M. Millar. 2002. in Malawian Children after Treatment for Malaria with Sulfadoxine/ "Superbugs II: How Should Economic Evaluation Be Conducted for Pyrimethamine." Journal of Infectious Diseases 181 (4): 1501­5. Interventions Which Aim to Contain Antimicrobial Resistance?" Fidler, D. P. 1998. "Legal Issues Associated with Antimicrobial Drug Health Economics 11 (7): 637­47. Resistance." Emerging Infectious Diseases 4: 169­77. Cohn, M. L., C. Kovitz, U. Oda, and G. Middlebrook. 1954. "Studies on Fischl, M. A., R. B. Uttamchandani, G. L. Daikos, R. B. Poblete, J. N. Isoniazid and Tubercle Bacilli: II. The Growth Requirements, Catalase Moreno, R. R. Reyes, and others. 1992. "An Outbreak of Tuberculosis Activities, and Pathogenic Properties of Isoniazid-Resistant Mutants." Caused by Multiple-Drug-Resistant Tubercle Bacilli among Patients American Review of Tuberculosis 70 (4): 641­64. with HIV Infection." Annals of Internal Medicine 117: 177­83. Cohn, M. L., G. Middlebrook, and W. F. Russell Jr. 1959. "Combined Drug Fujiwara, P. I., S. V. Cook, C. M. Rutherford, J. T. Crawford, S. E. Glickman, Treatment of Tuberculosis: Prevention of Emergence of Mutant B. N. Kreiswirth, and others. 1997. "A Continuing Survey of Drug- Populations of Tubercle Bacilli Resistant to Both Streptomycin and Resistant Tuberculosis, New York City, April 1994." Archives of Internal Isoniazid in Vitro." Journal of Clinical Investigation 38 (8): 1349­55. Medicine 10 (157): 531­36. Cunin, P., E. Tedjouka, Y. Germani, C. Ncharre, R. Bercion, J. Morvan, and Garcia-Garcia, M. L., A. Ponce-de-Leon, M. E. Jimenez-Corona, A. P. M. V. Martin. 1999. "An Epidemic of Bloody Diarrhea: Escherichia Jimenez-Corona, M. Palacios-Martinez, S. Balandrano-Campos, and Coli O157 Emerging in Cameroon?" Emerging Infectious Diseases 5 (2): others. 2000. "Clinical Consequences and Transmissibility of Drug- 285­90. Resistant Tuberculosis in Southern Mexico." Archives of Internal Curtis, V., and S. Cairncross. 2003. "Effect of Washing Hands with Soap Medicine 160: 630­36. on Diarrhoea Risk in the Community: A Systematic Review." Lancet Gerding, D. N. 2000. "Antimicrobial Cycling: Lessons Learned from the Infectious Diseases 3 (5): 275­81. Aminoglycoside Experience." Infection Control and Hospital Denis, M. B. 1998. "Improving Compliance with Quinine Tetracycline Epidemiology 21 (Suppl.): S10­12. for Treatment of Malaria: Evaluation of Health Education Interven- Goble, M., M. D. Iseman, L. A. Madsen, D. Waite, L. Ackerson, and C. R. tions in Cambodian Villages." Bulletin of the World Health Organization Horsburgh Jr. 1993. "Treatment of 171 Patients with Pulmonary 76 (Suppl 1): 43­49. Tuberculosis Resistant to Isoniazid and Rifampin." New England Ding, J., Y. Ma, Z. Gong, and Y. Chen. 1999. "A Study on the Mechanism Journal of Medicine 328 (8): 527­32. of the Resistance of Shigellae to Fluoroquinolones" (in Chinese). Goff, B. J., J. M. Koff, and J. A. Geiling. 2002. "Obtaining Antibiotics with- Zhonghua Nei Ke Za Zhi 38: 550­53. out a Prescription." New England Journal of Medicine 347 (3): 223. Drobniewski, F. A., and Y. M. Balabanova. 2002. "The Diagnosis and Gonzalez Ochoa, E., L. Armas Perez, J. R. Bravo Gonzalez, J. Cabrales Management of Multiple-Drug-Resistant-Tuberculosis at the Begin- Escobar, R. Rosales Corrales, and G. Abreu Suarez. 1996. "Prescription ning of the New Millenium." International Journal of Infectious Diseases of Antibiotics for Mild Acute Respiratory Infections in Children." 6 (Suppl 1): S21­31. Bulletin of the Pan American Health Organization 30 (2): 106­17. Duse, A. G., and R. A. Smego. 1999. "Challenges Posed by Antimicrobial Resistance in Developing Countries." Baillière's Clinical Infectious Gove, S. 1997. "Integrated Management of Childhood Illness by Diseases--Antibiotic Resistance, 5 (2): 193­201. Outpatient Health Workers: Technical Basis and Overview. The WHO Working Group on Guidelines for Integrated Management of the Sick Dye, C., and M. A. Espinal. 2001. "Will Tuberculosis Become Resistant to Child." Bulletin of the World Health Organization 75 (Suppl. 1): 7­24. All Antibiotics?" Proceedings of the Royal Society of London, Series B, Biological Science 268 (1462): 45­52. Greenberg, A. E., M. Ntumbanzondo, N. Ntula, L. Mawa, J. Howell, and F. Davachi. 1989. "Hospital-Based Surveillance of Malaria-Related Dye, C., C. J. Watt, D. M. Bleed, and B. G. Williams. 2003. "What Is the Paediatric Morbidity and Mortality in Kinshasa, Zaire." Bulletin of the Limit to Case Detection under the DOTS Strategy for Tuberculosis World Health Organization 67 (2): 189­96. Control?" Tuberculosis (Edinburgh) 83 (1­3): 35­43. Gupta, R., J. Y. Kim, M. A. Espinal, J. M. Cauldron, B. Pecoul, P. E. Farmer, Dye, C., B. G. Williams, M. A. Espinal, and M. C. Raviglione. 2002."Erasing and M. C. Raviglione. 2001. "Responding to Market Failures in the World's Slow Stain: Strategies to Beat Multidrug-Resistant Tuberculosis Control." Science 293 (5532): 1049­51. Tuberculosis." Science 295 (5562): 2042­46. Hadiyono, J. E., S. Suryawati, S. S. Danu, Sunartono, and B. Santoso. 1996. Elbasha, E. 1999. "Deadweight Loss of Bacterial Resistance Due to "Interactional Group Discussion: Results of a Controlled Trial Using Overtreatment." U.S. Centers for Disease Control and Prevention, a Behavioral Intervention to Reduce the Use of Injections in Public Atlanta. Health Facilities." Social Science and Medicine 42 (8): 1177­83. Emborg, H. D., J. S. Andersen, A. M. Seyfarth, S. R. Andersen, J. Boel, and H. C. Wegener. 2003. "Relations between the Occurrence of Resistance Hayes, D. J., and H. H. Jensen. 2003. "Lessons from the Danish Ban on to Antimicrobial Growth Promoters among Enterococcus Faecium Feed-Grade Antibiotics." Choices (3rd quarter): 1­6. Isolated from Broilers and Broiler Meat." International Journal of Food Helitzer-Allen, D. L., D. A. McFarland, J. J. Wirima, and A. P. Macheso. Microbiology 84 (3): 273­84. 1993. "Malaria Chemoprophylaxis Compliance in Pregnant Women: A Espinal, M. A., C. Dye, M. Raviglione, and A. Kochi. 1999."Rational `DOTS Cost-Effectiveness Analysis of Alternative Interventions." Social Science Plus' for the Control of MDR-TB." International Journal of Tuberculosis & Medicine 36 (4): 403­7. and Lung Disease 3 (7): 561­63. Ho, P. L., W. C. Yam, T. K. Cheung, W. W. Ng, T. L. Que, D. N. Tsang, and Espinal, M. A., S. J. Kim, P. G. Suarez, K. M. Kam, A. G. Khomenko, G. B. others. 2001. "Fluoroquinolone Resistance among Streptococcus Miglori, and others. 2000. "Standard Short-Course Chemotherapy for Pneumoniae in Hong Kong Linked to the Spanish 23F Clone." Drug-Resistant Tuberculosis: Treatment Outcome in Six Countries." Emerging Infectious Diseases 7 (5): 906­8. Journal of the American Medical Association 283 (19): 2537­45. Hoge, C. W., L. Bodhidatta, C. Tungtaem, and P. Echeverria. 1995. Farmer, P., M. Becerra, and J. Kim, eds. 1999. The Global Impact of Drug- "Emergence of Nalidixic Acid Resistant Shigella Dysenteriae Type 1 in Resistant Tuberculosis. Boston: Harvard Medical School and Open Thailand: An Outbreak Associated with Consumption of a Coconut Society Institute. Milk Dessert." International Journal of Epidemiology 24 (6): 1228­32. 1048 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, and others Holmberg, S. D., S. L. Solomon, and P. A. Blake. 1987. "Health and Kublin, J. G., J. F. Cortese, E. M. Njunju, R. A. Mukadam, J. J. Wirima, P. N. Economic Impacts of Antimicrobial Resistance." Reviews of Infectious Kazembe, and others. 2003. "Reemergence of Chloroquine-Sensitive Diseases 9 (6): 1065­78. Plasmodium falciparum Malaria after Cessation of Chloroquine Use in Holmstrom, K., S. Gräslund, A. Wahlström, S. Poungshompoo, B.-E. Malawi." Journal of Infectious Diseases 187 (12): 1870­5. Bengtsson, and N. Kautsky. 2003. "Antibiotic Use in Shrimp Farming Kurlat, I., G. Corral, F. Oliveira, G. Farinella, and E. Alvarez. 1998. and Implications for Environmental Impacts and Human Health." "Infection Control Strategies in a Neonatal Intensive Care Unit in International Journal of Food Science and Technology 38 (3): 255­66. Argentina." Journal of Hospital Infection 40 (2): 149­54. Hu, S., W. Chen, X. Cheng, K. Chen, H. Zhou, and L. Wang. 2001. Laing, R., H. Hogerzeil, and D. Ross-Degnan. 2001. "Ten Recom- "Pharmaceutical Cost-Containment Policy: Experiences in Shanghai, mendations to Improve Use of Medicines in Developing Countries." China." Health Policy and Planning 16 (Suppl. 2): 4­9. Health Policy and Planning 16 (1): 13­20. Igun, U. A. 1994. "Reported and Actual Prescription of Oral Rehydration Laxminarayan, R. 2002. "How Broad Should the Scope of Antibiotics Therapy for Childhood Diarrhoeas by Retail Pharmacists in Nigeria." Patents Be?"American Journal of Agricultural Economics 84 (5): 1287­92. Social Science and Medicine 39 (6): 797­806. Laxminarayan, R., and M. L. Weitzman. 2002. "On the Implications of INCLEN (International Clinical Epidemiology Network). 1999. Endogenous Resistance to Medications." Journal of Health Economics "Prospective Multicentre Hospital Surveillance of Streptococcus 21 (4): 709­18. Pneumoniae Disease in India. Invasive Bacterial Infection Surveillance le Grand, A., H. V. Hogerzeil, and F. M. Haaijer-Ruskamp. 1999. (IBIS) Group, International Clinical Epidemiology Network "Intervention Research in Rational Use of Drugs: A Review." Health (INCLEN)." Lancet 353 (9160): 1216­21. Policy and Planning 14 (2): 89­102. Indalo, A. A. 1997. "Antibiotic Sale Behaviour in Nairobi: A Contributing Legros, D., D. Ochola, N. Lwanga, and G. Guma. 1998. "Antibiotic Factor to Antimicrobial Drug Resistance." East African Medical Journal Sensitivity of Endemic Shigella in Mbarara, Uganda." East African 74 (3): 171­3. Medical Journal 75 (3): 160­1. Iseman, M. D., D. L. Cohn, and J. A. Sbarbaro. 1993. "Directly Observed Levy, S. B. 1992. The Antibiotic Paradox: How Miracle Drugs Are Destroying Treatment of Tuberculosis: We Can't Afford Not to Try It." New the Miracle. New York: Plenum Press. England Journal of Medicine 328 (8): 576­78. Lwilla, F., D. Schellenberg, H. Masanja, C. Acosta, C. Galindo, J. Aponte, Ison, C. A., P. J. Woodford, H. Madders, and E. Claydon. 1998. "Drift in and others. 2003. "Evaluation of Efficacy of Community-Based vs. Susceptibility of Neisseria Gonorrhoeae to Ciprofloxacin and Institutional-Based Direct Observed Short-Course Treatment for the Emergence of Therapeutic Failure." Antimicrobial Agents and Control of Tuberculosis in Kilombero District, Tanzania." Tropical Chemotherapy 42 (11): 2919­22. Medicine and International Health 8 (3): 204­10. IUATLD (International Union against Tuberculosis and Lung Disease). Madhi, S. A., K. Petersen, A. Madhi, M. Khoosal, and K. P. Klugman. 2000. 1998. "Guidelines for Surveillance of Drug Resistance in Tuberculosis." "Increased Disease Burden and Antibiotic Resistance of Bacteria International Journal of Tuberculosis and Lung Diseases 2: 72­89. Causing Severe Community-Acquired Lower Respiratory Tract Infections in Human Immunodeficiency Virus Type 1-Infected Jacobs, M. R., D. Felmingham, P. C. Appelbaum, R. N. Grünebera, Children." Clinical Infectious Diseases 31 (1): 170­6. and the Alexander Project Group. 2003. "The Alexander Project 1998­2000: Susceptibility of Pathogens Isolated from Marin, G., L. Burhansstipanov, C. M. Connell, A. C. Gielen, D. Helitzer- Community-Acquired Respiratory Tract Infection to Commonly Allen, K. Lorig, and others. 1995. "A Research Agenda for Health Used Antimicrobial Agents." Journal of Antimicrobial Chemotherapy Education Among Underserved Populations." Health Education 52: 229­46. Quarterly 22 (3): 346­63. Jones, N., R. Huebner, M. Khoosal, H. Crewe-Brown, and K. Klugman. Maxwell, C. A., E. Msuya, M. Sudi, K. J. Njunwa, I. A. Carneiro, and C. F. 1998. "The Impact of HIV on Streptococcus Pneumoniae Bacteraemia Curtis. 2002. "Effect of Community-Wide Use of Insecticide-Treated in a South African Population." AIDS 12 (16): 2177­84. Nets for 3­4 Years on Malarial Morbidity in Tanzania." Tropical Medicine and International Health 7 (12): 1003­8. Kam, K. M., and C. W. Yip. 2001. "Surveillance of Mycobacterium Tuberculosis Drug Resistance in Hong Kong, 1986­1999, after the McCaig, L. F., R. E. Besser, and J. M. Hughes. 2002. "Trends in Implementation of Directly Observed Treatment." International Antimicrobial Prescribing Rates for Children and Adolescents." Journal Journal of Tuberculosis and Lung Diseases 5 (9): 815­23. of the American Medical Association 287 (23): 3096­102. McGee, L., L. McDougal, J. Zhou, B. G. Spratt, F. C. Tenover, R. George, and Kettler, H. 2000. Narrowing the Gap between Provision and Need for others. 2001. "Nomenclature of Major Antimicrobial-Resistant Clones Medicines in Developing Countries. London: Office of Health of Streptococcus Pneumoniae Defined by the Pneumococcal Economics. Molecular Epidemiology Network." Journal of Clinical Microbiology Khan, M. A., J. D. Walley, S. N. Witter, A. Imran, and N. Safdar. 2002."Costs 39 (7): 2565­71. and Cost-Effectiveness of Different DOT Strategies for the Treatment McGowan, J. E. 1986. "Minimizing Antimicrobial Resistance in Hospital of Tuberculosis in Pakistan: Directly Observed Treatment." Health Bacteria: Can Switching or Cycling Drugs Help?" Infection Control 7: Policy and Planning 17 (2): 178­86. 573­76. Khatri, G. R., and T. R. Frieden. 2002. "Rapid DOTS Expansion in India." Mitema, E. S., G. M. Kikuvi, H. C. Wegener, and K. Stohr. 2001. "An Bulletin of the World Health Organization 80 (6): 457­63. Assessment of Antimicrobial Consumption in Food Producing Klugman, K. P. 2001. "Efficacy of Pneumococcal Conjugate Vaccines and Animals in Kenya." Journal of Veterinary Pharmacology and Their Effect on Carriage and Antimicrobial Resistance." Lancet Therapeutics 24 (6): 385­90. Infectious Diseases 1 (2): 85­91. Monnet, D. L., and T. L. Sorensen. 2001. "The Patient, Their Doctor, the Klugman, K. P., H. J. Koornhof, V. Kuhnle, S. D. Miller, P. J. Ginsburg, and Regulator, and the Profit Maker: Conflicts and Possible Solutions." A. C. Mauff. 1986. "Meningitis and Pneumonia Due to Novel Multiply Clinical Microbiology and Infection 7 (Suppl. 6): 27­30. Resistant Pneumococci." British Medical Journal (Clinical Research Ed.) Mthwalo, M., A. Wasas, R. Huebner, H. J. Koornhof, and K. P. Klugman. 292 (6522): 730. 1998. "Antibiotic Resistance of Nasopharyngeal Isolates of Kochi, A., B. Vareldzis, and K. Styblo. 1993. "Multidrug-Resistant Streptococcus Pneumoniae from Children in Lesotho." Bulletin of the Tuberculosis and Its Control."Research in Microbiology 144 (2): 104­10. World Health Organization 76 (6): 641­50. Drug Resistance | 1049 Mukherjee, J. S., M. L. Rich, A. R. Socci, J. K. Joseph, F. A. Viru, S. S. Shin, Quale, J., D. Landman, J. Ravishankar, C. Flores, and S. Bratu. 2002. and others. 2004. "Programmes and Principles in Treatment of "Streptococcus Pneumoniae, Brooklyn, New York: Fluoroquinolone Multidrug-Resistant Tuberculosis." Lancet 363 (9407): 474­81. Resistance at Our Doorstep." Emerging Infectious Diseases 8 (6): Nosten, F., M. van Vugt, R. Price, C. Luxemburger, K. L. Thway, A. 594­97. Brockman, and others. 2000. "Effects of Artesunate­Mefloquine Rhodes, G., G. Huys, J. Swings, P. McGann, M. Hiney, P. Smith, and R. W. Combination on Incidence of Plasmodium Falciparum Malaria and Pickup. 2000. "Distribution of Oxytetracycline Resistance Plasmids Mefloquine Resistance in Western Thailand: A Prospective Study." Between Aeromonads in Hospital and Aquaculture Environments: Lancet 356 (9226): 297­302. Implication of Tn1721 in Dissemination of the Tetracycline Resistance Office of Technology Assessment. 1995. Impact of Antibiotic-Resistant Determinant Tet A." Applied and Environmental Microbiology 66 (9): Bacteria: A Report to the U.S. Congress. Washington, DC: U.S. Govern- 3883­90. ment Printing Office. Rimon, M. M., S. Kheng, S. Hoyer, V. Thach, S. Ly, A. E. Permin, and S. Okeke, I., and A. Lamikanra. 2001. "Quality and Bioavailability of Pieche. 2003. "Malaria Dipsticks Beneficial for IMCI in Cambodia." Ampicillin Capsules in a Nigerian Semi-Urban Community." African Tropical Medicine and International Health 8 (6): 536­43. Journal of Medicine and Medical Sciences 30: 47­51. Rowe, A. K., M. S. Deming, B. Schwartz, A. Wasas, D. Rolka, H. Rolka, and Okeke, I. N., A. Lamikanra, and R. Edelman. 1999. "Socioeconomic and others. 2000. "Antimicrobial Resistance of Nasopharyngeal Isolates Behavioral Factors Leading to Acquired Bacterial Resistance to of Streptococcus Pneumoniae and Haemophilus Influenzae from Antibiotics in Developing Countries." Emerging Infectious Diseases Children in the Central African Republic." Pediatric Infectious Disease 5 (1): 18­27. Journal 19 (5): 438­44. Oluwole, D., E. Mason, and A. Costello. 2000. "Management of Childhood Rubin, R. J., C. A. Harrington, A. Poon, K. Dietrich, J. A. Greene, and A. Illness in Africa: Early Evaluations Show Promising Results." British Moiduddin. 1999. "The Economic Impact of Staphylococcus Aureus in Medical Journal 320 (7235): 594­95. New York City Hospitals." Emerging Infectious Diseases 5 (1). Pablos-Méndez, A., D. K. Gowda, and T. R. Frieden. 2002. "Controlling Saez-Llorens, X., M. M. Castrejon de Wong, E. Castano, O. de Suman, D. Multidrug-Resistant Tuberculosis and Access to Expensive Drugs: A de Moros, and I. de Atencio. 2000. "Impact of an Antibiotic Restriction Rational Framework." Bulletin of the World Health Organization 80 (6): Policy on Hospital Expenditures and Bacterial Susceptibilities: A 489­95. Lesson from a Pediatric Institution in a Developing Country." Pediatric Palmer, H. M., J. P. Leeming, and A. Turner. 2001. "Investigation of an Infectious Disease Journal 19 (3): 200­6. Outbreak of Ciprofloxacin-Resistant Neisseria Gonorrhoeae Using a Sa-Leao, R., S. E. Vilhelmsson, H. de Lencastre, K. G. Kristinsson, and A. Simplified Opa-Typing Method." Epidemiology and Infection 126 (2): Tomasz. 2002. "Diversity of Penicillin-Nonsusceptible Streptococcus 219­24. Pneumoniae Circulating in Iceland after the Introduction of Penicillin- Paredes, P., M. de la Pena, E. Flores-Guerra, J. Diaz, and J. Trostle. 1996. Resistant Clone Spain(6B)-2." Journal of Infectious Diseases 186 (7): "Factors Influencing Physicians' Prescribing Behaviour in the 966­75. Treatment of Childhood Diarrhoea: Knowledge May Not Be the Clue." Santoso, B., S. Suryawati, and J. E. Prawaitasari. 1996. "Small Group Social Science and Medicine 42 (8): 1141­53. Intervention vs. Formal Seminar for Improving Appropriate Drug Pechere, J. C. 2001. "Patients' Interviews and Misuse of Antibiotics." Use." Social Science and Medicine 42 (8): 1163­68. Clinical Infectious Diseases 33 (Suppl. 3): S170­73. Sarkar, R., A. N. Chowdhuri, J. K. Dutta, H. Sehgal, and M. Mohan. 1979. Philipson, T., and S. Mechoulan. 2003. "Intellectual Property and External "Antibiotic Resistance Pattern of Enteropathogenic E. coli Isolated Consumption Effects: Generalizations from Pharmaceutical Markets." from Diarrhoeal Disease in Children in Delhi." Indian Journal of NBER Working Paper 9598, National Bureau of Economic Research, Medical Research 70: 908­15. Cambridge, MA. Schrag, S. J., C. Pena, J. Fernandez, J. Sanchez, V. Gomez, E. Perez, and Pichichero, M. E., and R. Cohen. 1997. "Shortened Course of Antibiotic others. 2001. "Effect of Short-Course, High-Dose Amoxicillin Therapy Therapy for Acute Otitis Media, Sinusitis and Tonsillopharyngitis." on Resistant Pneumococcal Carriage: A Randomized Trial." Journal Pediatric Infectious Disease Journal 16: 680­95. of the American Medical Association 286 (1): 49­56. Ponce-de-Leon, R. S., and M. S. Rangel-Frausto. 1993. Organising for Seppala, H., T. Klaukka, J. Vuopio-Varkila, A. Muotiala, H. Helenius, K. Infection Control with Limited Resources. In Prevention and Control of Lager, and P. Huovinen. 1997. "The Effect of Changes in the Nosocomial Infections, ed. R. P. Wenzel. Baltimore: Williams & Wilkins. Consumption of Macrolide Antibiotics on Erythromycin Resistance in Group A Streptococci in Finland. Finnish Study Group for Anti- Prazuck, T., I. Falconi, G. Morineau, V. Bricard-Pacaud, A. Lecomte, and microbial Resistance." New England Journal of Medicine 337 (7): F. Ballereau. 2002. "Quality Control of Antibiotics before the 441­46. Implementation of an STD Program in Northern Myanmar." Sexually Transmitted Diseases 29 (11): 624­27. Siddiqi, S., S. Hamid, G. Rafique, S. A. Chaudhry, N. Ali, S. Shahab, and R. Pungrassami, P., and V. Chongsuvivatwong. 2002. "Are Health Personnel Sauerborn. 2002. "Prescription Practices of Public and Private Health the Best Choice for Directly Observed Treatment in Southern Care Providers in Attock District of Pakistan." International Journal of Thailand? A Comparison of Treatment Outcomes among Different Health Planning and Management 17 (1): 23­40. Types of Observers." Transactions of the Royal Society of Tropical Smith, D. L., A. D. Harris, J. A. Johnson, E. K. Silbergeld, and J. G. Morris Medicine and Hygiene 96 (6): 695­99. Jr. 2002. "Animal Antibiotic Use Has an Early but Important Impact Pungrassami, P., S. P. Johnsen, V. Chongsuvivatwong, and J. Olsen. 2002. on the Emergence of Antibiotic Resistance in Human Commensal "Has Directly Observed Treatment Improved Outcomes for Patients Bacteria." Proceedings of the National Academy of Sciences, U.S.A. 99 with Tuberculosis in Southern Thailand?" Tropical Medicine and (9): 6434­39. International Health 7 (3): 271­79. Song, J. H., N. Y. Lee, S. Ichiyama, R. Yoshida, Y. Hirakata, W. Fu, and oth- Qingjun, L., D. Jihui, T. Laiyi, Z. Xiangjun, L. Jun, A. Hay, and others. 1998. ers. 1999."Spread of Drug-Resistant Streptococcus pneumoniae in Asian "The Effect of Drug Packaging on Patients' Compliance with Countries: Asian Network for Surveillance of Resistant Pathogens Treatment for Plasmodium Vivax Malaria in China." Bulletin of the (ANSORP) Study." Clinical Infectious Diseases 28 (6): 1206­11. World Health Organization 76 (Suppl. 1): 21­27. Su, Xin-Zhuan, L. A. Kirkman, H. Fuzioka, and T. E. Wellems. 1997. "Complex Polymorphisms in an 330 kDa Protein are Linked to 1050 | Disease Control Priorities in Developing Countries | Ramanan Laxminarayan, Zulfiqar A. Bhutta, Adriano Duse, and others Chloroquine-Resistant P. Falciparum in Southeast Asia and Africa." WHO (World Health Organization). 1993. WHO-RBM Africa Malaria Cell 91: 593­603. Report. Suarez, P. G., K. Floyd, J. Portocarrero, E. Alarcon, E. Rapiti, G. Ramos, and ------. 1994. Tuberculosis Program: Framework for Effective Tuberculosis others. 2002. "Feasibility and Cost-Effectiveness of Standardised Control. Geneva: WHO. Second-Line Drug Treatment for Chronic Tuberculosis Patients: A ------. 1996. Assessment of Therapeutic Efficacy of Antimalarial Drugs for National Cohort Study in Peru." Lancet 359 (9322): 1980­89. Uncomplicated Falciparum Malaria in Areas of Intense Transmission. Tapsall, J. 2002. Current Concepts in the Management of Gonorrhoea. Geneva: WHO. Expert Opinion on Pharmacotherapy 3: 147­57. ------. 1997. Guidelines for the Management of Drug-Resistant Taylor, R. B., O. Shakoor, R. H. Behrens, M. Everard, A. S. Low, J. Tuberculosis. Geneva: WHO. Wangboonskul, and others. 2001. "Pharmacopoeial Quality of Drugs ------. 2000. "WHO Global Principles for the Containment of Anti- Supplied by Nigerian Pharmacies." Lancet 357 (9272): 1933­36. microbial Resistance in Animals Intended for Food." Report of a WHO Telzak, E. E., K. Sepkowitz, P. Alpert, S. Mannheimer, F. Medard, W. el-Sadr, Consultation, Geneva, WHO. and others. 1995."Multidrug-Resistant Tuberculosis in Patients without ------. 2001. WHO Global Strategy for Containment of Antimicrobial HIV Infection." New England Journal of Medicine 333 (14): 907­11. Resistance. Geneva: WHO. Trape, J. F. 2001. "The Public Health Impact of Chloroquine Resistance in ------. 2002a. Surveillance Standards for Antimicrobial Resistance. Africa." American Journal of Tropical Medicine and Hygiene 64 (Suppl. Geneva: WHO. 1­2): 12­17. ------. 2002b. The World Health Report 2002: Reducing Risks, Promoting United Kingdom Government. 2001. "International Action against Child Healthy Life. Geneva: WHO. Poverty--Meeting the 2015 Targets." Proceedings of Conference on ------. 2004. The WHO/IUATLD Global Project on Anti-Tuberculosis Drug Elimination of Child Poverty. London, February 26, 2001. Resistance Surveillance 1999­2002: Anti-Tuberculosis Drug Resistance in Wammanda, R. D., C. L. Ejembi, and T. Iorliam. 2003. "Drug Treatment the World. Report 3. Geneva: WHO. Costs: Projected Impact of Using the Integrated Management of Wierup, M. 2001. "The Swedish Experience of the 1986 Year Ban of Childhood Illnesses." Tropical Doctor 33 (2): 86­88. Antimicrobial Growth Promoters, with Special Reference to Animal Wang, H., R. Huebner, M. Chen, and K. Klugman. 1998. "Antibiotic Health, Disease Prevention, Productivity, and Usage of Susceptibility Patterns of Streptococcus pneumoniae in China and Antimicrobials." Microbial Drug Resistance 7 (2): 183­90. Comparison of MICs by Agar Dilution and E-Test Methods." Wilson, J. M. and G. N. Chandler. 1993. "Sustained Improvements in Antimicrobial Agents and Chemotherapy 42 (10): 2633­36. Hygiene Behaviour amongst Village Women in Lombok, Indonesia." Wegener, H. C., F. M. Aarestrup, L. B. Jensen, A. M. Hammerum, and F. Transactions of the Royal Society of Tropical Medicine and Hygiene 87 Bager. 1999."Use of Antimicrobial Growth Promoters in Food Animals (6): 615­16. and Enterococcus faecium Resistance to Therapeutic Antimicrobial Wongsrichanalai, C., J. Sirichaisinthop, J. J. Karwacki, K. Congpuong, R. S. Drugs in Europe." Emerging Infectious Diseases 5 (3): 329­35. Miller, L. Pang, and K. Thimasarn. 2001. "Drug Resistant Malaria on Wellems, T. E., and C. V. Plowe. 2001. Chloroquine-Resistant Malaria. the Thai-Myanmar and Thai-Cambodian Borders." Southeast Asian Journal of Infectious Diseases. 184: 770­76. Journal of Tropical Medicine and Public Health. 32: 41­49. White, D. G., S. Zhao, R. Sudler, S. Ayers, S. Friedman, S. Chen, and others. Xu, Q., S. G. Jin, and L. X. Zhang. 2000. "Cost Effectiveness of DOTS and 2001. "The Isolation of Antibiotic-Resistant Salmonella from Retail Non-DOTS Strategies for Smear-Positive Pulmonary Tuberculosis in Ground Meats." New England Journal of Medicine 345 (16): 1147­54. Beijing." Biomedical and Environmental Sciences 13 (4): 307­13. White, N. J. 1998. "Preventing Antimalarial Drug Resistance Through Yang, F., Y. Zhang, and L. McGee. 2001. "Population Biology of Combinations." Drug Research, Updates 1: 3­9. Streptococcus Pneumoniae Carried By Healthy Children in ------. 1999. "Antimalarial Drug Resistance and Combination Shanghai"(in Chinese). Zhonghua Yi Xue Za Zhi 81 (10): 589­92. Chemotherapy." Philosophical Transactions of the Royal Society of Zaidi, M. B., E. Zamora, P. Diaz, L. Tollefson, P. J. Fedorka-Cray, and London B Series 354: 739­49. M. L. Headrick. 2003. "Risk Factors for Fecal Quinolone-Resistant Whitney, C. G., M. M. Farley, J. Hadler, L. H. Harrison, N. M. Bennett, R. Escherichia Coli in Mexican Children." Antimicrobial Agents and Lynfield, and others. 2003. "Decline in Invasive Pneumococcal Disease Chemotherapy 47 (6): 1999­2001. after the Introduction of Protein-Polysaccharide Conjugate Vaccine." New England Journal of Medicine 348 (18): 1737­46. Drug Resistance | 1051 Chapter 56 Community Health and Nutrition Programs John B. Mason, David Sanders, Philip Musgrove, Soekirman, and Rae Galloway Rapid improvements in health and nutrition in developing their health and nutrition and to facilitate access to treatment countries may be ascribed to specific, deliberate, health- and of sickness. Mothers and children are the primary focus, but nutrition-related interventions and to changes in the underly- others in the household should participate. Commonly, people ing social, economic, and health environments. This chapter go regularly to a central point in their community--for exam- is concerned with the contribution of specific interventions, ple, for growth monitoring and promotion--or are visited at while recognizing that improved living standards in the long home by a health and nutrition worker. The existence, training, run provide the essential basis for improved health. support, and supervision of the community worker--based in Consideration of the environment as the context for interven- the community or operating from a nearby health facility--are tions is crucial in determining their initiation and in modifying indispensable features of these programs. Thus community their effect, and it must be taken into account when assessing organizations are a key aspect of community-based health and this effect. nutrition programs (CHNPs). Undoubtedly much change has stemmed from scientific This chapter focuses on large-scale (national or state) pro- advances, immunization being a prominent case. However, the grams. Although these programs are primarily initiated and organizational aspects of health and nutrition protection are run at the local level, links with the national level and levels in equally critical. In the past several decades, people's contact between are necessary. Both horizontal and vertical organiza- with trained workers has been instrumental in improving tions are needed. Local organizations make action happen, but health in developing countries. This factor applies particularly they need input and resources, such as training, supervision, to poor people in poor countries but is relevant everywhere; and supplies, from more central levels. indeed, it is a reason that social services have essentially elimi- The experience on which this chapter is based comes from a nated almost all occurrences of child malnutrition in Europe considerable number of national and large-scale programs. (where, when malnourished children are seen, it is caused by Most of these programs include both nutrition and health neglect). activities, aimed particularly at the health and survival of Community-based programs under many circumstances reproductive-age women and children. We draw on these expe- provide this crucial contact. Their role is partly in improving riences as we try to put forward principles on which future pro- access to technology and resources, but it is also important in grams can be based--programs that may have broader health fostering behavior change and, more generally, in supporting objectives for other population groups and diseases. caring practices (Engle, Bentley, and Pelto 2000; UNICEF As of 2001, some 19 percent of global deaths were among 1990). Such programs may also play a part in mobilizing social children--and 99 percent of all child deaths took place in demand for services and in generating pressure for policy low- and middle-income countries. The disability-adjusted life change. years (DALYs) lost attributed to zero- to four-year-olds--plus In community-based programs, workers--often volunteers maternal and perinatal conditions, nutrition deficiencies, and and part-time workers--interact with households to protect endocrine disorders--amount to 42 percent of the total disease 1053 Table 56.1 Estimated Contributions to the Disease Burden in context, how much was programmatic, and what were the Developing Countries interactions. The contrasts between these three countries are instructive in part because they have several similar contextual DALYs lost (percentage) factors; for instance, the status of women is relatively good, and Factor Direct effect As risk factor Total social exclusion1 is not extensive (compare both of these in, for General malnutrition 1.0 14.0 15.0 example, South Asia). Thus programs may account for a signif- Micronutrient deficiencies 9.0 8.5 17.5 icant part of the differences seen in improvement. Total 10.0 22.5 32.5 The benefits from CHNPs extend well beyond child nutri- tion (which is used as a summary measure). These benefits Source: Mason, Musgrove, and Habicht 2003, table 10. have not been quantified but would include improved educa- bility (see chapter 49) and probably increased earning capacity burden (all ages, both sexes) from all causes for developing associated with it and with physical fitness. regions. CHNPs address about 40 percent of the disease bur- den. In terms of prevention, Mason, Musgrove, and Habicht (2003) estimated that eliminating malnutrition would remove WHAT IS KNOWN ABOUT EFFICACY one-third of the global disease burden. Comparative studies by AND EFFECTIVENESS Ezzati, Lopez, and others (2002) and Ezzati, Vander Hoorn, and others (2003) have reemphasized malnutrition as the predom- The efficacy of health and nutrition interventions in develop- inant risk factor and improvement of nutrition as playing ing countries has been established for decades (for example, a potentially major role in reducing the burden. Clinical Gwatkin, Wilcox, and Wray 1980). Prospective studies in sev- deficiencies contribute directly to malnutrition, but even more, eral settings showed that health interventions with or without malnutrition is a risk factor for infectious diseases (table 56.1). supplementary foods caused children to thrive and survive Furthermore, changes in child malnutrition levels in develop- better: studies in Narangwal, India (Kielmann and others 1978; ing countries are closely related to the countries' mortality Taylor, Kielmann, and Parker 1978); by the Institute for trends (Pelletier and Frongillo 2003). Nutrition for Central America and Panama (Delgado and Dealing with women and children's health and nutrition others 1982); in Jamaica (Waterlow 1992); and in The Gambia addresses a substantial part of global health problems. (Whitehead, Rowland, and Cole 1976) are examples.2 These Moreover, the experience of community-based programs linked studies showed the effect of interventions on growth and (usu- to nutrition constitutes a significant part of the body of knowl- ally) mortality but did not generally factor out the relative con- edge on ways of improving it.A number of large-scale, sustained tributions of health and nutrition. In fact, results from health interventions, such as those described by Sanders and Narangwal showed similar mortality effects from food or health Chopra (2004), use a mix of improved access to facilities and care; results from The Gambia indicated interaction such that community health workers. These interventions include the sick children did not grow even with adequate food intake Comprehensive Rural Health Project, Jamkhed, India; com- (appetite also playing an important role), and well children did munity health projects in Brazil (Ceará, Pelotas); and the work not grow with inadequate food intake (Gillespie and Mason of the Bangladesh Rural Advancement Committee (BRAC). 1991, annex 2). Table 56.2 describes the program experiences drawn on. By the early 1980s, the conclusion, based on data at the The evidence is clear that significant differences occur experimental level (not from routine large-scale programs), between countries in the rates of change in health and nutri- was that better health and better nutrition are both required for tional status. Figure 56.1 shows a comparison of Indonesia, the child survival and development. This conclusion remains gen- Philippines, and Thailand. As is common, the indicator used is erally agreed on today; furthermore, concern exists that health underweight children, which is likely to reflect broader condi- interventions may become less effective unless nutrition is con- tions of health and survival. For Thailand, the figure shows the currently addressed (Measham and Chatterjee 1999; Pelletier now-well-known rapid improvement in the 1980s and 1990s. and Frongillo 2003). In their chapter on malnutrition in the For Indonesia, it shows slower but consistent improvement. first edition of this book, Pinstrup-Andersen and colleagues The Philippines had little progress until recently, and the start (1993) drew largely on efficacy findings, with an emphasis on of an improving trend coincided with increases in the number food supplementation. Those studies are not revisited here, but of village health workers and implementation of high-coverage we can continue to build on their conclusions. interventions such as iodized salt and vitamin A supplementa- The efficacy studies were followed by a number of national tion (FNRI 2004). A crucial issue is how much of the improve- or other large-scale programs in several countries. Some of ments was caused by interventions that could be replicated-- those were a direct follow-on; for example, the World Bank and within that issue is subsumed how much was because of Tamil Nadu Integrated Nutrition Program (TINP) followed the 1054 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others Table 56.2 Country Experiences in Community-Based Programs Country and program Program experience Africa Tanzania: Iringa Area program with UNICEF and WHO inputs, 1984­91. After rapid initial drop in child malnutrition, moderate steady improvement. Program not sustained. Tanzania: Child Survival and 1985­95, World Bank support. Results similar to Iringa. Development Program Zimbabwe: Supplementary Wide-scale program following independence, 1980­90; infant mortality rate (IMR) dropped from 110 to 53 (1988). Not sustained. Feeding Programme Asia Bangladesh: Bangladesh BINP: area targeted covering 7 percent of population. Rapid improvement at start (1997); final evaluation not seen. Integrated Nutrition Program National: program coverage expanding from 2000 on. Substantial improvement in anemia and child underweight seen in and national Bangladesh starting 1995. Bangladesh: Bangladesh Rural Community-based health services with village health workers. Wide coverage since 1980s; particular focus on diarrhea. Advancement Committee India: Integrated Child Implemented since 1976. Village program with community health (anganwadi) worker. Accelerated improvement reported in Development Services some states. India: Tamil Nadu Integrated Implemented 1980 to mid 1990s. Village program in Tamil Nadu with World Bank support; growth monitoring, supplementary Nutrition Program feeding, and so on. Substantial improvement in underweight reported. Indonesia Massive expansion of village programs 1975­90, covering all villages by 1990. Steady decrease in underweight during this time. Program not sustained in 1990s; now planned to restart. Philippines: national No wide CHNPs despite national decree in 1974. No significant improvement in child nutrition. Thailand National program from late 1970s; 600,000 village health volunteers trained (1 percent of population). Rapid improvement 1980­90; for example, 36 percent to 13 percent underweight children. Americas Costa Rica Expanding rural health services from 1970s following malaria control. Rapid fall in IMR, 1965­80; in stunting, 1979­89. Jamaica Expanded health services with community health aides from mid 1970s. Rapid fall in underweight, 1985­89. Nicaragua Community health movement, 1979­90, reduced IMR, eliminated polio; about 1 percent of population as village health volunteers. Source: Authors, from data derived as follows: Tanzania--Gillespie and Mason 1991; Gillespie, Mason, and Martorell 1996; Jennings and others 1991, 117; Kavishe and Mushi 1993; Pelletier 1991; Sanders 1999; Zimbabwe--Sanders 1999; Tagwireyi and Greiner 1994; Werner and Sanders 1997; Bangladesh--BINP and UNICEF 1999; BRAC 2004; Chowdhury 2003; INFS and Department of Economics, University of Dhaka 1998; Mason and others 1999, 2001; Save the Children U.K. 2003; India--Administrative Staff College of India 1997; Mason and others 1999, 2001; Measham and Chatterjee 1999; Reddy and others 1992; Shekar 1989; Indonesia--Berg 1987; Jennings and others 1991, 108; Rohde 1993; Soekirman and others 1992; the Philippines--Guillermo-Tuazon and Briones 1997; Heaver and Hunt 1995; Heaver and Mason 2000; Mason 2003; Thailand--Kachondam, Winichagoon, and Tontisirin 1992; Tontisirin and Winichagoon 1999; Winichagoon and others 1992; Costa Rica--Horwitz 1987; Jennings and others 1991, 77­81; Muñoz and Scrimshaw 1995; Honduras--Fiedler 2003; Jamaica--ACC/SCN 1989, 1996; P. Samuda personal communication, 2004; Robinson per- sonal communication, 2004; Nicaragua--Sanders 1985; Werner and Sanders 1997. Narangwal study, which was supported by the U.S. Agency for from an earlier study (Gillespie and Mason 1991, 76), com- International Development (USAID). A number of overviews bined with the estimated improvements from large-scale and analyses of these programs have been conducted--for programs, led to the assertion that "there seems to be some example, Allen and Gillespie (2001); Berg (1981, 1987); convergence on around $5 to $10 per head (beneficiary) per Gillespie, Mason, and Martorell (1996; includes a summary of year being a workable, common level of expenditure in nutri- overviews, 60); Gillespie, McLachlan, and Shrimpton (2003); tion programmes, though not generally including supplemen- Jennings and others (1991); Mason (2000); Sanders (1999); and tary food costs . . . effective programmes, with these levels of Shrimpton (1989). These plus some newer examples provide expenditure, seem to be associated with reducing underweight case studies for this chapter, and the sources for the case stud- prevalences by around 1­2 percentage points per year" ies are included in table 56.2. (Gillespie, Mason, and Martorell 1996, 69­70). Underweight prevalences are improving at about 0.5 per- A further important consideration is that the effect is likely centage points (ppts) per year except in Sub-Saharan Africa, to be nonlinearly related to the expenditure, showing the famil- which is largely static (ACC/SCN 1989, 1992, 1996, 1998, iar dose-response S-shaped curve. Thus, the first expenditures 2004). Programs are needed to accelerate this trend. Cost data produce little effect on the outcome, and one needs a minimum Community Health and Nutrition Programs | 1055 Prevalences of underweight children (percent) preventive services, notably immunization, access to basic 50 drugs, and management of the most serious threats to health, such as some access to emergency care. Moving up the devel- opment scale, starting community-based activities may soon Indonesia 40 become cost effective for prevention, referral, and management of some diseases (notably diarrhea) when coverage of health services is poor. Community-based programs continue to play Philippines 30 a key role until health services, education, income, and com- munications have improved to the point that maternal and child mortality has fallen substantially and malnutrition is Thailand 20 much reduced; at this intermediate development level, the needs are less felt, and health services again take on a more prominent role. In this scheme, the widely felt need for better access to emergency obstetric services is problematic, requiring 10 1970 1980 1990 2000 2010 a well-developed human and physical infrastructure, yet Sources: ACC/SCN 2004; FNRI 2004; Mason, Rivers, and Helwig 2005. arguably being one of the highest priorities. Note: 2 standard deviations NCHS/WHO standards; ages 0­60 months. Facility-based programs can be seen either as linking with the community program (referrals, home visits from clinics, Figure 56.1 Comparison of Trends in Underweight Children in and so forth) or as actually being part of the same enterprise. A Indonesia, the Philippines, and Thailand distinction is that community-based activities take place out- side the health facility, in the home or at a community central input level of resource use before a worthwhile response is point, even if they may be supported by health personnel based achieved (Habicht, Mason, and Tabatabai 1984). This factor in health facilities. The local workers in community-based pro- generally applies to drawing inferences from cost-effectiveness grams may be drawn from the community itself, may be home ratios, which often assume linearity. If the relation is S-shaped, visitors from a health center or clinic, or may sometimes be vol- the implication is important: applying too few resources does unteers supervised by these home visitors. Many community- not simply solve the problem more slowly but does not solve it based programs come under the health sector, whatever the at all and is a waste. Therefore, program intensity (resources exact arrangements with local health services. Regarding spe- per person) is a critical measure. cific program components, we return to the relative role of Effective interventions must include a range of activities community programs and facilities later. relating to health and nutrition. They should be multifaceted, The integrated management of infant and childhood illness not just for effectiveness but also for organizational efficiency. (IMCI) program provides guidance mainly on the curative The structure needed for community-based programs could health aspects and contains a number of nutrition activities never make sense or be sustainably set up for single interventions (for example, administration of vitamin A capsules). Links to alone. One often-argued case (for example, by Save the Children local health facilities are essential for the maintenance of the U.K. 2003) concerns children's growth monitoring: evidently community activities and for referral in cases of illness (see growth monitoring in isolation from activities that improve chil- chapter 63). As the IMCI training and implementation pro- dren's growth is not going to achieve anything (or worse, con- gresses, it should integrate directly with CHNPs (in fact, sidering the opportunity cost); however, weighing children and become part of the same exercise), which will add treatment of charting their weight can be a useful part of broader programs additional diseases. IMCI addresses diarrhea, acute respiratory (for example, as growth monitoring and promotion). infection (ARI), malaria, nutrition, immunization, safe moth- erhood, and essential drugs (WHO 1997). The 16 key practices COMMUNITY- AND FACILITY-BASED PROGRAMS for child survival defined in the context of IMCI (Kelley and Black 2001, S115) are exactly those to be promoted within Protecting and improving health, especially in poor communi- CHNPs, and most are already included (four are nutritional). ties, requires a combination of community- and facility-based Decentralization should be considered in this context. activities, with support from central levels of organization, as Although decentralized systems might be thought to be more well as some centrally run programs (for example, food fortifi- effective in supporting CHNPs, the evidence for this assump- cation). The place of these activities in a strategy is likely to tion is scarce. Decentralization can reduce resources available at vary, depending on level of development (of infrastructure, the local level if it involves devolving responsibility without the health services, and socioeconomic status) and on many local concomitant budgetary resources (Mills 1994). For example, in factors. For the poorest societies, the first priorities are basic Kenya, decentralization did not accompany devolving authority 1056 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others for raising revenue locally. In other cases (for example, the programmatic (directly under the influence of the intervention Philippines), decentralization has involved a shifting of itself); others are contextual. resources, but with priorities set in the local government units The importance of context, within which programs are ini- by locally elected officials (municipal and city mayors), these tiated and run, thus emerged as crucial, and priority factors resources may be used for shorter-term priorities than under were proposed from studies of community-based programs in previous, centrally decided, policies. Asia (Gillespie, Mason, and Martorell 1996, 67; Jonsson 1997). Sanders (1999) described similar concepts under the headings SUCCESS FACTORS of community participation and political will. This distinction and interplay between context and program factors is helpful in A number of useful concepts grew in the 1990s in relation to identifying required supporting policies to improve the context effective community-based programs. The concept of success to make programs work. Details are in the later section titled factors helped sort out complex interactions: when numerous "Contextual Factors." possibilities exist, understanding the successful pathway to An overall framework (figure 56.2) for causal links to child effectiveness is more important than trying to disentangle what survival and nutrition, put forward by the United Nations did not work. Focusing on successful programs helps simplify Children's Fund (UNICEF 1990), gave a basis for a common complexity and identify success factors, only some of which are language--even if the details might be questioned--revolving Malnutrition Outcomes and death Inadequate Immediate Disease dietary intake causes Insufficient health Inadequate Inadequate care for Underlying services and unhealthy access to food mothers and children causes environment Inadequate education Formal and nonformal Basic institutions causes Political structure Economic structure Potential resources Source: Redrawn from UNICEF 1990. Figure 56.2 Conceptual Framework for the Causes of Malnutrition in Society Community Health and Nutrition Programs | 1057 around food, health, and care as proximal causes to be addressed planning, community participation, and spiritual and ethical through programs. Improving these factors attacks hunger, dis- development. ease, and neglect, which are the converse of food, health, and A diagram of the structure, derived from Thailand's program care. Basic causes are, like context, open to influence through (figure 56.3), shows the relations between services that provide policy decisions and acting through directly influencing food, supervision and contacts with the community ("facilitators") health, and care and by modifying the effect of programs. Here and with community workers, referred to as "mobilizers." malnutrition is seen as the outcome of processes in society, and The activities undertaken in CHNPs--the program direct interventions are seen as both shortcutting the needed content--are familiar and are described here only briefly. basic improvements in living conditions and being dependent Program components, implemented by village workers or in on these improvements in the long run for sustainability. facilities, come under the following headings, which form a menu, with the actual mix depending on local capabilities and conditions (UNICEF 1998, 84; see chapter 24): COMMUNITY-BASED PROGRAMS--WHAT ARE THEY? · Prenatal care includes checking weight gain in pregnancy, prepregnancy weight, anemia, and blood pressure; Community health and nutrition programs are often initiated providing multiple micronutrient supplementation and and run by the health sector, but sometimes a separate ministry immunization (tetanus); counseling on diet, workload, (for example, in India and Indonesia) or service (for example, breastfeeding; and predicting and arranging for delivery. in Bangladesh) is set up. Attempts to use a national coordinat- · Women's health and nutrition entails counseling on health ing body appear to be less effective in leading to widespread and nutrition and checkups, promoting improved status community programs; an example existed in the Philippines and resource allocation in home and outside, promoting until approximately 2000 (Heaver and Mason 2000). This inef- improved access to health services, and often offering fectiveness stems from the tendency of the coordinating body family-planning services (these services may even be an not to have direct authority over fieldworkers or the budget to initiating factor for CHNPs, for example, in Indonesia). create a national program with sufficient coverage and intensity · Breastfeeding includes providing knowledge on practices (ini- to have a measurable effect. In some other cases, the services tial, exclusive, continued); arranging mutual support; build- linked to poverty alleviation and social welfare programs can ing confidence; preventing misinformation and undermin- play this role (for example, the Samurdhi program in Sri ing factors; facilitating time for breastfeeding; and providing Lanka). Involvement of the health services remains crucial, information along the lines of the infant formula code. sometimes as the operational agency responsible for the pro- · Complementary feeding includes providing knowledge and grams and certainly always for referral. counseling (timing of introduction, type, energy density, CHNPs have so far been much more relevant to communi- frequency, and so on); sometimes promoting village or cable diseases than to noncommunicable diseases in conditions urban area production of weaning foods; sometimes of poverty and where undernutrition is common. (An excep- marketing inexpensive food; facilitating mother's time allo- tion occurs if CHNPs help prevent intrauterine growth cation; and promoting technology--storage, preservation, retardation with later risks of noncommunicable diseases.) hygiene methods (fermentation, even refrigerators). However, in areas where diet-related chronic diseases are · Growth monitoring and promotion requires equipment developing in conditions of poverty (for example, much of (scales, charts, manuals); training and supervision; needs Latin America and the Caribbean) and obesity is rising rapidly, training of weigher to interpret charts and counsel mother; the promotion of behavior change through counseling in and a referral system for problems (for treatment, counsel- CHNPs may become increasingly important. Promoting ing, or other preventive intervention if growth is faltering). healthier diets requires access to outlets for fruit and vegetables, Weighing at birth and monthly weighing should be often displaced by fast foods, which should be a concern of included, if possible, and adequate weight gain (rather than community activities, as should lifestyle improvements such as achieved weight or any gain) should be used for guidance on use of exercise and recreational facilities. counseling or other intervention. CHNPs often include activities well beyond direct preven- · Micronutrient supplementation should include vitamin A tion and behavior change. As envisaged with primary health for nonpregnant and pregnant women (low dose weekly, care, water, sanitation, and other aspects of environmental preferably as part of multinutrients); for women within one health are frequently included, as well as agricultural interven- month of delivery (massive dose to protect infant through tions (for example, Zimbabwe in the 1980s). In Thailand, the breast milk); for infants and children (massive dose at nine village programs are part of the "Basic Minimum Needs" months immunization contact and thereafter every six approach, which includes housing and environment, family months and when medically indicated). It should also 1058 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others Services Government, NGO--health, education, agriculture, and so forth Facilitators Interface 1:10­20 mobilizers Community Supervision, training, Plan, implement, monitor, ... information, support 1:10­20 Mobilizers families Families Counseling, organization, supplies, and referral for prenatal care, child care practices, growth monitoring, micronutrients ... Source: Adapted from Tontisirin (1996, personal communication). NGO = nongovernmental organization Figure 56.3 General Structure for Community-Based Programs, Based on Thailand's Program include vitamin A--daily or weekly, with immunization · Oral rehydration includes highly effective local preparations campaigns, and so forth--and iron--daily or weekly for for dehydration in acute diarrhea, as well as (or better than) women (especially during pregnancy) as well as for children oral rehydration salts. These preparations require counsel- and adolescents. Iron is usually provided together with folic ing of mothers and take a lot of parents' time. Persistent acid and may also be provided as part of multiple micronu- diarrhea requires other intervention, especially nutritional. trient supplementation. Iodine is usually provided by forti- Care of children during sickness--especially continued fication and can be an infrequent (six-monthly) oral sup- breastfeeding and other foods--needs to be stressed plement, if necessary, but it should be part of multiple (applies also to other illnesses). micronutrients for pregnancy. · Immunization includes informing, referring, and facilitating. · Micronutrient fortification is not usually included locally, · Deworming requires distribution and dosage supervision of although it is an important central program, but local mon- mebendazole every few months, a highly effective nutrition itoring is a coming opportunity, especially of iodized salt intervention. Distribution methods are an issue. testing kits. · Supplementary feeding, using external supplies may some- The relative suitability of community- and facility-based times be appropriate in emergencies and in conditions of operations for the different components again depends on extreme poverty (for example, the Bangladesh Integrated local conditions, and these operations should be complemen- Nutrition Program, or BINP), providing 200 to 500 kilo- tary. Community activities are essential for infant and child calories per person per day, but otherwise it is to be avoided feeding, other caring practices, environmental sanitation, and as costly, with high opportunity cost, and not very effective; the like. Facilities have a key role in immunization, prenatal moreover, it can distort programs, which come to be seen care, and--of course--referral for treatment. Growth monitor- largely as a source of free food. ing, micronutrient interventions, oral rehydration, and similar · Supplementary feeding, using local supplies can be useful for activities may be focused in either. Because it has more regular complementary feeding (weaning) if carefully organized contact with clients, a community-based program may be (which requires some resources). Village community pro- more effective in actually reaching mothers and children with duction and processing are useful, if feasible (for example, the component interventions than one that is facility based. in Zimbabwe), and the system can move to coupon method Box 56.1 compares two programs in Honduras that offered the (for example, in Thailand). same content but differed in where the programs were based. Community Health and Nutrition Programs | 1059 Box 56.1 Differential Effectiveness of Community- and Facility-Based Programs Effectiveness is more likely to be possible through age participating in the community-based program were community-based programs because contact with care- 1.6 times more likely to be appropriately fed than were givers is typically more frequent and consistent. For ex- children not enrolled in growth monitoring and promo- ample, 83 percent of children enrolled in a community- tion. Children participating in the community program based growth monitoring and promotion program in also were more likely to have received vitamin A and iron Honduras (Atención Integral a la Niñez Comunitaria, or supplements than children participating in the facility- AIN-C) were weighed two or more times in a given based program. Results show that consistent participa- three-month period, whereas only 70 percent of children tion in the community-based program was associated were weighed with the same frequency in a facility-based with better weight for age. When a range of maternal and program. Workers visited 30 percent of mothers partici- socioeconomic factors were taken into account, children pating in the community program in their homes at least participating fully in the community program were once for follow-up when their children were sick, were 435 grams heavier than children who were enrolled but not growing, or had missed a weighing session. participated infrequently. In the facility-based program, Controlling for a range of maternal and socioeconomic there was little difference in weight for children based on factors, researchers found that children 6 to 24 months of levels of participation. Source: Plowman and others 2002. PROGRAMMATIC FACTORS the programs (a form of targeting), based on nutritional status, as in growth monitoring and promotion, as well as on a one- Programmatic factors are considered first in terms of the char- time basis (for example, thin children in Zimbabwe). Recent acteristics of the activities--their population coverage and tar- thinking suggests that because mortality risk, growth failure, geting, how much resources are applied per head (intensity), and morbidity are concentrated in children less than two or and the technologies used. Then the needs for initiating and three years of age, in contrast to an earlier focus on children sustaining these activities are discussed--the training needs, under five, these younger children should increasingly be a supervision methods, and (importantly) incentives and remu- focus of CHNPs. A common policy observed in practice, there- neration for field workers. fore, is to aim for complete coverage within the areas partici- pating, adding new sites until the entire region is covered. Coverage, Targeting, Resource Intensity, and Technology Relatively untargeted expansion to universal coverage may have Even effective programs improve the health and nutrition only been at the expense of establishing adequate resources and of those they reach, so achieving as complete coverage as quality in the areas initially covered. In at least one case possible of those at risk is a major determinant of the effect. (Thailand), having achieved broad coverage and reduced mal- Although variations in the content of programs are seen in nutrition, the program became more targeted to areas in which different circumstances, most activities are common to most progress was lagging. The coverage figures in table 56.3, programs. Variations in effect stem from factors such as cover- although approximate, demonstrate considerable success in age and adequacy of resources. How have CHNPs fared in initiating and implementing CHNPs on a large scale--usually reaching large sections of the population with adequate enough to have a substantial effect if the other factors needed resources--and, indeed, what is the gap that would need to be for success were met. filled? The achievements of the 14 programs drawn on here as How complete a coverage of the population should one rec- case studies are summarized in table 56.3. ommend? This factor relates to targeting, to the additional The programs expanded to include most of the communi- resource requirements to reach the nonparticipants, and to ties within the areas targeted. The common evolution was to their level of risk. Usually risk is spread throughout the popu- target select areas and specific biological groups within those lation, although the extent varies considerably--at least a areas--generally women and children--but not to give priority doubling of indicators of risk is usually seen between better- to any great extent to poorer or less healthy communities. and worse-off areas or groups (for example, see Mason and Screening is sometimes done of individuals for admittance into others 2001, figures 1.4­1.7, 1.10­1.13). The remoter areas--or 1060 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others Table 56.3 Characteristics of Selected Programs Country Coverage, targeting Resources, intensity Africa Tanzania: Iringa Population served 250,000 in 6 districts, 610 US$8 to US$17/child/year (approximately US$30/child/year from total costs: F: ( ) villages, 46,000 children, of which 33,700 par- approximately US$6 million) ticipated (73 percent). Targeting: children 5 2 village health workers/village 1,220 total; approximately 1:40 children years and women; no socioeconomic selection [Volunteers] of communities. Progressed from 168 to 610 villages 1984­88. Tanzania: Child Survival 9 of 20 regions (population total approximately US$2 to US$3/child/year and Development Program 12 million; 2 million children). Aimed for com- [Volunteers] F: 0 plete coverage. Zimbabwe: Supplementary Population served: 56,000­96,000 with supple- External: US$3 million over 10 years Feeding Programme mentary feeding; up to 60 percent of all chil- For example, 1990, US$0.5 million, approximately US$0.50/child/year F: dren in community-based growth monitoring. (Approximately 1:10­200, based on numbers per project) [Extension agents] Asia Bangladesh: BINP BINP: in 6 thanas, or subdistricts (7 percent of US$14 million/year; approximately US$18/child/year F: population), children 2 years, 8 million preg- 1 community worker per 1,000 population nant and lactating women. Approximately 1:200 children [Project supported] Bangladesh: BRAC Health coverage 25 percent. Nutrition with 1 community health volunteer per 300 households; 1 community nutrition BINP, now expanding. promoter per 200 households; community nutrition centers, 1:120 mothers and children; supervision of community nutrition promoters by community nutrition organizer, 1:10 India: ICDS Children 0­6 years and pregnant and lactating Nonfood costs: approximately US$2/child/year. F: / women, in 3,900 of 5,300 blocks, or subdis- 1 community worker (anganwadi worker, or ANW) per 200 children; tricts; approximately 74 percent of population. 1 supervisor per 20 ANWs Coverage expanded without targeting except [ANW paid, at low rate] by area. India: TINP Children 6­36 months, pregnant and lactating US$9/child/year, plus approximately US$3 on food. F: women. Children with growth failure selected. 1 community nutrition worker per 300 children; 1 supervisor per 40 percent of blocks in Tamil Nadu; 20 percent 10 community nutrition workers of children in 1990. [Project supported] Indonesia By 1990, 60,000 villages (of 65,000: 92 per- US$2­11/child/year, depending on supplemental food; Rohde (1993) gives F: ( ) cent) had posyandus (village health/nutrition US$1 recurrent. center). Women and young children. Village workers (approximately 3 million total), 1 per 60 people, approximately 1 per 10 children; supervision 1 per 200. [Volunteer] Philippines: national Several programs, all targeted (for example, to US$0.40/child/year in targeted areas. F: 0 poorer areas), none with national coverage. Village workers (barangay nutrition scholars) approximately 1:300 [Low allowance given] Thailand: Primary Health Expanded over about 5 years to cover 95 per- Ministry of Public Health; approximately US$11/head/year (1990) Care Poverty Alleviation cent of villages. 600,000 village health commu- 1 village health communicator or volunteer per approximately 20 children; Program Basic Minimum nicators (1 percent of population) trained; 1 supervision extension worker per 24 village health communicators and Needs 60,000 village health volunteers. volunteers F: ( ) [Volunteer] (Continues on the following page.) Community Health and Nutrition Programs | 1061 Table 56.3 Continued Country Coverage, targeting Resources, intensity Americas Costa Rica Expanded rural health program coverage Rural health program: US$1.70/child/year F: to 0 19­67 percent (1974­89). Food and Nutrition Program: US$12.50/child/year 2 health workers (full time) per 5,000 population; approximately 1:350 children [Health worker] Honduras With community health volunteers, AIN-C Cost estimated as US$6/child/year F: 0 covers 50 percent of health areas (expanded Volunteer teams 3:25 children, about 3.5 hours/volunteer/week 1991 on), 90 percent of children 2 years in these; growth monitoring and home follow- up, plus referral and treatment. Jamaica Community health aides (CHAs), waged, cover CHAs (full time) 1:500 households; approximately US$7/household/year F: 0 most of country from health centers, with [Health worker] home visiting. Nicaragua Community health workers (brigadistas) with Volunteers, approximately 1:20 households F: 0 "multiplier" approach, training others; 1980 approximately 1 percent trained; many more for malaria control. Source: See sources for table 56.2. F role of supplementary feeding in the program; F: mainly a feeding program, or primary role; F: significant but not main role, often to selected children; F: ( ) existed but relatively minor; F: 0 none. Note: The status of community workers is given in brackets in the last column. groups that are hard to include for other reasons--may be the basis for estimating that US$5 to US$10 per child per year more expensive to reach. Clearly the calculations depend on may be needed for effective programs. The dollar figures vary conditions and have to be made on a case-by-case basis. The from less than US$1 to more than US$20. Probably the low end principle is obvious: only those areas and people included in of this range (say, less than US$1 per child per year) does explain CHNPs are going to benefit; so wherever need exists, programs low or doubtful effect. Both low coverage and low intensity are indicated. The implementation strategy, in theory, may may explain the unchanged underweight prevalences in the need to begin with the most urgent needs, although in practice, Philippines until 2000. Fund levels in Indonesia are unsure; programs may expand from the easier, more accessible areas; Rohde (1993) gave a figure of less than US$1, but others gave this practice seems reasonable, provided that the expansion higher estimates. Most would reckon the intensity in India too really occurs and leads to equitable use of resources. low (Measham and Chatterjee 1999) at about US$2 per child The program content is a mix of the components described per year. Looked at otherwise, the intensity planned for external earlier, varying with local priorities. The most crucial difference funding (even if part of such funding is international costs) is in is whether extensive supplementary feeding is included. In the US$10 to US$20 range (Bangladesh, India--Tamil Nadu, middle-income countries, supplementary feeding was less and Tanzania) and is the same as the estimate for Thailand. A prominent, often considered unnecessary, and because expen- level of US$10 to US$20 per participant per year is probably sive, perhaps counterproductive (for example, in Costa Rica; advisable for planning and sustaining effective programs. Mata 1991). At the other extreme, such as for the Integrated The intensity measures of workers per mother-child and the Child Development Services (ICDS) in India, food distribution supervision ratios are relevant in assessing needs. The sug- became the raison d'être of the program but, alone, was again gested norms, originating from the Thai experience are 1:10­20 probably not worthwhile. For some of the intermediate cases, for both. Since then, it has emerged that the full-time equiva- supplementary food played a supporting role, with varying lence of community workers must be taken into account; the results. Except in the very poorest societies, supplementary Thai workers are local volunteers, probably devoting 10 to feeding seems unlikely to be cost-effective. 20 percent of their time. In Honduras, Fiedler (2003) in a care- The resources used for the programs found in table 56.3 can ful cost study estimated that each volunteer spent 3.5 hours per be expressed per participant (referred to as intensity), as total week (less than 10 percent of full-time equivalent, or FTE), expenditures, and in terms of personnel; the latter figures may with a ratio of 1 volunteer to 8 children. The ratio of commu- be more generalizable. (The outcomes associated with these nity health and nutrition workers (CHNWs) to children may, resources are shown in table 56.5.) Data such as these have been therefore, be as low as 1:200 for FTEs and as high as 1:8 or 1:10 1062 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others for part-time volunteers. In Jamaica, where the community which CHNPs' main role is to provide information and to health aides work full time, the ratio is 1:500 households; in the ensure that children are taken for immunization (either to reg- BRAC program in Bangladesh, it is 1:300, about half-time work ular clinics or for National Immmunization Days and the like). (afternoons) (Chowdhury 2003). (An indication of the status Periodic deworming can be conducted by community pro- of community workers is shown in brackets in the last column grams (and hookworm vaccines currently under development of table 56.3.) In any event, these ratios provide some basis for may soon contribute). Supporting the use of insecticide-treated gauging the adequacy of personnel, and it seems that an effec- bednets could be fostered through CHNPs. By far the most tive ratio may be about 1:500 for community workers potentially important application of technology, certainly in employed full time and 1:10 or 1:20 for local volunteers work- Sub-Saharan Africa, will be the unprecedented effort to provide ing part time. millions of people with antiretroviral therapy and associated In reality, the ratios of community workers to children are care and support, as discussed later. probably--not surprisingly--on the low side. Thailand, which trained 600,000 village workers (1 percent of the population), operated at about 1:20 for part-time volunteers, with similar Training, Supervision, Incentives, and Remuneration supervision ratios. The Indonesian program was similar (or Community-based health and nutrition programs typically better) but had much less supervision. In contrast, the low involve community workers, who may be entirely part-time resourcing of the ICDS in India shows up in a ratio of 1:200 volunteers (for example, in Honduras and Thailand) or may (for part-time anganwadi workers, or ANWs), and in the receive some remuneration financially or in kind (for example, Philippines, the ratio has until recently been 1:300 (for essen- in India). Community workers may be part of the health sys- tially voluntary workers). tem, earning a wage and based in a local clinic (for example, in Increased application of technology can contribute to the Jamaica) or in the community itself (for example, in Costa organization and running of community-based programs. Rica); or they may be selected by and report to the community Technology can be applied easily to methods of assessment and (for example, in Tanzania and Thailand). Table 56.3 indicates monitoring of children's progress; improved weighing scales (or the status of community workers in the programs examined in some circumstances, where rapid assessment in remote areas here. The training, supervision, and incentives for community is important, using arm circumference) can simplify anthro- workers are critical aspects of successful programs. pometry. Modern computer technology for recordkeeping Inadequate training and supervisory support of community could be much more widely used, freeing staff time for home workers are common weaknesses. Considerable attention was visits (for example, in Jamaica); e-mail, which is being rapidly given to training for the Iringa project (Tanzania), with village adopted, has great potential for transferring information, trou- health workers trained for up to six months. In the Tamil Nadu bleshooting, and consultation. Cell phone use is beginning Integrated Nutrition Program in India, community workers to transform communications even in the poorest countries, received three months of training and participated in annual where it is leapfrogging landline installation and use; as cover- refresher trainings. ICDS (India) initially trained the ANWs for age expands, it will facilitate referral, for example, for emer- three months, with two annual refresher courses, but this gency obstetric care, the need for which may first be identified process declined. In Thailand, volunteers had two to five days by community workers. Coupled with improved transportation of initial training, with annual refresher courses; Indonesian and procedures to allow the use of such transportation in cases practice was similar. In Jamaica, where the community workers of urgent need, modern communications can link communities are employees of the health system, two months of initial train- to centers with advanced knowledge for information exchange ing is provided to recruits with significant prior educational and, by facilitating transportation when time is crucial, for requirements. In Bangladesh, the BRAC community health vol- referral. Modern communications may also provide more effi- unteers have four weeks of training. The quality of the training cient ways of providing training, retraining, and supervision. has varied, poor training having been blamed for inadequate Application of current research and resulting technologies implementation in cases such as ICDS in India (Measham and can improve many of the other interventions discussed earlier. Chatterjee 1999). Sanders (1985, 176­93) describes experiences In the micronutrient field, periodic supplementation (with in the 1980s of village health workers (and barefoot doctors) vitamin A in high doses) can be extended through community and their relation to the community. programs, and fortified foods and micronutrient "sprinkles" Supervision of community workers is generally done by can be promoted (see chapter 28). The prospect of enabling employees who are commonly in the sector. Training of communities to test their salt for iodine content with simple supervisors (who often take on the role in addition to many and cheap test kits is intriguing and has often been recom- other tasks) for these purposes is highly variable and not always mended but has not yet been widely applied. Improved immu- adequate. Providing resources for visits to provide supervision nization technology should continue to protect health, for to community workers is a further constraint. Supervision Community Health and Nutrition Programs | 1063 ratios in effective programs are about 1:20 (table 56.3, last col- frequently be an option. Hiring additional personnel as com- umn, when reported). Supervision and training of community munity health workers would consume a significant proportion workers are closely linked; indeed, supervision (which must be of typical health budgets (at 1:200 households for FTEs, this supportive rather than disciplinary) should include a substan- would amount to US$1 to US$2 per person per year, or about tial element of on-the-job training. 20 percent of public health budgets in low-income countries). Remuneration and incentives for sustaining motivation are A mix of redeployment of existing staff and new hiring from key issues in replicating the successful features of these pro- budget reallocations should, nonetheless, be cost-effective. grams, and the options vary with the culture. In Thailand, it is argued that village volunteers consider the prestige associated with the role of health worker preferable to getting a low wage. Organization In many cases, some right of access to health care is part of the Effective, respected, and socially inclusive organization at the incentive. For the ICDS in India, in contrast, the ANW receives community level seems to have been a key feature of the suc- a small financial remuneration, but the government (as else- cess in launching, expanding, and sustaining CHNPs. Most of where) will not grant formal employment status (and attempts the successful CHNPs drew and built on established community to form unions have been strongly discouraged). Direct com- procedures; where they did not, effect and sustainability were parisons of the options of paid remuneration and voluntary in doubt. In Thailand, the health services and the religious work are rare. One opportunity to study options for remuner- organization at village level were important. The health services ation is in the Philippines, where under a World Bank Early themselves play a key role in Costa Rica, Honduras, and Child Development project, the child development worker Jamaica. In Indonesia, it was the community organizations receives a wage, which could be compared to near-volunteers at (and women's groups) together with (initially) the family- the barangay (village) level. planning services. In Iringa, Tanzania, it was the local political When CHNWs are primarily voluntary, they are selected by party structure, with substantial input from UNICEF. In the community and report to community committees in some Zimbabwe, immediately after independence, it was the village form. CHNWs on government payrolls may come from the organizations that had fought the war, later helped by a con- communities and thus may be known to and identify with the sortium of national and international nongovernmental communities, but they may report to supervisors higher up in organizations (NGOs). The major part of the still-expanding the system. Both models can work, depending on the culture. program in Bangladesh is run by BRAC, an NGO, and has built What probably works least well is when the community worker on its links to the community for development, food security, is paid little and receives inadequate support and recognition and educational activities, as well as for health. In contrast, from the community or even comes from elsewhere. CHNPs that either failed to launch a wide program (for exam- Furthermore, as development progresses, reliance on volun- ple, in the Philippines) or had limited effect (in India, ICDS) teerism may become less useful. probably lacked some of these features. Inclusiveness is proba- For planning CHNPs in terms of community workers, the bly hard to achieve if not inherent. total numbers and resource implications can be estimated as Support from the central government is also crucial: CHNPs follows. A full-time equivalent CHNW might visit 5 to 10 need this support for training,supervision,wages,supplies,facil- households per day, averaging a visit to each household roughly ities, and the like. Where such support becomes a regular gov- every two months; a ratio of 1 CHNW to 200 households, there- ernment budget item, activities tend to become embedded and fore, seems to be in the range within which an effect in terms of are sustained, in contrast to where the support is from donors. improving child health and nutrition is expected. Calculations A further issue concerns maintaining the community pro- from salaries of community health aides (CHAs) in Jamaica gram's preventive orientation. In Indonesia, for example, work out to US$7 per household per year, within the usual according to Rohde (1993), the health services co-opted (and range for expected effect. An important factor in regard to medicalized) the posyandu (weighing post, or community financial resources, however, concerns the substantial cadre of health and nutrition center) system by adding a diagnostic and personnel who have training and job descriptions for commu- treatment module (in fact, a table in the meeting place). This nity work, are based in health centers, and for administrative module attracted most of the attention, to the disadvantage of and financial reasons seldom leave the health facility. Moreover, the preventive aspects of the program. Thus, if the extension of funds may not be released to allow travel to nearby villages. An IMCI into the community means treatment (by trained but not example is from Jamaica, where, because of clinic workloads, medically qualified people) in the community rather than CHAs spend time helping in clinics rather than on home visits; referral to facilities, treatment could become the main or even in fact, technology could free staff time for community work sole focus, shifting attention from prevention. Some parallels by automating tasks, such as record keeping, that detain the exist to the experience of ICDS in India, where, as noted earlier, CHAs. More effective deployment of existing personnel may food became the raison d'être. 1064 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others CONTEXTUAL FACTORS cult to distinguish from program effects (current examples are in Bangladesh and Vietnam, both showing rapid improvement Community-based programs can work usefully, bringing in nutrition). Moreover, the same factors (again, such as edu- steady progress; whether they do depends on myriad factors cation) may both produce endogenous change and increase the relating to the context. Three different concerns are (a) factors effect of program activities. affecting widespread initiation of CHNPs of potentially ade- Five contextual factors have been suggested as priorities (in quate coverage, intensity, and content; (b) factors that lead to Asia; Mason and others 2001): sustainability; and (c) factors that allow activities to be effective · women's status and education in improving health and nutrition--at best, when they, them- · lack of social exclusion selves, also contribute to a rapid transitional improvement, as · community organization in Thailand, Costa Rica, and Jamaica. · literacy Contextual factors may bring about improvements in health · political commitment. and nutrition without any additional direct action--through improving living conditions, education, and so forth. Often, Table 56.4 shows estimates of the positions of countries with the changes caused by such nonprogrammatic factors are diffi- case study programs in regard to these factors. The levels of Table 56.4 Context in Which Selected CHNPs Start and Run Level of Women's Lack of health and Total minus Approximate status and social Community administrative Political political Country period educationa exclusion organization Literacy infrastructure commitment commitment Total Tanzania Iringa starts 1984­90 2 4 4 3 2 5 15 20 Iringa declines 1990­ 2 4 2 3 2 2 13 15 Zimbabwe Supplementary 1981­90 2 4 5 2 2 5 15 20 Feeding Programme starts Supplementary 1990­ 2 2 2 2 2 2 10 12 Feeding Programme declines Bangladesh BINP 1997­ 1 3 2 2 3 3 11 14 India ICDS 1975­ 1 1 2 2 2 3 8 11 TINP 1980­9 2 2 3 3 3 4 13 17 Indonesia UPGK starts 1975­ 2 4 3 2 2 4 13 17 UPGK declines 1990­ 2 4 2 2 3 2 13 15 Philippinesb 1974­2000 4 4 3 4 3 1 18 19 Thailand 1982­ 4 3 4 4 3 4 18 22 Costa Rica RHP 1973­ 4 4 4 3 4 4 19 23 Jamaica 1985­ 4 4 3 4 4 4 19 23 Nicaragua 1979­90 3 2 3 3 3 4 14 18 Source: Authors. a. Women's status and education can be quantified by indicators such as adult literacy rates, females as percentage of males, and secondary school enrollment for girls. b. Since 2000, the Philippines has begun a significantly improving trend, one factor being increased implementation of programs (CHNPs, as well as others, such as salt iodization); this increase is caused in part by increased political commitment, both as new legislation and resource allocations. Note: 0: worst; 5: best. Community Health and Nutrition Programs | 1065 health and administrative infrastructure have been added. The prerequisite for progress and as a responsibility for those table also shows changes in these factors that may help explain working for health, especially of the poor and of the desti- why the CHNPs declined in three cases. tute sick. This investment may be long term and difficult-- Political commitment can lead to initiating community as in Kerala, India, for instance--but must be seen as programs and mobilizing resources. It may also respond to integral to the struggle for health (Sanders 1985). emerging community mobilization, as seems to be the case Operationally, this commitment to human rights puts when programs start after political upheavals, as in Zimbabwe greater responsibility on advocates and investors in health to and Nicaragua. Declining political commitment accounts for broaden the dialogue and scope for allocating resources and loss of interest by the government in CHNPs; economic decline to avoid committing resources regardless of the prospect of undermining resource availability may cause a shift away from success as influenced by the social and human rights con- financial support of CHNPs (for example, in Tanzania). In text. In health and nutrition, as in other areas, adjustment of table 56.4, estimates of levels of contextual factors are totaled policies to support the success of interventions would be both without and including political commitment (last two pragmatic as well as the right thing to do. columns). The total without commitment may indicate how · Second, even if the context is more favorable, genuine polit- favorable the context is if commitment is then made. Costa ical commitment is still essential. Excessive donor input may Rica, Jamaica, and Thailand had a favorable context and, with inhibit this commitment. It is striking that Thailand had commitment, succeeded. The Philippines had comparable to reject donor influence and make its internal decisions favorable conditions--the position of women is generally before its programs became successful (Tontisirin and good, there is limited social division (exclusion), and so on. Winichagoon 1999), Costa Rica had to fight and overcome However, the necessary commitment (of resources, in particu- a medicalized approach (Muñoz and Scrimshaw 1995), and lar) was made only recently, with new legislation, adherence to Indonesia's posyandu system was undermined when treat- regulations (for example, iodized salt went from 25 to 65 per- ment displaced prevention (Rohde 1993). cent coverage), and increased resource allocation and assign- · Third, it is clear that severe economic stress, political pres- ment of community workers. This new commitment may well sure, or both have caused unsustainability (Indonesia, explain the recently resumed decrease in child malnutrition Nicaragua, Tanzania, and TINP). (figure 56.2). In other examples--such as Indonesia and · Fourth, if the context is unfavorable, it might be better to Tanzania--the conditions were moderately favorable, and work on improving the context than to commit resources to while political support and finance existed, progress was made. programs that may not succeed--but, of course, success in In Tanzania, financial crisis denied the programs sustained improving context itself depends on circumstances, notably support; in Indonesia and Zimbabwe, bureaucratization and political commitment. centralization of the political process, followed by political turmoil, contributed to a similar outcome (Sanders 1993). The Considerations like these should contribute to identifying situations in India and Bangladesh have not been very favor- supporting policies needed for programs to be effective and able. The position of women and social rifts, amounting to modifications to interventions in particular conditions. For exclusion, have probably inhibited effective programs, even example, it is often observed that a particular factor--say, with political commitment. This context may now be changing access to health services--is more strongly related to improve- in Bangladesh, as seen in the activities of BRAC. Finally, this ment among the better off (for example, the educated) analysis demonstrates the relation of decline in programs to population. This interaction of program with context leads to falling political commitment in Tanzania, Zimbabwe, and identifying new needs--in this example, perhaps facilitating Indonesia. access for the illiterate. In the longer run, resources or legisla- If this analysis approximates the truth, the forward-looking tion (for example, to combat social exclusion or discrimination policy implications may be important: against women) may be highlighted as prerequisites before a program can be expected to work. Often failure to take account · First, investing initially in a favorable context makes sense of context when trying to transfer experiences from a pilot trial (as does possibly committing resources preferentially to ("scaling up") may explain why efficacious interventions prove interventions in the more favorable contexts). Supporting ineffective in a larger population. policies can address social constraints--such as improving This analysis of contextual circumstances indicates that tar- education for women--and (relatedly) seek to improve geting the poor may not always be cost-effective, and some human rights. In many cases, human rights may be of over- interventions may not be feasible in certain contexts. An exam- riding importance for health: Farmer (2003) has made a ple is when the health infrastructure and services are almost compelling case for rethinking health and human rights as a nonexistent; under those conditions, it can be argued that 1066 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others emergency treatment (especially for the diseases addressed by RESULTS IMCI) should be established and reliable resources put in place first. A similar difficulty, often seen in food security, is that Indicators of progress in implementation--process indicators-- most interventions may not work for the poorest of the poor. referring to coverage, intensity, and so on, are shown in table For instance, supporting food (or cash crop) production in 56.3. As discussed earlier, most programs expanded population low-potential areas may not be realistic; nonagricultural coverage without much targeting. But usually the level of employment may be better. resource application (intensity) was on the low side. More Thus, community-based programs work in a specific time research is needed on the contribution of CHNPs to health and place: programs may start, work for a time, and then evolve process indicators, such as immunization coverage rates, as well or fade away. Even if they fade away, some useful effect may as to nonhealth activities, for instance, in agriculture and com- be achieved: sustainability need not be forever. At the same munity development. time, short project cycles (three years for many donors) can act Impact evaluation, which refers to the net effects of inter- against sustained programs. Some compromise in donor poli- ventions on changing health outcomes, is sorely lacking. The cies to allow assurance of continuity for reasonable periods efficacy of most of the component parts of CHNPs, when (such as 10 years) could do a lot to increase the effectiveness of resources are adequate and the problems are correctly identi- donor support to CHNPs. fied, is established, but in routinely administered large-scale The essence of time and place is not fully understood. programs, the changes in outcome that can be ascribed to pro- Werner and Sanders (1997) give examples of favorable times, gram actions are less known. Although controlled trials by def- as when the old order is changing (for example, after internal inition are not applicable, plausible evidence can be obtained by conflict, as in Nicaragua and Zimbabwe) and when there is careful attention to research design, measurement, and analysis renewed vigor and some new organization is in place. Another (Habicht, Victora, and Vaughan 1999). Some form of "with and generalization of a favorable context is when energy and inter- without" and "before and after" comparisons is needed; for connectedness exist in society. Thailand illustrates both: the instance, such methods as staggered implementation, natural Thais needed to change the approach, moving away from experiments, and selection of comparison groups with some donor influence, in order to initiate the successful community statistical control can yield valuable information now lacking programs that helped transform health and nutrition and should be more widely attempted. In this context, it should throughout the country, and that worked in part because of be noted that because of the timing and level of effort necessary cohesive features of Thai society (Tontisirin and Winichagoon for the evaluation, the impact evaluation results (changes in 1999). outcome ascribed to the program) may be more important for In these examples, programs that continued on a large policy decisions on future programs than for the program that scale--either until the problem was largely resolved, as in Costa has been evaluated. Moreover, not all programs require detailed Rica, Jamaica, and Thailand, or as it was expanding, as in BRAC evaluation. Thus, financial support for such policy-relevant in Bangladesh or AIN-C (Atención Integral a la Niñez evaluations may come from budgets other than that of the pro- Comunitaria) in Honduras--clearly had supportive context, gram to be evaluated. The evaluations should also be prospec- but their specific common features (and hence how they could tive as far as possible, so decisions on evaluation design and be replicated) are elusive. Perhaps one crucial condition for finance are needed earlier rather than later. success is that circumstances are such that people and commu- nities begin to have the sense that they can take responsibility for--and control of--their health and quality of life. Impact Responsibility comes with the emancipation of societies from For the examples used here, inferences were drawn from piec- colonial or other repressive conditions and possibly when ing together results either from ad hoc surveys or from pro- grassroots attention becomes widespread, as it did in gram and administrative data; occasionally such inferences Bangladesh through an NGO that identified with the people. were made from the comparison of baseline estimates with Evidence is growing that, among the poor in the United States, midterm or final assessments, but the comparison groups, if this sense of control is directly related to better health and any, were imperfectly matched. Thus, the conclusions on reduced exposure to HIV and AIDS; Sampson, Raudenbusch, impact now put forward are tentative and based on judgments and Earls (1997) call the concept collective efficacy. Cohen and from available information. Some of these conclusions were others (2000) show that health conditions improve when com- drawn from trend assessments, details for which are in Mason munities themselves fix up their environment--the "broken (2000, annex 5). windows" theory. Such ideas may equally apply to poor com- The most widely available indicators are mortality rates munities, especially urban ones, in developing countries too. (infant, child, and to a lesser extent, maternal; reliable data on Community Health and Nutrition Programs | 1067 age zero and cause-specific mortality rates are not usually Interestingly, the Child Survival and Development Projects available from most developing countries); prevalences of (supported by the World Bank, among others), which covered underweight in children from national surveys (often a much larger population (but with less intensity than in supported by demographic and health surveys or UNICEF Iringa), appeared to show almost the same pattern as in Iringa: Multi-Indicator Cluster Surveys); and indicators of health serv- a rapid initial fall (as much as 8 ppts per year, for one to two ices (notably immunization coverage rates). Estimates of mor- years), continuing at 1 to 2 ppts per year. bidity, even of the common diseases (such as diarrhea and In Costa Rica, the child mortality rate plummeted in the late ARI), are not available systematically enough to judge trends in 1960s, well before stunting fell in the 1970s (Saenz 1995, 129; relation to programs. Child underweight (or stunting) has a Vargas 1995, 111). A lag was seen in Thailand, where the child particular value, because it measures an attribute of all children mortality rate started to fall rapidly in 1977, and both severe (age and weight or height), rather than assessing a relatively and moderate malnutrition appeared to start their fall in rare event, as in mortality estimates. Moreover, experience is 1983­84. Both these improvements preceded the major growth well established of how underweight prevalences behave as a in gross national product, which began in 1987 (Kachondam, robust indicator, having a useful degree of responsiveness but Winichagoon, and Tontisirin 1992, tables 8 and 33). In analyz- not being subject to wide fluctuations with transient events. ing Indian experience, where the IMR has fallen faster than Under controlled conditions, improving health and nutri- child malnutrition, Measham and Chatterjee (1999) suggest tion allows rapid catch-up in bodyweight and fast rates of that further improvements in child survival may be con- reduction in underweight prevalence (for example, Pinstrup- strained by the high rates of child malnutrition. Andersen and others 1993, 405). But in the real-world condi- The sustained effects are generally of about an additional tions of CHNPs, the expected rate is slower. As examples, 1 ppt per year improvement (table 56.5). For Bangladesh (the Thailand maintained a reduction in underweight of about BINP), Tanzania, and Thailand, it is possible to distinguish 2.9 ppts per year in the 1980s (see figure 56.1); the 22 projects the sustained rate from the initial rapid fall. In Bangladesh, the reviewed as reported by Jonsson (1997) ranged between about BINP started during a period of rapid improvement overall, so 1 and 3 ppts per year. A reduction rate of 2 ppts per year, sug- extracting the underlying trend is especially important to give gested earlier as an expectation from successful programs, a plausible view of the "with-project" rate: about 1.6 ppts per would lead to very significant improvement if achieved at year again seems a reasonable estimate. A similar extraction of national levels: for South Asia, it would mean going from a likely with-project changes allowing for underlying trends was prevalence of underweight of about 60 percent in 1980 to reported previously (Mason 2000, annex 5) for Tamil Nadu, 20 percent in 2000; for Africa, from 30 percent in 1990 to 10 Andhra Pradesh (ICDS), and Orissa, indicating plausible percent in 2000. improvements for the first two states. Detecting this rate can be difficult with the noise of sam- In sum, these results support the contention that after an pling and nonsampling errors and with the common seasonal initial rapid fall, the sustained rates of improvement in child changes, which can amount to 5 ppts fluctuations or more, cer- underweight prevalence settle down to about an additional tainly in Africa. The potential program-ascribed trend needs to reduction of 1 or 2 ppts per year. This conclusion is the same as be separated from the underlying secular trend for the country, previously reached (Gillespie and Mason 1991), now supported roughly 0.5 ppts per year (from 1985 to 1995; ACC/SCN 1996). by some new results. Clearly the longer the program and the observing periods, the easier it is to assess trends. Where the data are detailed enough,an initial rapid fall is seen Cost-Effectiveness in severe malnutrition--and probably in mortality,--followed Therefore, if we use prevalences of underweight children as the by a slower fall in mild to moderate malnutrition. The reasons basis for calculation, US$10 per child per year gives a reduction for the initial rapid fall are presumed to be immediate effects of of 2 ppts per year. If we are to translate this cost into an implied improved health care, immunization, and the use of oral rehy- effect on health and survival, underweight must be related to dration therapy. The outcomes estimated for the programs con- the measure of disease burden, DALYs lost. Then the resources sidered here concentrate on the sustained trend--after a year or needed per DALY saved--dollars per DALY--can be estimated. two of implementation--as summarized in table 56.5. A 32.5 percent reduction in the loss of DALYs is associated with In Zimbabwe, from 1980 to 1988, the infant mortality rate eliminating general plus micronutrient malnutrition as both (IMR) fell from 110 to 53 per 1,000 live births, and severe mal- direct effects and risk factors (see table 56.1, discussed earlier); nutrition fell from 17.7 to 1.3 percent. However, stunting fell as a first approximation, the average prevalence for developing only in 1982­88, from 35.6 to 29 percent (1.1 ppts per year). countries of 30 percent underweight can be applied. We can Tanzania shows a similar effect in Iringa, with severe and mod- calculate the associated DALYs gained from reducing malnutri- erate malnutrition falling much faster for the first two years. tion at this rate (and assume that loss of DALYs from all 1068 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others Table 56.5 Outcomes and Resources in Selected Programs Country Outcomes Resources Africa Tanzania: Iringa Underweight: 50 to 35 percent (1984­88) US$8 to US$17/person/year (US$34/child/year from total costs); Immunization: 50 to 90 percent 2 village health workers/village 1,220 total; Rates in underweight: initial 2 years, Approximately 1:40 children 8 ppts/year; first 4 years, 4.5 ppts/year; sustained (years 2­7), 0.8 ppts/year Tanzania: Child Survival Underweight reduction rates similar to Iringa US$2 to US$3/child/year and Development Program Zimbabwe: Supplementary Stunting: 35.6 to 29 percent (1982­88) External funds, approximately US$0.50/child/year Feeding Programme 1.1 ppts/year IMR: 1980­88: 110 to 53 Asia Bangladesh: BINP BINP, first 6 thanas, initial effect (1997): 1 community worker per 1,000 population; approximately 11 ppts/year; then (to Approximately 1:200 children; US$14 million/year, approximately February 1999) approximately 1.6 ppts/year US$18/child/year additional Underlying (nonprogram) trend: national approximately 1.7 ppts/year, program area approximately 2.4 ppts/year Bangladesh: BRAC No program-specific data, but child under- Over all programs, US$196 million in 2003 (approximately US$8/household weight and anemia in women have substantial over all households); health program covered 31 million people, over 20 percent falling trend in recent years. India: ICDS Overall underweight prevalence declining only 1 supervisor to 20 ANWs slowly; some states reported faster, but link to ICDS not shown. India: TINP 1979­90: 1.4 ppts/year in TINP districts; US$7­12/child/year 0.7 in non-TINP districts: increased improve- ment of approximately 0.7 ppts/year (Reddy and others 1992, 45). From other data, increased improvement of 1.0 ppts/year. Indonesia Probably about 1 ppt/year; underlying trend US$2 to US$11/child/year, depending on supplementary food. Rohde (1993) unknown gives US$1 recurrent. IMR: 1970, 1980, 1990: 118, 93, 61, respectively Village workers (about 3 million total) 1:60 people; approximately 1:10 children; supervision 1:200 Philippines: national No change found in underweight. Low coverage and intensity IMR: 1960, 1996: 77, 32, respectively Thailand Approximately 2.9 ppts/year improvement in Ministry of Public Health, approximately US$11/head/year (1990) child underweight. Breaks down to 1982­84: Village health communicator or volunteer approximately 1:20 children; supervi- 7.8 ppts/year; 1985­90, 1.9 ppts/year. sion by extension workers: village health communicators and volunteers IMR: 1970, 1980, 1990: 73, 55, 27, respectively approximately 1:24 Americas Costa Rica Stunting improved by approximately Rural health program: US$1.7/child/year 1­1.5 ppts/year (estimated from Muñoz Food and nutrition program: US$12.50/child/year and Scrimshaw 1995, 111), 1979­89. 2 health workers (full time)/5,000 population, approximately 1:350 children IMR: 1970, 1980, 1988; 62, 19, 16, respectively Jamaica 1.9 ppts/year 1985­89 IMR: 1960, 1996: 58, 10, respectively Nicaragua IMR fell from (at least) 92 to 80 Large numbers community health volunteers trained and supported Source: See sources for table 56.2. Community Health and Nutrition Programs | 1069 malnutrition comes down at this rate; CHNPs include some and Vietnam and up to 60:1 in the Philippines. Further research attention to micronutrients). This reduction is then cumulated has stressed the variation in time commitment of CHNWs in through time (five years here) and assumes a linear relation different places--hence the need to convert to FTEs. The ratio between cost, underweight reduction, and disease burden used as the norm, derived from experience in Thailand and avoided. The calculation also assumes a persistent effect of Indonesia, of about 1:20 is probably equivalent to 1:200 in FTEs. reducing malnutrition. In India, opinion has been that about a doubling of the ANW Using these assumptions gives an estimate of US$200 to numbers in the ratios is needed to get more effect (Measham and US$250 per DALY saved in sustained programs. This estimate Chatterjee 1999). From this perspective, these estimates indicate does not include gains in DALYs from diseases that do not that both coverage and intensity are low, although intensity may show up as underweight, which must be substantial. Moreover, be half that needed, whereas coverage (except in India) is far too if this calculation is applied just to the first rapid fall, typically small. Supervision ratios are estimated as about 1:20 and higher. (in the three cases examined) about 8 ppts per year, the ratio Expanding the numbers of CHNWs also means increasing the might fall by a factor of four, to US$50 to US$60 per DALY number of supervisors (usually from the health system), with saved (but start-up costs are higher too). The sustained figure associated costs. should be the more generalizable. Calculations from scarce financial resource data show that Many further provisos exist. Much of the effect here is on most government programs cost about US$1 per participant a risk factor--malnutrition--reducing which, in turn, makes child per year or less, whereas Bangladesh (BINP, with donor other interventions more effective; hence, the comparison of support and in line with other donor-supported programs) CHNPs with more direct interventions may not be valid. But reached costs of US$15 to US$20 per child per year. By this cal- conversely (or perversely) improving nutrition could actually culation, too, the resources per head, as well as the coverage, reduce the cost-effectiveness of other interventions--such as were in most cases too low for widespread effect. measles immunization--by reducing the mortality risk of chil- The estimates of coverage and intensity can be combined to dren who are not immunized. calculate the extent of current programs in relation to that needed for full coverage at adequate intensity. The results based on a 1:20 ratio of CHNW to children suggest that less than FUTURE APPLICATIONS 1 percent of the need was currently available; at 1:200 (which would cost more, because the CHNW would work full time) The experience so far in CHNPs can be applied more broadly, perhaps 10 percent of the need would be covered. Either way, a especially where community organizations can sustain support massive expansion would be called for if CHNPs were to be for CHNWs. CHNPs have worked best so far in Asia and Latin used as a means for widely improving health (but still calling America. However, with the HIV/AIDS epidemic in Sub- for only about 20 percent of the public budget for health). Saharan Africa needing high-priority attention, application of Expansion requires major resources, and not only financial CHNP experience to the HIV/AIDS crisis should be explored. ones. Thailand trained 1 percent of the population as commu- nity health workers (part time) and established an extensive supervision and support structure, including retraining. The Extending CHNPs' Coverage and Intensity estimates for the ADB-UNICEF project in financial terms were, In a project sponsored by the Asian Development Bank (ADB) for Bangladesh, Cambodia, Pakistan, Sri Lanka, and Vietnam, and UNICEF that was aimed at identifying ways of investing in some US$190 million to US$280 million per year for improve- improved child nutrition, Mason and others (1999, 2001) have ment of underweight by an additional 1.5 ppts per year (Mason reviewed the extent of CHNPs in Asian countries. Resources and others 2001, 64­68). were estimated in terms of annual expenditures per child and of ratios of population to community workers ("mobilizers"). The project addressed the needs of eight countries (Bangladesh, The Potential Role of CHNPs in Combating HIV and AIDS Cambodia, China, India, Pakistan, the Philippines, Sri Lanka, in Sub-Saharan Africa and Vietnam), and previous experience in Indonesia and Controlling the epidemic of HIV and AIDS in Sub-Saharan Thailand provided additional guidance. Africa will take an unprecedented effort. As antiretroviral ther- The population coverage of CHNPs was estimated as about apy becomes available there will be new opportunities to turn 5 to 20 percent, except for India with the ICDS, which reports the tide. Supply of antiretroviral therapy drugs, although essen- about 70 percent coverage. The next indicators refer to estimates tial and the cutting edge of new programs, is only part of the within programs. The calculated intensity was commonly 200 need. Food and income support, care for children (orphans children to 1 community worker (for example, Bangladesh, and others affected), counseling, support to promote and India, Sri Lanka); ratios of up to 100:1 were reported in Pakistan sustain behavior change, and rehabilitation of people and 1070 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others communities are needed (see chapter 18). Many of these activ- for community workers, and this issue may need some research ities have precedents in the types of CHNPs run by community and testing of different approaches. The activities of com- health workers that are discussed here. What lessons are munity workers in dealing with treatment (and prevention) of transferable? HIV and AIDS have parallels to malnutrition and would prob- One concern is that CHNPs have a greater history of success ably include the following items: in developing countries outside Africa. Those within Africa seem to have been sustained for limited periods, often linked to donor · social support and facilitating access to resources (possibly interests. Reasons may have to do with lower levels of adminis- including food aid) trative infrastructure, different existing community organiza- · counseling tion, and varying political commitment (see table 56.4). These · treatment and referral for opportunistic infections factors may now be weakened as the AIDS epidemic reduces the · promoting rehabilitation to productive life (which may numbers of qualified people and undermines community benefit from improved nutrition) as antiretroviral therapy organizations. It will be urgent to work on such contextual fac- progresses. tors to create conditions in which community organizations can be refurbished and built on. Schools too have an extremely important role in the fight Community organizations can work in Africa, as elsewhere, against HIV and AIDS and should be linked to, or part of, when they have a real function with activities perceived as use- CHNPs. Schools provide a refuge and a means of providing ful to pursue and some resources (including mobilizing their help for orphans and vulnerable children, and they also provide own) to use. Some transferable lessons are that such local a crucial opportunity for preempting and combating high-risk organizations are crucial; that in regard to supervision and behavior. access to certain resources, they need to work with the govern- ment structure--often through health system employees; and that they need sustained resource support, much of which RESEARCH NEEDS must come from donors. Treatment and rehabilitation of people with AIDS will be The question of incentives, training, and support for commu- home based in most cases and will depend substantially on nity workers urgently needs research, both from current expe- community support. Nutrition is an important component; rience and with prospective designs. The issues include the improved food intake is likely to enhance the effect of anti- following: retroviral therapy, and when treatment progresses, nutrition will help get sick people back on their feet and returned to a · What is the CHNW's status, relative to the community or to productive life. A village health worker could play a key role in the government (or NGO) hierarchy? this process. In much of Africa, HIV and AIDS affect many · How are CHNWs selected and to whom do they report (for communities, and in southern Africa, where HIV prevalences example, community health committees, supervisors reach 30 to 40 percent, almost all communities have chronically employed in the health or other system)? sick adults. This fact means that most communities need pro- · What educational background and how much training and grams: the problem is not highly concentrated. On the positive retraining--and by what methods--are needed for side, the more developed and accessible communities are those CHNWs? most affected by AIDS (Mason and others forthcoming; · What ratios of CHNWs to households are effective (or most UNICEF 2004), where establishing programs may be easier. cost-effective), both as part-time workers (volunteer or HIV and AIDS are affecting children both directly, as pediatric otherwise) and as full-time employees? AIDS, and indirectly, through the impoverishment and destitu- · What supervision ratios work? tion of affected households. This effect is seen in worsening · What remuneration and incentives are effective? child malnutrition. Here, too, support through CHNPs could · How can these efforts be financed? play a useful role. The characteristics of CHNPs elsewhere--in terms of inten- Enough programs have been in operation for long enough sity, training, supervision, and so forth--may provide some that researchers could base on them much of the needed guidance for establishing or extending them in Africa. research on processes of implementation, launching new trials Mechanisms for identifying, supporting, and training village or only for processes for which sufficient information does not community health workers in this context can draw on experi- exist. In contrast, impact evaluation requires new and prefer- ence with CHNPs; such issues as their identification in the ably prospective studies. community and links with community and facility programs A major gap in research is the application of community- will arise. A key issue will be the remuneration and incentives based programs to urban areas. Urban communities are Community Health and Nutrition Programs | 1071 conceived differently from the rural areas of most CHNPs, Cohen, D, S. Spear, R. Scribner, P. Kissinger, K. Mason, and J. Wildgen. organizations run along different lines, and so forth. Yet popu- 2000. "`Broken Windows' and the Risk of Gonorrhea." American Journal of Public Health 90 (2): 230­36. lation growth is in urban areas, and some problems, notably Delgado, H. L., V. E. Valverde, R. Martorell, and R. E. Klein. 1982. HIV and AIDS, are worse there. "Relationship of Maternal and Infant Nutrition to Infant Growth." Finally, the cost-effectiveness analysis results given in an ear- Early Human Development 6 (3): 273­86. lier section are based on rather few and approximate results. Engle, P. L., M. Bentley, and G. Pelto. 2000. "The Role of Care in Nutrition CHNPs may well provide a viable and cost-effective approach Programmes: Current Research and a Research Agenda." Proceedings of the Nutrition Society 59 (1): 25­35. under many circumstances in poor countries, and it may be necessary to demonstrate this viability better and more Ezzati, M., A. Lopez, A. Rodgers, S. vander Hoorn, C. J. L. Murray, and the Comparative Risk Collaborating Group. 2002. "Selected Major Risk quantitatively for support to CHNPs to compete with more Factors and Global and Regional Burden of Disease." Lancet 360: traditional service delivery interventions. That, too, would 1347­60. constitute worthwhile research. Ezzati, M., S. Vander Hoorn, A. Rodgers, A. Lopez, C. D. Mathers, C. J. L. Murray, and the Comparative Risk Collaborating Group. 2003. "Estimates of Global and Regional Potential Health Gains from Reducing Multiple Major Risk Factors." Lancet 362: 271­80. NOTES Farmer, P. 2003. Pathologies of Power: Health, Human Rights, and the New War on the Poor.Berkeley and Los Angeles: University of California Press. 1. Social exclusion refers to the exclusion of groups from the main- stream of public actions: lower castes in India, poorer groups in Pakistan, Fiedler, J. L. 2003. "A Cost Analysis of the Honduras Community-Based many indigenous ethnic groups throughout Asia and the Americas, and Integrated Child Care Program (Atención Integral a la Niñez- migrant workers in China and elsewhere; the result for public health is Comunitaria, AIN-C)." Health, Nutrition, and Population Discussion that excluded people do not participate in programs even if they are Paper, World Bank, Washington, DC. available. FNRI (Food and Nutrition Research Institute). 2004. "Results of FNRI 2. Pinstrup-Andersen and others (1993) provide a more complete list. National Nutrition Surveys, 2001 and 2003." Department of Science and Technology, Food and Nutrition Research Institute, Government of the Philippines, Manila. Gillespie, S., and J. Mason. 1991. "Nutrition-Relevant Actions--Some REFERENCES Experiences from the Eighties and Lessons for the Nineties." ACC/SCN State-of-the-Art Series Nutrition Policy Discussion Paper 10, United ACC/SCN (United Nations Administrative Committee on Nations Administrative Committee on Coordination/Sub-Committee Coordination/Sub-Committee on Nutrition). 1989. Update on the on Nutrition, Geneva. http://www.unsystem.org/scn/archives/npp10/ World Nutrition Situation: Recent Trends in Nutrition in 33 Countries. index.htm. Geneva: ACC/SCN. Gillespie, S., J. Mason, and R. Martorell. 1996. "How Nutrition Improves." ------. 1992. Global and Regional Results. Vol. 1 of Second Report on the ACC/SCN Nutrition Policy Discussion Paper 15, United Nations World Nutrition Situation. Geneva: ACC/SCN. Administrative Committee on Coordination/Sub-Committee on ------. 1996. Update on the Nutrition Situation, 1996. Geneva: ACC/SCN. Nutrition, Geneva. http://www.unsystem.org/scn/archives/npp15/ ------. 1998. Third Report on the World Nutrition Situation. Geneva: index.htm. ACC/SCN. Gillespie, S., M. McLachlan, and R. Shrimpton. 2003. Combatting ------. 2004. Fifth Report on the World Nutrition Situation. Geneva: Malnutrition: Time to Act. Washington, DC: World Bank. ACC/SCN. Guillermo-Tuazon, A., and R. M. Briones. 1997. "Comprehensive Administrative Staff College of India. 1997. "National Strategy to Reduce Assessment of Nutrition Interventions." Report prepared for Asian Child Malnutrition; Investment Plan; Final Report." Asian Devel- Development Bank­UNICEF Regional Technical Assistance project. opment Bank, Manila, December. University of Philippines at Los Baños, Regional Training Programme Allen, L. H., and S. R. Gillespie. 2001. "What Works? A Review of the on Food and Nutrition. UNICEF, Manila. Photocopy. Efficacy and Effectiveness of Nutrition Interventions." ACC/SCN Gwatkin, D., J. Wilcox, and J. Wray. 1980. Can Health and Nutrition Nutrition Policy Paper 19; Asian Development Bank (ADB) Nutrition Interventions Make a Difference? Monograph 13, Overseas Devel- and Development Series 5, United Nations Administrative Committee opment Council, Washington, DC. on Coordination/Sub-Committee on Nutrition, Geneva; ADB, Manila. Habicht, J.-P., J. Mason, and H. Tabatabai. 1984. "Basic Concepts for the Berg, A. 1981. Malnourished People: A Policy View. Poverty and Basic Needs Design of Evaluation during Programme Implementation." In Methods Series. Washington, DC: World Bank. for the Evaluation of the Impact of Food and Nutrition Programmes, ed. ------. 1987. Malnutrition: What Can Be Done? Washington, DC: World D. Sahn, R. Lockwood, and N. Scrimshaw, 1­25. Tokyo: United Nations Bank; Baltimore: Johns Hopkins University Press. University. BINP and UNICEF (Bangladesh Integrated Nutrition Program and Habicht, J.-P., C. G. Victora, and J. P. Vaughan. 1999. "Evaluation Designs United Nations Children's Fund). 1999. "Preliminary Results." for Adequacy, Plausibility, and Probability of Public Health UNICEF. Programme Performance and Impact." International Journal of Epidemiology 28 (1): 10­18. BRAC (Bangladesh Rural Advancement Committee). 2004. "BRAC Health Programme," http://www.brac.net/health_programme.pdf, and Heaver, R. A., and J. M. Hunt. 1995. Improving Early Child Development: "BRAC at a Glance," http://www.brac.net/aboutb.htm. An Integrated Program for the Philippines. Washington, DC: World Chowdhury, M. 2003. "Health Workforce for TB Control by DOTS: The Bank and Asian Development Bank. BRAC Case." Joint Learning Initiative Working Paper 5-2. http://www. Heaver, R., and J. B. Mason. 2000. Making a National Impact on globalhealthtrust.org/doc/JLI%20WG%20Paper%205-2.pdf. Malnutrition in the Philippines: You Can't Get There from Here--A Case 1072 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others Study of Government Policies and Programs, and the Role of UNICEF ACC/SCN State-of-the-Art Series, Nutrition Policy Discussion and the World Bank. New York: UNICEF. Paper 8. Geneva: United Nations Administrative Committee on Horwitz, A. 1987. "Comparative Public Health: Costa Rica, Cuba, and Coordination/Sub-Committee on Nutrition. Chile." Food and Nutrition Bulletin 9 (3): 19­29. Measham, A. R., and M. Chatterjee. 1999. Wasting Away: The Crisis of INFS (Institute of Nutrition and Food Science) and Department of Malnutrition in India. Washington, DC: World Bank. Economics, University of Dhaka. 1998. "Strategies for Bangladesh." Mills, A. 1994. "Decentralization and Accountability in the Health Sector Dhaka Urban Community Health Program. Asian Development Bank, from an International Perspective: What Are the Choices?" Public Manila. Administration and Development 14: 281­92. Jennings, J., S. Gillespie, J. Mason, M. Lotfi, and T. Scialfa. 1991."Managing Muñoz, C., and N. S. Scrimshaw. 1995. The Nutrition and Health Transition Successful Nutrition Programmes." ACC/SCN State-of-the-Art Series, of Democratic Costa Rica. Boston: International Nutrition Foundation Nutrition Policy Discussion Paper 8, United Nations Administrative for Developing Countries. Committee on Coordination/Sub-Committee on Nutrition, Geneva. Pelletier, D. L. 1991. "The Uses and Limitations of Information in the http://www.unsystem.org/scn/archives/npp08/index.htm. Iringa Nutrition Program, Tanzania." Cornell Food and Nutrition Jonsson, U. 1997. "Success Factors in Community-Based Nutrition- Policy Program, Working Paper 5, Cornell University, Ithaca, NY. Oriented Programmes and Projects." In Malnutrition in South Asia: Pelletier, D. L., and E. A. Frongillo. 2003. "Changes in Child Survival Are A Regional Profile, ed. S Gillespie, 161­89. ROSA Publication 5. Strongly Associated with Changes in Malnutrition in Developing Kathmandu: UNICEF, Regional Office for South Asia. Countries." Journal of Nutrition 133 (1): 107­19. Kachondam, Y., P. Winichagoon, and K. Tontisirin. 1992. "Nutrition Pinstrup-Andersen, P., S. Burger, J.-P. Habicht, and K. Peterson. 1993. and Health in Thailand: Trends and Actions." ACC/SCN case study, "Protein-Energy Malnutrition." In Disease Control Priorities in Institute of Nutrition at Mahidol University, Bangkok, and United Developing Countries, ed. D. T. Jamison, W. H. Mosley, A. R. Measham, Nations Administrative Committee on Coordination/Sub-Committee and J. L. Bobadilla, 391­420. New York: Oxford University Press. on Nutrition, Geneva. Plowman, B., J. I. Picado, M. Griffiths, K. Van Roekel, and V. Vivas de Kavishe, F. P., and S. S. Mushi. 1993. Nutrition-Relevant Actions in Alvarado. 2002. BASICS II Evaluation of the AIN Program in Honduras. Tanzania. ACC/SCN case study. Tanzania Food and Nutrition Centre, Arlington, VA: Basic Support for Institutionalizing Child Survival Monograph Series 1. Geneva: United Nations Administrative Project (BASICS II) for the U.S. Agency for International Development. Committee on Coordination/Sub-Committee on Nutrition. http:// www.unsystem.org/scn/archives/tanzania/index.htm. Reddy, V., M. Shekar, P. Rao, and S. Gillespie, S. 1992. "Nutrition in India." ACC/SCN case study. National Institute of Nutrition, Hyderabad, Kelley, L. M., and R. E. Black. 2001. "Research to Support Household and India. http://www.unsystem.org/scn/archives/india/index.htm. Community IMCI. Report of a Meeting, 22­24 January 2001, Baltimore, Maryland, USA." Journal of Health, Population, and Rohde, J. 1993. "Indonesia's Posyandus: Accomplishments and Future Nutrition 19 (Suppl. 2): S111­54. Challenges." In Reaching Health for All, ed. J. Rohde, M. Chatterjee, and D. Morley, 135­57. Oxford, U.K.: Oxford University Press. Kielmann, A. A., C. E. Taylor, C. DeSweemer, I. S. Uberoi, H. S. Takulia, N. Masih, and S. Vohra. 1978. "The Narangwal Experiment: II. Morbidity Saenz, L. 1995. "Evolution of an Epidemiological Profile." In The Nutrition and Mortality Effects." Indian Journal of Medical Research 68 (Suppl.): and Health Transition of Democratic Costa Rica, ed. C. Muñoz and N. S. 42­54. Scrimshaw, 119­43. Boston: International Nutrition Foundation for Developing Countries. Mason, J. B. 2000. "How Nutrition Improves, and What That Implies for Policy Decisions." Paper prepared for World Bank­UNICEF Nutrition Sampson, R. J., S. E. W. Raudenbusch, and F. Earls. 1997. "Neighborhoods Assessment, UNICEF, New York, and World Bank, Washington, DC. and Violent Crime: A Multilevel Study of Collective Efficacy." Science http://www.tulane.edu/~internut/publications/WB_Bckgrd_Pprs/ 277: 918­24. Narrative/NarrativethreeMason.doc. Sanders, D. 1985. With R. Carver. The Struggle for Health. London: ------. 2003. "Philippines Case Study." In Combating Malnutrition: Time Macmillan. to Act, ed. G. Gillespie, M. McLachlan, and R. Shrimpton, 85­101. ------. 1993. "The Potential and Limits of Health Sector Reform in Washington, DC: World Bank. Zimbabwe." In Reaching Health for All, ed. J. Rohde, M. Chatterjee, and Mason, J., A. Bailes, K. Mason, O. Yambi, U. Jonsson, C. Hudspeth, and D. Morley, 239­66. Oxford, U.K.: Oxford University Press. others. Forthcoming. AIDS, Drought, and Child Malnutrition in ------. 1999. "Success Factors in Community-Based Nutrition Pro- Southern Africa. Public Health Nutrition. grammes." Food and Nutrition Bulletin 20 (3): 307­14. Mason, J. B., J. Hunt, D. Parker, and U. Jonsson. 1999. "Investing in Child Sanders, D., and M. Chopra. 2004. "Child Health." Paper prepared for Save Nutrition in Asia." Asian Development Review 17 (1, 2): 1­32. the Children U.K., London. ------. 2001. "Improving Child Nutrition in Asia." Food and Nutrition Save the Children U.K. 2003. Thin on the Ground: Questioning the Evidence Bulletin 22 (3 Suppl.): 5­80. behind World Bank Funded Community Nutrition Projects in Mason, J. B., P. Musgrove, and J.-P. Habicht. 2003. "At Least One-Third of Bangladesh, Ethiopia, and Uganda. London: Save the Children U.K. Poor Countries' Disease Burden Is Due to Malnutrition." Disease http://www.savethechildren.org.uk/temp/scuk/cache/cmsattach/ Control Priorities Project Working Paper 1, Fogarty International 666_ThinOnTheGround.pdf. Center/National Institutes of Health, Washington, DC. http://www. Shekar, M. 1989. "The Tamil Nadu Integrated Nutrition Programme: A fic.nih.gov/dcpp/wps/wp1.pdf. Review of the Project with Special Emphasis on the Monitoring and Mason, J., J. Rivera, and C. Helwig, eds. 2005. "Recent Trends in Information System." Paper prepared for the Rockefeller Foundation Malnutrition in Developing Regions: Vitamin A Deficiency, Anemia, and the Food and Nutrition Policy Programme, Cornell University, Iodine Deficiency, and Child Underweight." Food and Nutrition Ithaca, NY. Bulletin 26 (1) Special Issue: 1­110. Shrimpton, R. 1989. "Community Participation in Food and Nutrition Mata, L. 1991. "National Nutrition and Holistic Care Programme Programmes: An Analysis of Recent Governmental Experiences." (NNHCP), Costa Rica." In Managing Successful Nutrition Programmes, Paper prepared for the Food and Nutrition Policy Programme, Cornell ed. J. Jennings, S. Gillespie, J. Mason, M. Lotfi, and T. Scialfa, 77­81. University, Ithaca, NY. Community Health and Nutrition Programs | 1073 Soekirman, I. Tarwotjo, I. Jus'at, G. Sumodiningrat, and F. Jalal. 1992. Vargas, W. 1995. "Development and Characteristics of Health and Economic Growth, Equity and Nutritional Improvement in Indonesia. Nutrition Services for Urban and Rural Communities of Costa Rica." ACC/SCN case study. http://www.unsystem.org/scn/archives/ In The Nutrition and Health Transition of Democratic Costa Rica, ed. C. indonesia/index.htm. Muñoz and N. S. Scrimshaw, 68­117. Boston: International Nutrition Tagwireyi, J., and T. Greiner. 1994. Nutrition in Zimbabwe. Washington, Foundation for Developing Countries. DC: World Bank. Waterlow, J. C. 1992. Protein Energy Malnutrition. London: Edward Taylor, C. E., A. A. Kielmann, and R. L. Parker. 1978. "The Narangwal Arnold. Nutrition Study: A Summary Review." American Journal of Clinical Werner, D., and D. Sanders. 1997. Questioning the Solution: The Politics Nutrition 31: 1040­52. of Primary Health Care and Child Survival. Palo Alto, CA: Tontisirin, K., and P. Winichagoon. 1999. "Community-Based Programs: HealthWrights. Success Factors for Public Nutrition Derived from Thailand's Whitehead, R. G., M. G. Rowland, and T. J. Cole. 1976. "Infection, Experience." Food and Nutrition Bulletin 20 (3): 315­22. Nutrition, and Growth in a Rural African Environment." Proceedings UNICEF (United Nations Children's Fund). 1990. "Strategy for Improved of the Nutrition Society 35 (3): 369­75. Nutrition of Children and Women in Developing Countries." Policy Winichagoon, P., Y. Kachondam, G. Attig, and K. Tontisirin. 1992. Review Paper E/ICEF/1990/1.6, UNICEF, New York. Integrating Food and Nutrition into Development: Thailand's ------. 1998. State of the World's Children: Progress against Worms for Experiences and Future Visions. UNICEF/EAPRO and Institute of Pennies. New York: UNICEF. Nutrition at Mahidol University, Bangkok, Thailand. ------. 2004. Drought, AIDS, and Child Malnutrition in Southern Africa: WHO (World Health Organization). 1997. Improving Child Health. Preliminary Analysis of Nutritional Data on the Humanitarian Crisis. IMCI: The Integrated Approach. WHO/CHD/97.12 Rev 2. Geneva: Nairobi: UNICEF, Eastern and Southern Africa Regional Office. WHO. 1074 | Disease Control Priorities in Developing Countries | John B. Mason, David Sanders, Philip Musgrove, and others Chapter 57 Contraception Ruth Levine, Ana Langer, Nancy Birdsall, Gaverick Matheny, Merrick Wright, and Angela Bayer The use of modern contraception to prevent pregnancies is a NATURE, CAUSES, AND BURDEN OF THE unique health intervention because, in many ways, it is not a CONDITIONS THAT CAN BE ADDRESSED health intervention at all. In general, couples in sexual relation- ships use contraception because, at the time the decision is Three rationales--each one involving controversy and touch- made, one or both members do not wish to conceive a child, ing on deeply held political and cultural perspectives--have rather than because they wish to become healthier or to pre- underlain policy and programmatic interest in contraception, vent a risk to health. Governments also may have an interest in since the 1960s. promoting particular patterns of childbearing to meet social and economic objectives. This is most often the case when · The demographic rationale, typically framed around lower- rapid population growth is seen as a drag on economic growth; ing the rate of population growth to achieve broad in contrast, however, governments in low-fertility countries economic, social, and environmental aims, was most promi- may promote pro-natalist policies to increase the labor supply. nently applied in the 1960s and 1970s. Thus, the effectiveness of contraception has to be seen, first and · The fertility rationale, which emerged in the 1980s, promoted foremost, in terms of the effectiveness in permitting couples to lower fertility under the assumption that smaller families achieve their desired family size. are better off in terms of economic and health conditions. Although contraception is rarely used to improve health, it · The human rights rationale, which surfaced in 1994 at the does have health consequences. On the negative side, conse- International Conference on Population and Development quences may include the potential health risks of hormonal in Cairo, served as a major force in the 1990s to focus atten- contraception or surgery. On the positive side, health may tion on women's rights to autonomy and empowerment in benefit from fewer pregnancies, lower exposure to sexually childbearing and on female and male reproductive health. transmitted infections (STIs), and protection against ovarian The Cairo Programme of Action stressed the political and cancer through the use of some types of contraception. Some of cultural dimensions of contraception, including gender the consequences affect the users, some affect their sexual part- issues. ners, and some affect their children. Contraceptives affect a user's sexuality by changing menstrual patterns and, therefore, particularly in some cultures, sexual activity. Also, by eliminat- Population Growth and Fertility ing fear of unwanted pregnancy, contraceptives may enhance In part because of the demand for and availability of modern the quality of sexual experience. Finally, condoms may decrease contraception, the worldwide rate of annual population sexual pleasure for men; true or not, this explanation is one of growth has declined from just over 2 percent in the late 1960s the most commonly cited to account for why some men (or to 1.5 percent during 1980­2001. It is projected to decrease couples) do not use condoms. to 1 percent during 2001­15. Although the growth rate has 1075 slowed, population growth in absolute terms is unprecedented. America and the Caribbean and Central Asia have unmet need World population increased from 2.5 billion people in 1950 to for contraception. Whereas women in the other regions of the 6.3 billion in 2003 and is expected to rise to 7.1 billion by 2015 world have an equally distributed unmet need for spacing and (UN 2003; World Bank 2003). limiting births, the majority of unmet need in Sub-Saharan Fertility rates in developing countries have declined rapidly Africa is for spacing (Ross and Winfrey 2002). Unmet need is in the past 50 years, from more than 6.0 children per woman in highest in countries where growing numbers of women want the 1950s to about 2.8 children per woman today. Fertility rates to avoid pregnancy but contraceptive prevalence is low. So, for remain high, however, in the 49 least developed countries, example, among developing countries for which data are avail- which had an average total fertility rate of 5.46 children per able from USAID's Demographic and Health Surveys, unmet woman during 1995­2000 (UN 2003). need is currently highest in Haiti, where it nears 40 percent of Fertility levels and trends vary greatly between regions. all married women; it is more than 30 percent in Cambodia, Fertility rates are lowest in low- and middle-income countries Nepal, Pakistan, Rwanda, Senegal, Togo, Uganda, and the in East Asia and the Pacific, at 2.1 children per woman. Republic of Yemen; and it is lowest, at less than 7 percent, in Countries in Central Asia and Latin America and the Brazil, Colombia, and Vietnam. Caribbean also have relatively low total fertility rates, at 2.5 and 2.6 children per woman, respectively. The Middle East and Total Potential Demand for Contraception. A rough measure North Africa and South Asia follow, with average total fertility of the total potential use of modern contraception in a country rates of 3.4 and 3.3 children per woman in 2001, respectively. can be estimated by combining the measure of unmet need Fertility rates are highest in Sub-Saharan Africa, at 5.2 children with the current proportion of women using contraception. per woman (World Bank 2003).1 Regional averages conceal Brazil, Colombia, and Vietnam all have demand for contracep- substantial variation among and within countries. tion greater than 80 percent of currently married women. They Because of a legacy of high fertility and corresponding have satisfied most of this demand, with contraceptive preva- young population structures, population momentum ensures lence rates above 75 percent, resulting in both low fertility rates that many developing countries will continue to grow at a rel- and low unmet need (Westoff 2001). In contrast, in most Sub- atively high rate for many decades, even as fertility rates con- Saharan African countries, the unmet need percentage exceeds tinue to decline. Population momentum alone will account for the percentage of women currently using contraception (see almost three-fourths of the population growth in developing table 57.1). countries in the next quarter-century. The largest growth at present is in Asia and Sub-Saharan Africa. Total population increase in these regions is now twice what it was in 1950. By Health Consequences 2015, population growth is expected to be substantially lower Excess fertility is responsible for between 12 and 30 percent of in all regions except Sub-Saharan Africa. the maternal burden of disease (see table 57.2)2, although this is clearly an underestimate of the present and future burden of disease that can be prevented through investments in family Demand for Contraception planning. These estimates include only the direct health bene- If a woman wishes to postpone or avoid childbearing and is not fits of family planning for women by preventing unwanted using contraception (including use by her partners), she is said births, decreasing the number of abortions, and increasing the to have an unmet need for contraception. The most com- length of birth intervals. Because of data limitations, these esti- monly reported reasons for unmet need are lack of knowledge, mates exclude the potential effect of family-planning programs health concerns, and social disapproval (Casterline and Sinding on children's long-term nutritional status and education; 2000). women's status and the household economy; and public In 2003, an estimated 122.7 million women in developing savings from reduced fertility, AIDS, and other STIs through countries had an unmet need for contraception, including condom promotion and prevention of mother-to-child trans- 105.2 million married women, 8.4 million unmarried women, mission (PMTCT). They also exclude the effect of such pro- and 9.1 million women of all marital status in the states of the grams on environmentally related disease owing to population former Soviet Union. This figure represents 17 percent of all growth. Last, the estimates do not consider the disabling effects married women, a 2 percent decline from the late 1990s that is of unwanted pregnancies, despite the effect those pregnancies due to increasing contraceptive use. or their termination under unsafe conditions can have on Overall, the highest unmet need is in Sub-Saharan Africa, women's welfare and productivity. where 19.4 percent of all women have unmet need. About Each year, 585,000 women die and more than 54 million 13 percent of women in Asia, 10.6 percent of women in North women suffer from diseases or complications caused by preg- Africa and the Middle East, and 8.5 percent of women in Latin nancy and childbirth (WHO 1997). 1076 | Disease Control Priorities in Developing Countries | Ruth Levine, Ana Langer, Nancy Birdsall, and others Table 57.1 Total Potential Demand for Family Planning (percentage of currently married women) Current Total Current Total Unmet contraceptive potential Unmet contraceptive potential Country need use demand Country need use demand East Asia and the Pacific Sub-Saharan Africa Cambodia 32.6 23.8 56.4 Benin 25.7 16.4 42.1 Indonesia 9.2 57.4 66.6 Burkina Faso 25.8 11.9 37.7 Philippines 18.8 47.8 66.6 Cameroon 19.7 19.3 39.0 Vietnam 6.9 75.3 82.2 Central African Rep. 16.2 14.8 31.0 Central Asian republics Chad 9.7 4.1 13.8 Kazakhstan 8.7 66.1 74.8 Côte d'Ivoire 27.7 15.0 42.7 Kyrgyz Republic 11.6 59.5 71.1 Eritrea 27.5 8.0 35.5 Turkmenistan 10.1 61.8 71.9 Ethiopia 23.1 5.9 29.0 Uzbekistan 13.7 55.6 69.3 Gabon 28.0 32.7 60.7 Latin America and the Caribbean Ghana 23.0 22.0 45.0 Argentina -- -- -- Guinea 24.2 6.2 30.4 Belize -- -- -- Kenya 23.9 39.0 62.9 Bolivia 26.1 48.3 74.4 Madagascar 25.6 19.4 45.0 Brazil 7.3 76.7 84.0 Malawi 29.7 30.6 60.3 Colombia 6.2 76.9 83.1 Mali 25.7 6.7 32.4 Dominican Republic 11.8 69.5 81.3 Mozambique 6.7 7.3 14.0 Guatemala 23.1 38.2 61.3 Namibia 21.9 28.9 50.8 Haiti 39.8 28.1 67.9 Niger 16.6 8.2 24.8 Nicaragua 14.7 60.3 75.0 Nigeria 17.5 15.3 32.8 Paraguay 15.0 48.4 63.4 Rwanda 35.6 13.2 48.8 Peru 10.2 68.9 79.1 Senegal 34.8 12.9 47.7 Middle East and North Africa South Africa 15.0 56.3 71.3 Algeria -- -- -- Tanzania 21.8 25.4 47.2 Egypt, Arab Rep. of 11.2 56.1 67.3 Togo 32.3 23.5 55.8 Jordan 14.2 52.6 66.8 Uganda 34.6 22.8 57.4 Morocco 16.1 50.3 66.4 Zambia 26.5 25.9 52.4 Yemen, Rep. of 38.6 20.8 59.4 Zimbabwe 12.9 53.5 66.4 South Asia Bangladesh 15.3 53.8 69.1 India 15.8 48.2 64.0 Nepal 31.4 28.5 59.9 Pakistan 31.8 11.8 43.6 Source: Demographic and Health Survey, various years. -- not available. Risks Associated with Unwanted Pregnancies. Unwanted Ambivalence toward pregnancy also is associated with less early pregnancies expose women to additional health risks by and continuous prenatal care and lower use of professional increasing the number of lifetime pregnancies and deliveries. delivery care (Gage 1998; Joyce and Grossman 1990; Weller, Because the lifetime risk of maternal mortality is a function of Eberstein, and Bailey 1987). the number of pregnancies and the quality and utilization of Many women who have unintended pregnancies turn to health care, reducing the number of pregnancies can lower induced abortion, both in countries where abortion is legal maternal mortality rates (Koenig and others 1988). and safe and in those where it is illegal and too often unsafe. Contraception | 1077 Table 57.2 Maternal Burden of Disease Associated with Unwanted Fertility and Unsafe Abortions Disability- Percentage Years lived adjusted of all Years of with life years maternal Region Deaths life lost disability (DALYs) DALYs East Asia and the Pacific 3,637 107,795 380,255 420,030 17 Latin America and the Caribbean 6,323 190,544 298,390 429,399 30 Middle East and North Africa 8,428 244,461 256,742 395,368 12 South Asia 30,074 878,028 1,079,661 1,669,727 20 Sub-Saharan Africa 43,077 1,270,598 764,012 1,821,721 18 Sources: Adapted from WHO 2002a. Maternal morality related to unsafe abortions from WHO 1998. Percentage of infant disease preventable by family planning from Rutstein 2003. Notes: Because of data limitations, estimates are not available for Eastern Europe and Central Asia. Burden-of-disease estimates include a 3 percent discount rate without age weighting. WHO regions are not identical to the World Bank Regions used here; however, a very close approximation was made by excluding WHO regions AMRO A, WPRO A, and EURO A. These estimates assume that 90 percent of abortion-related disease burden is preventable by family planning and that the percentage of other preventable maternal disease is equal to the percentage of all births that are unwanted. Accurate measures of abortion are difficult to obtain in most Legalizing abortion, improving the quality of abortion care, parts of the world. In countries where abortion is illegal, data and increasing access to safe abortion can have profound are lacking or incomplete, and even where it is legal, abortions impacts on the health consequences of abortion. When abor- may be underreported because of societal attitudes (Bongaarts tion was illegal in Romania in 1988, complications from unsafe 1997; Henshaw, Singh, and Haas 1999). It is estimated that abortion caused 86 percent of maternal deaths. After abortion about one-fourth of the 210 million pregnancies each year end was legalized in 1989, the frequency of abortion persisted in abortion. because of contraceptive supply shortages, but the number of In 1995, approximately 35.5 million abortions were per- maternal deaths fell by 50 percent (Hord and others 1991). formed in developing countries. The large majority of legal abortions, 10.6 million, occurred in China. Most of the remain- Risks Associated with Pregnancy and Birth. All pregnancies ing legal abortions took place in other parts of Asia (5.7 mil- and births involve some health risks to women, so prevent- lion) and in the Caribbean (0.2 million). Developing countries, ing any pregnancy reduces women's health risks. Higher mor- which experienced an estimated 19 million illegal abortions in tality and morbidity of women, infants, and children are 1995, account for 95 percent of illegal abortions worldwide. positively associated with the risk factors of giving birth when Nearly 10 million illegal abortions occurred in Asia, followed a woman is too young or too old, the births are too close by an estimated 5 million in Africa and 4 million in Latin together, there are too many births, or a woman has a preexist- America. These figures mark a particularly large increase for ing health condition. Births in most of these groups--women Africa, which was estimated to have only 1.5 million illegal who are older (over age 35), births that are spaced too closely abortions in 1987 (Henshaw, Singh, and Haas 1999). (24 months or less after the preceding birth), and births that Unsafe abortion, typically associated with illegality, has are higher order (fifth or higher)--are also more likely to be large impacts on both maternal mortality and maternal mor- reported as unintended, making their prevention doubly bidity. Each year, unsafe abortion is believed to account for important (Tsui, Wasserheit, and Haaga 1997). 80,000 maternal deaths, or 13 percent of the burden of disease An estimated 15 million women under age 20 give birth in women of reproductive age (WHO 2002a). Deaths related each year. Women age 15 to 19 are twice as likely to die from to unsafe abortions are estimated at 100 to 600 death per childbearing as women in their 20s, and women under age 17 100,000 abortions, compared with the mortality rate from are at even greater risk (Starrs 1997). Adolescent mothers are legal abortions of 0.6 deaths per 100,000 abortions (Salter, more likely to suffer from obstetric complications if they lack Johnston, and Hengen 1997). Survivors of unsafe abortions physical maturity or are malnourished (Aitken and Walls also experience consequences; unsafe abortion causes disabili- 1986). They are also less likely to receive adequate prenatal or ty in an additional 5 million women (WHO 2002a). Treatment obstetric care, which may cause negative outcomes for them of complications from unsafe abortions constitutes a large and for their infants (McDevitt and others 1996). In many con- proportion of emergency gynecological hospital admissions texts, negative social consequences are profound, including loss (Konje, Obisesan, and Ladipo 1992) and requires substantial of school and employment opportunities. resources (Kinoti and others 1995; Salter, Johnston, and Children born to adolescent mothers face a number of risks. Hengen 1997). Research has demonstrated that infants of teenage mothers are 1078 | Disease Control Priorities in Developing Countries | Ruth Levine, Ana Langer, Nancy Birdsall, and others more likely than those born to mothers in their 20s to die assumptions and have varied widely. The World Bank (1993) before they reach their first birthday (Hobcraft 1992; McDevitt estimated that family-planning programs could prevent and others 1996; Ross and Frankenberg 1993; Starrs 1997; between 20 and 40 percent of all infant deaths by preventing Sullivan, Rutstein, and Bicego 1994). Children of mothers mistimed and underspaced births. In a study of 25 countries, under age 20 may have a 20 to 30 percent higher risk of death Hobcraft (1994) estimated that if all birth intervals of less than than children of older mothers (Hobcraft 1992; Sullivan, two years were prevented, child mortality levels would be Rutstein, and Bicego 1994). Infants of teenage mothers are also reduced by an average of 17 percent and up to one-third in at higher risk of being of low birthweight, small for gestational several countries. Rutstein (2003) found that birth spacing age, or premature. Finally, adolescent women are less likely to of three to five years alone could prevent up to 46 percent of provide adequate care for their infants and children, because infant mortality. Muhuri and Menken (1997) found that in they often lack the maturity, education, and resources to do so rural Bangladesh poor spacing and timing accounted for (Geronimus 1987; Govindasamy and others 1993). 25 percent of child mortality. Trussell and Pebley (1984) esti- Women over age 35 face an increased risk of maternal mor- mated that spacing could decrease infant mortality by 10 per- tality. Mothers age 40 to 44, for example, are five times more cent and child (age 1 to 5) mortality by 21 percent. Another likely to die during pregnancy or childbirth than mothers in study found that, even in Latin America, which has lower child their 20s (Royston and Lopez 1987). Mothers in their late 30s mortality rates, spacing could reduce perinatal mortality by and 40s may also face additional negative consequences, 14 percent (Conde-Agudelo and Belizan 2000). because they may have preexisting health problems owing to In a study of 19 African countries, Rafalimanana and age or previous births. Westoff (2001) found that median actual birth intervals in As with children of adolescent mothers, children of women every country were shorter than the preferred birth intervals over age 35 also suffer negative consequences. Children born to reported by women, reflecting the substantial unmet need for mothers over age 40 are more likely than those born to younger birth spacing. Achieving preferred intervals would decrease mothers to die before age 5 (Bicego and Ahad 1996; Sullivan, neonatal mortality by only 6 percent on average, and infant Rutstein, and Bicego 1994). Older women are also more likely mortality by a comparable amount, whereas removing all short to have stillbirths or to bear children with congenital abnor- intervals would decrease both by 13 percent. malities who may not survive childhood (Cnattingius and Women giving birth for the fourth or higher time are at others 1992; Fretts and Usher 1997). much higher risk of maternal complications and death. Longer birth intervals reduce women's risks of death and ill Independent of a woman's age, her risk of dying when giving health during pregnancy and childbirth. One study assessed the birth for the fourth or higher time is 1.5 to 3 times greater than effects of birth spacing in 450,000 women on the basis of hos- during a second or third birth (Winikoff and Sullivan 1987). pital records from 1985 to 1997 in 19 Latin American and Women who have had at least three births are also more likely Caribbean countries. The study found that women who have to experience hemorrhage, uterine rupture or prolapse, or kid- their babies 27 to 32 months after a preceding birth are more ney disease (Maymon and others 1998). than two times more likely to survive pregnancy and childbirth Children born to mothers who have had many births face than women who have short intervals of 9 to 14 months. Birth risks similar to those of children born to older mothers; they are intervals of 27 to 32 months are also associated with lower inci- often the same women. Children born to mothers who have had dence of third trimester bleeding, premature rupture of mem- three or more births are more likely than those born to younger branes, anemia, and other negative outcomes (Conde-Agudelo mothers (those under age 20) to die before age 5 (Bicego and and Belizan 2000). Ahad 1996; Sullivan, Rutstein, and Bicego 1994). Women who Recent research suggests that birth intervals of three to five have had many births are also more likely to have stillbirths or to years provide even greater benefits than the two-year intervals bear children with congenital abnormalities who may not sur- that were previously promoted. One study assessed the effect of vive childhood (Cnattingius and others 1992; Fretts and Usher this longer birth interval in more than 430,000 pregnancies in 1997). Children from larger families often receive lower levels of 18 countries and found that children who are born three to five education and health care than children from smaller families years after the preceding birth are more likely to survive from because of competition for finite family resources (Blake 1981). the perinatal period through age five. Children born at intervals Women with preexisting health conditions often face greater of three to five years are also 1.2 to 1.4 times less likely to be risks in pregnancy and childbirth. Pregnancy can aggravate malnourished or underweight or to experience stunting than conditions such as high blood pressure, heart disease, malaria, those born at intervals shorter than two years (Rutstein 2003). anemia, tuberculosis, hepatitis, and STIs, including HIV. Putting together a range of patchy data on the effect of fam- Indirect causes, including these preexisting conditions, ac- ily planning on child mortality can yield estimates of the total count for an estimated 20 percent of maternal deaths each year global impact, but those estimates are highly dependent on (WHO 1997). Contraception | 1079 INTERVENTIONS Failure rates for all copper-bearing IUDs are usually less than 1 per 100 women in the first year of use. The "intervention" of contraception can be considered as the method itself and as the means by which family-planning Temporary Methods. By far the most popular temporary con- clients obtain services (including counseling) and contracep- traceptive method is the oral contraceptive, commonly known tive commodities. Both the methods and the types of services as "the pill," which has a failure rate typical use of less than are diverse. 10 percent over a year. Among 67 developing countries for which survey data are available (not including China or India), about 50 percent of married women who have ever used contra- Contraceptive Methods ception have used the pill at some point. The pill has been most Contraceptive methods can be classified as permanent and popular in Latin America; there, about 55 percent of all married long term--primarily for those women and couples who have women have used the pill. In the Near East and North Africa, completed childbearing--or temporary--primarily for those about one-third of married women have used the pill, and in women and couples who wish to delay pregnancy. Sub-Saharan Africa, about 15 percent have used it at some time (Johns Hopkins Population Information Program 2000). Permanent and Long-Term Methods. Female sterilization, or More than 20 million women use systemic contraceptives tubal ligation, used by about 187 million women worldwide containing only progestins. These contraceptives include (WHO 2002a), is the most popular and effective contraceptive subdermal implants such as Norplant, injectable products, available. The most effective types of female sterilization have a IUDs, and vaginal rings. These products have high rates of con- 10-year cumulative pregnancy rate of 7.5 per 1,000 procedures traceptive efficacy (0.3 to 1.0 percent failure rate over 12 (Peterson and others 1996). months). Their long duration of action allows for a relatively Sterilization accounts for one-third of all contraceptive infrequent dose. Their main drawbacks are their tendencies to practice. Because sterilization is considered a permanent form cause highly irregular endometrial bleeding and amenorrhea. of contraception, some women may regret their decision dur- Although altered bleeding does not have any ill health effects, it ing ensuing years. Some dissatisfaction with sterilization is does pose a problem for women in societies that bar or restrict expected and is always observed among sterilized populations; women from certain social and religious activities during men- in most cases, the proportion of women regretting sterilization struation. The World Health Organization estimates that falls below 10 percent. Regret is higher when sterilization was a between 10 and 30 percent of women abandon their progestin- woman's first and only contraceptive method, when a woman only methods for this reason (WHO 2002a). was sterilized at or under age 30, or when a woman has fewer Barrier methods, although less effective than hormonal than four living children (Loaiza 1995). methods, IUDs, or sterilization, can offer effective contracep- Between 40 million and 50 million men worldwide have tion when used consistently and correctly. Barrier methods, undergone a vasectomy, a figure representing 8 percent of the and particularly condoms, are the only type of contraception world's couples of reproductive age. This method comes in that offers additional protection against STIs. fourth in contraceptive popularity, after female sterilization When used correctly during every act of sexual intercourse, (19 percent), the intrauterine device (IUD; 13 percent), and the male latex condom is effective against both unwanted preg- oral contraceptives ("the pill"; 8 percent), and right ahead of nancy and HIV infection and other STIs. Typical use results in the male condom (4 percent; WHO 2002a). The method is as pregnancy rates of 3 to 14 percent per year. If a condom breaks effective as female sterilization (failure rate of less than 1 per- or tears during intercourse, emergency contraception can be cent) and much simpler and safer than tubal ligation. used to reduce the risk of pregnancy. The IUD is now used by 150 million women worldwide, or The female condom, made of soft, pliable polyurethane and about 13 percent of the world's women of reproductive age, prelubricated with a silicone-based substance, is inserted into because of its efficacy, safety, and convenience. After female the vagina before sexual intercourse. The female condom is sterilization, it is the most popular method of contraception. slightly less effective than the male condom, with a failure rate The 5-year life span of the IUD means fewer visits to health of 5 to 21 percent. Unlike the male condom, the female condom providers and less expenditure of money, time, and effort. can be inserted up to eight hours before intercourse. The IUDs prevent pregnancy through several mechanisms: they female condom adds to the arsenal of weapons in the fight alter sperm migrations, inhibit fertilization, and generate a against STIs; offers women more control in sexual negotiations; foreign-body reaction in the endometrium. Progestogen- can be used in conjunction with the IUD, hormonal methods, releasing IUDs cause changes in the amount and viscosity of and sterilization; and has no special storage requirements. cervical fluid, altering sperm penetration. In a small percentage The diaphragm, although not a popular method in develop- of women, ovulation is inhibited in the first two years of use. ing countries, is being studied as a means of preventing not 1080 | Disease Control Priorities in Developing Countries | Ruth Levine, Ana Langer, Nancy Birdsall, and others only pregnancy but also bacterial STIs. Results from those health services, particularly other types of reproductive health randomized trials are pending. services. Under this arrangement, family-planning services are delivered in conjunction with routine primary care--a recog- Emergency Contraception. Since the mid 1960s, the use of nition that women's health needs are in no way confined to certain oral contraceptives has been shown to be effective in contraception and that a broad range of reproductive and other preventing pregnancy. Two hormonal regimens have proved to health services must be provided. be both safe and effective for emergency contraception: com- bined oral contraceptives and progestogen-only pills. Both Social Marketing can be taken for up to 120 hours after unprotected intercourse. Social marketing refers to a variety of strategies using tradi- Emergency contraception represents a second chance to prevent tional commercial-marketing techniques to promote socially an unwanted pregnancy after unprotected sex, and it is particu- beneficial behaviors, products, and services. In family planning, larly responsive to the needs of youths and of women who have social marketing has focused on making supplies of methods been coerced into intercourse. of contraception widely available in existing commercial retail Despite the demonstrated safety and efficacy of emergency outlets and on promoting these contraceptives to consumers contraception,its acceptance by providers and the public,and its through mass media. In recent years, social-marketing pro- inclusion on the WHO's essential drug list, emergency contra- grams have expanded their focus to behavior change and the ception is not widely available in many developing countries delivery of clinical services through social franchises. (Langer and others 1999). In countries with low contraceptive prevalence, social- marketing programs for contraceptives typically import ORGANIZATION OF FAMILY-PLANNING donated contraceptives and then repackage and promote them PROGRAMS with their own sales force. In higher-prevalence countries, pro- grams may partner with commercial manufacturers to market Because of the variety of users and methods of contraception, existing brands, often subsidizing promotion in exchange for a range of quite distinct ways to deliver needed goods and price guarantees. In the Dominican Republic, for instance, the services has been developed and tried over the past several oral contraceptive manufacturer Schering agreed to cut the decades. Some programs are dedicated to providing only family- price of its pills in half in exchange for advertising by a social- planning services, often referred to as vertical programs; some marketing organization. provide a range of reproductive and other health services, Most social-marketing programs for contraceptives distrib- loosely termed integrated programs; and some try to reach ute commodities through existing commercial outlets, such as current and potential clients through social marketing and shops, pharmacies, and tobacco shops. Retailers make a small community-based distribution methods. profit, which increases their incentive to stock and promote the Vertical versus Integrated Service Delivery products. Because such programs do not spend money build- ing their own distribution network, they typically enjoy low The original large-scale family-planning programs in develop- costs per output. ing countries, many of which were launched on the basis of the Social-marketing programs have achieved dramatic contra- demographic rationale, tended to be organized around a verti- ceptive sales throughout the developing world, providing cal structure with central management and logistics. Family- contraception to about 10 million couples in 60 countries, in planning workers based at fixed sites--whether run by addition to promoting a variety of other health products. government agencies or by nongovernmental organizations However, evidence from cross-national studies suggests that (NGOs)--were dedicated exclusively to providing informa- social marketing's effect on contraceptive use is driven less by tion, services, and contraceptive commodities related to family its brand sales--because users often switch brands--and more planning. Funding, often from external donor agencies, was by its promotional activity (Bulatao 2002). earmarked for family-planning activities. In many instances, the supervisory, budgeting, training, and logistics systems were all separate from those of other health services. Some of the INTERVENTION COST AND COST-EFFECTIVENESS largest vertical programs, such as the India program, promoted sterilization on a large scale, sometimes provided through rudi- Synthesizing data on family-planning costs, outputs, and out- mentary health facilities. Others, such as the Bangladesh pro- comes to inform resource allocation has distinct challenges. gram, relied on the provision of information and commodities First, nearly all existing studies have estimated average--not through female outreach workers going house to house. marginal--costs. The dearth of marginal cost data, general to In the past two decades, increasing emphasis has been most of the cost-effectiveness studies presented in this volume, placed on integrating family-planning programs into other severely weakens the ability to recommend interventions. Contraception | 1081 Second, cost data based on units of output that cannot be developing countries. Across regions, costs per CYP were com- linked to health outcomes have limited use. Most studies on parable in Asia, Latin America, and the Middle East, at US$4.00 effectiveness have not included costs, and most studies on costs to US$5.00. Costs were considerably greater in Africa, where have not included effectiveness. In the family-planning litera- the average cost per CYP was US$14.00. ture, cost per output (such as the cost per couple-year of pro- The least expensive mode of service delivery was clinical tection, or CYP) has often been called cost-effectiveness. Here, provision of sterilization, with a weighted average cost of we reserve the term cost-effectiveness for measures of cost per US$2.34 per CYP. The second most economical mode was unit of health impact (Gift, Haddix, and Corso 2003). Those social marketing of contraceptives, although those costs are studies that have measured cost-effectiveness typically come highly dependent on setting. In African countries, where such from mature programs, where both marginal costs and mar- programs were less developed, the costs were high--up to ginal effects are likely to be low. US$19.00 per CYP. The weighted average cost of social market- Third, costs vary considerably according to the accounting ing of contraceptives was US$3.00. method used, the program setting and scale, the level of latent Costs of community-based distribution programs ranged demand among couples targeted, the method mix, the quality, from US$4.85 to US$35.37 per CYP, with a weighted average of and the existing supply and service infrastructure (Dayaratna US$12.55. Costs of clinic-based services, excluding steriliza- and others 2000; Janowitz and Bratt 1992). Clinical costs are tion, ranged from US$4.44 to US$16.65 per CYP, with a particularly sensitive to setting, given the broad differences in weighted average of US$7.93. Clinic-based services supple- local salaries. mented with community-based distribution were the most Finally, most evaluations of family planning estimate effects expensive mode of service delivery, ranging from US$4.44 to within a single generation. However, averting a birth also averts US$19.38, with a weighted average of US$18.21. As noted grandchildren, great-grandchildren, and so on. For economic above, these estimates have typically not included costs to users cost-benefit analyses, the multigenerational effects of a single (see table 57.3). birth averted can be significant. However, this consideration can be addressed through the use of an appropriate discount Marginal Costs rate. Knowles and Wagman (1991) found that unit costs decline as the number of contraceptive users increases. Larger programs Cost per Output enjoy an economy of scale in procurement; average costs The most common measure of output in family-planning pro- grams is the CYP. Stover and others (1996) estimated the num- Table 57.3 Weighted Average Program Costs per CYP, ber of contraceptives of various methods needed to provide by Region and Mode of Delivery (2001 U.S. dollars) 1 CYP, given typical use: 15 cycles of pills, 120 condoms, 120 foaming tablets, or 4 three-month or 12 one-month injecta- Region Mode of service delivery Cost per CYP bles. An IUD provides 3.5 CYPs; male and female sterilization, Africa Social marketing 15.95 11 CYPs; and an individual trained in natural methods, 2 CYPs. Community-based distribution 20.32 IUDs and voluntary sterilization have the lowest cost per Clinics 16.65 CYP, although they have a high up-front cost. Oral contracep- tives are usually the least costly methods. Condoms and injecta- Clinics with community-based distribution 8.02 bles are more expensive than the pill, and implants are often the Asia Social marketing 2.59 most expensive method per CYP. Mauldin and Miller (1994) Community-based distribution 6.50 estimated that, based on the method mix in developing coun- Clinics 5.07 tries, the weighted average commodity cost of contraception Clinics with community-based distribution 19.38 was US$1.55 per CYP.3 Latin America Social marketing 1.64 Commodity costs are relatively constant across programs, Community-based distribution 35.37 although there is some variation owing to bulk procurement. Clinics 6.40 When program costs are added, costs vary considerably by the Clinics with community-based distribution 6.47 program setting and mode of delivery. A review of programs in Middle East Social marketing 3.82 Africa, for instance, found method costs per CYP could vary by Community-based distribution 4.85 an order of magnitude, with the greatest variation in the cost of Clinics 4.44 condoms. Clinics with community-based distribution 9.03 Barberis and Harvey (1997) found considerable variation in costs per CYP by region and mode of service delivery in 14 Source: Adapted from Barberis and Harvey 1997. 1082 | Disease Control Priorities in Developing Countries | Ruth Levine, Ana Langer, Nancy Birdsall, and others decrease as fixed costs from training and from information, acteristics such as fecundity, coital frequency, use effectiveness, education, and communication (IEC) programs are distrib- or frequency of contraceptive use--one couple using condoms uted over more units; labor may be more efficiently used at all the time has a greater effect on fertility than two couples higher volumes; and larger programs may be further ahead on using condoms half the time. the learning curve. However, as marginal costs diminish with Most critically, even if CYP were an appropriate measure of size, so may marginal returns in mature programs (Haaga and protection, CYPs do not account for method substitution. Many Tsui 1995). clients who obtain contraceptives from a family-planning pro- High rates of underutilization suggest marginal costs may be gram, at some program expense, were buying or would other- close to zero for many clinics. Knowles and Wagman (1991) and wise have bought contraceptives from commercial providers or Janowitz and others (1996) found underutilization of clinical would have relied on natural methods. Community-based dis- capacity in Morocco and Bangladesh. In Morocco, more than tribution programs, for instance, may be more expensive per a third of all labor costs were spent waiting for patients. Foreit CYP than social-marketing programs, but if such programs and others (1992) found that administrative costs accounted reach more nonusers than social-marketing programs, they may for 16 to 65 percent of fixed costs for a sample of clinics, whereas be more cost effective in preventing unwanted pregnancies. capacity utilization ranged from only 32 to 76 percent. Similarly, sterilization is the cheapest method per CYP, but in One study of a community-based distribution project in many countries, it has a limited impact on fertility. Bangladesh found that marginal costs represented 20 to 40 per- Three examples demonstrate the hazard of confusing out- cent of average costs and decreased with scale (Attanayake, put, or even intermediate outcomes, with impact. Jensen (1996) Fauveau, and Chakraborty 1993). In social-marketing pro- found that public providers in Indonesia, while more expensive grams, marginal costs are close to commodity costs less per CYP than private providers, had a greater effect on fertility revenue (Bulatao 1993). A weighted average of current com- per unit output. Public programs may have reached more modity costs based on the existing method mix in developing nonusers. Janowitz and others (1992) found that a Honduran countries suggests a marginal cost of about US$1.55 per CYP social-marketing program distributed more than 40,000 CYPs (Mauldin and Miller 1994). but had no net effect on contraceptive prevalence. Users substi- tuted one contraceptive brand or method for another. Bertrand and others (1986) found that a community-based distribution Cost-Effectiveness Outside of Programs program in then Zaire increased modern contraceptive preva- The estimates of cost-effectiveness included here show only the lence without affecting total contraceptive prevalence. Modern direct health benefits of family planning for women and chil- methods were substituted for traditional ones, such as pro- dren from increasing birth intervals and reducing teenage preg- longed lactation, periodic abstinence, or withdrawal. (Such nancies. As noted earlier, these estimates ignore important ben- substitution can be an improvement when the traditional efits of family planning for the health of families. Because of method replaced is rhythm or withdrawal, but there is no more data limitations, most of these omissions cannot be corrected protective contraception than postpartum abstinence.) here. Some are a general consequence of using the global Cost-effectiveness analysis thus requires true estimates of burden-of-disease framework, which takes disease, rather than project effect--the difference between what happens in a pro- health interventions, as the starting point. This orientation ject's presence and what happens in that project's absence. A makes results less useful for the purpose of setting priorities in number of studies have estimated effect, measuring births health care, particularly for interventions, such as family plan- averted, total or unwanted pregnancies averted, unsafe abor- ning, that affect a number of sequelae (Williams 1999). tions averted, maternal and child deaths averted, and measures Despite the abundance of cost-per-output data in family of health utility, such as the disability-adjusted life year (DALY). planning, these data cannot be used to set priorities in health As table 57.4 shows, costs vary significantly within regions. funding because the protection offered by a unit of contracep- In some regions, maximum and minimum costs differ by two tion is related to the behavior of a particular user. Several stud- orders of magnitude. Our sensitivity analysis found that the ies have sought to estimate a program's impact on fertility as a cost-effectiveness of programs was most sensitive (as a percent- simple division of CYPs distributed in a population by the age of swing) to the existing level of unmet need for birth mean birth interval in the population (Cakir, Fabricant, and spacing and limiting. Kircalioglu 1996). However, such adjustments are not substi- tutes for actual measures of impact. Costs per Births Averted. In a review, Pritchett (1994) found A considerable literature has developed around the prob- that costs per birth averted ranged from US$37 in Jamaica, lems of using CYP as a measure of protection (for reviews, see the Philippines, and Thailand to US$415 in Nepal, with the Fort 1996 and Shelton 1991). The typical calculation of CYPs median value of 12 studies being US$82. Cochrane (1988) does not account for use-failure rates, wastage, and client char- estimated US$78 per averted birth in a cross-national analysis. Contraception | 1083 Table 57.4 Average Costs per Benefit of Family Planning (2001 U.S. dollars) Disability- Years Years Infant Maternal adjusted of life lived Births deaths deaths life years lost with Region averted averted averted saved averted disability East Asia and the Pacific 163 4,907 12,880 60 110 103 Latin America and the Caribbean 87 2,316 34,564 53 66 187 Middle East and North Africa 97 1,989 18,917 49 55 209 South Asia 113 1,577 5,172 30 37 98 Sub-Saharan Africa 131 1,367 10,231 34 37 194 Source: Authors, based on a model by AGI and others (2000). Note: The model used country-level data for 68 developing countries. Output costs were based on Population Action International estimates in 1994 of the public sector cost per user. Estimates are not available for Eastern Europe and Central Asia. Experimental studies in Bangladesh estimated the cost per Costs per DALY. The 1993 report on world development averted birth at US$95 (Attanayake, Fauveau, and Chakraborty (World Bank 1993) estimated that family-planning programs 1993), US$281 (Pritchett 1994), US$293 (Balk and others in low-income countries cost from US$40 to US$60 per DALY. 1988), and US$296 (Simmons, Balk, and Faiz 1991), varying by In the first edition of Disease Control Priorities in Develop- the accounting method used. The Bangladesh experiment was ing Countries, Jamison (1993) estimated the costs of IEC or likely more expensive than other programs, as it involved a fre- behavior-change communication (BCC) programs promoting quent schedule of visitation for community-based distribution. condom use for family planning at between US$20 and The model we developed for the Disease Control Priorities US$100 per DALY, depending on child and maternal mortality Project produced costs between US$87 and US$163 per birth rates. averted, with costs lowest in Latin America and the Caribbean Average costs for regions per YLL ranged from US$37 to and in the Middle East and North Africa. US$110. Costs per DALY and year of life lost (YLL) were lowest in South Asia and Sub-Saharan Africa, whereas costs per year Costs per Death Averted. Walsh and others (1993) estimated lived with disability (YLD) were lowest in South Asia and in that in a typical high-mortality country with a 20 percent con- East Asia and the Pacific. As with costs per death averted, costs traceptive prevalence rate, family-planning programs cost per DALY varied within regions by as much as two orders of US$8,261 per maternal death averted and US$1,276 per peri- magnitude. natal infant death averted. In a setting with high mortality and Family-planning programs that target HIV-positive women low contraceptive prevalence rates, offering family planning to prevent mother-to-child transmission may be even more alone was considerably more cost-effective in averting both cost-effective. According to an analysis of Stover's (2003) maternal and infant deaths than offering an integrated pro- estimates, such programs cost about US$5 per DALY. In com- gram including prenatal care and birth attendant training. The parison, traditional PMTCT services, including antiretrovirals model developed for the Disease Control Priorities Project pro- and replacement feeding, cost US$37 per DALY, and nevirapine duced regional average costs between US$5,172 and US$34,564 regimens cost US$5 to US$12 per DALY (Marseille and others per maternal death averted and between US$1,367 and 1999; UNAIDS 1999). Kumar (2000) found that adding family- US$4,907 per infant death averted. Costs were lowest in South planning and abortion services to PMTCT programs increased Asia and Sub-Saharan Africa. Costs within regions varied by as their cost-effectiveness from US$124 per DALY for a short- much as two orders of magnitude. course antiretroviral drug regimen to US$93 per DALY for an Stover (2003) estimated that by offering family-planning integrated strategy. Despite its cost-effectiveness, family plan- services at PMTCT and voluntary counseling and testing sites in ning is not currently included in most policies addressing countries with high HIV prevalence, child HIV infections could mother-to-child transmission of HIV. be averted at a cost of US$489, and child deaths could be averted at a cost of US$278 per event--well below the costs of averting these events using traditional PMTCT services. In addition, Cost-Effectiveness within Programs these family programs would avert orphans at a cost of US$278, Few studies compare the cost-effectiveness of program designs and maternal deaths at a cost of US$1,824 per event. and elements--for instance, between social marketing and 1084 | Disease Control Priorities in Developing Countries | Ruth Levine, Ana Langer, Nancy Birdsall, and others community-based distribution, or between price subsidies and increases the likelihood of contraceptive use (Dayaratna and promotional spending. Thus, although donors or government others 2000; Westoff and Bankole 1997). Social-marketing ministers may have enough information to justify funding programs appear to succeed in increasing contraceptive use family-planning programs, managers of such programs have principally through their promotional activities (Bulatao much less information to use in deciding how best to spend 2002). IEC and BCC activities can also increase the efficiency of money within their programs. clinic programs by increasing caseloads. Last, among all family- planning activities, IEC and BCC activities may be the most Program Inputs. Roughly half of program funds are spent prone to market failure, providing a strong rationale for public subsidizing price (Sanderson and Tan 1996), but there is little investment. evidence that price subsidies significantly affect contraceptive use, even among the poor (Matheny 2004). In every demo- Integration. Integrating family planning with other reproduc- graphic and health survey over the past decade, price has been tive health services was a major goal of the 1994 International reported as a barrier to contraceptive use by fewer than 2 per- Conference on Population and Development. Although the cent of women with unmet need. Several studies suggest that logic of integration is compelling, little research has been done even the poorest urban users are typically willing to pay com- on the costs of integrated programs. In theory, integrated pro- mercial contraceptive prices--once they intend to contracept. grams should be more efficient because they distribute joint This finding is not entirely surprising, because the economic costs across more services. In clinics, integration should costs of childbearing dwarf those of contraception (Pritchett increase cost-effectiveness if services such as Papanicolaou 1994). Particularly if profits are reinvested in a program's qual- smears, ultrasounds, pregnancy testing, abortions, and post- ity of services, higher prices may permit the provision of abortion care are used to cross-subsidize less profitable contra- amenities that attract more clients (Foreit and Levine 1993; ceptive services. Mitchell, Littlefield, and Gutter (1999) found Litvack and Bodart 1993). relatively low family-planning costs in two integrated clinics The most commonly reported barriers to contraceptive use and concluded that adding clients has a larger effect on costs in are lack of knowledge, health concerns, and social disapproval a clinical setting than adding services does. Mancini, Stecklov, (Casterline and Sinding 2000). It is likely, then, that the most and Stewart (2003), however, found that vertical programs that cost-effective inputs are those that address these barriers--by focused exclusively on family planning offered contraceptive increasing accurate knowledge about and social acceptance of supplies and services at a significantly lower cost per CYP than contraception, as well as by improving the quality of and access integrated programs. to a variety of contraceptive methods. Delivering contraception services alongside abortion or Evidence suggests that quality has improved contraceptive postabortion care can be cost-effective. Few clinics and hospi- use for birth spacing in Tunisia (Cochrane and Guilkey 1995). tals that treat women suffering from abortion complications However, results are mixed in Peru (Mensch, Arends-Kuenning, offer contraceptive counseling and services as part of their and Jain 1996) and the Philippines (DeGraff, Bilsborrow, and postabortion care, yet these services are effective in increasing Herrin 1993). Schultz (1994) found that, on average, improved contraceptive use. The key obstacle is finding support from quality does not affect the contraceptive prevalence rate at the abortion providers--especially private providers, who may not national level. see profit potential in providing family planning. The same is Some improvements in quality require additional true of prenatal care. One study in Kenya showed that exposing resources--expanding facilities, adding equipment and staff, women to information about family planning during prenatal and diversifying services. Quality improvements are necessary care doubled the likelihood that they would use contraception to respond to ethical imperatives. Beyond the ethical ration- (Ndhlovu 1997). A study of 27 countries found that two-thirds ale, quality improvements can reduce the cost per unit of of women had an unmet need for contraception within one output--improving the flow of clients and shifting service year of their last childbirth (Ross and Winfrey 2001). In fact, delivery from physicians to paramedics. Improvements in qual- about one-third of all unmet need was among women post- ity can also increase utilization and allow higher prices to be partum. This need could be satisfied with greater integration of charged. maternal and child health services and family planning, partic- Unfortunately, the relative cost-effectiveness of investments ularly during prenatal visits, delivery care, and six-week post- in these inputs has not been established. However, there is rea- partum visits. son to believe that, among all inputs, investments in IEC and Finally, integrating STI services with family planning can BCC activities are the most cost-effective--especially those that help identify women with STIs who wish to avoid risky preg- encourage users to purchase contraceptives or services from the nancies and prevent fetal and neonatal complications. private sector. Studies in many countries show that exposure to Integrating STI services may also draw more men to family- IEC and BCC messages through television, radio, and print planning clinics. Contraception | 1085 EQUITY: DISTRIBUTING THE BENEFITS contraception can produce economic benefits by reducing OF SUBSIDIES maternal mortality and morbidity and improving child health--benefits that can be captured through estimates of It is the poorest whose health and prosperity are most threat- savings to the health system. ened by unwanted pregnancies, and who are least able to pay Second, contraception can have economic payoffs if it oc- for family-planning services. Yet a significant portion of public curs in a setting where high fertility is constraining economic subsidies for family planning benefit the wealthy. This outcome growth. Longstanding arguments in development economics is in part a natural consequence of the demand for family plan- have centered on whether rapid population growth acts as a ning. Wealthier couples typically want smaller families than drag on economic growth. Some camps argued that high fertil- poorer couples. If subsidies follow demand rather than need, ity, in particular, would condemn a country to slow (or even they will concentrate among the wealthy. However, because negative) per capita income growth. Others argued that the there is substantial unmet need for family planning among negative effects would be mild and short lived, as households poor couples, an efficient segmentation of the market would adapt to existing resource constraints by reducing childbearing have subsidized providers target poorer clients and unsubsi- and as technological changes, such as those that result in higher dized providers target wealthier clients. agricultural yields, alter the productivity equation. By the late In a study of 10 countries, Sine (2002) found that nonpoor 1990s, however, a consensus emerged from examination of users obtained 45 percent of subsidized oral contraceptives and more than 40 years of experience and data: rapid population 56 percent of subsidized condoms supplied by governments growth, in general, and high fertility, in particular, typically and nonprofit organizations. An analysis in Indonesia found lead to slower economic growth and higher levels of poverty that only 16 percent of public subsidies for family planning than would otherwise be realized (Birdsall and Sinding 2003). accrued to the poorest quartile. In the Philippines, 64 percent Third, when contraception reduces the rate of population of subsidies accrued to the richest half of households, whereas growth, it can have multiple effects (economic and other) on the 17 percent accrued to the poorest quarter. In the Philippines, environment. High fertility, coupled with rural-to-urban the distribution of subsidies varied widely by contraceptive migration, has led to rapid urban growth, often outpacing the method, with vasectomy subsidies concentrated among the provision of clean water and sanitation. In an analysis of 42 cities poor and condom subsidies concentrated among the middle in Latin America, Asia, and Africa, Brockerhoff and Brennan class (Haaga and Tsui 1995). (1998) found that urban growth rates were positively correlated Behrman and Knowles (1998) found that the family- with infant mortality rates, likely owing to crowding, a weaken- planning program in Vietnam was only weakly pro-poor. The ing of the public infrastructure, and increased air pollution. main source of inefficiency stemmed from the greater use of Ambient air pollution is now emerging as a serious threat to more heavily subsidized government providers, particularly human health in virtually all the large cities in the world, respon- hospitals, by wealthier clients. One study in Bangladesh found sible for 1 percent of global deaths and DALYs (WHO 2002b).4 that, although family-planning workers had the largest effect Population growth is also responsible for much of the on the contraceptive behavior of uneducated women, the increase in carbon emissions that contribute to global climate workers--who typically come from the educated class--were change. Although developing countries account for only more likely to visit educated women (Arends-Kuenning 1997). 20 percent of carbon emissions today, some researchers have For similar reasons, clinics are often located in wealthier areas. estimated that these countries will account for 50 percent by Whereas a significant share of wealthier users obtain subsi- 2050, given current rates of development (Bongaarts 1992). dized goods, a large share of poor users obtain goods from the Bongaarts estimated that population growth would account for commercial sector. In a study of 12 countries, more than half of 48 percent of the growth in carbon emissions in developing even the poorest condom users purchased condoms from the countries between 1985 and 2100. Birdsall (1994) estimated commercial sector in all countries but the Philippines (Foreit that realistic reductions in fertility could reduce emissions in 1999). Sine (2002), in turn, found that 20 percent of poor and 2050 by as much as 15 percent in developing countries; more near-poor users obtained their oral contraceptives from the important, however, she estimated that by reducing fertility, commercial sector, even in countries where free or subsidized family-planning programs could reduce emissions more cost- brands were available. effectively than taxes on carbon emissions could. Similarly, Brinkley, Potts, and Walsh (2003) found that family-planning ECONOMIC BENEFITS OF INTERVENTION programs are likely more cost-effective than any emissions pol- icy. (At the same time, it is important to note that the largest Contraception or family planning can yield long-term eco- per capita consumption of nonrenewable natural resources nomic benefits through three main routes. One is reasonably occurs in the low-fertility settings of wealthy countries, where direct and easy to quantify; the others are more complex. First, policies promoting conservation are overdue.) 1086 | Disease Control Priorities in Developing Countries | Ruth Levine, Ana Langer, Nancy Birdsall, and others RESEARCH AND DEVELOPMENT AGENDA particular program to the state of the world without that program. However, effectiveness analyses using disabilty-adjusted life years (DALYs) The agenda for research and development in the field of con- are typically not conducted this way. The number of DALYs averted by family planning is not the number of DALYs without a family-planning traceptives and family planning is ample. It benefits from a program minus the total number of DALYs with a family-planning pro- strong tradition of data- and research-driven policy and pro- gram. Given the negative nature of DALYs as a measure, such an exercise grammatic decisions. Priorities suggested below fall into the would lead to an absurd result. Nearly all lives have some number of DALYs attached to them. These DALYs may be averted most cost- realms of science, operations (or program), and policy. effectively by preventing all births, reducing the burden of disease to zero. On the scientific agenda, the top priority must be the devel- To avoid this result, we cannot measure the effect of family planning by opment of contraceptive products that protect women and comparison with a pure counterfactual. Rather, it is measured by ignoring men against both pregnancy and HIV infection in ways that are the DALYs that would have been attributed to contingent persons. By the definition of DALYs, no individual can have any expectation of healthy life more acceptable from a user perspective than male or female until the moment of live birth. This, however, is not an ideal solution. Both condoms. Two other high-priority topics are the development parents and health practitioners typically want to avoid perinatal deaths, of male contraceptives and of microbicides, a female-controlled even if doing so means preventing the birth of an infant. This fact suggests that one aim of health care is to prevent the ill health of contingent method that protects against both STIs (especially, HIV) and persons--in which case we ought, at least in some circumstances, to meas- unwanted pregnancy. ure the potential DALYs averted by preventing a birth. With respect to operations research, high-priority questions 3. All dollar amounts are expressed in 2001 U.S. dollars. In a few cases, the year was not specified in the original source, and currency was con- include how to reach adolescents effectively and cost-effectively; verted from the year of publication. what to do in settings where progress has stalled (for example, 4. Because of data limitations, these DALYs include no morbidity. Bangladesh); and how to stimulate demand for family planning in Sub-Saharan African countries, where the demand is rela- tively low. REFERENCES Another important and neglected topic is how to introduce the sexual health dimension into contraceptive services. In gen- AGI (Alan Guttmacher Institute), Futures Group International, Population Action International, Population Reference Bureau, and Population eral, providers do not take into account sexuality and sexual Council. 2000. The Potential Impact of Increased Family Planning health during counseling or service delivery, in spite of the Funding on the Lives of Women and Their Families.Washington,DC:AGI. mutual influence between sexual behavior and preferences and Aitken, I. W., and B. Walls. 1986. "Maternal Height and Cephalopelvic contraception. On the one hand, women and couples' sexual Disproportion in Sierra Leone." Tropical Doctor 16 (3): 132­4. activities strongly influence the adoption of contraception, and Arends-Kuenning, M. 1997. "The Equity and Efficiency of Doorstep the preference of these women and couples for a certain Delivery of Contraceptives in Bangladesh." Policy Research Division Working Paper 101, Population Council, New York. method, in turn, influences its effective use and continuity. On Attanayake, N., V. Fauveau, and J. Chakraborty. 1993. "Cost-Effectiveness the other hand, particular characteristics of the different meth- of the Matlab MCH-FP Project in Bangladesh." Health Policy and ods affect women's and couples' perceptions of their own sexu- Planning 8 (4): 327­38. ality (Bruce 1987). In spite of these obvious links, neither the Balk, D., K. K. Faiz, U. Rob, J. Chakraborty, and G. Simmons. 1988. "An family-planning literature nor family-planning services have yet Analysis of Costs and Cost-Effectiveness of the Family Planning Health Services Project in Matlab, Bangladesh." International Centre for paid enough attention to the matter (Dixon-Mueller 1993). Diarrhoeal Disease Research, Dhaka. On the policy research agenda, further research is required to Barberis, M., and P. D. Harvey. 1997. "Costs of Family Planning understand the relationship between fertility and other repro- Programmes in Fourteen Developing Countries by Method of Service ductive health outcomes and economic outcomes at microeco- Delivery." Journal of Biosocial Science 29 (2): 219­33. nomic and macroeconomic levels. Such research includes, for Behrman, J. R., and J. C. Knowles. 1998. "Population and Reproductive Health: An Economic Framework for Policy Evaluation." Population example, a careful analysis of how access to contraception may and Development Review 24 (4): 697­737. affect women's productivity, as well as human capital invest- Bertrand, J. T., N. Mangani, M. Mansilu, M. McBride, and J. Tharp. 1986. ments in the next generation. In addition, better understanding "Strategies for Family Planning Service Delivery in Bas Zaire. of the distributional effect of family planning and other repro- International Family Planning Perspectives 12 (4): 108­15. ductive health services would inform policies about how best to Bicego, G., and O. B. Ahad. 1996. Infant and Child Mortality. Demographic target public resources. and Health Surveys Comparative Studies 20. Calverton, MD: Macro International. Birdsall, N. 1994. "Another Look at Population and Global Warming." In Population, Environment, and Development, ed. United Nations, 39­54. NOTES New York: United Nations. 1. These are 2001 averages; country-by-country figures are 2000 and Birdsall, N., and S. Sinding, eds. 2003. Population Matters: Demographic earlier. Change, Economic Growth, and Poverty in the Developing World. New 2. One problem in applying the burden-of-disease framework to fam- York: Oxford University Press. ily planning is unique to interventions that affect the size of a population. Blake, J. 1981. "Family Size and the Quality of Children." Demography In most effectiveness analyses, one compares the state of the world with a 18 (4): 421­42. Contraception | 1087 Bongaarts, J. 1992. "Do Reproductive Intentions Matter?" International Gage, A. 1998."Premarital Childbearing, Unwanted Fertility and Maternity Family Planning Perspectives 18 (3): 102­8. Care in Kenya and Namibia." Population Studies 52 (1): 21­34. ------. 1997. "Trends in Unwanted Childbearing in the Developing Geronimus, A. T. 1987. "On Teenage Childbearing and Neonatal Mortality World." Studies in Family Planning 28 (4): 267­77. in the United States." Population and Development Review 13 (2): 245­79. Brinkley, G., M. Potts, and J. Walsh. 2003."Reducing Global CO2 Emissions through Family Planning Tradable Permits." School of Public Health, Gift, T. L., A. C. Haddix, and P. S. Corso. 2003. "Cost-Effectiveness University of California, Berkeley. Analysis." In Prevention Effectiveness: A Guide to Decision Analysis and Economic Evaluation, ed. A. C. Haddix, S. M. Teutsch, and P. S. Corso, Brockerhoff, M., and E. Brennan. 1998. "The Poverty of Cities in New York: Oxford University Press. Developing Regions." Population and Development Review 24 (1): 75­114. Govindasamy, P., M. K. Stewart, S. O. Rutstein, J. T. Boerma, and A. E. Sommerfelt. 1993. High-Risk Births and Maternity Care. Columbia, Bruce, J. 1987. "Users' Perspectives on Contraceptive Technology and MD: Macro International. Delivery Systems: Highlighting Some Feminist Issues." Technology in Society 9: 359­83. Haaga,J.G.,andA.O.Tsui,eds.1995.Resource Allocation for Family Planning in Developing Countries. Washington, DC: National Academy Press. Bula tao, R. 1993. Effective Family Planning Programe. Washington, DC: World Bank. Henshaw, S. K., S. Singh, and T. Haas. 1999. "The Incidence of Abortion Worldwide." International Family Planning Perspectives 25 (Suppl.): Cakir, H. V., S. J. Fabricant, and F. N. Kircalioglu. 1996."Comparative Costs S30­38. of Family Planning Services and Hospital-Based Maternity Care in Turkey." Studies in Family Planning 27 (5): 269­76. Hobcraft, J. 1992. "Fertility Patterns and Child Survival: A Comparative Analysis." Population Bulletin of the United Nations 33: 1­31. Casterline, J. B., and S. W. Sinding. 2000. "Unmet Need for Family Planning in Developing Countries and Implications for Population ------. 1994. The Health Rationale for Family Planning: Timing of Births Policy." Population and Development Review 26 (4): 691­723. and Child Survival. New York: United Nations. Cnattingius, S., M. R. Forman, H. W. Berendes, and L. Isotalo. 1992. Hord, C., H. P. David, F. Donnay, and A. Koblinsky. 1991. "Reproductive "Delayed Childbearing and Risk of Adverse Perinatal Outcome. A Health in Romania: Reversing the Ceausescu Legacy." Studies in Family Population-Based Study." Journal of the American Medical Association Planning 22 (4): 231­40. 268 (7): 886­90. Jamison, D. T. 1993. "Disease Control Priorities in Developing Countries." Cochrane, S. H. 1988. "The Effects of Education, Health, and Social In Disease Control Priorities in Developing Countries, ed. D. T. Jamison, Security on Fertility in Developing Countries." Policy, Planning, and W. H. Mosley, A. R. Measham, and J. L. Bobadilla, 1­34. New York: Research Working Paper 93, World Bank, Washington, DC. Oxford University Press. Cochrane, S. H., and D. K. Guilkey. 1995. "The Effects of Fertility Janowitz, B., and J. Bratt. 1992. "Costs of Family Planning Services." Intentions and Access to Services on Contraceptive Use in Tunisia." International Family Planning Perspectives 18 (4): 137­44. Economic Development and Cultural Change 43: 779­804. Janowitz, B., K. Jamil, J. Chowdhury, B. Rahman, and D. Hubacher. 1996. Conde-Agudelo, A., and J. M. Belizan. 2000. "Maternal Morbidity and "Productivity and Costs for Family Planning Service Delivery in Mortality Associated with Interpregnancy Interval: Cross Sectional Bangladesh." Technical Report, Family Health International, Research Study." British Medical Journal (Clinical Research Ed.) 321 (7271): Triangle Park, NC. 1255­59. Janowitz, B., M. Suazo, D. B. Fried, J. Bratt, and P. Bailey. 1992. "Impact of Dayaratna, V., W. Winfrey, K. Hardee, J. Smith, E. Mumford, W. Social Marketing on Contraceptive Prevalence and Cost in Honduras." McGreevey, and others. 2000. "Reproductive Health Interventions: Studies in Family Planning 23 (2): 110­7. Which Ones Work and What Do They Cost?" Policy Project Occasional Jensen, E. R. 1996. "The Fertility Impact of Alternative Family Planning Paper 5, Futures Group International, Washington, DC. Distribution Channels in Indonesia." Demography 33 (2): 153­65. DeGraff, D., R. Bilsborrow, and A. Herrin. 1993."The Implications of High Johns Hopkins Population Information Program. 2000. "Oral Fertility for Children's Time Use in the Philippines." In Fertility, Family Contraceptives." Population Reports Series A-9, Johns Hopkins Size, and Structure: Consequences for Families and Children, ed. C. University, Baltimore. Lloyd, 297­329. New York: Population Council. Joyce, T. J., and M. Grossman. 1990. "Pregnancy Wantedness and the Dixon-Mueller, R. 1993. "The Sexuality Connection in Reproductive Initiation of Prenatal Care." Demography 27: 1­17. Health." Studies in Family Planning 23 (5): 330­35. Kinoti, S. N., L. Gaffikin, J. Benson, and L. A. Nicholson. 1995. Monograph Foreit, J. R., M. R. Garate, A. Brazzoduro, F. Guillen, M. C. Herrera, and on Complications of Unsafe Abortion in Africa. Arusha, Tanzania: F. C. Suarez. 1992. "A Comparison of the Performance of Male and Commonwealth Regional Health Community Secretariat. Female CBD Distributors in Peru." Studies in Family Planning 23 (1): Knowles, J. C., and A. Wagman. 1991. The Relationship between Family 58­62. Planning Costs and Contraceptive Prevalence: Will FP Costs per User Foreit, K. G. 1999. "Use of Commercial and Government Sources of Decline over Time? Chapel Hill, NC: Futures Group International. Family Planning and Maternal and Child Health Care." Policy Project Koenig, M. A., V. Fauveau, A. I. Chowdhury, J. Chakraborty, and M. A. Working Paper Series 4, Futures Group International, Washington, Khan. 1988. "Maternal Mortality in Matlab, Bangladesh, 1976­1985." DC. Studies in Family Planning 19 (2): 69­80. Foreit, K. G., and R. E. Levine. 1993. "Cost Recovery and User Fees in Konje, J. C., K. A. Obisesan, and O. A. Ladipo. 1992."Health and Economic Family Planning." Options for Population Policy, Policy Paper Series 5, Consequences of Septic Induced Abortion." International Journal of Futures Group International, Washington, DC. Gynecology and Obstetrics 37 (3): 193­7. Fort. A. L. 1996. "More Evils of CYP." Studies in Family Planning 27 (4): Kumar, M. 2000. "Cost-Effectiveness of Prevention of Mother-to-Child 228­31. HIV Transmission in Kerala, India." Paper presented at the Inter- Fretts, R. C., and R. H. Usher. 1997. "Causes of Fetal Death in Women of national AIDS Economic Network Symposium, Durban, South Africa, Advanced Maternal Age." Obstetrics and Gynecology 89 (1): 40­45. July 7­8. 1088 | Disease Control Priorities in Developing Countries | Ruth Levine, Ana Langer, Nancy Birdsall, and others Langer, A., C. Harper, C. Garcia-Barrios, R. Schiavon, A. Heimburger, B. Royston, E., and A. D. Lopez. 1987. "On the Assessment of Maternal Elul, and others. 1999. "Emergency Contraception in Mexico City: Mortality." World Health Statistics Quarterly 40 (3): 214­24. What Do Health Care Providers and Potential Users Know and Think Rutstein, S. O. 2003. Effect of Birth Intervals on Mortality and Health. about It?" Contraception 60: 233­41. Calverton, MD: Measure/DHS+/Macro International. Litvack, J. L., and C. Bodart. 1993. "User Fees plus Quality Equals Salter, C., H. B. Johnston, and N. Hengen. 1997. "Care for Postabortion Improved Access to Health Care: Results of a Field Experiment in Complications: Saving Women's Lives." Population Reports Series L-10, Cameroon." Social Science and Medicine 37 (3): 369­83. Johns Hopkins University, Baltimore. Loaiza, E. 1995. "Sterilization Regret in the Dominican Republic: Sanderson, E. C., and J. Tan. 1996. Population in Asia. Washington DC: Looking for Quality of Care Issues." Studies in Family Planning 26 (1): World Bank. 39­48. Schultz, T. P. 1994. Human Capital, Family Planning, and Their Effects on Mancini, D. J., G. Stecklov, and J. F. Stewart. 2003. "The Effect of Population Growth. New Haven, CT: Yale University. Structural Characteristics on Family Planning Program Performance Shelton, J. D. 1991. "What's Wrong with CYP?" Studies in Family Planning in Côte d'Ivoire and Nigeria." Social Science and Medicine 56 (10): 22 (5): 332­35. 2123­37. Simmons, G. B., D. Balk, and K. K. Faiz. 1991. "Cost Effectiveness Analysis Marseille, E., J. G. Kahn, F. Mmiro, L. Guay, P. Musoke, M. G. Fowler, and of Family Planning Programs in Rural Bangladesh: Evidence from J. B. Jackson. 1999. "Cost Effectiveness of Single-Dose Nevirapine Matlab." Studies in Family Planning 22 (2): 83­101. Regimen for Mothers and Babies to Decrease Vertical HIV-1 Transmission in Sub-Saharan Africa." Lancet 354: 803­9. Sine, J. 2002. How Much Is Enough? Estimating Requirements for Subsidized Contraceptives. Washington, DC: Commercial Market Strategies. Matheny, G. 2004. "Family Planning Programs: Getting the Most for the Money." International Family Planning Perspectives 30 (3): 134­8. Starrs, A. 1997. The Safe Motherhood Action Agenda: Priorities for the Next Decade. New York: Family Care International. Mauldin, W. P., and V. C. Miller. 1994. Contraceptive Use and Commodity Costs in Developing Countries, 1994­2005. New York: United Nations Stover, J. 2003. "Costs and Benefits of Providing Family Planning Services Population Fund. at PMTCT and VCT Sites." Futures Group International, Washington, DC. Maymon, E., F. Ghezzi, I. Shoham-Vardi, R. Hershkowitz, M. Franchi, M. Katz, and M. Mazor. 1998. "Peripartum Complications in Grand Stover, J., J. Bertrand, S. Smith, N. Rutenberg, and K. Meyer-Ramirez. 1996. Multiparous Women." European Journal of Obstetrics, Gynecology, and "Empirically Based Conversion Factors for Calculating Couple-Years Reproductive Biology 81 (1): 21­25. of Protection." Futures Group, Chapel Hill, NC. McDevitt, T. M., A. Adlakha, T. B. Fowler, and V. Harris-Bourne. 1996. Sullivan, J. M., S. O. Rutstein, and G. T. Bicego. 1994. Infant and Child Trends in Adolescent Fertility and Contraceptive Use in the Developing Mortality. Demographic and Health Surveys Comparative Studies 15. World. Washington, DC: Bureau of the Census. Calverton, MD: Macro International. Mensch, B., M. Arends-Kuenning, and A. Jain. 1996. "The Impact of the Trussell, J., and A. Pebley. 1984. "The Potential Impact of Changes in Quality of Family Planning Services on Contraceptive Use in Peru." Fertility on Infant, Child, and Maternal Mortality." Studies in Family Studies in Family Planning 27 (N2): 59­75. Planning 15 (6): 267­80. Mitchell, M. D., J. Littlefield, and A. Gutter. 1999."Costing of Reproductive Tsui, A. O., J. N. Wasserheit, and J. G. Haaga, eds. 1997. Reproductive Health Health Services." International Family Planning Perspectives 25 in Developing Countries: Expanding Dimensions, Building Solutions. (Suppl.): S17­21. Washington, DC: Panel on Reproductive Health, Committee on Population, and Commission on Behavioral and Social Sciences and Muhuri, P. K., and J. Menken. 1997. "Adverse Effects of Next Birth, Gender, Education, National Research Council. and Family Composition on Child Survival in Rural Bangladesh." Population Studies 51 (3): 279­94. UN (United Nations). 2003. State of the World's Population. New York: UN. Ndhlovu, L. 1997. "The Route from MCH to Family Planning: Why Do UNAIDS (Joint United Nations Programme on HIV/AIDS). 1999. Women Switch Health Facilities in the Kenyan MCH/FP Program?" Prevention of HIV Transmission from Mother to Child: Strategic Options. Paper presented at the 125th annual meeting of the American Public Geneva: UNAIDS. Health Association, Indianapolis. Walsh, J. A., C. N. Feifer, A. R. Measham, and P. J. Gertler. 1993. "Maternal Peterson, H. B., Z. Xia, J. M. Hughes, L. S. Wilcox, L. R. Tylor, and and Perinatal Health." In Disease Control Priorities in Developing J. Trussell. 1996. "The Risk of Pregnancy after Tubal Sterilization: Countries, ed. D. T. Jamison, W. H. Mosley, A. R. Measham, and J. L. Findings from the U.S. Collaborative Review of Sterilization." Bobadilla, 363­90. New York: Oxford University Press. American Journal of Obstetric Gynecology 174 (4): 1161­68. Weller, R. D., I. W. Eberstein, and M. Bailey. 1987. "Pregnancy Wantedness Pritchett, L. H. 1994. "Desired Fertility and the Impact of Population and Maternal Behavior during Pregnancy." Demography 24: 407­12. Policies." Population and Development Review 20 (1): 1­56. Westoff, C. F. 2001. Unmet Need at the End of the Century. Demographic Rafalimanana, H., and C. F. Westoff. 2001. Gap between Preferred and and Health Surveys Comparative Reports 1. Calverton, MD: ORC Actual Birth Intervals in Sub-Saharan Africa: Implications for Fertility Macro International. and Child Health. Demographic and Health Surveys Analytical Westoff, C. F., and A. Bankole. 1995. Unmet Need 1990­1994. Studies 2. Calverton, MD: Macro International. Demographic and Health Surveys Comparative Studies 16. Calverton, Ross, J. A., and E. Frankenberg. 1993. Findings from Two Decades of Family MD: ORC Macro. Planning Research. New York: Population Council. ------. 1997. Mass Media and Reproductive Health. Demographic and Ross, J. A., and W. L. Winfrey. 2001. "Contraceptive Use, Intention to Use, Health Surveys Analytical Reports 2. Calverton, MD: Macro and Unmet Need during the Extended Postpartum Period." Inter- International. national Family Planning Perspectives 27 (1): 20­27. WHO (World Health Organization). 1997. "Maternal Health around the ------. 2002. "Unmet Need for Contraception in the Developing World World." Wall Chart. Department of Reproductive Health and Research, and the Former Soviet Union: An Updated Estimate." International Geneva. Family Planning Perspectives 28 (3): 138­43. ------. 1998. Abortion in the Developing World. Geneva: WHO. Contraception | 1089 ------. 2002a. Research on Reproductive Health at WHO: Biennial Report World Bank. 1993. Investing in Health: World Development Report 1993. 2000­2001. Special Programme of Research, Development, and New York: Oxford University Press. Research Training in Human Reproduction. Geneva: UNDP/UNFPA/ ------. 2003. World Development Indicators 2003. New York: Oxford WHO/World Bank. University Press. ------. 2002b. World Health Report. Geneva: WHO. Yun, S. H., D. L. Kincaid, Y. Yaser, and G. Ozler. 1990."The National Family Williams, A. 1999. "Calculating the Global Burden of Disease: Time for a Planning IEC Campaign of Turkey." Population Communication Strategic Reappraisal?" Health Economics 8 (1): 1­8. Services, Johns Hopkins University, Baltimore. Winikoff, B., and M. Sullivan. 1987."Assessing the Role of Family Planning in Reducing Maternal Mortality." Studies in Family Planning 18 (3): 128­43. 1090 | Disease Control Priorities in Developing Countries | Ruth Levine, Ana Langer, Nancy Birdsall, and others Chapter 58 School-Based Health and Nutrition Programs Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, Kathleen Beegle, Amaya Gillespie, Lesley Drake, Seung-hee Frances Lee, Anna-Maria Hoffman, Jack Jones, Arlene Mitchell, Delia Barcelona, Balla Camara, Chuck Golmar, Lorenzo Savioli, Malick Sembene, Tsutomu Takeuchi, and Cream Wright The paradigmatic shift in the past decade in our understanding · Second, operations research shows that the preexisting of the role of health and nutrition in school-age children has infrastructure of the educational system can often offer a fundamental implications for the design of effective programs. more cost-effective route for delivery of simple health inter- Improving the health and nutrition of schoolchildren through ventions and health promotion than can the health system. school-based programs is not a new concept. School health Low-income countries typically have more teachers than programs are ubiquitous in high-income countries and most nurses and more schools than clinics, often by an order of middle-income countries. In low-income countries, these pro- magnitude. grams were a common feature of early, particularly colonial, · Third, empirical evidence shows that good health and nutri- education systems, where they could be characterized as heavily tion are prerequisites for effective learning. This finding is focused on clinical diagnosis and treatment and on elite not simply the utopian aspiration for children to have schools in urban centers. This situation is changing as new healthy bodies and healthy minds, but also the demonstra- policies and partnerships are being formulated to help ensure tion of a systemic link between specific physical insults and that programs focus on promoting health and improving the specific cognitive and learning deficits, grounded in a new educational outcomes of children, as well as being socially pro- multisectoral approach to research involving public health gressive and specifically targeting the poor, girls, and other dis- and epidemiology, as well as cognitive and educational advantaged children. This evolution reflects five key changes psychology. in our understanding of the role of these programs in child · Fourth, the provision of quality schools, textbooks, and development. teachers can result in effective education only if the child is present, ready, and able to learn. This perception has addi- · First, ensuring good health at school age requires a life cycle tional political momentum as countries and agencies seek to approach to intervention, starting in utero and continuing achieve Education for All (EFA) by 2015 and address the throughout child development. In programmatic terms this Millennium Development Goals of universal basic educa- requirement implies a sequence of programs to promote tion and gender equality in education access. If every girl maternal and reproductive health, management of child- and boy is to be able to complete a basic education of good hood illness, and early childhood care and development. quality, then ensuring that the poorest children, who suffer Promoting good health and nutrition before and during the most malnutrition and ill health, are able to attend and school age is essential to effective growth and development. stay in school and to learn while there is essential. 1091 Morbidity as a proportion of peak value and important infections are caused by geohelminths (the 1 roundworm Ascaris, the whipworm Trichuris, and the two 0.9 species of hookworms Ancylostoma and Necator) and by the schistosomes (Schistosoma spp.), which give rise to a wide range 0.8 Ascaris of chronic but largely nonspecific symptoms. The most intense 0.7 worm infections and related illnesses occur at school age 0.6 (Partnership for Child Development 1998b, 1999) and account 0.5 School-age for some 12 percent of the total disease burden and 20 percent children of the loss of disability-adjusted life years (DALYs) from com- 0.4 Schistosoma Cerebral malaria municable disease among schoolchildren (World Bank 1993). 0.3 haematobium Infected schoolchildren perform poorly in tests of cognitive 0.2 function; when they are treated, immediate educational and 0.1 cognitive benefits are apparent only for children with heavy Diarrhea 0 worm burdens or with concurrent nutritional deficits. 0 5 10 15 20 25 30 Treatment alone cannot reverse the cumulative effects of life- Age (years) long infection or compensate for years of missed learning, but Source: Bundy and Guyatt 1996. studies suggest that children are more ready to learn after treat- ment for worm infections and may be able to catch up if this Figure 58.1 Age Distribution of Infection-Specific Morbidity learning potential is exploited effectively in the classroom (Grigorenko and others forthcoming). In Kenya, treatment reduced absenteeism by one-fourth, with the largest gains for · Finally, education, including education that promotes posi- the youngest children who suffered the most ill health (Miguel tive health behaviors, contributes to the prevention of and Kremer 2004). HIV/AIDS--the greatest challenge for generations to come. School health and nutrition programs that help children complete their education and develop knowledge, practices, Malaria and behaviors that protect them from HIV infection as they Up to 5 percent of children infected with malaria early in life mature have been described as a "social vaccine" against the have residual neurological sequelae (Snow 1999). In areas of disease. unstable transmission, malaria accounts for 10 to 20 percent of all-cause mortality among school-age children (Bundy and Because of the success of child survival programs, the num- others 2000), and those who have suffered repeated attacks have ber of children reaching school age (defined as 5 to 14 years of poorer cognitive abilities. In Kenya, primary school students age) is increasing and is estimated to be 1.2 billion children, miss 11 percent of school days because of malaria, equivalent to with 88 percent living in less developed countries (U.S. Census 4 million to 10 million days per year (Brooker and others 2000). Bureau 2002). As figure 58.1 illustrates, the pattern of disease is Oral antimalarial treatment reduced school absenteeism by age specific. A large body of evidence shows that these condi- 50 percent in Ghana (Colbourne 1955); the use of insecticide- tions affect cognition, learning, and educational achievement treated bednets in Tanzania reduced malaria and increased (see Jukes, Drake, and Bundy forthcoming; Pollitt 1990 for attendance (Shiff and others 1996). Girls in The Gambia were reviews of this extensive literature). more than twice as likely to enroll in primary school if they had This chapter focuses on the health, nutrition, and education received malaria prophylaxis in early childhood (Jukes and of the school-age child and on the programs that can be imple- others submitted). mented at school age to promote positive outcomes. HIV/AIDS INFECTIOUS DISEASE AND SCHOOL-AGE CHILDREN Although school-age children have the lowest infection preva- lence of any age group (figure 58.2), an estimated 3.8 million A range of infectious diseases affect school-age children. children under 15 years of age have been infected with HIV and more than two-thirds have died (UNAIDS 2002). Even unin- Helminth Infections fected children suffer physically, socially, and psychologically Between 25 and 35 percent of school-age children are esti- through death or illness in their family (World Bank 2002). The mated to be infected with one or more of the major species of proportion of orphans, most of whom are of school age, has worms (Bundy 1997; see also chapter 24). The most common risen from 2 to 15 percent in some African countries, with 1092 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others a. Male Cases Acute Respiratory Infection Percent infected in each age group, as a percentage of cases Acute respiratory infection, the most common acute infection 45 in school-age children globally, is a significant cause of absen- 40 teeism. Research in industrial countries (Cohen and Smith 1996) finds that flu infection affects attention and reaction 35 time; colds primarily affect hand-eye coordination, as well as 30 reduce the ability to tolerate high levels of noise and other dis- tractions common to the classroom. 25 20 15 MALNUTRITION, NONINFECTIOUS DISEASE, AND HEALTH AND EDUCATION 10 5 Malnutrition and noninfectious disease also affect school-age children. 0 0­4 5­14 15­19 20­29 30­39 40­49 50 Age group Malnutrition b. Female Cases Stunting (low height for age) is a physical indicator of chronic or long-term malnutrition, whereas underweight (low weight Percent infected in each age group, as a percentage of cases 50 for age) is an indicator of both chronic and acute malnutrition. Both are common in school-age children (figure 58.3). 45 Girls who are better nourished are more attentive and more 40 involved during class, and boys have improved classroom 35 behavior and increased activity levels. One Z-score increase in 30 height for age is associated with an increase of 0.1 standard deviation (SD) in tests of arithmetic and language. Stunted 25 children enroll in school later than other children. School food- 20 service programs have been successful in improving school 15 attendance. 10 Z-score 5 0 0 0­4 5­14 15­19 20­29 30­39 40­49 50 0.5 Age group 1.0 Botswana Côte d'Ivoire Malawi Tanzania Zimbabwe 1.5 Source: UNAIDS epidemiological fact sheets 2000. 2.0 Note: Figure shows percentage of males (top) and females (bottom) infected with HIV in each age group (as a percentage of all HIV-infected males and females, respectively), for five countries in Africa. Infection peaks at a younger age in women 2.5 than in men, and the lowest prevalence of infection occurs in school-age children. 3.0 Figure 58.2 Age Prevalence of HIV/AIDS 3.5 6 7 8 9 10 11 12 13 14 15 16 17 18 Age (years) AIDS accounting for 50 percent of this increase. The number of Ghana India Indonesia orphans is expected to reach more than 25 million by 2010. Tanzania Vietnam School-age children with HIV infections have lower IQ levels and poorer academic achievement, language, and visual Source: Data from Partnership for Child Development 1998a. Note: Z-scores of less than 2 indicate stunting. motor functioning. These deficits can be reduced or reversed with antiretroviral therapy. The improvement is greater for Figure 58.3 Mean Z-Scores of Height-for-Age of Boys in Five children of school age than for younger children. Countries School-Based Health and Nutrition Programs | 1093 Short-Term Hunger the children of many middle-income countries is similar to Hunger, which reduces ability to perform school tasks, is read- that in the United States. ily reversed by feeding. Children age 11 to 13 years in Jamaica improved their scores on arithmetic tests after one semester of receiving breakfast at school because they attended more regu- ESTIMATING THE BURDEN OF DISEASE larly and studied more effectively (Simeon 1998). Missing breakfast impairs performance to a greater extent for children The cost per DALY of school health programs has been esti- of poor nutritional status, who also benefit most from food mated at US$20 to US$34, implying that the programs are at intervention (Pollitt, Cueto, and Jacoby 1998; Simeon and least as cost-effective as many other public health "best buys" Grantham McGregor 1989). (Bobadilla and others 1994). However, current methods of estimating the burden for school-age children result in a signif- icant underestimation of both the developmental conse- Micronutrient Deficiency quences of disease and malnutrition at school age and the over- all benefits for health and development of school health and Micronutrient deficiencies may take several different forms, nutrition programs. each with negative impacts on children's ability to perform well There are two key reasons for this underestimation. The first in school. issue relates to time scales. Many serious diseases in adulthood, including heart disease and carcinomas, are a consequence of Iron Deficiency. Iron deficiency, the most common form of unhealthy practices established in early life. This later burden micronutrient deficiency in school-age children, is caused by can be substantially and cost-effectively averted by early inter- inadequate diet and infection, particularly by hookworm and vention, particularly by school-based life-skills programs. For malaria (Hall, Drake, and Bundy 2001). More than half the example, in the United States (Del Rosso and Marek 1996), school-age children in low-income countries are estimated to US$1 invested can avert US$18.80 spent on the later problems suffer from iron deficiency anemia (Partnership for Child caused by tobacco and US$5.70 on problems of drug and alco- Development 2001). Children with iron deficiency score 1 to 3 hol abuse. DALY estimates cannot capture these downstream SD worse on educational tests and are less likely to attend consequences of upstream intervention and instead attribute school. Iron supplementation reduces these deficits. the disease burden to the adult age group in which it appears. This kind of estimate is particularly misleading in the case of Iodine Deficiency. Iodine deficiency affects an estimated HIV/AIDS, for which prevention education at school age is 60 million school-age children; studies indicate prevalence effective in averting later infection and disease (World Bank rates between 35 and 70 percent. Iodine deficiency is related to 2002), and in the case of estimates of intergenerational effects, lowered general cognitive abilities and tests scores. No conclu- in which ensuring the health of an adolescent girl may help sive evidence shows that iodine supplementation improves secure the health of her baby born a few years later. cognitive abilities in this age group (Huda, Grantham- The second issue is illustrated by experience with helminth McGregor, and Tomkins 2001). infections. In 1990, the burden was first estimated at 18 million DALYs, close to the value for tuberculosis, measles, and malaria. Vitamin A Deficiency. Vitamin A deficiency affects an esti- This estimate reflected the ubiquity of infection and the long- mated 85 million school-age children. The deficiency, which term consequences of cognitive impacts. In 2001, the estimate causes impaired immune function and increases risk of was only 4.7 million DALYs (WHO 2003), and during the inter- mortality from infectious disease, is an important cause of vening years one estimate put the value as low as 2.6 million. blindness. Recent studies suggest that this deficiency is also This extraordinary variability is caused in part by different a major public health problem in school-age children. emphases on the cognitive and health impacts and illustrates Multiple-micronutrient supplements have improved cogni- how, for very common conditions, even minor changes in dis- tive function and short-term memory in schoolchildren and ability weight can affect the overall values. This variability also have reduced absenteeism caused by diarrhea and respiratory reflects the importance of a sectoral perspective, because the infections. low estimates reflect a focus on health, whereas the higher estimates include impact on educational achievement and child development. Obesity The scale of the burden of disease in terms of cognition is An estimated 17.6 million children worldwide are overweight. illustrated by estimating the impact of stunting, anemia, and Obesity is associated with underperformance in education. In helminths on the cognition of the estimated 562 million school- low-income countries obesity is still rare, but the prevalence in age children in developing countries. According to typical 1094 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others deficits in test scores attributable to these diseases, the total glob- The United Nations Population Fund (UNFPA) has pio- al loss of points ranges from 600 million to 1.8 billion IQ points, neered population and family life education (PopEd) as an an additional 15 million to 45 million cases of mental retarda- intrinsic part of school curricula. In 1994, the International tion (defined here as IQ less than 70), and a loss of between 200 Conference on Population and Development placed specific million and 524 million years of primary schooling (Jukes, emphasis on school health, including reproductive and sexual Drake, and Bundy forthcoming). Although the precision of health. Efforts at country level have addressed PopEd both these striking figures may be open to debate, they clearly show within the school system and outside, and the concept has that even minor cognitive deficits resulting from ubiquitous evolved to include references to family life education, sex edu- conditions can result in an extraordinarily large scale of effect. cation, HIV/AIDS awareness and prevention, and life-skills programs. Today, approximately 84 countries have UNFPA- supported school health programs. INTERVENTIONS In 1995, the World Health Organization (WHO) launched its Global School Health Initiative to foster the development of In light of the significant effects of ill health and malnutrition health-promoting schools (HPSs) (WHO 1996). The concept on educational outcomes, the role of effective health promo- started in Europe in the early 1990s, based on the Ottawa tion and simple school-based programs to deliver low-cost Charter of Health Promotion (WHO 1986; European interventions becomes increasingly important (Bundy and Commission 1996), which recognized that health is created by others 1992). Other chapters provide information on the caring for oneself and others, by being able to make decisions integrated management of childhood illness, early child devel- and have control over one's life and circumstances, and by creat- opment, and adolescent health (see chapters 63, 27, and 59, ing conditions that support health for all. WHO's European respectively). The focus here is on ill health and malnutrition at Regional Office, the Council of Europe, and the Commission of school age and the role of the formal and nonformal education the European Communities widely promoted the concept of sector in delivering interventions. HPSs to foster healthy lifestyles and develop environments con- ducive to health (European Commission, WHO Europe, and Council of Europe 1996). Although definitions vary among Developing a Programmatic Approach regions, countries, and schools, an HPS may be characterized as The focus of school health and nutrition programs in low- one that is constantly strengthening its capacity as a healthy set- income countries has shifted significantly over the past two ting for living, learning, and working. The initiative fosters the decades away from a medical approach that favored elite development of HPSs by the following: schools in urban centers and toward an approach that improves health and nutrition for all children, particularly the · consolidating research and expert opinion to describe the poor and disadvantaged. This change began in the 1980s, when nature and effectiveness of school health programs research showed not only that school health and nutrition pro- · building capacity to advocate for the creation of HPSs and grams were important contributors to health outcomes but to apply the components to priority health issues also that they were essential elements of efforts to improve edu- · strengthening collaboration and national capacities to assess cation access and completion, particularly for the poor. the prevalence of important health-related behaviors and In an effort to reconceptualize the relationship between conditions and to plan and implement policies and pro- health and education, the United Nations Education, Scientific, grams that improve health through schools and Cultural Organization (UNESCO) hosted a series of work- · creating networks and alliances,including regional networks. shops on this topic in the 1980s (Bundy 1989; Halloran, Bundy, and Pollitt 1989) and supported one of the first authoritative The key elements of how this approach is interpreted today reviews of the area (Pollitt 1990). Similarly, the United Nations are listed in table 58.1. Development Programme, in conjunction with the Rockefeller, In the mid 1990s, the United Nations Children's Fund Edna McConnell Clark, and J. S. McDonnell Foundations sup- (UNICEF) began promoting the Child-Friendly Schools ported the creation of the Partnership for Child Development framework as a holistic way to promote children's rights as to strengthen the evidence base across the education and health expressed in the Convention on the Rights of the Child sectors and to support the dissemination of information (UNICEF 1990) and children's access to education as stated in (Berkley and Jamison 1990; Bundy and Guyatt 1996). This par- the World Declaration of Education for All (UNESCO 1990). adigm shift coincided with the World Conference on Education This approach included a gender-sensitive component, which for All in Jomtien, Thailand, in 1990 and led to renewed efforts was further strengthened when girls' education became the by countries and agencies to develop more effective program- first priority in UNICEF's Medium Term Strategic Plan, matic approaches to school health and nutrition. 2002­5. Another key element is skills-based health education, School-Based Health and Nutrition Programs | 1095 Table 58.1 Characteristics of Agency-Specific School Health and Nutrition Programs, within the FRESH framework FRESH Health-promoting Child-friendly Global school feeding framework schools (WHO) schools (UNICEF) PopEd (UNFPA) campaign (World Food Program) Policy Respects an individual's Respects and realizes the rights of Creates a supportive and enabling Focuses on the poorest and most well-being and dignity every child policy environment for reproductive food-insecure communities. Provides multiple opportunities Acts to ensure inclusion, respect, health and HIV prevention for Gives priority to girls and for success and equality of opportunity for all young people AIDS-affected children Acknowledges good efforts and children intentions as well as personal Is gender sensitive and girl friendly achievements Is flexible and responds to diversity Sees and understands the whole child in a broad context Enhances teacher capacity, morale, commitment, and status School Is healthy Is healthy, safe, and secure Protects young people from early Serves as platform for essential environment Provides opportunities for Is protective emotionally and and unwanted pregnancy, sexually package approach that includes physical education and psychologically transmitted diseases, sexual water, sanitation, and recreation abuse, and violence environmental measures Education Provides skills-based health Promotes quality learning outcomes Strengthens HIV/AIDS and sexual Supports learning through good education Provides education that is affordable and reproductive health education nutrition Fosters health and learning and accessible programs Promotes access to education Provides skills-based health education, including life skills relevant to children's lives Services Provides school health services Promotes physical health Ensures access to youth-friendly Provides food Provides nutrition and food- Promotes mental health sexual and reproductive health Promotes and supports safety programs services deworming Provides programs for counsel- ing, social support, and mental health promotion Provides health promotion programs for staff Includes school and community projects and outreach Supportive Engages health and education Is child centered Targets young people in school Promotes community and school partnerships officials, teachers, teachers' Is family focused and out of school partnerships unions, students, parents, health Is community based Ensures active participation of providers, and community parents, youths, community leaders in efforts to make the leaders, and organizations school a healthy place Source: Summarized from World Bank Fresh Toolkit (2000), WHO (1996), and personal communications from Arlene Mitchell and Sheldon Shaeffer (May 2005). including life skills, which has been promoted through required to promote health and learning (UNICEF, WHO, UNICEF with partner organizations as part of HPSs, child- World Bank, UNFPA, UNESCO 2003). The approach is inter- friendly schools, and the framework for Focusing Resources on active, activity based, and flexible so that it can be used to Effective School Health (FRESH). Research shows that this address a range of health and social issues, including approach is more effective than traditional strategies, which HIV/AIDS, sanitation, drug use, violence and bullying, nutri- tend to be didactic and to focus on scientific information tion, and cross-cutting issues such as gender and culture. Some alone. In contrast, skills-based health education uses the expe- key elements of how the child-friendly schools approach is riences of students as the starting point and explores the links interpreted currently, including its focus on healthy and pro- between knowledge, attitudes, and the interpersonal skills tective learning environments, are listed in table 58.1. 1096 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others Also during the 1990s, the World Bank Human Develop- · School environment: access to safe water and provision of ment Network sought to support countries in implementing separate sanitation facilities for girls, boys, and teachers school health and nutrition programs (Del Rosso and Marek · Education: skills-based education, including life skills, that 1996; World Bank 1993) and launched an International School addresses health, nutrition, HIV/AIDS prevention, and Health Initiative with the aim of raising awareness among deci- hygiene issues and that promotes positive behaviors sion makers in the education sector. · Services: simple, safe, and familiar health and nutrition serv- Thus, the 1990s were characterized by the creation of a ices that can be delivered cost-effectively in schools (such as number of apparently separate programs to promote and sup- deworming services, micronutrient supplements, and nutri- port school health. However, analysis at the country level tious snacks that counter hunger) and increased access to revealed that although the various agency initiatives used youth-friendly clinics. different "prisms" to view school health--public health for WHO, quality education for UNESCO, and child rights for The FRESH framework further proposes that these four UNICEF--the core activities for all the programmatic core components can be implemented effectively only if they approaches were essentially the same. are supported by strategic partnerships between the following groups: FRESH Framework A major step forward in international coordination and cohe- · health and education sectors, especially teachers and health sion was achieved when the FRESH framework was launched workers at the World Education Forum in Dakar in April 2000 (World · schools and the community Bank FRESH Toolkit 2000). Among the early partners in this · children and others responsible for implementation. effort were the Education Development Centre, Education International, the Partnership for Child Development, Adopting this framework does not imply that these core UNESCO, UNICEF, the World Food Programme (WFP), components and strategies are the only important elements; WHO, and the World Bank. This partnership recognizes that rather, implementing all of these in all schools would provide a the goal of universal education cannot be achieved while the sound initial basis for any pro-poor school health program. health needs of children and adolescents remain unmet and The common focus has encouraged concerted action by the that a core group of cost-effective activities can and must be participating agencies. It has also provided a common plat- implemented across the board to meet those needs and to form on which countries, agencies, donors, and civil society deliver on the promise of EFA. can support all programs, including agency-specific programs The expanded commentary on the Dakar Framework for (table 58.1). Another important consequence of the FRESH Action reflects the recommendations of this partnership and consensus framework has been to offer a common point of describes three ways in which health relates to EFA: as an input entry for new efforts to improve health in schools, as illus- and condition necessary for learning, as an outcome of effective trated by the three examples in box 58.1. quality education, and as a sector that must collaborate with This consensus approach has increased significantly the education to achieve the goal of EFA. In the follow-up to the number of countries implementing school health reforms. The Dakar Forum, UNESCO designated FRESH as an interagency simplicity of the approach, combined with the enhanced flagship program that will receive international support as a resources available from donor coordination, has helped strategy to achieve EFA. ensure that these programs can go to scale. Annual external The FRESH framework, which is based on good practice support from the World Bank for these actions approaches recognized by all the partners, provides a consensus approach US$90 million, targeting some 100 million schoolchildren. for the effective implementation of health and nutrition services within school health programs. The framework proposes four core components that should be considered in Common Interventions designing an effective school health and nutrition program Table 58.2 lists some specific interventions commonly com- and suggests that the program will be most equitable and bined within the school health intervention package, but it cost-effective if all of these components are made available, should be recognized that not all of these interventions will be together, in all schools: needed or be appropriate for all locations. Some interventions are synergistic: for example, worm infection will be addressed · Policy: health- and nutrition-related school policies that are by the provision of latrines, the promotion of hand washing, nondiscriminatory, protective, inclusive, and gender sensi- relevant health and hygiene education, and deworming tive and that promote the nutrition and physical and psy- services. Similarly, HIV/AIDS infection among youths will be chosocial health of staff, teachers, and children addressed by ensuring girls' participation in school, offering School-Based Health and Nutrition Programs | 1097 Box 58.1 Three Efforts to Improve Health in Schools The Multiagency Effort to Accelerate the Education UNICEF, WHO, the World Bank, UNESCO, and the Food Sector Response to HIV/AIDS in Africa and Agriculture Organization, the campaign promotes the This effort, coordinated by a Working Group of the following: UNAIDS Inter-Agency Task Team on HIV/AIDS and · policies that make food aid conditional on girls' partic- Education, promotes the FRESH framework specifically ipation in education and helps education systems do the following: · an essential package that includes school sanitation and · adopt policies that avoid HIV/AIDS discrimination water and environmental improvement and stigmatization · nutrition education that improves the quality of stu- · provide a safe and secure school environment dents' diets and HIV prevention education · provide skills-based health education, including life · nutrition services that include food, deworming, and skills, in schools to promote positive behaviors and alleviation of short-term hunger. healthy lifestyles Some 70 countries have begun to implement these · improve access to youth-friendly health services. principles and activities since 2002. More than 36 countries and a similar number of agen- cies, bilateral donors, and nongovernmental organiza- The Partnership for Parasite Control tions have collaborated in this effort since November Led by WHO and involving a broad range of development 2002. partners, this initiative promotes public and private efforts to include deworming in school health services, following a The Global School Feeding Campaign of the WFP resolution of the 54th World Health Assembly to provide by This campaign has gone beyond providing food aid to 2010 regular deworming treatment to 75 percent of school- develop a programmatic link between nutrition and educa- age children at risk (an estimated target population of 398 tion. Working with partners, including national govern- million). Of 41 target countries in Africa, 19 have begun ments, parent-teacher and other community organizations, school-based deworming programs since 2001. Source: Authors. skills-based health education (including life skills), offering Nevertheless, it is apparent that out-of-school children can- peer education, providing access to health clubs, and providing not benefit from many of the important components of access to treatment for sexually transmitted infections (STIs) at school-based programs, such as skills-based health education clinics. It is also apparent that whereas some interventions and life-skills development programs to prevent HIV/AIDS. promote multiple outcomes--for example, skills-based health Reaching these children requires more flexible approaches that education and life-skills development can help promote posi- combine the best of nonformal, informal, and community- tive behaviors that prevent STIs and substance abuse--other based approaches (see chapter 59). interventions may have a single focus, such as iron supplemen- tation to avoid anemia. COST-EFFECTIVENESS OF INTERVENTION Out-of-School Children A key issue in addressing the costs of the new approach to More than 100 million school-age children are out of school; school health and nutrition programs is the significant savings 60 percent are girls (UNESCO 1993). School health programs offered by using the school system infrastructure rather than in Guinea and Madagascar have demonstrated that many of that of the health system as the key delivery mechanism. The these children will take advantage of simple services, such as school system provides not only a preexisting mechanism, so deworming, provided in schools (Del Rosso and Marek 1996); costs are at the margins, but also a system that aims at being the school acts essentially as a community center. It also has pervasive and socially progressive. Some important interven- been demonstrated that deworming programs in schools ben- tions, especially in terms of health education, may be virtually efit out-of-school children by reducing disease transmission in cost free; they require only policy changes that result in doing the community as a whole (Bundy and others 1990). things differently. 1098 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others Table 58.2 Common Interventions within a School Health Program FRESH category Intervention Expected outcome Policy 1. Child rights, avoidance of discrimination and stigmatization, gender 1. Inclusion of all children sensitive, child centered 2. Inclusion of pregnant girls and mothers in education 2. Specific inclusion of girls 3. Enforcement of code of practice for teacher behavior zero tolerance policy 3. Avoidance of harassment and abuse 4. Collaboration between health and education sectors 4. Effective implementation Environment 1. Access to safe water 1,2,3,5. Reduced infection 2. Hand washing 4. Reduced drop out of adolescent girls 3. Provision of sanitation 4. Gender-separate sanitation 5. Garbage disposal Education 1. Curriculum addressing health, hygiene, and nutrition 1. Improved knowledge and skills to promote good 2. Life-skills program health, hygiene, and nutrition 3. Peer education program 2. Lifelong positive behaviors such as avoidance of HIV/AIDS and substance abuse 4. Health-promoting clubs 3, 4. Reinforcement of positive behaviors Services 1. Deworming for intestinal worms and schistosomiasis 1. Reduction in worm infection 2. Prompt recognition and treatment of malaria 2. Reduction in impact of malaria 3. Insecticide-treated nets 3. Reduction in incidence of malaria 4. Micronutrient supplements 4. Reduction in anemia and malnutrition 5. Breakfast, snacks, and meals 5. Avoidance of hunger 6. First-aid kits 6. Management of injuries 7. Referral to youth-friendly clinics 7. Access to specific treatment 8. Counseling and psychosocial support 8. Mental health Source: Authors. Table 58.3 Annual per Capita Costs of School-Based Health which is affected by factors such as local capacity, location and and Nutrition Interventions Delivered in Schools remoteness of communities, and community values and opin- ions; hence, these factors must be borne in mind when identi- Condition Intervention Cost (US$) fying a school health package. (See chapter 41 for details of the Intestinal worms Albendazole or mebendazole 0.03­0.20 costs of sanitation provision.) Schistosomiasis Praziquantel 0.20­0.71 Not illustrated in the table is the cost advantage of using the Vitamin A deficiency Vitamin A supplementation 0.04 existing school infrastructure for delivery. Estimates for deliv- Iodine deficiency Iodine supplementation 0.30­0.40 ery of simple interventions (such as anthelmintic pills or Iron deficiency and anemia Iron folate supplementation 0.10 micronutrient supplements) suggest that the teacher-delivery Refractive errors of vision Spectacles 2.50­3.50 approaches listed here may be one-tenth of the cost of the more Clinically diagnosed Physical examination 11.50 traditional mobile health teams and yet equally effective conditions (Guyatt 2003). As with all education innovations, however, Undernutrition, hunger School feeding 21.60­151.20, the additional cost of teacher orientation and training (in- 21.26­84.50a service as well as preservice) needs to be factored into the costs of using the education system for delivery of health services. Sources: Del Rosso and Marek 1996; Partnership for Child Development 1999; WHO 2000. a. For South America and Africa, costs are standardized for 1,000 kilocalories for 180 days. ECONOMIC BENEFITS OF INTERVENTION Annual costs of providing some common school-based interventions to students are given in table 58.3. This table The most obvious benefit of school health interventions is illustrates two important points. First, some of the most widely arguably through the economic returns of improved adult health needed interventions can be provided at remarkably low cost. outcomes. Studies have increasingly documented a causal effect Second, significant diversity exists in the cost of interventions, of adult health (broadly defined) on labor force participation, School-Based Health and Nutrition Programs | 1099 wages, and productivity in developing countries; Strauss and schooling. In preschool children in Delhi, iron supplementa- Thomas (1995) present an overview of economic studies in this tion was associated with an increase of 5.8 percent in rates of area. For example, height has been shown to affect wage-earning participation at the preschool level (Bobonis, Miguel, and capacity as well as participation in the labor force for both Sharma 2004). In western Kenya, deworming treatment women and men (Haddad and Bouis 1991). The effect of health improved primary school participation by 9.3 percent, with an on productivity and earnings may be strongest where low-cost estimated 0.14 additional years of education per pupil treated health interventions produce large effects on health,such as low- (Miguel and Kremer 2004). On the basis of crude estimates of income settings where physical endurance yields high returns in returns to schooling, an increase of 9.3 percent in participation the labor market. For a 1 percent increase in height, Thomas and rates results in a return of US$44. Miguel and Kremer (2004) Strauss (1997) find a 7 percent increase in wages in Brazil com- conclude that these benefits still outweigh the costs even if pared with a 1 percent increase in the United States. increased school participation leads to greater costs in teacher However, the apparent benefits of school health and nutri- compensation through the need for additional teachers. They tion programs will be underestimated when measured using note that the benefit-cost ratio remains over 10 even if the rate only mortality or health-related disability metrics because these of return to an additional year of schooling is as low as 1.5 per- measures do not capture the impact of ill health on cognitive cent. These results suggest that for realistic estimates of returns development or educational outcomes. Evidence over the past to schooling, the net present discounted value of lifetime earn- decade suggests these impacts have effect sizes in the range 0.25 ings is likely to be high compared to the costs of treatment even to 0.4 SD and have implications for the child's education and for for small gains in school participation.1 life beyond school, including future earning potential.We inves- In the absence of studies estimating the direct link between tigate those implications by considering the economic benefits school health interventions and school participation, the rela- in terms of IQ and school attendance and by comparing school tionship can be estimated indirectly by considering the effect of health programs with traditional education interventions. interventions on test scores and the implications that improved test scores have for school participation. Improvements in cog- Economic Benefits of Long-Term Improvements in IQ nitive function can be converted into an equivalent number of years of schooling. For example, Jukes and others (2002) found School health interventions can yield considerable economic that heavy schistosomiasis was (nonsignificantly) associated benefits through returns to wages and productivity if they with a decrease in arithmetic scores of 1.35 marks (0.25 SD). translate into improved cognitive functioning and IQ in adult- An extra year of schooling was associated with an increase in hood. arithmetic scores of 2.24 marks (0.42 SD). Thus, the negative For the United States, Zax and Rees (2002) estimate conser- effect of heavy schistosomiasis was equivalent to missing just vatively that an increase in IQ of 1 SD is associated with an over half a year of schooling. The cognitive gains from an extra increase in wages of more than 11 percent, falling to 6 percent year of schooling can also be estimated retrospectively: in a when controlling for other covariates. Similar estimates for the study of adults in South Africa, each additional year of primary relationship between IQ and earnings have been made for schooling was associated with a 0.1 SD increase in cognitive test Indonesia (Behrman and Deolalikar 1995) and Pakistan scores (Moll 1998). According to these estimates, a typical (Alderman and others 1997) and in a review of developing increase of 0.25 SD associated with school health and nutrition countries (Glewwe 2002). In South Africa, an increase of 1 SD programs is equivalent to an additional 2.5 years of schooling. in literacy and numeracy scores was associated with a 35 percent Liddell and Rae (2001) assessed the direct effect of test increase in wages (Moll 1998). Extrapolating these results, a 0.25 scores on grade progression in Africa. Each additional SD SD increase in IQ, which is a conservative estimate of the bene- scored in first-grade exams resulted in children being 4.8 times fit resulting from a school health intervention, would lead to an as likely to reach seventh grade without repeating a year of increase in wages of from 5 to 10 percent. schooling.2 According to these estimates, an increase of 0.25 SD in examination scores, which is typically achieved by school Economic Benefit of Improved School Attendance health and nutrition programs, will make children 1.48 times3 School health interventions can raise adult productivity not as likely to complete seventh grade, which implies that the extra only through higher levels of cognitive ability, but also through cumulative years of schooling attributable to the school health their effect on school participation and years of schooling intervention average 1.19 years per pupil. The previous esti- attained. Healthier children are more likely to attend, and mod- mates for added years of schooling owing to school health est improvements in examination scores can be associated with interventions range from seven months to two years. Increased continuation in schooling. years of schooling are associated with, among other outcomes, Malaria chemoprophylaxis given in early childhood in The higher worker productivity and generally higher productivity Gambia led to an increase of more than one year in primary in nonmarket production activities, including greater farmer 1100 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others efficiency and productivity (Jamison and Lau 1982; tional cash-transfer approach is, in both cases, apparently at the Psacharopoulos and Woodhall 1985; Strauss and Thomas lower end of effectiveness and the higher end of cost. 1995). Psacharopoulos and Patrinos (2002) summarize a wide range of studies that focus on individual wage earnings. For IMPLEMENTATION OF PROGRAMS AND LESSONS Sub-Saharan Africa, they find a 12 percent rate of return to one FROM EXPERIENCE additional year in school, compared with 10 percent for Asian countries. These returns are very high, even allowing for a por- The FRESH framework provides strategic guidance, but the tion of the return to years of schooling to be capturing ability practical design of actual programs reflects differences in local and factors other than schooling itself (Card 2001). needs and capacity. Successful and equitable programs in low- Education brings benefits beyond improved earnings. One and middle-income countries are characterized by a focus on year of extra education for girls can lead to a reduction of from school-based delivery, on a public health paradigm that mini- 5 to 10 percent in infant mortality (Schultz 1993). Five extra mizes the need for clinical intervention and reliance on health years of education for women in Africa could reduce infant service facilities, and on participation of the public sector and mortality by up to 40 percent (Summers 1994). civil society locally. Economic Benefits of Programs Policy and Economic Issues in Defining Sectoral Roles The educational gains from school health and nutrition in Intervention programs should be considered in the context of alternative A negative correlation between income level and both ill health educational inputs, such as improving teacher salaries and and malnutrition is clearly demonstrated both in cross-country qualifications, reducing class size, improving school facility comparisons and within countries (see de Silva and others infrastructure, and providing instructional materials. Many 2003), partly because poverty promotes both disease and an studies relate student outcomes to school characteristics, but inadequate diet. Similarly, children who are not enrolled in few of these studies provide information on the relative or school come from households with lower income levels (Filmer actual costs of the educational inputs. The costs, however, are and Pritchett 2001). This fact suggests that school health services substantially greater than for the school health interventions that are pro-poor and specifically linked to efforts to achieve considered here. Despite the higher costs, the evidence from the universal participation in education will have a greater return. few randomized evaluations that have been conducted suggests Early school health programs, particularly in colonial Africa, that the scale of effect of additional education inputs is typically were intended to serve the minority of children who had access low (see discussion in Miguel and Kremer 2004). A review of to school in urban centers or elite boarding facilities. They relied studies showed that instructional materials (such as additional on specific infrastructures and services--such as mobile health textbooks) had the highest productivity, raising student test teams, school visits, school nurses, and in-school clinics--that scores significantly more than other inputs for each dollar were additional to the normal range of health service provision. spent. However, even these interventions have only a weak This approach has proven difficult to make universally available, effect. In a randomized experiment in Kenya, for example, pro- even in middle-income countries. A school nurse program in viding textbooks had no effect on the bottom three quintiles of KwaZulu-Natal, for example, achieved inadequate coverage (18 students and raised test scores by only 0.2 SD for the upper two percent of the target population) and little referral or follow-up quintiles. Relating these results to the findings in the previous treatment of cases of ill health detected, despite a relatively high section and to the annual per pupil costs, school health inter- investment of US$11.50 per student targeted per year (World ventions appear very cost-effective compared to the highest- Bank FRESH Toolkit 2000). As shown in the following exam- productivity, more traditional education inputs. ples, using the FRESH framework approach reduces costs sig- Recently, conditional cash-transfer programs have been nificantly and enhances both coverage and outcomes. viewed as potentially very cost-effective methods to increase An important element of the new approach to school health school enrollment. These programs are generally large in scope, is a focus on minimizing the need for clinical diagnosis. Mass representing a commitment of between 0.1 and 0.2 percent of delivery of services, such as deworming and micronutrient sup- gross national income. The Progresa program in Mexico is plementation, is preferable on efficacy, economic, and equity estimated to have increased enrollment by 3.4 percent and to grounds to approaches that require diagnostic screening have increased schooling by 0.66 years, with an average cash (Warren and others 1993). transfer (for grades 3 to 8) of about US$136 per child per school year (assumed to be 180 days). Gains from a similar Sectoral Roles in Implementation program in Nicaragua were estimated at 0.45 years of school at Table 58.4 gives examples from low- and middle-income a cost of US$77 per year. If we compare these results with those countries of how the four core components of FRESH are presented for school health and nutrition programs, the condi- being supported by different approaches. In about 85 percent School-Based Health and Nutrition Programs | 1101 1102 Table 58.4 Nine Low- and Middle-Income Countries and How They Use FRESH | Disease Program Country Outcomes (Costs approach examples Policy Environment Health education Health services per child per year) Control Public sector: Guinea, In all three countries, the Ministry of Education (or in Separate sanitation facili- Health, hygiene, and nutri- Deworming (for both schis- In three years, in Guinea-- public sector­ Ghana, and Ghana, its executive body, the Ghana Education ties for girls and boys in tion education as part of the tosomiasis and intestinal 1.1 million students, in Priorities supported and Tanzania Service) implements the program under the guidance all new schools; access to formal curriculum. worms) provided by teach- Ghana--577 schools and ­implemented of the Ministry of Health, on the basis of a formal potable water in all ers twice a year; in Guinea, 83,000 students (US$0.54), in in policy agreement. In Tanzania, the Ministries of schools. this service is followed Tanzania--353 schools and Developing Community Development and of Local Government by iron folate 113,000 students (US$0.89). are also parties to the agreement. The existing in- supplementation. service teacher training and supply-line infrastruc- Countries tures are used to prepare teachers and supply the necessary materials. | Parastatal support Madagascar The Community Nutrition Programme provides train- Access to potable water A formal health education Twice-yearly deworming In three years, 14,000 teach- Donald for public sector ing and support to the Ministry of Education on the and hand-washing facili- curriculum, supported by and iron folate (for three ers trained in 4,585 schools, intervention basis of a formally agreed-on health policy for the ties, in all schools; where community information, months) delivered by 430,000 students (US$0.78 A. P.Bundy education sector. In all schools in the 43 poorest requested by PTAs, education, and teachers; test kits to to US$1.08 per capita per districts (44 percent of all districts), the program construction of latrines, communication (IEC). confirm iodization of local year). prepares teachers and provides materials. In wells, fences, and sports sources of salt; where ,Sheldon addition, the program also provides Parent-Teacher facilities. requested by PTAs, Associations (PTAs) with access to a social fund to provision of food Shaeffer support construction of facilities. Each PTA can preparation facilities. request up to US$500, with a 20 percent community ,Matthew contribution based on an annual parental contribution of US$0.16. Social fund: public Tajikistan The Ministry of Labor and Social Protection, with the Provision of sanitation Training of teachers in Training of teachers to The program targets the Jukes, sector support for Ministries of Education and of Health, have devel- facilities, potable water, health promotion. provide first aid, micronu- 100,000 neediest children in community oped a memorandum of understanding that sets out and sports facilities. trients, and deworming; all 200 schools in the six and intervention health policies for the education sector. The program provision of food prepara- poorest districts of Tajikistan others channels resources through PTAs, which identify and tion facilities. (US$1 per capita per year). assist needy children. A training program, delivered by NGOs, prepares PTA members to develop propos- als of up to US$5,000 for their school to support activities selected from a menu of items. Private sector: Indonesia The NGO Yayasan Kusuma Buana has a formal agree- Not included in program. Nutrition and hygiene Stool examination by the The program has been in community ment with the education department in Jakarta and education as part of the laboratory and deworming existence for 17 years and payment for three other major cities to train teachers, perform curriculum. by teachers as necessary currently reaches 627 NGO-implemented diagnostic tests, and provide medicines and materi- twice a year; iron folate schools and 161,000 stu- intervention als. The NGO offers Papanicolaou smear tests and provided by teachers twice dents, at a cost to parents of referral services to teachers. Unit costs are low a year (for three months). US$0.10 annually. because parasite diagnosis involves mass screening in a central laboratory (approximately 2,500 diag- noses per day) and medicines are obtained at prefer- ential rates from two commercial partners. NGO implementa- Burkina Faso, The international NGO Save the Children U.S.A. Separate sanitation facili- Health, hygiene, and nutri- Deworming and micronutri- In three years, in Burkina tion with financial Malawi, and implements school health and nutrition activities in ties for girls and boys and tion education as part of the ent supplementation Faso, 42,000 students plus support from the Philippines nonformal schools created with support from govern- access to potable water. curriculum supported by (vitamin A and iron) nonenrolled children in 171 public sector ment, local communities, and private donations. extracurricular IEC activities provided by teachers schools (US$2). In four years, annually. in Malawi, 122,000 children in 181 schools (US$3). In four years in the Philippines, 23,000 children in 53 schools (US$6). Source: Authors. School-Based Health and Nutrition Programs | 1103 of programs reviewed, school health and nutrition programs particular, the consistency in the roles played by government are delivered and funded by the public education sector, with a and nongovernmental agencies as well as other partners and formal role for the health sector in design and supervision. stakeholders (table 58.5). In nearly every case, the Ministry of Although this public sector "mainstream" model has proven Education is the lead implementing agency, reflecting both the the most popular approach, it is not the only successful one. In goal of school health programs in improving educational some cases, the public sector has identified appropriate options achievement and the fact that the education system provides and developed operational manuals but then has used a social the most complete existing infrastructure for reaching school- fund to provide direct support to communities and has used age children. However, the education sector must share this schools to select and implement the most relevant actions responsibility with the Ministry of Health, particularly because locally, often with the assistance of nongovernmental organiza- the latter has the ultimate responsibility for health of children. tions (NGOs). In other cases, services have been contracted out It is also apparent that the program's success depends on the by the public sector, and in some middle-income countries, the effective participation of numerous other stakeholders, includ- move toward a demand-led approach has resulted in a private ing civil society, and especially the beneficiaries and their par- sector service. ents or guardians. The children and their families are the clients The private sector approach has proven sustainable over of these programs, and their support for program implementa- nearly two decades in urban Indonesia but may require a tech- tion is critical to the program's success. nical infrastructure and local market base that are inappropri- ate for predominantly rural low-income countries. The approach is modeled on a program initiated in Japan in 1948, Key Issues in Designing Effective Programs which relied on private sector technicians, working independ- The diverse experiences of school health programming suggest ently at first but later formalized within the Japan Association some key elements that are common contributors to success in of Parasite Control, who conducted stool examinations and many programs. then treated infected individuals for a per capita fee equivalent to approximately US$0.74 in 2004. At its peak, the private · Focus on education outcomes. Making explicit links among sector program conducted some 12 million examinations school health programs and learning and education sector annually, implying a turnover of nearly US$9 million at today's priorities (especially EFA and gender equity) helps ensure prices. The prevalence of roundworm infection fell from a the commitment of the sector to program support and high of 73 percent in 1949 to less than 0.01 percent by 1985. implementation. Although a private sector response is effective in some cir- · Develop a formal, multisectoral policy. Education sector cumstances, overall the characteristics of school health and actions in health require the explicit agreement of the health nutrition programs make a compelling case for public sector sector. This potential tension can be resolved by defining intervention. First, treatment externalities may create external sectoral responsibilities at the outset; failure to enter into benefits to others in addition to the benefit for the treated dialogue has led, in Africa and Central Asia, to some health individual. This situation is clearly the case for communicable sectors resisting teacher delivery of deworming drugs, disease interventions,especially against worm infection.Second, despite WHO recommendations. some forms of intervention (such as vector control, health · Initiate a process of wide dissemination and consultation. education campaigns, epidemiological surveillance, and inter- Because there are multiple stakeholders, implementers, ventions that have strong externalities) are almost pure public enablers, and gatekeepers, a process of consultation is neces- goods; that is, no one can be excluded from using the goods or sary to establish ownership and to identify obstacles before service they deliver,and thus the private sector is unlikely to com- they constrain progress. The process should involve at least pete to deliver these goods. Finally, there is typically little private community-based organizations, NGOs, faith-based organ- demand for general preventive measures,such as information on izations, pupils, and teacher associations. In one country the value of washing hands. None of these factors is an argument in East Africa, lack of prior agreement on the content of against a private sector role in service delivery, but they do sexuality education delayed implementation for more than suggest that private sector demand is likely to be greater in three years. middle-income populations and where public sector actions · Use the existing infrastructure as much as possible. Building have created a demand. on existing curriculum opportunities and the network of formal and nonformal teachers will accelerate implementa- tion and reduce costs. Programs that rely on the develop- Roles of Key Stakeholders in Implementation ment of new delivery systems--mobile school health teams, There are many ways to approach the delivery of school health, a cadre of school nurses--take longer to establish and are but these diverse experiences suggest common features--in expensive and complicated to sustain and take to scale. 1104 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others Table 58.5 Roles of Agencies, Partners, and Stakeholders in School Health and Nutrition Programs Partner Roles Comments Ministry of Education Lead implementing agency Health and nutrition of schoolchildren is a priority for EFA. Lead financial resource Education policy defines school environment, curriculum, duties of teachers. Education sector policy Education system has a pervasive infrastructure for reaching teachers and school-age children. Ministry of Health Lead technical agency Health of school-age children has lower priority than clinical services, infant Health sector policy health. Health policy defines role of teachers in service delivery, procurement of health materials. Other public sector agencies Support for education and health systems Ministries of local government are often fund holders for teachers and schools (for example, Welfare, Social Fund holder and for clinics and health agents. Affairs, local government) Ministries of Welfare and Social Affairs provide mechanisms for providing social funds. Private sector (for example, Specialist service delivery Sector has major role in drug procurement and training materials production. health service, pharmaceuti- Material provision Sector has specialist roles in health diagnostics. cals, publications) Civil society (for example, Training and supervision At the local level, organizations serve as gatekeepers and fund holders and NGOs, faith-based Local resource provision may target implementation. organizations, PTAs) Organizations provide additional resource streams, particularly international NGOs. Teacher associations Definition of teachers' roles School health programs demand an expanded role for teachers. Community (children, Partners in implementation Communities are gatekeepers for the content of health education (especially teachers, parents) Definition of acceptability of curriculum and moral and sexual content) and for the role of nonhealth agents (especially teachers' roles teachers) in health service delivery; pupils are active participants in all aspects of the process at the school level. Supplementation of resources Communities supplement program finance at the margins. · Use simple, safe, and familiar health and nutrition interven- RESEARCH AND DEVELOPMENT AGENDA tions. Success in rapidly reaching all schools depends on stakeholder acceptance, which is more likely if the interven- Reliable evidence suggests that ill health and malnutrition tions are already sanctioned by local and international agen- affect education access, participation, completion, and achieve- cies and are already in common use by the community. ment, and that school-based health and nutrition programs · Provide primary support from public resources. Compelling can provide a cost-effective and low-cost solution. This evi- arguments exist for public investment in school health pro- dence does not imply, however, that no uncertainties exist. grams: the contribution to economic growth, the high rate of return, the large externalities, and the fact that the major- Cost-Effectiveness of School-Based HIV/AIDS Prevention ity of interventions are public goods. Substantial evidence suggests that skills-based health education, · Be inclusive and innovative in identifying implementation including life-skills development programs, can promote posi- partners. Although public resources are crucial for school tive behaviors and reduce the risks of exposure to HIV infection, health programs, contributions from outside the public sec- and that girls' education programs have similar effects (Kirby tor can be vital. NGOs have proven effective in supporting 2002). Evidence also exists for a positive effect of completing public sector programs through training and supervision, education on HIV prevalence (de Walque 2004; World Bank particularly at local levels.Although market failure appears to 2002). What is lacking is direct evidence about the contribution have largely precluded the private sector from effectively that school-based prevention programs can make in reducing implementing national programs in low-income countries, the incidence of HIV infection, as well as evidence for the rela- examples of successful contributions do occur, particularly tive cost-effectiveness of such programs compared with existing in dense urban populations and in middle-income countries. efforts to promote education completion and girls' education. School-Based Health and Nutrition Programs | 1105 Cost-Effectiveness of Malaria Programs In consequence, the clearest benefit of school health and nutri- Malaria occurs commonly in schoolchildren, particularly in tion programs is measurable in terms of education outcomes areas of unstable transmission in Africa and Asia. It is a leading and their economic returns. The scale of benefit is significant: source of mortality in this age group and adversely affects edu- school health and nutrition interventions can add four to six cation by reducing school attendance, cognition, learning, and points to IQ levels, 10 percent to participation in schooling, and school performance. Current school-based approaches focus one to two years of education. This scale of benefit can add 8 to on knowledge of the disease and the use of impregnated bed- 12 percent to labor returns and provide a rate of return that nets but do not address the need for treatment of affected chil- offers a strong argument for public sector investment. dren. Yet presumptive treatment by teachers has been shown to Compelling evidence suggests that education qua education significantly reduce mortality (Pasha and others 2003), and can help protect individuals from HIV infection. Achieving EFA intermittent preventive treatment also shows considerable goals and combining this outcome with school health programs promise (Brooker and others 2000). There is a need to confirm that help establish lifelong positive behaviors are now recognized the success of school-based treatment in different epidemio- as essential to the multisectoral prevention response to logical settings and to address questions about the cost and sus- HIV/AIDS. tainability of this approach. The scale of the education benefit and the role of education in the fight against HIV/AIDS mean that school health and nutri- Cost-Effectiveness of Targeting Food Aid tion programs are today seen as a priority for both the education The high prevalence of malnutrition in children continues to and the health sectors. This focus, in turn, has resulted in a shift be a major challenge for low-income countries. Providing food toward public health rather than clinical intervention and to children at school is often seen as an important part of the toward school-based delivery rather than health system solution and is a major focus for food aid. However, the nutri- approaches. These policy changes enhance cost-effectiveness tion literature suggests that ensuring good nutrition earlier in and social progressiveness, because delivery through the school life--certainly before 3 years of age, but perhaps earlier--is system is an order of magnitude less costly than using health sys- essential to ensuring an appropriate development trajectory tems and in low-income countries is better targeted to the poor. throughout life (see chapter 27). Where food is limiting, it raises These changes in emphases have coincided with significant the question whether the first target should be preschool rather technical and political policy reform. Technical consensus than school-age children. This debate has been blurred by around the FRESH framework has encouraged countries and admixing the nutrition outcomes with broader social and edu- agencies to develop programs around a common coordinating cation issues. Clearly, providing a meal at school is socially principle, while the political imperative has been strengthened desirable and can offer education benefits for children who by the recognition that school health and nutrition programs otherwise would have to walk often long distances home to eat are essential to achieving EFA and the Millennium Develop- or remain hungry. It is also clear that schools represent an ment Goals and are at the center of the preventative response extensive and established network for providing nutrition to the HIV/AIDS pandemic. interventions to very large numbers of children at a low cost Although much of this change has evolved over the past two per child. No comparable network exists to reach preschool decades, significant acceleration has occurred since the World children. However, from a nutritional perspective, it remains Education Forum in 2000. Today, a majority of low-income unclear whether ensuring good nutrition early in life has more countries have recognized the need for school health and nutri- effect on subsequent development--including educational tion programs and are seeking to implement them. achievement--than providing food at school age. NOTES CONCLUSIONS 1. These calculations assume the following: a return to an additional year of school is 7 percent; wage gains are earned over 40 years in the The rationale for school-based health and nutrition programs workforce, discounted at 5 percent per year with no wage growth; annual and the approach to their implementation have undergone a wage earnings are US$400 per year, which is below the estimated agricul- paradigm shift over the past two decades. tural and nonagricultural annual wages for low-income countries (World Bank 2003). The opportunity costs of the additional schooling (child The traditional perception of these programs as seeking to labor) have not been considered but are likely to be negligible. improve the health of schoolchildren cannot be justified on the 2. These calculations assume that a pupil's falling behind the equiva- basis of mortality or public health statistics alone. Instead, it is lent of one year in test scores has the same effect on earnings as losing increasingly recognized that a major--perhaps the major-- one year of schooling; that the advantage that third graders have over second graders, for example, is the same as the advantage someone who impact of ill health and malnutrition on this age group is that on has studied for a total of three years has over someone who has studied cognitive development, learning, and educational achievement. for two years; and that the impact of first-grade examination scores on 1106 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others the probability of transition from one class to the next is the same at Filmer, D., and L. H. Pritchett. 2001. "Estimating Wealth Effects without each grade level. Expenditure Data--or Tears: An Application to Educational 3. If an increase of 1 SD in exam scores leads to children being 4.8 times Enrolments in States of India." Demography 38: 115­32. as likely to reach seventh grade, the increased likelihood of reaching sev- Glewwe, P. 2002. "Schools and Skills in Developing Countries: Education enth grade because of a 0.25 SD increase can be calculated as EXP (0.25 Policies and Socioeconomic Outcomes." Journal of Economic Literature LN(4.8)). 40: 436­82. Guyatt, H. L. 2003. "The Cost of Delivering and Sustaining a Control REFERENCES Program for Schistosomiasis and Soil-Transmitted Helminthiasis." Acta Tropica 86: 267­74. Alderman, H., J. R. Behrman, S. Khan, D. R. Ross, and R. Sabot. 1997. "The Grigorenko, E., R. Sternberg, D. Ngorosho, C. Nokes, M. Jukes, K. Alcock, Income Gap in Cognitive Skills in Rural Pakistan." Economic and D. Bundy. Forthcoming. "Effects of Antiparasitic Treatment Development and Cultural Change 46: 97­122. on Dynamically-Assessed Cognitive Skills." Journal of Applied Behrman, J. R., and A. B. Deolalikar. 1995. "Are There Differential Returns Developmental Psychology. to Schooling by Gender? The Case of Indonesian Labor-Markets." Haddad, L., and H. Bouis. 1991. "The Impact of Nutritional Status on Oxford Bulletin of Economics and Statistics 57: 97­117. Agricultural Productivity: Wage Evidence from the Philippines." Berkley, S., and D. Jamison. 1990. A Conference on the Health of School Age Oxford Bulletin of Economics and Statistics 53: 45­68. Children. New York: United Nations Development Programme and Hall, A., L. J. Drake, and D. A. P. Bundy. 2001. "Public Health Measures to Rockefeller Foundation. Processed. Control Helminth Infections." In Nutritional Anemias, ed. Bobadilla, J. L., P. Cowley, P. Musgrove, and H. Saxenian. 1994. "Design, U. Ramakrishnan. Boca Raton, FL: CRC Press. Content, and Financing of an Essential National Package of Health Halloran, M. E., D. A. P. Bundy, and E. Pollitt. 1989."Infectious Disease and Services." Bulletin of the World Health Organization 72: 653­62. the UNESCO Basic Education Inititative."Parasitology Today 5: 359­62. Bobonis, G., E. Miguel, and C. Sharma. 2004."Iron Deficiency Anemia and Huda, S. N., S. M. Grantham-McGregor, and A. Tomkins. 2001. "Cognitive School Participation." Working paper, University of California at and Motor Functions of Iodine-Deficient but Euthyroid Children in Berkeley. Bangladesh Do Not Benefit from Iodized Poppy Seed Oil (Lipiodol)." Brooker, S., H. Guyatt, J. Omumbo, R. Shretta, L. Drake, and J. Ouma. Journal of Nutrition 131: 72­77. 2000. "Situation Analysis of Malaria in School-Aged Children in Jamison, D., and L. J. Lau. 1982. Farmer Education and Farm Efficiency. Kenya: What Can Be Done?" Parasitology Today 16: 183­86. Baltimore: Johns Hopkins University Press. Bundy, D. A. P. 1989. "New UNESCO International Project." Parasitology Jukes, M. C. H., L. L. Drake, and D. A. P. Bundy. Forthcoming. Leveling the Today 5: 168. Playing Field: The Importance of School Health and Nutrition in ------. 1997."This Wormy World: Then and Now." Parasitology Today 13: Achieving Education for All. Washington, DC: World Bank. 407­8. Jukes, M., C. A. Nokes, K. L. Alcock, J. Lambo, C. Kihamia, A. Mbise, and Bundy, D. A. P., and H. L. Guyatt. 1996. "Schools for Health: Focus on others. 2002. "Heavy Schistosomiasis Associated with Poor Short-Term Health, Education, and the School-Aged Child." Parasitology Today 12: Memory and Slower Reaction Times in Tanzanian School Children." 1­16. Tropical Medicine and International Health 7: 104­17. Bundy, D. A. P., A. Hall, G. F. Medley, and L. Savioli. 1992. "Evaluating Jukes, M. C. H., M. Pinder, E. L. Grigorenko, H. Banos Smith, E. Measures to Control Intestinal Parasitic Infections." World Health Bariau-Meier, G. Walraven, and others. Submitted. "The Impact of Statistics Quarterly 45: 168­79. Malaria Chemoprophylaxis in Early Childhood on Cognitive Abilities Bundy, D. A. P., S. Lwin, J. S. Osika, J. McLaughline, and C. O. Pannenborg. and Educational Outcomes 14 Years Later: Follow-up of a Randomised 2000."What Should Schools Do about Malaria?" Parasitology Today 16: Controlled Trial in The Gambia." 181­82. Kirby, D. 2002. "The Impact of Schools and School Programs upon Bundy, D. A. P., M. S. Wong, L. L. Lewis, and J. Horton. 1990. "Control of Adolescent Sexual Behavior." Journal of Sexual Research 39: 27­33. Geohelminths by Delivery of Targeted Chemotherapy through Liddell, C., and G. Rae. 2001. "Predicting Early Grade Retention: A Schools." Transactions of the Royal Society of Tropical Medicine and Longitudinal Investigation of Primary School Progress in a Sample of Hygiene 84: 115­20. Rural South African Children." British Journal of Educational Card, D. 2001. "Estimating the Return to Schooling: Progress on Some Psychology 71: 413­28. Persistent Econometric Problems." Econometrica 69: 1127­60. Miguel, E., and M. Kremer. 2004. "Worms: Identifying Impacts on Cohen, S., and A. Smith. 1996. "Psychology of Common Colds and Other Education and Health in the Presence of Treatment Externalities." Infections." In Viral and Other Infections of the Human Respiratory Econometrica 72: 159­217. Tract, ed. S. Myint and D. Taylor-Robinson. London: Chapman and Moll, P. G. 1998. "Primary Schooling, Cognitive Skills and Wages in South Hall. Africa. " Economica 65: 263­84. Colbourne, M. J. 1955. "The Effect of Malaria Suppression in a Group of Partnership for Child Development. 1998a. "The Anthropometric Status Accra School Children." Transactions of the Royal Society of Tropical of School Children in Five Countries in the Partnership for Child Medicine and Hygiene 4: 356­69. Development." Proceedings of the Nutrition Society 57: 149­58. Del Rosso, J. M., and T. Marek. 1996. Class Action: Improving School ------. 1998b: "The Health and Nutritional Status of School Children in Performance in the Developing World through Better Health and Africa: Evidence from School-Based Health Programmes in Ghana and Nutrition School. Washington, DC: World Bank. Tanzania." Transactions of the Royal Society of Tropical Medicine and de Walque, D. 2004. "How Does the Impact of an HIV/AIDS Information Hygiene 92: 254­61. Campaign Vary with Educational Attainment? Evidence from Rural ------. 1999. "The Cost of Large-Scale School Health Programmes Uganda." Working paper, World Bank, Washington, DC. Which Deliver Anthelmintics to Children in Ghana and Tanzania." European Commission, WHO (World Health Organization) Europe, and Acta Tropica 73: 183­204. Council of Europe. 1996. Facts about the European Network of Health ------. 2001. "Anaemia in Schoolchildren in Eight Countries in Africa Promoting Schools. and Asia." Public Health Nutrition 4: 749­56. School-Based Health and Nutrition Programs | 1107 Pasha, O., J. Del Rosso, M. Mukaka, and D. Marsh. 2003. "The Effect of UNESCO (United Nations Educational, Scientific, and Cultural Organi- Providing Fansidar (Sulfadoxine-Pyrimethamine) in Schools on zation). 1990. "World Declaration on Education for All." http://www. Mortality in School-Age Children in Malawi." Lancet 361: 577­78. unesco.org/education/information/nfsunesco/pdf/JOMTIE_E.PDF. Pollitt, E. 1990. Malnutrition and Infection in the Classroom. Paris: United ------. 1993. World Education Report. Paris: UNESCO. Nations Educational, Scientific, and Cultural Organization. ------. 2003. EFA Global Monitoring Report 2003/4: Gender and Pollitt, E., S. Cueto, and E. R. Jacoby. 1998. "Fasting and Cognition in Well- Education for All: The Leap to Equality. Paris: UNESCO. and Undernourished Schoolchildren: A Review of Three Experimental UNICEF (United Nations Children's Fund). 1990. "Convention on the Studies." American Journal of Clinical Nutrition 67: 779S­84S. Rights of the Child." http://www.unicef.org/crc/crc.htm. Psacharopoulos, G., and H. Patrinos. 2002. Returns to Investment in UNICEF, WHO (World Health Organization), World Bank, UNFPA Education: A Further Update. Vol. 2881. Washington, DC: World Bank. (United Nations Population Fund), and UNESCO. 2003. Skills for Psacharopoulos, G., and M. Woodhall. 1985. Education for Development: Health: Skills-Based Health Education, Including Life Skills--An An Analysis of Investment Choices. New York: Oxford University Important Component of a Child Friendly/Health Promoting School. Press. U.S. Census Bureau. 2002. "Global Population Profile." http://www. Schultz, T. P. 1993. "Returns to Women's Schooling." In Women's Education census.gov. in Developing Countries: Barriers, Benefits, and Policy, eds. E. King and Warren, K. S., D. A. P. Bundy, R. M. Anderson, A. R. Davis, D. A. M. Anne Hill. Baltimore: Johns Hopkins University Press. Henderson, D. T. Jamison, and others. 1993. "Helminth Infection." In Shiff, C., W. Checkley, P. Winch, Z. Premji, J. Minjas,and P. Lubega. 1996. Disease Control Priorities in Developing Countries, ed. D. T. Jamison, "Changes in Weight Gain and Anaemia Attributable to Malaria in W. H. Mosley, A. R. Measham, and J. L. Bobadilla, 131­60. New York: Tanzanian Children Living under Holoendemic Conditions." Oxford University Press. Transactions of the Royal Society of Tropical Medicine and Hygiene 90: WHO (World Health Organization). 1986. Ottawa Charter. Geneva: WHO. 262­65. ------. 1996. Global School Health Initiative. WHO/HPR/98.4. Geneva: Simeon, D. T. 1998. "School Feeding in Jamaica: A Review of Its WHO. Evaluation." American Journal of Clinical Nutrition 67: 790S­94S. ------. 2000. Elimination of Avoidable Disability Due to Refractive Errors. Simeon, D. T., and S. Grantham McGregor. 1989. "Effects of Missing Report of an informal planning meeting. WHO/PBC/00.79. WHO: Breakfast on the Cognitive Functions of School Children of Differing Geneva Nutritional Status." American Journal of Clinical Nutrition 49: 646­53. ------. 2003. World Health Report. http://www.who.int/whr/2003. Snow, R. W. 1999. "Estimating Mortality, Morbidity, and Disability Due to World Bank. 1993. World Development Report: Investing in Health. New Malaria among Africa's Non-Pregnant Population." Bulletin of the York: Oxford University Press. World Health Organization 77: 624­40. ------. 2002. Education and HIV/AIDS: A Window of Hope. Washington, Strauss, J., and D. Thomas. 1995. "Human Resources: Empirical Modeling DC: World Bank. of Household and Family Decisions." In Handbook of Development ------. 2003. World Development Indicators. Washington, DC: World Economics, ed. J. Behrman and T. N. Srinivasan, vol. IIIA. Amsterdam: Bank. Elsevier. ------, with the Partnership for Child Development. 2000. "FRESH: Summers, L. H. 1994. Investing in All the People: Educating Women in Focusing Resources on Education and School Health. A FRESH Start Developing Countries. Vol. 45. Washington, DC: World Bank. to Improving the Quality and Equity of Education." http://www. Thomas, D., and J. Strauss. 1997."Health and Wages: Evidence on Men and schoolsandhealth.org/FRESH.htm. Women in Urban Brazil." Journal of Economics 77: 159­85. World Bank 2000. The FRESH Framework: A Toolkit for Task Managers. UNAIDS (Joint United Nations Programme on HIV/AIDS). 2000. Human Development Network, World Bank, Washington, DC. ------. 2002. Report on the Global HIV/AIDS Epidemic: The Barcelona Zax, J. S., and D. I. Rees. 2002. "IQ, Academic Performance, Environment, Report. Geneva: UNAIDS. and Earnings." Review of Economics and Statistics 84: 600­16. 1108 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others Chapter 59 Adolescent Health Programs Elizabeth Lule, James E. Rosen, Susheela Singh, James C. Knowles, and Jere R. Behrman This chapter reviews the main direct causes of loss of productive and young adults engage in a range of behaviors that can affect life years among adolescents and the range of interventions to the quality of their health and the probability of their survival in address these causes.It pays special attention to sexual and repro- the short term as well as affect their lifetime health and survival. ductive health because adolescence is when important transi- tions occur that can have a direct effect on young people's health Health Challenges of Adolescents as well as potential long-term consequences. In addition, a num- If we look only at disability-adjusted life years (DALYs) for the ber of interventions have focused on this aspect of young peo- adolescent age group, adolescents appear to be relatively ple's behavior. The discussion centers on defined interventions healthy. Nonetheless, more than 33 percent of the disease bur- that have some relatively rigorous evaluation component. There den and almost 60 percent of premature deaths among adults are some limitations to this chapter, however. First, it reviews can be associated with behaviors or conditions that began or existing research and does not carry out new research; however, occurred during adolescence--for example, tobacco and alco- it points to gaps in research and areas needing more work. hol use, poor eating habits, sexual abuse, and risky sex (WHO Second, the chapter uses a focused definition of health that 2002). Adolescence-related risk factors are a greater problem includes the most basic health outcomes (death,illness,timing of in wealthier countries, largely because of the relatively greater transitions,or specific risky or protective behaviors that research impact of smoking and diet-related risks in those countries, has attempted to measure). It does not attempt to cover aspects though the prevalence of these risks is expanding rapidly in that would be included in a broad definition of health and well- many low- and middle-income countries (LMICs). Thus, being (for example, potential for growth, creativity, or participa- although adolescents are apparently healthy, they are practicing tion) that are important but are poorly researched to date. unhealthy behaviors that will ultimately result in much death Finally, although the broader environment in which adolescents and disability. This is an immense public health issue. live influences their behavior and responses to programs, this Therefore, focusing attention both on diseases experienced chapter does not attempt to review that important group of fac- during adolescence and on risk factors with their roots in ado- tors or the broader set of programs that have a more indirect lescence makes sense. Adolescent health efforts should empha- effect on the health of young people (for example, school quality size prevention because so much of the disease burden is pre- or public health promotion activities at the societal level). ventable and because prevention is a particularly cost-effective strategy in relation to adolescents, given the long duration over NATURE AND CAUSES OF THE BURDEN which benefits will be reaped and adolescents' greater openness OF DISEASE IN YOUNG PEOPLE to change than adults. At first glance, adolescence appears to be a relatively healthy-- Burden of Disease in Adolescence although not hazard-free--period of life, given the relatively The global burden-of-disease approach used to calculate DALYs low mortality rates of young people.1 Nevertheless, adolescents is an imperfect representation of the prevalence, morbidity, and 1109 Table 59.1 Worldwide Distribution of DALYs for Major Categories of Diseases and Conditions by Age Group and Sex, 2002 (percent) Women Men Category Age 5­14 Age 15­29 Age 5­14 Age 15­29 Sexual and reproductive causes 4.6 33.4 3.9 9.5 HIV/AIDS 3.8 11.9 3.7 8.6 Maternal conditions 0.4 16.3 0.0 0.0 Other sexual and reproductive conditions 0.4 5.2 0.2 0.9 Respiratory conditions 11.9 3.7 9.9 3.7 Cardiovascular conditions 2.2 2.9 2.0 3.3 Neuropsychiatric conditions 15.5 33.8 14.9 32.0 Unipolar depressive disorders 5.4 10.1 5.6 7.1 Bipolar disorders 0.6 4.1 0.7 4.2 Schizophrenia 0.7 4.4 2.3 4.2 Other mental health conditions 1.0 4.4 0.6 2.6 Alcohol use disorders 0.1 1.1 0.3 5.7 Drug use disorders 0.1 0.6 0.2 2.3 Other 7.6 9.1 5.2 5.9 Injuries 25.0 14.0 32.4 33.1 Unintentional, road traffic accidents 5.7 2.2 7.7 7.5 Unintentional, other 16.8 7.7 21.8 12.9 Intentional, self-inflicted 1.0 2.9 0.9 3.8 Intentional, not self-inflicted (war, violence) 1.5 1.2 2.0 8.9 Other communicable diseases 31.1 7.1 27.7 11.1 Tuberculosis 1.5 3.0 1.4 3.9 Other noncommunicable diseases 9.9 9.6 9.2 10.0 Total 100.2 104.5 100.0 102.7 Source: WHO 2004a. mortality of conditions that adolescents face. DALYs fail to Using 2002 data, WHO has made more detailed calcula- capture fully the complexity of adolescent health concerns. tions of DALYs by sex for the 5 to 14 and 15 to 29 age groups Nonetheless, no better comprehensive and comparative meas- (table 59.1). These age ranges overlap adolescence and are, ure currently exists; thus, the discussion in this section will rely thus, broadly indicative of the 10 to 19 age group. Notably, primarily on available DALY data. table 59.1 shows large differences by sex in the pattern and level The World Health Organization (WHO), in 1999, commis- of DALYs. These differences are important, because they relate sioned a special analysis of the burden of disease in adoles- to the different needs of young women and young men for cence, which examined the 10 to 14 and the 15 to 19 age interventions and services. Particular interventions also poten- groups. The study found that young people age 10 to 19, who tially have different costs and benefits because of the different constitute 19 percent of the world's population, account for proportions of females and males. 15 percent of the disease and injury burden worldwide. It also Worldwide, among young men age 15 to 29, injuries and found that more than 1 million people in that age group die neuropsychiatric illnesses account for a high proportion of each year (WHO 1999). The top three causes of DALYs were DALYs (33 percent and 32 percent, respectively). By compari- found to be unipolar major depression, transportation son, among young women age 15 to 29, injuries account for accidents, and falls. The profile of disease burden was signifi- 14 percent of DALYs, and neuropsychiatric illnesses account cantly different for younger and older adolescents. In the 10 to for about the same percentage of DALYs as among young men. 14 age group, injuries and communicable diseases were promi- However, sexual and reproductive health conditions account nent causes of DALYs. For the 15 to 19 age group, the disease for 33 percent of young women's DALYs, much higher than the burden shifted to outcomes of sexual behaviors and mental 10 percent for young men. For both young men and young health. women, all other communicable and noncommunicable 1110 | Disease Control Priorities in Developing Countries | Elizabeth Lule, James E. Rosen, Susheela Singh, and others diseases account for moderate proportions of DALYs (7 to Smoking, Alcohol, and Drug Use. Most adult smokers world- 11 percent, depending on sex and disease group). wide begin smoking in adolescence or earlier (World Bank The disease burden among 5- to 14-year-olds is markedly 1999a). An estimated 15 percent of young men and 7 percent of different from that for the 15- to 29-year-olds, and differences women age 13 to 15 are currently smoking cigarettes, according between males and females are quite small. Communicable dis- to more than 100 surveys that have been conducted since 1999 eases and respiratory illnesses account for much larger propor- by the Global Youth Tobacco Survey Program and carried out tions of DALYs for this age group compared with the 15 to under the auspices of WHO and the U.S. Centers for Disease 29 age group, whereas neuropsychiatric and sexual and repro- Control and Prevention (National Research Council and ductive conditions account for much smaller proportions. Institute of Medicine 2005). Every day, worldwide, almost HIV/AIDS accounts for less than 4 percent of DALYs for both 100,000 young people start smoking, more than two-thirds of males and females age 5 to 14. them in LMICs (World Bank 1999a). Of the 300 million young people smoking today, half will eventually die from tobacco use (WHO 2001b). By 2030, tobacco is expected to be the single Health Risk Behaviors among Adolescents biggest cause of death worldwide, accounting for about 10 mil- and Young People lion deaths per year (World Bank 1999a). Young people's vulnerability to risky or unwanted sex and Although discouraging young people from starting to other unhealthy behaviors is tied to a host of individual, family, smoke and providing means for them to quit is extremely and community factors that influence their behavior and important, deaths caused by tobacco tend to occur many years that are closely related to their economic and educational later. Therefore, tobacco use as an underlying risk factor opportunities. Good health and other physical, moral, and accounts for very few DALYs in the 5 to 29 age group (WHO intellectual development outcomes are often mutually rein- 2002). Alcohol and drug use account for 8 percent of all DALYs forcing. For example, healthy children do better in school. for young men age 15 to 29 but for only 2 percent for young Similarly, having more years of schooling provides essential women. Evidence indicates that young people are starting to information and skills that are linked to more protective and drink at earlier ages. Longitudinal studies have found that the less risky behaviors. earlier young people start drinking, the more likely they are to experience alcohol-related injuries and alcohol dependence Injuries. Violence and war account for more than a quarter later in life (WHO 2001a). of injury-related deaths among young men age 15 to 29. Adolescent boys and men in their 20s are an important part of Nutrition and Exercise. Nutritional deficiencies such as ane- the military forces in all countries that have such forces. As mia are widespread in both young men and women. such, they are at high risk, particularly in areas where armed Worldwide, these conditions account for almost 5 percent of conflict is occurring. The United Nations Children's Fund esti- DALYs among girls age 5 to 14 and almost 4 percent among mates that approximately 300,000 soldiers under the age of 18 boys of the same age, with anemia being an important compo- are involved in armed conflicts worldwide (National Research nent for both girls and boys. Although nutritional deficiencies Council and Institute of Medicine 2005). Homicide is also an are relatively less important among 15- to 29-year-olds (just important cause of death for young men, in particular, and it over 1.0 percent among young men and about 1.5 percent is the leading cause of death for young men in some Latin among women), anemia accounts for the bulk of these defi- American countries (WHO 2001b). In addition, road accidents ciencies. Chronic undernutrition that causes stunting among account for significant proportions of injuries and deaths young people delays growth and physical maturation, increases among young people. Self-inflicted injuries, including suicide, risks to pregnant mothers and their newborns, and decreases which are often related to mental illness, are also a major health the capacity to learn and to work (Behrman and others 2004; problem for young people, accounting for 4 percent of DALYs Hoddinott and Quisumbing 2003). Malnutrition can take in men age 15 to 29 and 3 percent of DALYs in women age 15 other forms, some of which lead to being overweight or obese, to 29. thereby increasing the risks for diseases such as diabetes. Such forms of malnutrition are of increasing relevance in middle- Mental Health. Depression, schizophrenia, and other mental income countries such as Brazil, China, the Arab Republic of illnesses are important causes of illness and death among Egypt, Mexico, and South Africa and at times coexist with young men and women and account for a significant propor- undernutrition (see, for example, Doak and others 2000). tion of DALYs for both men (18 percent) and women (23 per- Nutritional deficiencies increase the risks that girls and cent) age 15 to 29; for 5- to 14-year-olds, it is about 9 percent young women face during pregnancy and childbirth (Delisle, for boys and 8 percent for girls. The relative importance of Chandra-Mouli, and de Benoist 2001), and evidence is emerg- mental illnesses is much greater in the high-income countries. ing about the connection between poor maternal nutrition and Adolescent Health Programs | 1111 greater risk of transmission of HIV from mothers to their In the most recent surveys carried out in LMICs, high pro- infants (Piwoz and Greble 2000). portions of adolescents report that they have heard of contra- Diet and lifestyle-related chronic diseases--many with their ceptive methods; however, little is known about the quality and roots in childhood and adolescence--are emerging as one of accuracy of young people's knowledge of contraception. the most important health problems in LMICs. Cardiovascular Moreover, substantial proportions of young women appear diseases, which are responsible for 10 percent of DALYs lost in to have an unmet need for contraception; they are not using LMICs, typically occur in middle age or later; however, risk fac- contraception even though they are sexually active and do not tors are determined to a great extent by behaviors learned want to have a child (CDC and ORC Macro 2003; Westoff and during childhood and adolescence and continued into adult- Bankole 1995). hood, such as dietary habits and smoking. Throughout the In addition to having a risk of early and unwanted preg- world, these risks are starting to appear earlier. Physical activity nancy, adolescents are also at risk of acquiring sexually trans- has decreased markedly in adolescence, particularly in girls, and mitted infections (STIs), including HIV. HIV/AIDS accounts obesity has increased substantially (MacKay and Mensah 2004). for most of the sexual and reproductive health DALYs lost by young men age 15 to 29 (almost 9 percent). Among young Sexual and Reproductive Behaviors. Worldwide, the majority women age 15 to 29, HIV/AIDS accounts for a higher propor- of young people initiate sexual activity during adolescence. tion of DALYs than for young men (almost 12 percent) because Significant proportions--in some regions and countries, the of their higher levels of susceptibility. STIs and other sexual and majority--marry and become parents (table 59.2). Globally, reproductive health disorders together account for just over the age of onset of puberty has been decreasing progressively 5 percent of young women's DALYs, much more than among for both boys and girls (National Research Council and young men. About half of all HIV infections occur in people Institute of Medicine 2005). The age at first marriage has also under age 25, and for biological, social, and economic reasons, increased in most parts of the world over recent decades, except young women are disproportionately affected, especially in in Latin America (Mensch, Singh, and Casterline 2003). The Sub-Saharan Africa, where young women have twice the preva- decline in the age at puberty, combined with the general trend lence rate of young men (UNAIDS 2003). toward later marriage, increases the period of time during which adolescents may be sexually active before marriage and may result in sexual initiation at an earlier age (National Poverty and Adolescent Health Research Council and Institute of Medicine 2005). Poverty and inadequate health systems compound adolescents' Young women typically make the transition to marriage and vulnerability to sickness and early death. At the same time, parenthood at an earlier age than young men, and early mar- poor health exacerbates poverty by disrupting and cutting riage predisposes girls to HIV infection through unprotected short school opportunities, by weakening or killing young sex, because the partner, by virtue of age, has an elevated risk of people in the prime of their working lives, or by placing heavy being HIV positive. In addition, marriage changes adolescent financial and social burdens on families. girls' support systems, thereby limiting their access to knowl- Poor adolescents bear a disproportionate burden of the edge about HIV/AIDS (Bruce and Clark 2003). health problems in their age group. An analysis of data from All these key transitions to adulthood bring with them the demographic and health surveys (Macro International potential for risks to health that may have both immediate and 1990­98, unpublished raw data) indicates a strong association longer-term effects. Among young women age 15 to 29, illnesses between poverty and the health status of adolescents and related to pregnancy and childbearing account for 16 percent between poverty and adolescents' use of health services. For of their DALYs. Some have unwanted pregnancies, and in example, the poorest 20 percent of young women are between countries where abortion is legally restricted, unsafe abortion 1.7 and 4.0 times as likely to have an early birth as the richest is an important source of mortality and morbidity for young 20 percent of young women. Similar disparities between rich women, with abortion complications accounting for almost and poor adolescents are seen for indicators such as early mar- 3 percent of DALYs worldwide among females age 15 to 29. riage, skilled attendance at birth, use of contraception, and (WHO 2004c). knowledge of HIV/AIDS transmission, and these disparities Even though adolescent childbearing has declined in recent tend to be greater for adolescents than for older women. For years, the proportion of young women who become mothers example, surveys in 45 countries show that the poorest 20 per- during adolescence remains high in most LMICs, and very cent of women age 15 to 49 have a total fertility rate almost early childbearing remains an issue in some regions (table 59.2). double that of the richest 20 percent, whereas among adoles- Childbearing before age 16 also brings with it a high risk of cents age 15 to 19, total fertility among the poorest 20 percent health consequences, both for the mother and for the newborn is more than triple that of the richest 20 percent (Macro (Save the Children U.S.A. 2004; WHO forthcoming-b). International 1990­98, unpublished raw data). 1112 | Disease Control Priorities in Developing Countries | Elizabeth Lule, James E. Rosen, Susheela Singh, and others Table 59.2 Indicators of Sexual and Reproductive Behaviors among Adolescents and Youth by Gender and Age Group, Late 1990s to Early 2000s a. Sexual Activity Percentage of females age 20­24 who Percentage of males age 20­24 who became sexually active before age became sexually active before age Region 15 18 20 15 18 20 East and Southern Africa 17 57 77 14 45 65 West and Central Africa 21 59 77 12 40 61 Caribbean and Central America 13 44 62 31 70 84 South America 9 41 61 31 73 87 Former Soviet Asia 1 20 53 -- -- -- Middle East -- -- -- -- -- -- South and Southeast Asia -- -- -- -- -- -- b. Marriage Percentage of females age Percentage of males age 20­24 who married before age 20­24 who married before age Percentage of men age Region 18 20 18 20 20­24 ever married East and Southern Africa 37 55 14 28 West and Central Africa 45 60 12 27 Caribbean and Central America 35 53 22 38 South America 23 38 14 29 Former Soviet Asia 16 50 -- 24 Middle East 23 40 -- 17 South and Southeast Asia 42 60 -- 41 c. Childbearing Percentage of females age 20­24 Percentage of males who whohad a child before age ever fathered a child at age Region 16 18 15­19 20­24 East and Southern Africa 9 27 2 24 West and Central Africa 13 31 2 13 Caribbean and Central America 7 22 2 27 South America 4 16 3 23 Former Soviet Asia 0 4 -- -- Middle East 3 11 -- -- South and Southeast Asia 9 24 -- -- d. Contraceptive Use Percentage of sexually active females Percentage of sexually active females age 15­19 using contraception age 20­24 using contraception Region All Unmarried All Unmarried East and Southern Africa 21 28 30 42 West and Central Africa 20 26 23 35 Caribbean and Central America 24 -- 36 -- South America 28 38 33 59 Former Soviet Asia 25 -- 43 -- Middle East -- -- -- -- South and Southeast Asia -- -- -- -- Source: National Research Council 2005. -- not available. Adolescent Health Programs | 1113 INTERVENTIONS such as tobacco and drug abuse. The lack of age-specific data and traditional reliance on mortality and morbidity statistics Improving the health of young people is complex and difficult, contribute to the unbalanced attention. Another factor may be arguably more so than for other age groups. Compared with that such behaviors tend to have longer-term health conse- children, adolescents are less protected by their families and quences that are not reflected in standard DALY calculations. communities and less amenable to simple solutions to their An additional reason for the imbalance in adolescent data may health problems, many of which are behavior based. Compared be the significant social impacts of sexual and reproductive with adults, adolescents know less about how to stay healthy behavior (for example, the contribution of high fertility to and have fewer resources to prevent or treat health problems. rapid population growth); the social and economic implica- By contrast, their behaviors are less firmly entrenched, and they tions of large proportions of HIV-infected adolescents in many are often involved in institutional activities, such as schools, countries; the mortality implications of initiating tobacco use training programs, and the military, where programs with high during adolescence; and the antisocial behavior associated with coverage can be sustained. The influences on young people's substance abuse. Public health systems' efforts to address health behaviors are becoming better understood (Blum and Mmari problems associated with road safety, malaria, and mental 2004; Pitts and others 2004), but even given what is known health have devoted inadequate attention to developing and about such influences, the challenge of designing interventions implementing programs that target adolescents. to reinforce protective factors and mitigate risk factors remains. Many of the factors associated with less risky health behaviors, such as family connectedness and academic performance, go Evidence on Sexual and Reproductive Health Interventions far beyond the purview of health program managers. Two recent reviews summarize research on the effectiveness of Programs will have to seek multisectoral solutions that link adolescent sexual and reproductive health interventions. health sector interventions with other types of interventions Table 59.3 presents the results of the FOCUS on Young Adults delivered through other sectors, either at the program level or (2001) review. The FOCUS report looks at interventions in at the policy level. The difficulty in attributing improvements LMICs and is based on relatively rigorous evaluations of 40 in health outcomes among adolescents to interventions deliv- programs. The general findings from the FOCUS review, sup- ered in multiple settings or sectors reflects the challenges plemented by more recent research findings, are as follows: involved. Programs aimed at young people are relatively new and untested. Nevertheless, accumulated experience, backed by · Almost all programs are effective in promoting positive knowl- an increasing body of research, has created international con- edge and attitudes. Almost all the rigorously evaluated pro- sensus around a multi-intervention approach centered on the grams that FOCUS reviewed improved knowledge of repro- following (WHO 1999): ductive health and selected attitudes. · Most programs effectively influence behaviors. A majority of · Young people need information and skills to make the right programs significantly changed at least one important ado- decisions about behaviors that affect their health, such as lescent behavior pertaining to reproductive health. Often, whether and when to have sex and whether to use tobacco. however, programs tried and failed to improve many impor- · Young people need access to a broad range of health services tant behaviors. Where research has found programs to be that give them the means to act on their knowledge, includ- effective in changing behaviors, such changes have typically ing access to condoms. not been large. · Young people need a social, legal, and regulatory environ- · All six categories of interventions studied proved effective at ment that supports healthy behaviors and protects them influencing reproductive behavior in at least one study. How- from harm. ever, maintaining that certain models are more effective than others is impossible because the period of observation Interventions to improve adolescent health have typically and the behaviors that were influenced varied by study. reflected this consensus and are echoed in goals that have been Moreover, further replications in multiple settings are nec- adopted internationally. essary to provide a basis for identifying the key features or This section summarizes what is known about the effective- elements of successful interventions. ness of such interventions. Note that the health community's · The evidence base is limited in a number of ways. Few studies consideration of adolescent health has only occasionally advo- look at the effects on ultimate health outcomes, such as cated attention to those health conditions that are of relatively pregnancy rates or rates of HIV infection. Furthermore, greater importance to the adolescent age group, at least as many promising approaches have not been rigorously eval- measured by indicators such as DALYS. Much of the focus has uated. Only a few studies assessed effects on the use of been on sexual and reproductive health and on risky behaviors health services, and none examined the effect of creating a 1114 | Disease Control Priorities in Developing Countries | Elizabeth Lule, James E. Rosen, Susheela Singh, and others Table 59.3 Effectiveness of Adolescent Sexual and Reproductive Health Programs, LMICs Number of programs showing significant impact/Total number of programs studied Impact on key behaviors Improved Reduced Increased Increased Improved at Number of knowledge Delayed number of contraceptive service least one Type of program studiesa and attitudes sex partners use use behavior All programs 40 33/36 6/17 6/10 18/23 4/9 23/30 School 21 17/19 4/11 3/6 6/10 1/3 9/14 Mass media 6 5/6 1/4 2/3 5/5 1/2 5/5 Community 4 4/4 1/1 -- 4/4 -- 4/4 Youth development (1) 1/1 -- -- 1/1 1/1 1/1 Peer education (3) 3/3 1/1 -- 3/3 -- 3/3 Workplace 4 4/4 -- -- 2/2 -- 2/2 Health facility 4 2/2 -- -- 0/1 2/3 2/4 Youth-friendly services (3) 1/1 -- -- -- 2/3 2/3 Youth center (1) 1/1 -- -- 0/1 -- 0/1 Multicomponent 1 1/1 0/1 1/1 1/1 0/1 1/1 Sources: Multicomponent program: AMREF, LSHTM, and NIMR 2003; all other programs: FOCUS on Young Adults 2001. -- not available. a. Numbers of studies in parentheses are subsets. supportive environment on behaviors. Furthermore, much In synthesizing U.S. and international data, Kirby (2003) of the available evidence from strong studies is for small- finds that programs are effective with different groups of ado- scale programs that are carried out over short periods of lescents in different countries. Also, programs seem to be par- time, and little evidence is available on long-term effects on ticularly effective for adolescents who are at especially high risk behaviors (Speizer, Magnani, and Colvin 2003). A recent of negative sexual and reproductive behaviors. In addition, study in Tanzania provides the first rigorous evidence that programs do not hasten or increase sexual activity--a common the benefits of adolescent sexual and reproductive health criticism of opponents of adolescent programs. Of all the pro- programs in low-income countries can last for at least three grams that have been rigorously evaluated, none has reported years (AMREF, LSHTM, and NIMR 2003). a decrease in the age of sexual debut or an increase in sexual activity among young people. A recent effort to review the evi- Kirby's (2001) review covers roughly 70 rigorously evalu- dence on interventions for preventing HIV among young peo- ated programs in Canada and the United States. The review ple has made tentative conclusions about the effectiveness of divides programs into three categories: (a) programs that focus and subsequent support for wide-scale implementation (WHO on sexual antecedents, such as sexuality and HIV education 2004b). and clinical programs; (b) programs that focus on nonsexual On the basis of these international reviews, relatively strong antecedents; and (c) programs that incorporate both youth evidence of effectiveness on a range of outcomes has emerged development and reproductive health components. Kirby finds for the following interventions: that programs in all three categories proved effective in reduc- ing sexual risk taking, pregnancy, and childbearing among · Life-skills and health and sexuality education in schools. Well- teens. In relation to youth development programs, Kirby finds designed, well-implemented sexuality and reproductive that a type of intervention known as service learning, in which health education can provide young people with a solid foun- students work on community projects, had the strongest evi- dation of knowledge and skills to enable them to engage in dence of effectiveness. By contrast, other types of youth devel- safe and responsible sexual behavior. opment programs were not effective in improving reproductive · Peer education. Peer education programs are especially health outcomes. Programs that incorporate both youth devel- appropriate for young people who are not in school and for opment and reproductive health components were effective hard-to-reach, at-risk subsets of the youth population, over long periods. including sex workers and street children. Adolescent Health Programs | 1115 · Mass media and community mobilization. Mass media and a school. After two years, observed effects of deworming treat- community mobilization efforts that engage influential ment included fewer absences and lower dropout rates, but no adults, such as parents, teachers, community and religious effect on test scores (World Bank 2002a). The treatment also leaders, and music and sports stars, can help normalize pos- resulted in health and school participation benefits among itive adolescent behaviors and gender roles as well as direct untreated children in the same schools, as well as in neighbor- young people to appropriate health services. ing schools, suggesting that the deworming had positive · Youth development programs. Youth development programs externalities. typically address a range of key adolescent needs, including life skills, education, jobs, and psychosocial needs. U.S. pro- Tobacco. Price increases are the most effective tool for reduc- grams with a voluntary community service component have ing or deterring the use of tobacco products by young people. successfully improved key reproductive health behaviors, Studies in the United States have shown that price increases but no evidence is available for developing countries. have a greater effect on tobacco use by young people than on · Clinical health services. Although some young people seek use by older age groups (University of Illinois at Chicago care through the formal health system, many others are Health Research and Policy Centers 2001). Other interventions deterred by the often judgmental attitudes of health work- have also reduced tobacco use among young people, such as ers, particularly when seeking care and advice on matters comprehensive bans on all advertising, including bans on the related to sexuality. promotion of tobacco products and trademarks (World Bank · Social marketing. This approach involves the use of public 2002c). Programs that give young people the skills to resist peer health messages to promote healthy behaviors and the use of pressure and other social pressures to smoke have demonstrated condoms and other health products and services. Effective consistent and significant reductions or delays in adolescent programs bring products and services to places in the com- smoking. School-based programs are also more effective when munity that young people frequent, such as shops, kiosks, combined with communitywide supportive efforts. Informa- and pharmacies. tion campaigns that help young people see how the tobacco · Workplace and private sector programs. Programs that reach industry tries to manipulate their behavior through advertising young people do so at their places of work and through pri- have been highly effective in changing behavior and attitudes vate channels, such as pharmacies and for-profit medical toward smoking among young people in the United States services, where many young people prefer to seek care. (American Legacy Foundation 2002). Many successful U.S. youth development programs have a work component. Promising but Unproven Interventions Many promising adolescent-focused interventions have not Evidence of Other Adolescent Health Interventions yet been rigorously evaluated. These interventions include Data on the effectiveness of other adolescent health interven- programs aimed at providing young, newlywed couples with tions are more scattered, partly because some issues have not reproductive health information and services (Alauddin and been recognized as adolescent-specific problems that require MacLaren 1999); programs that combine livelihoods skills with youth-focused interventions. reproductive health information and services (Rosen 2001b); voluntary counseling programs on and testing for HIV Nutrition. Because anemia is a critical health problem in (YouthNet 2002); actions aimed at changing social norms such many countries, many efforts have focused on improving the as gender roles (Horizons 2004); and interventions that address iron intake of adult women. Interventions aimed at adolescent the political and social context (WHO forthcoming-a). A few girls have found that daily iron supplementation effectively studies of multipronged approaches are just becoming available lowers anemia and iron deficiency (Elder 2002; MotherCare and have shown mixed results. Findings from a four-country 2000). Obesity is rapidly becoming a serious health problem study found little or no effect of such an approach on key among adolescents in many middle-income countries and is reproductive health behaviors among adolescents (Frontiers, often also associated with loss of self-esteem among adoles- Horizons, and YouthNet 2004). By contrast, a study in Tanzania cents. A few studies show that preventing obesity is more suc- found that a multicomponent approach had a significant effect cessful among adolescents than among adults (Delisle, on key reproductive health behaviors but no effect on health Chandra-Mouli, and de Benoist 2001). outcomes (AMREF, LSHTM, and NIMR 2003). Other possibly promising efforts include suicide prevention Mass Deworming. In the Busia district of western Kenya, an programs, tuberculosis education linked with health education ongoing World Bank study is evaluating the effect on learning (interpersonally or through the mass media), and malaria outcomes of providing deworming treatment to all students in treatment programs that focus on young people. Adequate 1116 | Disease Control Priorities in Developing Countries | Elizabeth Lule, James E. Rosen, Susheela Singh, and others evaluation is available on the efficacy of programs promoting synergistic effects that promote overall economic development the use of seat belts and crash helmets through enforcement of (Birdsall, Kelley, and Sinding 2001). For example, shifts to related laws and the support of intensive publicity and infor- smaller family size and slower rates of population growth in mation campaigns, as well as on the efficacy of programs pre- East Asia appear to have played a key role in the creation of an venting alcohol use among adolescents. The evaluations are educated workforce, the accumulation of household and gov- sufficient for building confidence for investment, and they ernment savings, the rise in wages, and the spectacular growth serve as a basis for intervention design (National Research of investment in manufacturing technology. The shift to Council and Institute of Medicine 2004; WHO 2004d). smaller families that is taking place in many countries will open another window of opportunity as workers have proportion- ately fewer old and young dependents to support. COSTS AND COST-EFFECTIVENESS If societies invest in health, education, and job creation, the OF INTERVENTIONS resulting economic gains will improve their overall quality of life. Education, particularly for girls, is strongly related to Good cost studies of adolescent health programs in LMICs are reproductive behavior. In most countries, girls who are edu- rare. The reported cost of such programs varies greatly depend- cated are more likely to delay marriage and childbearing, ing on the country, type of intervention, target group, and whereas girls with less education are more likely to become so on. For example, such programs cost between US$0.03 per mothers as adolescents. Unfortunately, the causal relationships adolescent reached in a family life education radio program in involved are not clear. Are girls more likely to get married when Kenya and US$71.00 per year per adolescent reached in a they leave school, or do some girls prefer to terminate their school-based HIV prevention program in Zimbabwe. Of the schooling and marry early? The difference is critical. In the first 32 programs studied, 12 have a unit cost of less than US$10 per case, the appropriate policy response would focus on improv- year, and others have a unit cost of US$10 to US$25 ing schooling opportunities for girls. In the second case, (table 59.4). Cost estimates are available only for certain types research would need to be done first to determine the underly- of interventions, with most of the estimates being for repro- ing reasons for girls'--and their parents'--preferences for early ductive health and HIV education programs. A few studies marriage instead of additional schooling. The next step would have tried to measure cost per DALY of adolescent sexual and be to assess whether these reasons appear to reflect the girls' reproductive health interventions; however, the estimates vary (and society's) best interests and, if not, to find interventions to widely. For example, in India, a youth-focused HIV prevention address the root causes of this preference. program costs US$66.20 per DALY gained; in Honduras, the The micro approach to measuring the economic benefits of cost of a voluntary counseling and testing program aimed at interventions is to build on microeconomic estimates of direct youth is US$5,873 per DALY gained. productivity effects that can be measured in monetary terms. The estimates shown in table 59.4 should be interpreted For other effects that cannot readily be translated into monetary with great caution. Comparing costs across types of programs terms, analysts can use the cost of the most cost-effective alter- and countries is difficult, but so is choosing comparable effec- native to achieve the same effects. Knowles and Behrman tiveness measures.3,4 (2003a) use this approach to estimate the benefits of various youth-focused investments, including in health. They summa- rize the three types of effects: (a) those that can be directly valued ECONOMIC ANALYSIS in monetary terms, (b) those that may require indirect valua- tion, and (c) those that are particularly difficulty to monetize. Economic analysis of adolescent health programs can pro- Table 59.5 presents examples of these effects. vide important information on their value relative to other interventions. Cost-Benefit Analysis Cost-benefit analysis is well suited to the economic analysis of Economic Benefits of Interventions projects aimed at youth, in part because many investments in The macro approach to measuring the economic benefits of young people yield multiple benefits, such as additional school- interventions is to define the benefits of investing in youth in ing and improved health. Finding any effectiveness measure terms of the investments' effect on economic growth, which that adequately reflects the wide range of benefits obtained typically is measured in terms of growth in gross national from some types of investments in youth is difficult, but cost- product per capita. Some research suggests that investments benefit analysis has the advantage of allowing comparisons in young people--whether in access to reproductive health across a range of interventions that may vary considerably in care, in education, or in other key facets of their lives--have terms of type and effects. Adolescent Health Programs | 1117 Table 59.4 Cost and Cost-Effectiveness of Adolescent Health Programs Number Cost per Region and served Unit cost DALY country Program type (name) per year (US$) Unit gained (US$) Source Sexual and reproductive health interventions Latin America Brazil School-based HIV/AIDS prevention program -- 0.70 Per condom -- Antunes and others 1997 Honduras School-based reproductive health program -- 10.44 Per targeted 1,323 World Bank 2002b to prevent HIV/AIDS adolescent Honduras Social marketing of condoms to adolescents 904,612 10.20 Per targeted 3,292 World Bank 2002b adolescent Honduras Symptomatic treatment of STIs -- 48.97 Per adolescent 28,306 World Bank 2002b, as treated cited in Knowles and Behrman 2003b Honduras Voluntary counseling and testing for youth 1,000 18.29 Per adolescent 5,873 World Bank 2002b, as undergoing cited in Knowles and voluntary counseling Behrman 2003b and testing Honduras Workplace information, education, and 1,000 20.88 Per worker 2,623.77 World Bank 2002b communication Mexico Community peer educators 4,000 63.64 Per active user -- Townsend and others of contraception 1987 per year Mexico Youth center 4,000 203.47 Per active user -- Townsend and others of contraception 1987 per year Peru School-based sexuality and HIV/AIDS 604 3.00 Per student reached -- Caceres and others 1994 prevention education Europe and Central Asia Hungary School-based HIV/AIDS prevention program 41,250 1.40 Per student -- Soderlund and others per year 1993 Newly School-based HIV/AIDS prevention program -- 1.33 Per student -- Forrai, personal independent reached communication, 1992 states Asia India Information, education, and communication -- 1,324 Per HIV infection 66.20 World Bank 1999b programs targeted to youth averted Sub-Saharan Africa Africa School-based HIV/AIDS prevention program -- 75­200 Average unit cost of -- UNECA 2000 (primary teacher training and school); simple materials 121­241 (secondary school) Africa School-based HIV/AIDS prevention program -- 1.40­7.90 Per student reached -- Watts and others 2000 Africa Peer education -- 8.00­10.81 Per out of school -- Kumaranayake and Watts adolescent reached 2001 Cameroon School-based HIV/AIDS prevention program 10,000 6.72 Per student reached -- Kumaranayake and del Amo 1997 Kenya Radio program delivering family life 3,354,000 0.03 Per adolescent reached -- Knowles and Behrman education (Youth Initiatives Project) 2003b 1118 | Disease Control Priorities in Developing Countries | Elizabeth Lule, James E. Rosen, Susheela Singh, and others Table 59.4 Continued Number Cost per Region and served Unit cost DALY country Program type (name) per year (US$) Unit gained (US$) Source Mozambique Community based "stepping stones" 500,000 0.30 Per person per year -- World Bank 2003 approach (Action Aid) Mozambique Voluntary counseling and testing, 11,726 18.40 Per person per year -- World Bank 2003 peer education Senegal Mulitpronged school, clinic, and -- 68,215­ Total intervention cost -- RamaRao and Diop 2003 community interventions 111,714 over a two-year period South Africa Television show, mass media 6 million 0.38 Per person per year -- World Bank 2003 campaign (Soul Buddyz) South Africa Mass media campaign (LoveLife) -- 20 million Annual budget -- World Bank 2003 Tanzania Primary school peer education 2,850 1.37 Per person per year -- World Bank 2003; Ross, (MEMA kwa Vijana) personal communication. 2003 Tanzania Secondary school peer education 16,250 24.12 Per person per year -- World Bank 2003 (School Health Education Program) (over three years) Uganda Outreach program for street children 5,000 18.50 Per person per year -- World Bank 2003 (GOAL: Baaba Project) Uganda Newsletters, radio show (Straight Talk) Nationwide 630,000 Amount spent in 2001 -- World Bank 2003 program reaching all schools Zambia Life skills for teachers and pupils -- 0.16 Additional cost of -- Knowles and Behrman (AIDS Action Program for Schools) one child-year of 2003b AIDS education Zambia School clubs, health clinics, peer 53,000 (over 2.26 Per person per year -- World Bank 2003 education (Kafue Adolescent five years) Reproductive Health Program) Zimbabwe Secondary school clubs, income 35,200 (over 8.89 Per person per year -- World Bank 2003 generation, peer education (Africare) two years) Zimbabwe Secondary school clubs, counseling, 2,000 71.00 Per person per year -- World Bank 2003 peer education (Midlands AIDS Service Organisation) Other health interventions LMICs Tax on tobacco products -- -- -- 5­17 World Bank 1999a LMICs Iron supplementation for 13- to -- 0.18 Per child per year -- Knowles and Berhman 15-year-olds 2003a Source: Authors. -- not available. One of the few cost-benefit analyses specific to adolescent the least cost of investments currently made to obtain the same health is Knowles and Behrman's (2003a) study that examines benefit; for example, the cost per birth averted in a family plan- three interventions: a program to provide iron supplementa- ning program was used to value reduced fertility. Table 59.6 tion for secondary schoolchildren, a school-based program of summarizes the findings of these cost-benefit studies, together health education to prevent HIV/AIDS, and a tobacco tax. The with estimates of the benefit-cost ratios for selected other study estimates benefits and costs over a youth's life cycle, dis- youth-targeted interventions. counted back to the age of 18. The study uses direct estimates The examples of cost-benefit studies cited here and other of benefits that could be readily estimated in monetary terms calculations of benefit-cost ratios show that health interven- (such as gains in labor productivity) and indirect estimates of tions aimed at adolescents can be good public investments; other benefits, such as reduced fertility and improved health, however, the results must be interpreted with some caution. that could not be easily monetized. The latter were estimated as For example, the relatively low benefit-cost ratio of an HIV Adolescent Health Programs | 1119 Table 59.5 Types of Effects of Adolescent Health prevention program in Honduras was for a program in a coun- Interventions Categorized According to Ease of Monetization try where HIV incidence among young people is relatively low (0.1 percent). Where the incidence is much higher, as in many Type of effect Examples of the hardest hit countries in Africa (1 percent or more), this Directly monetizable effects Enhanced labor productivity ratio would be proportionately higher. In addition, in the of investments in youth Reduced underutilization of labor Honduran study, the benefits included were limited to the pre- Increased or decreased work effort vention of HIV/AIDS and did not include other possible bene- Expanded access to risk-pooling services fits, such as increased education, reduced STIs other than HIV, Reduced age at which children achieve and reduced teen pregnancies and abortions. The Honduran a given level of schooling study also assumed that the effects of the intervention would Reduced cost of medical care not continue beyond one year; however, if they were to continue at the same level to age 29 (assuming that any decrease in the Indirectly monetizable broad Increased education effects of investments in youth effect of the intervention over time would be offset by increases Averted teen pregnancy in the incidence of HIV infection with age), the benefit-cost Averted HIV infection ratio would increase from 0.5 to 4.6. More than anything else, Averted tuberculosis infections Knowles and Behrman's estimates demonstrate the sensitivity Improved health of the benefit-cost ratios of investments in youth to wide vari- Improved nutritional status ations in key assumptions, which may be equally plausible Delayed marriage because of the limited information available on the costs and Averted abortion effects of many investments in youth. Reduced tobacco use Beyond the question of how sensitive such estimates are to the Effects that are particularly Increased social capital underlying assumptions and the context, the basic question is difficult to monetize what guidance they provide for public policy. High benefit-cost Averted infertility ratios certainly point to areas that merit further consideration for Averted social exclusion possible policy interventions, but they do not indicate whether Improved self-esteem using public resources for interventions has an efficiency ration- Enhanced national security (an effect of ale, because they generally do not identify differences between military training) private and social benefit-cost ratios.If the purely private benefit- Source: Adapted from Knowles and Behrman 2003a. cost ratios for an investment are high, then presumably incentives to use private resources for this investment are high, but an efficiency rationale for using public resources does not exist unless the social benefit-cost ratio exceeds the private one Table 59.6 Estimated Benefit-Cost Ratios, Selected Investments in Youth because of factors such as spillovers or market imperfections. High benefit-cost ratios that do not distinguish between social Estimated Plausible and private returns, therefore, call for further investigation. benefit-cost range of Interventions may warrant the use of public resources on effi- ratio (assuming estimated ciency grounds, but they also must answer that important 3 percent annual benefit-cost Investment discount rate) ratio question of whether the benefits are social or private. Scholarship program (Colombia) 4.4 2.8­25.6 Adult basic education and 27.6 8.1­1,764.0 literacy program (Colombia) PROGRAM IMPLEMENTATION AND LESSONS School-based reproductive 0.5 0.1­4.6 OF EXPERIENCE health program to prevent HIV/AIDS (Honduras) Relatively few adolescent-focused programs have been tried on a large scale. Sexual and reproductive health interventions and Iron supplementation 45.2 25.8­45.2 administered to suicide prevention are some of the few that have gone to scale, secondary schoolchildren and even in those areas, large-scale interventions are relatively (hypothetical low-income infrequent. The vast majority of interventions have been in rel- country) atively small programs, often through nongovernmental Tobacco tax (hypothetical 20.2 7.0­38.6 organizations. middle-income country) Perhaps the main lesson learned from the experience to date Source: Adapted from Knowles and Behrman 2003a. is an obvious one: programs to reach young people are not 1120 | Disease Control Priorities in Developing Countries | Elizabeth Lule, James E. Rosen, Susheela Singh, and others simply programs for adults applied to a younger population; involvement in the community can either increase (risk fac- they require different thinking and a different approach. tor) or decrease (protective factor) the chances that a young person will engage in unhealthy behaviors. Determining Key Principles of Health Programming for Adolescents Making a Difference on a Large Scale Experience to date suggests that effective, youth-focused efforts Adolescent health programs are complex and may not be easy to share a set of common general principles. These principles scale up because of technical, management, and political chal- include the following: lenges. The following are examples of adolescent health pro- grams that are national in scope. Unfortunately, little is known · Recognize the diversity of the youth age group. A sexually about the costs and effectiveness of such large-scale efforts. inexperienced 11-year-old has vastly different needs than a married 20-year-old. Programs should apply different National Suicide Prevention Program in New Zealand. strategies to reach youth, who vary by age, sex, employment, Among industrial nations, New Zealand has one of the highest schooling, and marital status. suicide rates for both males and females age 15 to 24 (New · Involve young people. Policies and programs are more effective Zealand Ministry of Health 2002). In 1998, on the basis of inter- when young people are involved in all aspects of their design, national good practice, the government developed the National implementation, and evaluation. Involvement must go Youth Suicide Prevention Strategy. This strategy, which includes beyond tokenism and be genuine,meaningful,and sustained. a component for the general population and one that focuses · Make health services appealing to youth. A key to rapidly on the indigenous Maori community, provides a framework for expanding young people's access to health services is to understanding suicide prevention and signals the steps that make them more youth friendly by using specially trained government agencies, communities, and service providers must health workers and by bolstering the privacy, confidentiality, take to reduce suicide. Even though the national strategy has not and accessibility of care. been in place long enough to adequately gauge its effects, in · Address gender inequality. Gender inequalities expose young 1999, the first year following the adoption of the strategy, youth girls to coerced sex, HIV infection, unwanted pregnancy, suicide rates fell to their lowest levels since 1991. and poor nutrition. Efforts should focus on changing the factors that perpetuate gender inequalities. Sexuality Education in Mongolia. Mongolia has implemented · Address the needs of boys. Adolescence presents a unique a locally developed and tested sexuality education curriculum in opportunity to help boys form positive notions of gender 60 percent of schools nationwide. Current challenges include relations and to raise their awareness of health issues. At the increasing the number of hours allocated to sexuality education; same time, boys seem to be disproportionately exposed to a developing more and better written resources for adolescents, number of adolescent health risks, including accidents and including textbooks; developing materials that will help parents injuries, suicide, tobacco use, substance abuse, and violence. communicate better with their children on sexuality; expanding Program design should take into account the specific needs access to clinical services for adolescents through the public of boys and young men as well as of girls and young women. health system; reaching out-of-school youths and the broader · Design comprehensive programs. Comprehensive programs community with sexuality education; and monitoring and eval- that provide information and services while addressing the uating the program regularly to assess its weaknesses and social and political context are more effective than narrowly strengths and how it could be improved (Gerdts 2002). focused interventions. · Consider all important benefits. Many adolescent health Addressing the Health Needs of Poor Youth interventions focus on only one benefit. For example, a The following strategies, based on what is known about ser- school-based sex education program may focus exclusively vices for poor people more generally and about the specific on HIV prevention and may neglect other possible benefits needs of young people, show promise for meeting the needs of from the intervention, such as increased education, averted poor youth: teen pregnancy and abortions, and other averted STIs. · Address the many nonhealth factors that influence adolescent · Targeting out-of-school youth. Out-of-school youth of a health. Linking school and livelihood opportunities to ado- given age are likely to be more marginalized than those who lescent health programs, at either the policy or program are in school, and they are often those most in need of crit- level, is key to helping young people avoid risky behaviors. ical services, such as pregnancy prevention and prevention · Address underlying risk and protective factors. Factors such as of HIV/AIDS and other STIs. A number of countries, feelings of self-efficacy, attitudes and behaviors of friends, including Paraguay, South Africa, and Zimbabwe, have connectedness with parents and other influential adults, and launched effective programs targeting out-of-school youth Adolescent Health Programs | 1121 that combine the use of mass media, peer education, and · Strengthening the stewardship oversight function of govern- community-based efforts. For instance, the Arte y Parte ments. Governments have a key role to play in developing project targeted out-of-school youth in three cities in supportive policies, both within the health sector and across Paraguay using a booklet about adolescent sexuality, street sectors; in contributing to cross-sectoral policies such as drama, radio programming, newspaper columns, and dis- national youth policies; and in providing input into policy tribution of promotional items (Magnani and others 2000). making in other sectors, especially education and labor. · Focusing efforts on vulnerable youth. Young people who have Ideally, governments should have an overarching adolescent been orphaned or left vulnerable by AIDS typically rely first health policy with specific reference to adolescent health on their extended families and communities for support. in policy documents for specific programs or diseases, such Efforts to help vulnerable youth should strengthen those as for AIDS, tuberculosis, malaria, sexual and reproductive safety nets. One example is the COPE program in Malawi, health, and population (WHO forthcoming-a; POLICY where a nongovernmental organization­sponsored effort Project and YouthNet forthcoming). works through existing government structures to help orphans and other vulnerable children (Phiri, Foster, and Nzima 2001). RESEARCH AND DEVELOPMENT AGENDA · Tailoring subsidized programs to poor youth. Social market- ing of reproductive health products and services--such as The striking lack of good research and evaluation of adolescent contraceptives and condoms for pregnancy and disease pre- health interventions limits countries' ability to address serious vention or promotion of iron supplementation--often tar- health problems. At this juncture, research in the following gets young consumers, but such efforts should ensure that broad areas is critical: they reach the desired clients--namely, those who are poor and less likely to be able to afford market prices. The Social · Refining estimates of DALYs for adolescents. Available DALY Marketing for Adolescent Sexual Health Project in four information is inadequate to fully explore the burden of African countries combined the use of mass media with disease for adolescent age groups. Future DALY estimates peer education to encourage young people to practice safer should be made for five-year age groups in the 10 to 24 age sex, including condom use (Agha 2000). range. · Documenting the effectiveness of current approaches. This Improving Health Systems to Meet Adolescents' area includes better process evaluation to understand the Health Needs functioning of successful programs. Such evaluation neces- The shortcomings of health systems in LMICs are well known, sitates more rigorous research designs so that the effective- and adolescents in particular would benefit from the following ness of programs can be better documented, both in terms health system improvements: of health outcomes and in terms of DALYs saved. Another area in which more research could help is better documen- · Strengthening human resource capacities. The poor quality of tation of the nonhealth effects of adolescent health inter- the interaction with health workers is one of the main bar- ventions. Greater investment is also needed to evaluate the riers to adolescents' use of health care in public sector facil- effects of health promotion strategies on reducing smoking, ities. Through training, supervision, and other means, including the smoke-free spaces prevalent in the Americas health systems should encourage health workers to adopt a and life-skills education. more youth-friendly outlook. In addition, health systems · Testing new interventions. This area includes more research should integrate such training into the curricula of medical, on multicomponent programs and on new types of inter- nursing, and nurse auxiliary schools. ventions. In relation to sexual and reproductive health, new · Involving the private sector. Many young people already seek interventions include approaches such as providing anti- care from private doctors, nurses, and nurse-midwives or retroviral therapy to HIV-infected youth and voluntary from local pharmacies or other drug distribution outlets. counseling and testing for HIV, encouraging adolescents to Along with encouraging private for-profit health providers have fewer sexual partners, reducing the trafficking of young to serve youth, government policies should encourage people, preventing and addressing the health consequences efforts to tap into the private sector as a source of health care of early marriage, and reaching young married women with for adolescents by means of interventions such as social information and services. Research must better inform marketing, contracted services, youth-focused social fran- interventions so that they reach groups at particularly high chising, and programs that serve young people at their place risk of poor health outcomes, such as child prostitutes, child of work. (Carranza 2003; LaVake and Rosen 2003; Rosen workers, refugees, AIDS orphans, and street children. More 2001a; Senderowitz and Stevens 2001). research is also needed on a broad range of other adolescent 1122 | Disease Control Priorities in Developing Countries | Elizabeth Lule, James E. Rosen, Susheela Singh, and others health interventions, especially for those health problems 2. Regional data cited from this report here and elsewhere in the chap- that are among the biggest killers and disablers of young ter are based on nationally representative surveys carried out between the mid 1990s and 2001. people: HIV/AIDS and mental illnesses for both males and 3. Knowles and Behrman (2003b) find that many cost estimates incor- females, maternal conditions for females, and road accidents rectly treat income transfers as costs and frequently fail to include esti- for males. In addition, research is needed on programs that mates of administrative and distortionary costs (for example, the distor- tionary cost of financing programs through taxes). attempt to influence gender roles and social norms and 4. In cost-effectiveness analysis, as in cost-benefit analysis, estimates are investments designed to avert drug and alcohol abuse and to needed not only of what has actually happened, but also of what would improve mental health. have happened in the absence of the program or intervention (that is, an estimate of the counterfactual). There are no exceptions to this rule. · Enhancing understanding of the risk and protective factors influencing young people's behavior. Even though our under- standing of the major influences on youth behaviors has REFERENCES come far, more refinement of such understanding is needed, Agha, S. 2000. "An Evaluation of Adolescent Sexual Health Programs in along with a better understanding of how to incorporate Cameroon, Botswana, South Africa, and Guinea." Population Services such knowledge into the design of programs and policies. International Research Division Working Paper 29, Population Services International, Washington, DC. · Improving cost and cost-benefit analysis. Good cost estimates Alauddin, M., and L. MacLaren. 1999. InFOCUS: Reaching Newlywed and are rare, and more needs to be done to more fully estimate Married Adolescents. Washington, DC: FOCUS on Young Adults. the costs of the range of adolescent health interventions. American Legacy Foundation. 2002. "New American Legacy Foundation This effort means collecting more data on program costs Study Shows Truth® Campaign Helping to Drive Down Youth and more accurate data that include programs' nonmone- Smoking Rates." American Legacy Foundation, Washington, DC. tary costs. Few full cost-benefit analyses of youth programs http://www.americanlegacy.org. exist, and more need to be done to improve evaluations of AMREF (African Medical and Research Foundation), LSHTM (London School of Hygiene and Tropical Medicine), and NIMR (National the economic value of investments targeted at young people. Institute for Medical Research). 2003. "MEMA kwa Vijana: Randomised Controlled Trial of an Adolescent Sexual Health Programme in Rural Mwanza, Tanzania." Technical briefing docu- CONCLUSIONS ment, LSHTM, London, August 11. Antunes, M., R. Stall, V. Paiva, C. A. Peres, J. Paul, M. Hudes, and others. The health community has only recently recognized the impor- 1997. "Evaluating an AIDS Sexual Risk Reduction Program for Young tance of adolescent health problems. To address the unique Adults in Public Night Schools in São Paulo, Brazil." AIDS 11: S121­27. health problems associated with the adolescent years, policy Behrman, J. R., J. Hoddinott, J. A. Maluccio, A. Quisumbing, R. Martorell, and A. D. Stein. 2004. "The Impact of Experimental Nutritional makers and the health community must expand the knowledge Interventions on Education into Adulthood in Rural Guatemala: base on effectiveness, costs, and economic benefits and pay Preliminary Longitudinal Analysis." Paper presented to the 2004 more attention to areas such as road safety, nutrition, mental Population Association of America Annual Meeting, Boston, April 1­3. health, and malaria. Well-documented implementation experi- Birdsall, N., A. C. Kelley, and S. W. Sinding, eds. 2001. Population Matters: ences from mostly small programs have produced a sound Demographic Change, Economic Growth, and Poverty in the Developing World. New York: Oxford University Press. body of knowledge about how programs function. These expe- Blum, R., and K. Mmari. 2004. Risk and Protective Factors Affecting riences can provide the foundation for scaled-up efforts and Adolescent Reproductive Health in Developing Countries. An Analysis can help the health community improve health systems in ways of the World's Literature 1990­2004. Summary. Geneva: World Health that will benefit adolescent health efforts. Organization. Bruce, J., and S. Clark. 2003."Including Married Adolescents in Adolescent Reproductive Health and HIV/AIDS Policy." Paper prepared for ACKNOWLEDGMENTS technical consultation on married adolescents, World Health Organi- zation, Geneva. The authors acknowledge the generous help of the following Caceres, C. F., A. M. Rosasco, J. S. Mandel, and N. Hearst. 1994."Evaluating a School-Based Intervention for STD/AIDS Prevention in Peru." people: Peju Olukoya for sharing materials and studies and for Journal of Adolescent Health 15 (7): 582­91. providing early feedback; Kimberly Switlick for preparing Carranza, J. M. 2003. "What Do Salvadoran Teens Think? Determining the graphics and providing editorial assistance; and peer reviewers, Feasibility of Youth-Friendly Pharmacies: A Focus Group Report." U.S. including Maria Teresa Cinqueira, Jane Ferguson, Elena Agency for International Development and Commercial Market Nightingale, and Audrey Smith Rogers. Strategies Project, Washington, DC. http://www.cmsproject.com/ resources/PDF/CMS_ElSalvador_Youth.pdf. CDC (U.S. Centers for Disease Control and Prevention) and ORC Macro. NOTES 2003. Reproductive, Maternal, and Child Health in Eastern Europe and Eurasia: A Comparative Report. Atlanta: U.S. Department of Health 1. The United Nations defines youth as those age 15 to 24. The World and Human Services. Health Organization defines adolescence as age 10 through 19 and uses the Delisle, H., V. Chandra-Mouli, and B. de Benoist. 2001. "Should term young people to refer to those age 10 to 24. Adolescents Be Specifically Targeted for Nutrition in Developing Adolescent Health Programs | 1123 Countries? To Address Which Problems, and How?" http://www. Paper presented at the Annual Meeting of the Population Association who.int/child-adolescent-health/New_Publications/NUTRITION/ of America, Minneapolis, May 1­3. Adolescent_nutrition_paper.pdf. MotherCare. 2000. "Anemia and Iron Deficiency in Adolescent Students Doak, C., L. Adair, C. Monteiro, and B. M. Popkin. 2000. "Overweight and in Lima, Peru: Causes, Consequences, and Prevention." Reproductive Underweight Co-exists in Brazil, China, and Russia." Journal of Health Focus 14. http://www.jsi.com/intl/mothercare/rhfocus.htm. Nutrition 130: 2965­80. National Research Council and Institute of Medicine. 2004. Reducing Elder, L. 2002. "Adolescent Nutrition: Issues and Interventions." Underage Drinking: A Collective Responsibility. Washington, DC: Background paper prepared for the World Bank Learning Exchange on National Academies Press. Exploring Strategies for Reaching and Working with Adolescents, ------. 2005. Growing up Global: The Changing Transitions to Adulthood Washington, DC, June 5. in Developing Countries. Washington, DC: National Academies Press. FOCUS onYoung Adults.2001.AdvancingYoung Adult Reproductive Health: Phiri, S., G. Foster, and M. Nzima. 2001. Expanding and Strengthening Actions for the Next Decade. Washington, DC: FOCUS on Young Adults. Community Action: A Study of Ways to Scale Up Community Frontiers, Horizons, and YouthNet. 2004. "New Findings from Mobilization Interventions to Mitigate the Effect of HIV/AIDS on Intervention Research: Youth Reproductive Health and HIV Children and Families. Washington, DC: Displaced Children and Prevention." Meeting report, Washington, DC, September 9, 2003. Orphans Fund. http://www.usaid.gov/pop_health/dcofwvf/reports/ orphanreps/dcaction.html. Gerdts, C. 2002. "Universal Sexuality Education in Mongolia: Educating Pitts, M., G. Dowsett, M. Couch, D. Keys, and S. Dutertre. 2004. "Looking Today to Protect Tomorrow." Quality 12, Population Council, New for More: A Review of Social and Contextual Factors Affecting Young York. http://www.popcouncil.org/publications/qcq/QCQ12.pdf. People's Sexual Health." Document prepared for the World Health Hoddinott, J., and A. Quisumbing. 2003. "Investing in Children and Youth Organization, Department of Child and Adolescent Health and for Poverty Reduction." Unpublished paper, International Food Policy Development. La Trobe University, Melbourne, Australia. Research Institute, Washington, DC. Piwoz, E., and E. Greble. 2000. HIV/AIDS and Nutrition: A Review of the Horizons. 2004. "Promoting Healthy Relationships and HIV/STI Literature and Recommendations for Nutritional Care and Support in Prevention for Young Men: Positive Findings from an Interven- Sub-Saharan Africa. Washington, DC: USAID Support for Analysis and tion Study in Brazil." Population Council's Research Update Research in Africa (SARA) Project and the Academy for Educational series, Population Council and Horizons Communications Development. Unit, Washington, DC. http://www.popcouncil.org/pdfs/horizons/ POLICY Project and YouthNet. Forthcoming. Guide to Key Elements of brgndrnrmsru.pdf. Youth Reproductive Health Policy. Washington, DC: POLICY Project; Kirby, D. 2001. Emerging Answers: Research Findings on Programs to Reduce Arlington, VA: YouthNet. Teen Pregnancy. Washington, DC: National Campaign to Prevent Teen Rosen, J. 2001a. "A Dialogue on Social Marketing and Other Commercial Pregnancy. Approaches to Improving Adolescent Reproductive Health." Meeting ------. 2003. "Changing Youth Behaviors: Findings from U.S. and report. FOCUS on Young Adults, Washington, DC, February 15. Developing Country Research and Their Implications for A, B, and ------. 2001b. In FOCUS: Youth Livelihoods and HIV/AIDS. Washington, C." Paper presented at the meeting on HIV Prevention for Young DC: FOCUS on Young Adults. People in Developing Countries, Washington, DC, July 24. http://www. RamaRao, S., and N. J. Diop. 2003. "Serving the Reproductive Health fhi.org/en/Youth/YouthNet/NewsEvents/HIVprevenmeeting.htm. Needs of Adolescents in Senegal: Analysis of Costs." Frontiers Project Knowles, J. C., and J. R. Behrman. 2003a. "Assessing the Economic Benefits Report. Population Council, Washington, DC. of Investing in Youth in Developing Countries." Health, Nutrition, and Save the Children U.S.A. 2004. Children Having Children: State of the Population Discussion Paper, World Bank, Washington, DC. World's Mothers 2004. Westport, CT: Save the Children U.S.A. ------. 2003b. "The Economic Returns to Investing in Youth in Senderowitz, J., and C. Stevens. 2001. Leveraging the For-Profit Sector in Developing Countries: A Review of the Literature." Unpublished Support of Adolescent and Young Adult Reproductive Health Program- paper. World Bank, Washington, DC. ming. Washington, DC: Futures Institute for Sustainable Development. Kumaranayake, L., and J. del Amo. 1997. Resource Allocation for HIV Soderlund, N., J. Lavis, J. Broomberg, and A. Mills. 1993. "The Costs of Prevention: Cost, Epidemiological, and Behavioural Analysis-- HIV Prevention Strategies in Developing Countries." Bulletin of the HIV/AIDS Education in Schools. London: London School of Hygiene World Health Organization 71 (5): 595­604. and Tropical Medicine. Speizer, I. S., R. J. Magnani, and C. Colvin. 2003. "The Effectiveness of Kumaranayake, L., and C. Watts. 2001. Scaling-Up Priority HIV/AIDS Adolescent Reproductive Health Interventions in Developing Interventions: A Problem of Constrained Optimisation. London: London Countries: A Review of the Evidence." Journal of Adolescent Health 33 School of Hygiene and Tropical Medicine, Department of Public (5): 324­48. Health and Policy Publication. Townsend, J. W., E. Dias de May, Y. Sepulveda Santos de Garza, and S. LaVake, S., and J. Rosen. 2003. Private Sector Country Assessment Manual, Rosenhouse. 1987. "Sex Education and Family Planning Services for 2003: A Handbook for Assessing the Potential for Youth Reproductive Young Adults: Urban Strategies in Mexico." Studies in Family Planning Health and HIV/AIDS Interventions in the Private Sector. Arlington, VA: 18 (2): 103­8. YouthNet. UNAIDS (Joint United Nations Programme on HIV/AIDS). 2003. AIDS MacKay, J., and G. A. Mensah. 2004. The Atlas of Heart Disease and Stroke. Epidemic Update: December 2003. Geneva: UNAIDS. Geneva: World Health Organization. UNECA (United Nations Economic Commission for Africa). 2000. Costs Magnani, R., A. Robinson, E. Seiber, and G. Avila. 2000. Evaluation of "Arte of Scaling HIV Programme Activities to a National Level in Sub-Saharan y Parte": An Adolescent Reproductive Health Communications Project Africa: Methods and Estimates. http://www.uneca.org/adf2000/ Implemented in Asunción, San Lorenzo, and Fernando de la Mora, costsofaids.htm. Paraguay. Washington, DC: FOCUS on Young Adults. University of Illinois at Chicago Health Research and Policy Centers. 2001. Mensch, B. S., S. Singh, and J. Casterline. 2003. "Trends in the Timing of Cigarette Taxes and Kids. Policy Briefs, vol. 1, April. Chicago: Health First Marriage among Men and Women in the Developing World." Research and Policy Centers, University of Illinois at Chicago. 1124 | Disease Control Priorities in Developing Countries | Elizabeth Lule, James E. Rosen, Susheela Singh, and others Watts, C., P. Vickerman, L. Kumaranayake, C. Cheta, C. C. Nama, G. ------. 2004d. World Report on Road Traffic Injury Prevention: Summary. Kwenthieu, and J. Del Amo. 2000. "Impact and Cost-Effectiveness Geneva: WHO. Modelling of In-School Youth Programmes in Sub-Saharan Africa." ------. Forthcoming-a. Making Policy Happen: Lessons from Countries on Paper presented at the 13th International AIDS Conference, Durban, Developing National Adolescent Health and Development Policy. South Africa, July 9­14. Geneva: WHO, Department of Child and Adolescent Health and Westoff, C. F., and A. Bankole. 1995. Unmet Need: 1990­1994. Comparative Development. Studies 16, Demographic and Health Surveys. Calverton, MD: Macro ------. Forthcoming-b. Pregnant Adolescents: Delivering on Global International. Promises of Hope. Geneva: WHO. WHO (World Health Organization). 1999. Programming for Adolescent World Bank. 1999a. Curbing the Epidemic: Governments and the Economics Health and Development: Report of a WHO/UNFPA/UNICEF Study of Tobacco Control. Washington, DC: World Bank. Group on Programming for Adolescent Health. Technical Report 886. Geneva: WHO. ------. 1999b. Project Appraisal Document on a Proposed Credit in the Amount of SDR 140.82 Million to India for a Second National HIV/AIDS ------. 2001a. Global Status Report: Alcohol and Young People. Control Project. Report 18918-IN, Washington, DC: World Bank. WHO/MSD/MSB/01.1. Geneva: WHO. ------. 2002a. "The Impact of Deworming Treatment on Primary School ------. 2001b. The Second Decade: Improving Adolescent Health and Performance in Busia, Kenya." Abstract of Current Research. http:// Development. Geneva: WHO. www.worldbank.org/research. ------. 2002. World Health Report. Geneva: WHO. ------. 2002b. Optimizing the Allocation of Resources among HIV ------. 2004a. "Estimates of DALYs by Sex, Cause, and WHO Mortality Prevention Interventions in Honduras. Washington, DC: World Bank. Subregion, Estimates for 2002, 2003." WHO, Geneva. http://www3. ------. 2002c. Tobacco at a Glance. Washington, DC: World Bank. who.int/whosis/menu.cfm?path=evidence,burden,burden_ ------. 2003. Education and HIV/AIDS: A Sourcebook of HIV/AIDS estimates_2002,burden_estimates_2002_subregion&language= Prevention Programs. Washington, DC: World Bank. english. YouthNet. 2002. "VCT and Young People." YouthLens. http://www.fhi.org/ ------. 2004b. "Steady, Ready GO!" Information brief on the Talloire en/youth/youthnet/publications/youthlens+english.htm. consultation to review the evidence for policies and programmes to achieve the global goals on young people and HIV/AIDS. WHO, Geneva. ------. 2004c. Unsafe Abortion: Global and Regional Estimates of Incidence of Unsafe Abortion and Associated Mortality in 2000. 4th ed. Geneva: WHO. Adolescent Health Programs | 1125 Chapter 60 Occupational Health Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut Workers around the world--despite vast differences in their health in the workplace is integrated into all aspects of daily life physical, social, economic, and political environments--face for these often subsistence agricultural workers. For example, virtually the same kinds of workplace hazards. These hazards pesticide poisoning is a hazard faced by workers and their fam- are traditionally categorized into four broad types: chemical, ilies and communities. biological, physical, and psychosocial. What emerges from our The informal workforce, which in industrial countries is incomplete knowledge of their risk, however, is that the more rarely larger than 10 percent of total employment, looms large than 80 percent of the world's workforce that resides in the in developing countries. This workforce includes self- developing world disproportionately shares in the global bur- employed, household-based unpaid labor (family members, den of occupational disease and injury. Several classic occupa- for example) and independent service workers such as street tional diseases, such as silicosis and lead poisoning, that have vendors. In the developing world, employment in the informal been substantially eliminated in industrial countries remain sector may reach 70 percent, with the contribution to the gross endemic elsewhere in the world. Whether this high and pre- domestic product (GDP) ranging from 10 to 60 percent ventable burden of ill health faced by workers in the develop- (ILO 2002). ing world is the result of ignorance, inattention, or intent, com- Informal economy workers are often unprotected in the reg- pelling evidence indicates that work-related health conditions ulatory arena even in the industrial world. This circumstance could be substantially reduced, often at modest cost. is exacerbated when the vulnerable employment status in the developing world is coupled with problems of poverty and ill health. Cottage-industry workers abound in the informal sec- NATURE AND CAUSES OF OCCUPATIONAL tor, and home-based work can fully blur distinctions between HEALTH CONDITIONS IN THE DEVELOPING occupational and other environmental hazards. Not uncom- WORLD mon across the developing world are lead-poisoned adults who manufacture batteries in crude facilities at home and their Despite country-to-country differences, some commonalities lead-poisoned children, exposed to the lead while sleeping and exist within the workforce of the developing world that are playing in the next room. worth noting. Workforce distribution by economic sector is The migrant workforce, which is increasing worldwide, is different from that in the industrial world. Compared with estimated to be 120 million (ILO 2000). In the industrial world, industrial countries, where single-digit percentages prevail-- immigrant workers often perform work deemed unattractive for example, approximately 2 percent in the United (seasonal agricultural work in the United States, service sector Kingdom--developing countries employ about 70 percent of work in the United Kingdom), but the issues of a migrant their economically active population in the agricultural sector workforce in some parts of the developing world take on even (World Bank 2003). For many of these workers, the distinction greater import. In southern Africa, for example, migrant min- between health at work and health at home is blurred, because ing workers face the extraordinary burden of risk for the triad 1127 of silicosis, tuberculosis, and HIV/AIDS--diseases that are would be reduced by 90 percent; even if the smoking exposure inextricably linked to interactive determinants of workplace, continued, elimination of the asbestos exposure would reduce housing, and social and economic factors (Trapido and others the overall risk by 80 percent. Those considerations are not the- 1996). oretical but well supported by empirical data. In parts of China Workers in the developing world face different risks in the and elsewhere in the developing world, asbestos exposure health transition than do their counterparts in the industrial abounds as cigarette smoking is rising. Effective intervention world. They may be exposed to the combined and often syner- strategies will be those based on a comprehensive approach to gistic risks of both traditional and emerging hazards. Workers the overall burden rather than those addressing the individual may also face unregulated and unprotected exposures to burdens of specific exposures, recognizing that organizational known hazards just as those same hazards--silica and asbestos, or institutional interventions (such as eliminating asbestos for example--were faced decades ago by millions of workers in from the workplace) are far more effective than those targeting the industrial world. A significant difference, though, is that individual behaviors (such as smoking cessation). workers in the developing world are being exposed when wide- spread knowledge is available about the risks and effective pre- ventive measures (Kjellstrom and Rosenstock 1990). Even as GLOBAL BURDEN OF DISEASE FROM these workers are forced to replay history, despite the availabil- OCCUPATIONAL HEALTH RISKS ity of information and knowledge transfer unthinkable just a generation ago, they face other hazards, including the produc- The overall picture that emerges from all parts of the develop- tion, marketing, and importation of environmental hazards ing world is one of increased health and safety risks in all such as cigarettes. In the instance of asbestos and tobacco, both occupations for which data are available. products are being aggressively marketed and exported by the Dramatic changes in the global labor force will occur as industrial world (especially asbestos from Canada and tobacco globalization and population growth continue to affect the from the United States) to the developing world. global economy. For example, the labor force in Latin America A real example of hazards faced by developing workers in and the Caribbean is one of the fastest growing in the world, what might be called the risk transition is that posed by dual with 217 million workers in 2000; the number of workers is exposure to asbestos and cigarette smoke and risk for lung can- expected to reach 270 million in 2010 (PAHO 2002). The bur- cer. This example is especially troubling not only because the den of disease and injury attributable to workplace risks in the risk is dauntingly high but also because exposures to both are formal and informal sectors is grave and will continue to rise. occurring with full knowledge of their individual and cumula- Inadequate data and reporting systems make capturing the tive effects. As shown in table 60.1, against a background of rel- effect of workplace risks problematic. Nonetheless, several ative risk for lung cancer of 1 for a nonsmoking, nonasbestos- recent efforts by international bodies have shed some light on exposed population, a working population with significant the staggering burden, although in general attempts to derive asbestos exposure but no tobacco exposure may face a relative evidence-based estimates are likely to systematically and signif- risk of lung cancer of 5; a smoking population not exposed to icantly underrepresent the extent of the problem. asbestos faces a relative risk of 10; and rather than these risks The gravity of workplace risks is seen in the recent being additive (that is, 15) the smoking, asbestos-exposed pop- International Labour Organization (ILO) estimate that among ulation has the extraordinary relative risk of lung cancer of 50. the world's 2.7 billion workers, at least 2 million deaths per year Most important, in this well-recognized multiplicative-effect are attributable to occupational diseases and injuries. The ILO scenario, if the smoking exposure alone were eliminated among estimates for fatalities are the tip of the iceberg because data for the asbestos-exposed workers, the overall risk of lung cancer estimating nonfatal illness and injury are not available for most of the globe. The ILO also notes that about 4 percent of the GDP is lost because of work-related diseases and injuries Table 60.1 Relationship between Asbestos, Smoking, and (Takala 2002). Risk for Lung Cancer A recent effort of the World Health Organization (WHO) has provided insight into the global dimensions of several Asbestos exposurea selected occupational health risks. WHO included five occupa- Smokingb No Yes tional risk factors in its comparative risk assessment in a uni- No 1 5 fied framework of 26 major health risk factors contributing to Yes 10 50 the overall global burden of disease and injury (Ezzati and others 2004; WHO 2002). The WHO comparative approach Source: Kjellstrom and Rosenstock 1990. used a common statistical model that allows a reader to com- a. If asbestos exposure eliminated, eliminate 80 percent lung cancers in asbestos workers. b. If smoking eliminated, eliminate 90 percent lung cancers in asbestos workers. pare the contribution (attributable fraction) of several risk 1128 | Disease Control Priorities in Developing Countries | Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut factors to a single outcome--lung cancer, for example. The absence of data in much of the developing world Stringent requirements for consistency in describing risk fac- limited the range of occupational risk factors that WHO could tors limited the number of occupational risk factors that could measure, and the available data excluded children under age be included in the study. For all risk factors, it was necessary to 15 who work. The WHO comparative risk assessment also estimate an exposed population and exposure levels for 224 excluded important occupational risks for reproductive disor- age, sex, and country groups in the 14 WHO geographic ders, dermatitis, infectious disease, coronary heart disease, regions of the world. Where possible, data could be extrapo- intentional injuries, musculoskeletal disorders of the upper lated to age, sex, and country groups for which data were not extremities, and most cancers. Psychosocial risk factors such as available, based on similarities in demographic, socioeco- workplace stress could not be studied, nor could pesticide, nomic, or other relevant indicators. Because knowing the exist- heavy metal, or solvent exposures. The potential consequences ing burden of disease and injury globally was necessary, the of omitting just pesticides from the global burden analysis can only outcomes included were those for which WHO had rates be illustrated by the situation in Central America (PAHO of disease or injury for all regions calculated by International 2002). The isthmus is primarily an agricultural and forested Classification of Disease (ICD) codes. Finally, estimates of the area of .5 million square kilometers, of which 40 percent is risk factor­burden relationships by age, sex, and WHO subre- arable. Pesticide imports almost tripled from 15,000 metric gion were generated. Risk measures (relative risks or mortality tons in 1992 to 41,000 in 1998, and 35 percent of the pesticides rates) for the health outcomes resulting from exposure to the were restricted in the exporting countries. In 2000, the subre- risk factors were determined primarily from studies published gion imported some 1.5 kilograms of pesticides per inhabitant in peer-reviewed journals. Adjustments were made to account per year, a quantity 2.5 times greater than the world average for differences in levels of exposure; exposure duration; and estimated by WHO. Exposures in the formal and informal sec- age, sex, and subregion, as appropriate. The information about tors extend to the homes and families of the pesticide workers. each risk factor was entered into the WHO common model for Although this situation is common in developing nations, the comparative analysis. The resulting burden was described as WHO comparative risk assessment captured none of these the attributable fraction of disease or injury, using mortality exposures. and disability-adjusted life years (DALYs) lost, with one DALY The ILO and WHO data provide the most current, yet still being equal to the loss of one healthy life year--the common incomplete, picture of the overall problem of occupational currency measure that includes mortality and morbidity. health risks. Nonetheless, with just the few occupational risk Because of the requirements for global data, only five occu- factors studied in depth by WHO a picture emerges of the sig- pational risk factors could be described: risks for injuries, car- nificant effect of largely preventable conditions (Ezzati and cinogens, airborne particulates, ergonomic risks for back pain, others 2004). WHO found that occupational injuries result in and noise. The exposed worker populations were estimated about 312,000 deaths per year for the world's 2.7 billion work- using an approach based on the International Standard ers; this figure contrasts to the approximately 6,000 deaths per Industrial Classification of All Economic Activities (ISIC), an year caused by occupational injuries for the 150 million work- economic classification system of the United Nations that ers in the United States. As in the industrial world, high injury organizes all economic activities by economic sectors and rele- fatality rates in the developing world are clustered in certain vant subgroupings (UN 2000). The ISIC system is used almost sectors, including agriculture, construction, and mining. universally by national and international statistical services to Using this metric, occupational injuries account for more than categorize economic activity; therefore, it allows global com- 10 million DALYs and 8 percent of unintentional injuries parisons. The ILO has developed economically active popula- worldwide. tion (EAP) estimates by applying economic activity rates, by The second occupational factor WHO analyzed was the sex and by age group (older than age 15), to the population effect of exposure to workplace lung carcinogens (such as estimates and projections of the United Nations (ILO 1996). asbestos, diesel exhaust, and silica) and leukemogens (such as The EAP provides the most comprehensive global accounting benzene, ionizing radiation, and ethylene oxide). WHO con- of people who may be exposed to occupational risks because it cluded that occupational exposures account for about 9 per- includes people in paid employment, the self-employed, and cent of all cancers of the lung, trachea, and bronchus and about people who work to produce goods and services for their own 2 percent of all leukemias. Overall, about 102,000 deaths were household consumption, both in the formal and in the infor- attributable to these two occupational cancers and about 1 mil- mal sectors (ILO 2002). For the WHO comparative risk assess- lion DALYs. ment, the EAP was further divided into nine economic subsec- Estimates of the global burden of chronic lung disease tors (where people work) and seven occupational categories demonstrate the significant contribution of occupational expo- (what type of work people do), on the basis of country-level sures, which account for about 13 percent of all chronic data for 31 countries (ILO 1995). obstructive pulmonary disease (COPD) and about 11 percent Occupational Health | 1129 of asthma. In total, WHO found the annual worldwide burden countries that have engaged in serious prevention efforts. of work-related COPD to be about 318,000 deaths per year and Proper needle handling and waste management, substitutions about 3.7 million DALYs. The occupational risk contribution for sharps, hepatitis B virus (HBV) immunization, postexpo- to the worldwide asthma burden was about 38,000 deaths and sure prophylaxis, training, and legislative measures have been about 1.6 million DALYs, reflecting the fact that a great deal of successful. Beyond the personal and workplace consequences, asthma occurs at younger ages and is not fatal. WHO found the potentially devastating societal impact of loss of this criti- that 37 percent of all back pain worldwide is attributable to cal worker group can be anticipated if prevention measures are work, resulting in an estimated 800,000 DALYs, a significant not ensured in developing countries, where the proportion of loss of time from work, and a high economic loss. Worldwide, health care workers in the population is already small. 16 percent of all hearing loss is attributable to workplace expo- In total, the few occupational risk factors considered here sures, resulting in 4.2 million DALYs. were responsible for about 800,000 deaths worldwide in 2000. WHO made a special risk analysis of hepatitis B, hepatitis C, Not considered by WHO because of lack of global data are the and HIV infections among health care workers caused by con- additional 1.2 million deaths that ILO estimated are attributa- taminated sharps, such as syringe needles, scalpels, and broken ble to work-related risks (Takala 2002). The leading occupa- glass (WHO 2002). This analysis illustrates the general prob- tional cause of death was unintentional injuries, followed by lem of high risks existing in the small worker population hav- COPD and lung cancer. Workers who developed outcomes ing exposure. WHO found that, among the 35 million health related to these occupational risk factors lost about 25 million workers worldwide, there were 3 million percutaneous expo- years of healthy life. Among the occupational factors analyzed sures to bloodborne pathogens in 2000. This finding is equiva- in this study, injuries, hearing loss, and COPD accounted for lent to between 0.1 and 4.7 sharps injuries per year per health about 80 percent of years of healthy life lost. Low back pain worker. WHO concluded that of all the hepatitis B and hepati- and hearing do not directly produce premature mortality, but tis C present in health care workers, about 40 percent was they do result in substantial disability. This feature differenti- caused by sharps injuries, with wide regional variation. WHO ates these conditions from the others analyzed in the study. also found that between 1 and 12 percent of HIV infections in Figure 60.2 provides summary results for the occupational risk health care workers was caused by sharps injuries. The com- factors. parative risk assessment by region and type of infection indi- The WHO comparative risk assessment has accounted for cates where special emphasis is needed (see figure 60.1). only about 800,000 (40 percent) of the 2 million deaths Clearly, solutions exist to these problems, as shown by the estimated by ILO to occur each year because of occupational illness and injury. Deaths attributable to a wide range of occupational exposures could not be included because of the Percentage 100 90 HCV HBV HIV 32 80 Low back pain (37) 41 70 11 Hearing loss (16) 60 22 50 6 COPD (13) 18 40 30 7 Asthma (11) 14 20 Unintentional 2 10 injuries (8) 12 0 D E A B D B D A B C B D A B Trachea, bronchus, 5 Afr Afr or lung cancer (9) 10 Amr Amr Amr Emr Emr Eur Eur Eur Sear Sear Wpr Wpr Regions 2 Leukemia (2) Male Female 2 Source: WHO 2002, 130. HCV = hepatitis C virus; HBV = hepatitis B virus; HIV = human immunodeficiency virus; 0 10 20 30 40 50 Afr = Africa; Amr = America; Emr = Eastern Mediterranean; Eur = Europe; Sear = Southeast Asia; Wpr = Western Pacific. Source: Adapted from Nelson and others 2005. Figure 60.1 Fraction of Hepatitis C Virus, Hepatitis B Virus, and HIV Figure 60.2 Fraction of Global Disease and Injury Attributable to Infections in Health Care Workers Attributable to Injuries with Occupational Risk Factors Contaminated Sharps, Ages 20 to 65 (percent) 1130 | Disease Control Priorities in Developing Countries | Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut strict requirements for global data. Missing are deaths · Regional conflict, economic pressures, climatologic factors, attributable to asbestosis, silicosis, and other dust diseases; and lack of foreign exchange may make otherwise straight- infectious diseases; cardiovascular diseases; and violence. forward choices impractical. Deaths attributable to workplace exposures to pesticides, heavy · Supply of labor is often high, as is turnover, so economic metals, solvents, and other chemicals are not included. incentives for investment in health capital are lower than in Outcomes such as dermatitis, psychological disorders, and industrial countries. upper-extremity musculoskeletal disorders that cause little mortality but substantial disability are also not captured by the Strategies for Improving Working Conditions WHO comparative risk analysis. Additionally, the consequences With these differences in context in mind, we now consider the of underreporting in existing systems and the dearth of record- major types of intervention: international, national, workplace, keeping systems in the developing nations lead to substantial and individual. undercounting by both the ILO and WHO. Nonetheless, the analyses provide important insights into the immense global International Interventions. The ILO­WHO Joint Com- burden of disease and injury attributable to occupational risk mittee on Occupational Health was formed in 1950 to provide factors. guidance to the ILO and WHO regarding international occu- pational health issues. The committee meets periodically. At its 13th session, held in December 2003, the committee recom- INTERVENTIONS mended that WHO and ILO pursue the following priorities (ILO and WHO 2003): Strategies for controlling injury and occupational disease, developed by industrial hygienists and others over many · Guide and support national occupational safety and health decades in industrial countries, are as fully applicable in devel- programs. Such guidance and support includes providing oping countries. The strategies include a hierarchy of controls models for organizing at national or subnational levels; in the following decreasing order of preference: substituting providing basic occupational health services; promoting major hazards for less hazardous materials or processes; apply- management systems and tools, including control banding; ing engineering controls to separate workers from hazards that developing national profiles and indicators; assessing the remain; using administrative controls to minimize contact cost-effectiveness of interventions; and establishing effective uncontrollable by engineering; and, as the last line of defense, enforcement agencies. using personal protective equipment such as respirators and · Enhance regional collaboration and coordination, including protective garments. What differs in developing countries is the the development and dissemination of models for coopera- context in which the paradigm must be applied. Options are tion, such as the African Joint Effort. often sharply limited, and knowledge of them even more so; · Coordinate and enhance information and educational pro- economic and political factors may impede otherwise obvious grams and materials (for example, by developing a joint or desirable solutions; and the differing workplace context Internet-based global portal) and statistics. may demand that attention be paid to certain problems and · Provide awareness-raising activities and instruments concerns that would not be relevant in industrial countries in through campaigns, events, and special days. temperate climates. The following generic factors associated with work in devel- State or Government Interventions. The major role the gov- oping countries may alter industrial hygiene practice and must ernment can play is to establish workplace rules and provide be considered in every effort to intervene to improve working a system of dissemination and enforcement. Evidence from conditions and occupational health: industrial countries is overwhelming that conditions are sub- stantially improved when both a strong regulatory framework · Access to industrial hygiene consultation is limited; pro- and enforcement are achieved. An added benefit of govern- fessionals, sampling equipment, and laboratories are all in ment, rather than private sector, control is to "level the playing short supply. field": all employers in an economic sector carry the same bur- · Knowledge level about occupational health among man- den. Conversely, improved health of the workforce, achieved by agers and workers is often limited. developing strategies beyond the minimum required, could be · Markets for production materials as well as safety equip- used to confer competitive advantages, a message to reluctant ment may be limited and may include more hazardous employers that has been used in different parts of the world materials or less effective protective equipment "dumped" with some success. from industrial countries where they are no longer mar- Regulatory decisions, such as the choice of exposure limits ketable (Hecker 1991; Ives 1985; Jeyaratnam 1990). or allowable practices, often stimulate the biggest discussion-- Occupational Health | 1131 for example, the debate about dust levels to be allowed in South have markets in industrial countries that have banned or African mines--but the larger issue for most countries is gar- restricted their use--for example, solvent mixtures containing nering resources to ensure compliance, to attract adequately benzene and construction materials containing asbestos. trained personnel to conduct inspections, and to establish and Similarly, equipment such as machines that are well guarded to monitor laboratories to support regulatory efforts. The most prevent injury or well baffled to limit noise may be prohibi- stringent exposure levels (often referred to as threshold limit tively expensive in a marketplace geared to "hand-me-downs" values, or TLVs) are useless if the offending hazard cannot be compared with respirators or gloves. Unfortunately, even these routinely and accurately measured. Indeed, the South African last lines of defense may be difficult to obtain or relatively experience, despite the presence of excellent regulations, is not expensive unless local suppliers are available. encouraging in this regard (Joubert 2002). The single strategy for which no compelling economic dis- Other forms of government intervention may indirectly incentive exists--training--may also be difficult. Through the improve working conditions. Among these are workers' com- efforts of the ILO and numerous nongovernmental organiza- pensation regulations and stipulations that employers of cer- tions and with widening access to the Internet, vast resources tain sizes must engage professionals in health and safety (most have become available. Ample documentation from the indus- often nurses). Each of these interventions has the advantage of trial and developing world indicates that even rudimentary stimulating certain behaviors and practices without requiring knowledge by supervisors and workers about risks and risk- the government to maintain the elaborate and technically com- prevention measures is beneficial. Major impediments remain, plex machinery required for direct monitoring of workplace however, such as educational proficiency, language barriers, conditions. and the applicability of training materials--often developed in Constraints on governmental regulatory and other inter- other contexts--to local situations. Thus, for example, ventions are many. Occupational and environmental regula- although the ILO has recently reported success with informa- tions are often perceived as burdensome costs that impede tion programs in rural Thailand (Kawakami and Kogi 2001), investment and growth, perhaps creating what has been a report from Ghana (Smith-Jackson and Essuman-Johnson referred to as "the race to the bottom," in which threat of out- 2002) suggests that workers and supervisors were unable to migration of industry from one jurisdiction enhances reluc- correctly interpret four of the most common warning signs tance to regulate or enforce control strategies (Frumkin 1999). used for hazard identification, despite having been trained in Moreover, the costs to the government itself, notwithstanding their use. Worker training appears, on the whole, widely technical support from such agencies as the ILO, may be con- underused. siderable in terms of personnel and equipment, and occupa- Problems of infections in patients and health care workers tional health has to compete with other public health priorities from reused needles and needlestick injuries have prompted for scarce resources. The result may be the promulgation of the international organizations to develop model interventions minimal standards or emasculated enforcement of those that that can be transferred elsewhere. WHO initiated Project already exist. The general impression of those working Focus: Ensuring Immunization Safety in Burkina Faso in July throughout the developing world is that the level of regulation 2002 as a pilot project to use WHO materials in a focused effort and enforcement is woefully inadequate compared with that to address all issues related to injection and immunization in industrial countries. Detailed case examples from Brazil safety: availability of equipment and supplies (auto-disposable (Bedrikow and others 1997); Kenya (Mbakaya and others syringes, safety boxes, incinerators); safe injection practices; 1999); Nigeria (Asuzu 1996); and Taiwan, China (Chen and safe vaccine delivery; and safe waste management (WHO Huang 1997), underscore the ubiquity of this problem. 2002). In 2000, WHO conducted a survey to assess the safety of injections in a study group of a random sample of 80 health Workplace-Based Interventions. Issues beyond the economic centers. The situation was reassessed in June 2003 to evaluate and legal ones impede application of the principles of indus- the use of safety boxes (which had been provided in a WHO trial hygiene. A primary factor is ignorance; many employers immunization campaign in Burkina Faso in 2001) and the may be uninformed about available controls and their value. impact of Project Focus. Table 60.2 shows results of the Insurance agencies, local safety groups, and--in some regions reassessment. Dramatic reductions were found in needle recap- of the world--trade unions may serve as facilitators of positive ping, needlestick injuries, and misuse of safety boxes. influence. In general, however, such resources fall short of the Additionally, the number of clinics using safety boxes increased benefit of on-site industrial hygiene expertise that is lacking in from fewer than half to 86 percent. many regions of the world. Small enterprises present special challenges because they Economic factors often impede efforts to institute voluntary lack resources and expertise to address health and safety prob- controls. Materials used are frequently far cheaper than safer lems. Thailand's National Institute for the Improvement of substitutes, often precisely because these materials no longer Working Conditions and Environment has used the ILO 1132 | Disease Control Priorities in Developing Countries | Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut Table 60.2 Prevalence of Risk Factors and Injuries at Control of Nonoccupational Exposures 80 Observed Health Centers In industrial countries, a sharp demarcation exists between (percent) environmental risks associated with work and those associated Needle Needlestick Lack of Misuse of with home life. This differentiation is not the case in many Year recapping injuries safety boxes safety boxes developing countries, especially at large, remote industrial 2000 55 71 51 83 complexes and farms. Workers--with or without their 2003 17 32 14 18 families--often cohabit with the workplace--and often with many or all of its risks, including noise, chemicals, and biohaz- Source: S. Khamassi, WHO Mediterranean Centre, personal communication 2003. ards. The most dramatic examples of this situation were the industrial disasters at Chernobyl and Bhopal. It is not just in disaster, however, that risk occurs. Pesticides, for example, training approach called WISE (Work Improvement in Small result in hundreds of thousands of cases of poisoning a year, a Enterprises) with some success. In one example, six enterprises high fraction from the misuse of farm chemicals for nonwork in the metal industry in Bangkok with between 15 and 115 applications, such as the appropriation of empty (but not workers participated in the WISE program, in which practical clean) drums for transporting water or other household goods, workshops involved workers and management in deciding on a disturbingly common practice. In the industrial setting, car- changes to be made in the workplaces. A wide range of inex- cinogens, neurotoxins, and other hazardous chemicals often pensive changes were put in place, and a booklet to illustrate pollute homes, drinking supplies, and common areas for recre- good practices for others was prepared (Krungkraiwong 2000). ation on a daily basis, adding to the exposure of workers and placing nonworking family members at risk from what would Individual Interventions. The general principle that, for most normally be seen as workplace hazards. public health intervention, organization-level change is more The remedy is often complex and beyond traditional indus- effective than strategies targeting the individual is even more trial hygiene practice. Housing, which in any event may be true when it comes to the workplace. With the exception of substandard, needs to be modified to exclude the possibility self-employed workers, such as those in the informal sector and of contamination by effluent from farm or factory under any subsistence farmers, occupational health and safety does not foreseeable circumstances. Children and family members need lend itself readily to individual solutions, with the same factors to be apprised of the hazards of all materials used for work and limiting employees more likely to limit individuals. prevented from even accidental access, a situation complicated by the fact that children are themselves often inappropriately engaged in the workplace. Food and water supplies need to be Improvement of Access to Health Care secure and protected from cross-contamination, a particular In a few developing countries, workers enjoy broad access to problem in the farm setting. high-quality health care. Chile, for example, has a system of nonprofit employer mutual associations that provide advice on reducing risks in workplaces and medical treatment and sick pay for work-related illness and injury (Contreras and Surveillance and Reporting Dummer 1997). In most countries, the role of on-site services Even in industrial countries, the strategies for recording any is generally limited to emergency services for an injury or acci- aspect of workplace harm beyond acute injury has been an dental overexposure and the conduct of medical surveillance issue; in most developing countries, even injury reports are examinations for workers at risk for chronic conditions such as largely nonexistent. Still, broad agreement exists on the value of noise-induced hearing loss, pneumoconiosis, or cancer. statistical summaries of occurrences. In the developing world, access to health care is critical both Unfortunately, a strong disincentive exists for such report- for work-related and other health issues. In many areas, espe- ing unless it is required by law or by a parent company (as in cially remote or rural areas, on-site service may be the only the case of some multinationals). If reporting is required, as in health care services available to workers and their families. the formal laws of many countries, successful implementation Moreover, the blurred distinction between "general health" and calls for resources for systematic review, verification, and main- "occupational health" in societies where people live and work tenance of the information. Even records whose limitations are in the same community and environment, and where children otherwise legion, such as workers' compensation records or and spouses of workers may share common exposures and regional reporting schemes, have proven highly advantageous adverse conditions with workers, serves to confer some advan- to control efforts in industrial countries. These, too, have a role tage to a more holistic approach to health services often best in developing countries, helping target even rudimentary and provided at or near the workplace itself. limited control efforts. Occupational Health | 1133 Capacity Building occupational health that work together with WHO and ILO Human capital in the form of professional capacity is critical to headquarters and regional offices and three international non- improving working conditions. Professional capacity varies governmental organizations: the International Commission on greatly in developing nations but is higher where recognition of Occupational Health, the International Occupational Hygiene the field is high and the need for professionals and for workplace Association, and the International Ergonomics Association expertise is driven by occupational safety and health legislation (Fingerhut and Kortum-Margot 2002). These partners, located and enforcement. In Malaysia, for example, four decades of in approximately 40 countries, work together in 15 priority rapid industrialization have included a series of legislative acts; areas within a 2001­2005 Work Plan. More than 300 projects development of federal agencies; and inclusion of training at are under way, independently or jointly, to benefit workers in various levels in occupational health in universities, the public developing and industrializing nations in about 15 priority sector, and the private sector (Rampal, Aw, and Jefferelli 2002). areas (WHO 2003). Key international events, such as joining the World Trade Another strong regional coalition, coordinated with and Organization, encourage the development of an economic cul- benefiting from the Global Network of Collaborating Centers, ture that better recognizes the value of safe workplaces. is the WHO­ILO Joint Effort on Occupational Health and Enforcement of national regulations, adoption of International Safety in Africa (WHO and ILO 2002b). This partnering Standards Organization standards, and establishment of man- coalition--where centers outside Africa assist African agement systems lead to broadening of training for workers and partners--includes individual occupational safety and health managers, although the scarcity of trained professionals is a professionals, employers, labor unions, and governmental and major obstacle to adequately implementing regulations and academic institutions in all countries in Africa. policies and providing occupational health services (Christiani, Enlarging small but successful existing programs is one Tan, and Wang 2002; Wang, Cheng, and Guo 2002). approach to capacity building. The U.S. National Institutes of In countries with some capacity, the expertise tends to be Health Fogarty International Center, NIOSH, and the National medical, rather than in other areas, such as industrial hygiene, Institute of Environmental Health Sciences sponsor a suc- engineering, or ergonomics. In most countries, ministries of cessful program, International Training and Research in health and of labor have jurisdiction over working conditions Occupational and Environmental Health, which has developed but often have too few experts and inadequate coordination. small but strong programs between U.S. universities and insti- Moreover, the large percentage of work conducted in the infor- tutions in more than 30 developing nations (NIH 2003). mal sector presents a special challenge to these ministries. Capacity building requires high-quality educational op- Because globalization has brought the need for professionals in portunities. Advances in information technology over the past occupational health to a crisis level, it is appropriate for inter- decade are revolutionizing methods of education, and univer- national trade and development bodies to support national sities worldwide are developing large numbers of Internet- and international capacity-building programs. based courses. Fostering access of students from developing In 1970, when the United States enacted the Occupational nations to these courses in leading universities is now feasible, Safety and Health Act that established the National Institute for but a national or international program is needed to address Occupational Safety and Health (NIOSH), the country had issues such as tuition, competition, intellectual property, and little professional capacity in that field. The new law charged degree requirements. This effort might be called Access to NIOSH with ensuring an adequate number of trained profes- Universities, following the model WHO program Access to sionals and accomplished this task successfully by funding Biomedical Journals, through which WHO and the world's graduate programs in U.S. universities. A follow-up 25 years largest medical journal publishers have provided about 100 later found that 90 percent of NIOSH-supported trainees pur- developing countries with Internet access to journals at no cost sued careers in the field, with more than 50 percent working in or at deeply reduced rates (WHO 2001). private organizations and the balance in government and aca- Professional associations have a long history of assisting in demia (U.S. DHHS 1996). Similar results could be achieved by capacity building through training, research, and conferences. a determined, large-scale effort focused on assisting developing Recently, when the University of Witwatersrand in countries in achieving adequate professional capacity. Johannesburg developed the first Diploma Occupational Both infrastructure and programs are necessary to build Hygiene program in South Africa, the country had too few adequate capacity. In the international arena, a number of suc- industrial hygienists to provide mentors for the field research cessful coalitions exist that provide experienced institutions of the graduate students. The American Conference of and individuals. The WHO Global Network of Collaborating Governmental Industrial Hygienists responded to a request Centers in Occupational Health is a strong international coali- of the International Occupational Hygiene Association, and tion of 70 national, governmental, and academic centers of 11 U.S. industrial hygienists volunteered to be occupational hygiene field practitioner long-distance mentors for the 1134 | Disease Control Priorities in Developing Countries | Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut incoming students during the 2003 course year. This approach country in terms of legal liability and adverse publicity, a lesson will continue until there are adequate industrial hygienists in well taught by Union Carbide's experience in the aftermath of country to serve as mentors to future classes (WHO and ILO the Bhopal disaster in 1984. 2002a). One approach to align economic incentives is to use regula- The U.K. Health and Safety Executive developed a model tory and legal reform to shift the existing cost burden to those program that provides clear solutions to chemical control in a position to remedy the situation--that is, to employers. problems in workplaces. This Web-based, user-friendly prod- Increasing workers' compensation benefits, especially those for uct was launched to enable small business owners in the United long-term effects and disabilities, is an example of such an Kingdom to use information from the suppliers of chemicals to approach. Some evidence exists that, at least in southern Africa, proceed through a series of simple steps to identify practical this approach does stimulate preventive behavior by employers. control solutions that reduce worker exposures to levels that An alternative is to critically reexamine the assumption that present no danger to health (U.K. HSE 2002). This approach, employers do not harbor substantial underrecognized costs of which eliminates the need to measure exposures and meets the injury and illness even under the current situation, especially regulatory requirements of the United Kingdom, has immense in terms of indirect costs such as lowered productivity and potential value for employers in developing nations, who could morale. Harari and his colleagues in Ecuador (Cullen and devote scarce resources to controlling exposures rather than to Harari 1995) have been studying the effect of such exposures as measuring exposures. The approach has gained momentum solvents and organophosphate pesticides on production levels. through adoption by the International Program on Chemical They are attempting to make the case that relatively inexpen- Safety and through formation of an international workgroup sive strategies for exposure control are economically advanta- to advance the approach in developing nations. To enable geous to employers. global use of this approach, the ILO has translated the U.K.- specific system into a product called the ILO Chemical Control Banding Toolkit (ILO 2003). Intervention Costs and Cost-Effectiveness Workplace illness and injury produce personal suffering and high economic costs. The ILO estimates that about 4 percent of ECONOMIC ASPECTS OF INTERVENTION GDP worldwide is lost because of work-related diseases and injuries (Takala 2002). The European Agency for Occupational Measures to prevent occupational risks are not cost-free, and Safety and Health at Work (1998) indicates that the costs to where those costs ultimately come to rest affects the willingness society in European countries ranged from 0.4 percent to of employers to implement the preventive measures. 4 percent of gross national product. Examining Industrial Countries. Identifying interventions Who Bears the Costs of Preventive Measures? to successfully reduce or prevent workplace injuries and ill- In industrial countries, compelling economic incentives exist nesses will benefit society, employers, and workers. In 1996, for employers to control risks for injury and illness on the job, stakeholders in the United States identified intervention effec- especially those that result in demonstrable near-term lost tiveness research as one of 21 priority areas in occupational work or function. These include the high cost of medical care health research for the next decade (U.S. DHHS 1996). From (especially in the United States), the burden of workers' com- 1996 to 2002, research conducted or funded by NIOSH to pensation payments, high replacement costs for the labor, risk develop and evaluate the effectiveness of solutions to prevent of litigation and liability, and negative business consequences work-related injury and illness has increased nearly sixfold, of adverse publicity. Although these factors may differ by coun- from about US$5.5 million to US$33 million (U.S. DHHS try and sector, they are less likely in the developing world to 2003). confer on employers a strong economic imperative for Research studies of workplace interventions often use sur- prevention--labor is plentiful, its replacement cost is low, rogate or implied measures for economic evaluation. For and--most important--a high portion of the real cost of example, economic benefit is assumed to occur following an injury and illness will not be borne by the employer. The statis- intervention if symptoms of illness or injury in a workforce tics in Latin America are staggering: although an estimated 2 to decrease while productivity remains constant. Intensive data- 4 percent of the GDP of the region is lost because of occupa- entry workplaces are increasing rapidly in both industrial and tional fatalities alone, no evidence exists of private sector developing nations. Three studies of U.S. Internal Revenue investment to reduce the risk (Giuffrida, Iunes, and Savedoff Service data-entry clerks by NIOSH found that short, strategi- 2002). Multinational companies appear to be an exception. For cally placed rest breaks of 5 to 15 minutes during the regular many, the costs of injury and illness may accrue to the parent daily schedule reliably reduced eyestrain, fatigue, and Occupational Health | 1135 Box 60.1 Use of a Toolkit to Determine Return on Investment in Central American Garment Factories The Regional Occupational Safety and Health Center Change in Illness and Injury Indicators in the First (Centro Regional de Seguridad y Salud Ocupacional) Quarter of 2003 Compared with the First Quarter of 2002 at Confecciones La Palma Project in Central America developed an occupational safety and health toolkit to enable managers and line Indicator 2002 2003 Percent change workers in garment factories to self-diagnose plant and Number of accidents 63 36 40 workstation hazards and to estimate the costs and benefits of interventions (Amador and others 2003). Managers and Days of absenteeism 200 149 25 employees at more than 100 Central American garment Sick days 822 426 48 factories have been trained to use the toolkit. An in-depth Visits to factory clinic 2,716 2,163 20 evaluation of the use of the toolkit in three garment facto- Productive person-days Not given Not given 12 ries, each employing between 700 and 1,000 workers, in Source: Data provided by Confecciones La Palma. El Salvador, Guatemala, and Nicaragua found that within one year the factories generated savings that were four to eight times the costs of the interventions. A CD-ROM version of the tool kit in Spanish and The overall investment by Confecciones La Palma in English also contains the guide "How to Design and 2002 was US$6,360, and the savings attributed to that year Establish an Occupational Safety and Health Program in a were US$27,242 from reduced injury, illness, and absen- Garment Factory" and can be found on the NIOSH Web teeism and an increase in productive days (see table). site at http://www.cdc.gov/niosh. Source: Amador and others 2003. musculoskeletal discomfort for video-display terminal opera- rates for key measures of success were recorded for the three tors without decreasing productivity (Galinsky and others years before the intervention and for the three years after the 1999, 2000). Similar benefits of improved comfort and reduced intervention. The successful project reduced the frequency of stress from short rest breaks were observed among workers in back injuries in six nursing homes by 57 percent, lowered a meat-processing plant without affecting productivity injury rates by 58 percent, and decreased workers' compensa- (Dababneh, Swanson, and Shell 2001). tion expenses by 71 percent. Box 60.1 illustrates the value of Including economic costs of interventions is more difficult evaluating costs of interventions in garment factories in but is an important measure to allow employers to make deci- Central America. sions about interventions. A model intervention study, "Evaluation of a Best Practices Back Injury Prevention Reducing Risk for Back Pain. The WHO summary of a variety Program in Nursing Homes," received the 2003 National of cost-effectiveness studies for interventions addressing all Occupational Research Agenda Partnering Award for Worker aspects of global health illustrated that the cost-effectiveness of Safety and Health (APHA 2003; Collins and others 2004). interventions in some areas of personal health has been well Members of the partnership that carried out the intervention studied but that environmental and occupational topics have study included a large nonprofit U.S. health care system that had relatively few studies. The WHO comparative risk assess- owns nursing homes; manufacturers of lifting equipment; ment concluded that about 37 percent of back pain globally is researchers at Washington University, West Virginia University, attributable to occupational risk factors (WHO 2002). A cost- and NIOSH; and health care workers. The prevention program effectiveness study of interventions to reduce occupational back combined measures to reduce back injury by identifying the pain was also reported, using economic models developed to movements and postures that put nursing assistants at risk of calculate costs of interventions in three WHO geographic back strain, stress, and injury in lifting and moving residents. regions that illustrate different levels of development. Mechanical lifting devices for reducing those stresses and Intervention studies were obtained from the published and strains were evaluated in the laboratory and then in the nurs- unpublished literature. All costs of running the interventions ing homes. A best-practices training and lifting program was were measured in international dollars (not exchange rate put in place on the basis of researcher and employee input, and dollars, as in the analysis reported later in this chapter) and 1136 | Disease Control Priorities in Developing Countries | Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut Cost (international dollars, millions) Americas. The relative cost-effectiveness of the three interven- 6,000 tions would be unchanged in each region. Amr A EP EP Eur B 5,000 Sear D T&EC Reducing Risk for Silicosis. Silicosis is a disabling and often T&EC fatal workplace lung disease caused by inhalation of silica dust. 4,000 T = training EC EC = engineering control The high-risk sectors of the economy include construction, EP = ergonomics programme EC 3,000 mining and mineral processing, foundries, and manufacturing EP T&EC of pottery and glass. Large numbers of workers in both indus- 2,000 EC trial and developing nations are exposed. Box 60.2 illustrates T control of silica caused by grinding wheels in the agate cottage 1,000 industry in India. In the United States, more than 3.2 million T T workers are exposed to silica dust, even though methods exist 0 0 1 2 3 4 5 6 7 8 to eliminate exposure (Harley and Vallyathan 1996). DALYs averted (millions) A study was conducted to evaluate the cost-effectiveness of Source: Reprinted from WHO 2002, 130. alternative interventions to reduce silicosis in industrial and Amr = Americas; Eur = Europe; Sear = Southeast Asia developing nations (Lahiri and others 2005). The authors used the limited published and unpublished data on costs of the var- Figure 60.3 Cost-Effectiveness of Interventions for Lower Back Pain ious interventions and on the efficacy of exposure reduction. To analyze the cost-effectiveness of each intervention in reduc- effectiveness was measured as age-weighted DALYs gained by ing disease incidence, they used models developed for WHO the intervention. The interventions for the prevention of back (Murray and others 2000). The WHO DALY concept was used pain were grouped into three major categories: worker training to combine mortality and morbidity resulting from silicosis. (awareness education and hazardous job training); engineering Two WHO regions were studied: the highly developed America control (physical measures that control the exposure to the A region, represented by the United States and Canada; and the hazard, including equipment that assists lifting, pushing, and developing Western Pacific B1 region, represented by China, pulling); and the full ergonomics program (which includes both the Democratic People's Republic of Korea, the Republic of of the previous interventions and implementation procedures). Korea, and Mongolia. Exposure estimates were taken from the As shown in figure 60.3, the analysis found that the most WHO comparative risk assessment study (Ezzati and others effective intervention is the full ergonomics program, offering 2004). a 74 percent reduction in back-pain incidence. Lesser benefits The interventions included engineering control interven- are obtained by engineering control (56 percent reduction) and tions that protect many workers (use of the wet method--that training (20 percent reduction). The total costs of worker train- is, spraying a surface or wetting a blade to reduce dust; of local ing are largely labor related, the costs of engineering control are exhaust ventilation; and of total plant ventilation) and worker primarily capital costs, and the full ergonomics program costs training plus personal protective equipment, an intervention are equally shared between the two. Training was found to be that protects the individual worker. The training involved four the most cost-effective intervention, as indicated by the lower types of personal protective equipment: comfort masks, dust slopes of the lines from 0 to T, and was recommended as the masks, full-face respirators, and half-face respirators. Lahiri first choice when resources are scarce. However, the incremen- and others (2005) summarize the evidence of reduction in tal cost-effectiveness ratios for the other options (indicated by exposure through the use of selected interventions from the lit- higher slopes of the lines) demonstrated that both engineering erature. The difference in the health life years gained with and control and the full ergonomics program are attractive alterna- without the intervention represented the effectiveness of the tives. Thus, even the full ergonomics programs were found to intervention and was used as the denominator for the cost- be cost-effective in all three regions for their health effects effectiveness ratio. alone, without even considering the possible increase in pro- The engineering control interventions involve large capital ductivity that could be brought about by the interventions expenditures, whereas the implementation of personal protec- (WHO 2002). Recalculating these results according to Disease tive equipment requires ongoing large equipment costs (filters Control Priorities Project methods--using exchange rate and cartridges) as well as labor costs for training the workers. dollars and removing the age weights from DALYs (see chap- Costs of interventions vary from region to region, depending ter 15)--would increase effectiveness somewhat because back on wage rates and raw material costs, but the costs of equip- pain is more common at later ages and would reduce the ment seem not to vary. The authors found the least expensive apparent costs in both Europe and Southeast Asia. It would alternative is training associated with use of a comfort mask. have little effect on costs in the high-income countries of the However, that intervention has a relatively low efficacy of Occupational Health | 1137 Box 60.2 Economic Evaluation of an Engineering Control for Silica Dust in India The agate industry is a cottage industry concentrated in (Bhagia, Ramnath, and Saiyed 2003). Economic analysis residential settings in Khambhat and Dahegam, in the was based on 600 dust control devices that could be state of Gujarat, India, where 15,000 grinders and 60,000 installed in the communities. The total costs include the other workers, family members, and neighbors are initial one-time cost of the devices (Rs 8,000, or approxi- exposed to silica dust. The making of key chains, neck- mately US$92); depreciation (10 percent per year); main- laces, and art pieces involves baking, chipping, grinding, tenance of machines (equivalent to the costs saved by and polishing agate stones. The grinding-machine wheels recycling the dust to be used in polishing); and the cost of are driven at speeds of 1,440 rpm, generating large treating the diseases (about Rs 4,000, or approximately amounts of dust containing respirable silica. The table US$184 per year per case). The gains included annual shows the extraordinary prevalence in the total exposed income per avoided case of silicosis. Total savings per year populations (noted above) of silicosis and tuberculosis were estimated to be between Rs 23 million and Rs 29 mil- caused by silica exposure. lion (US$527,039 to US$664,528). The conclusion is that installation of dust control devices in all the agate- Prevalence of Diseases in Agate-Dominated Areas of grinding units of Gujarat would reduce silicosis and Khambat and Dahegam tuberculosis as well as yield financial and health benefits to the workers, families, and the greater society that bears the Silicosis Tuberculosisa cost of illness. Category Percent Number Percent Number Grinders 30 4,500 36 5,400 Nongrinding 8 4,800 16 9,600 workers, family, and neighbors Source: Bhagia, Ramnath, and Saiyed 2003. a. National tuberculosis prevalence in India of 4 percent and resulting cases have been subtracted. The National Institute of Occupational Health in India designed and distributed 10 dust control devices for the grinding machines to employers, who generally employ 5 to 10 workers (see figure). The efficacy of the devices was found to be 93 percent, and dust was greatly reduced Traditional Grinding Machine with Dust Control System Table 60.3 Average Cost-Effectiveness Ratio 30 percent exposure reduction. Although the initial capital (US$/DALY gained) expenditures are high for engineering controls, the annualized costs based on a 10-year horizon are encouraging, with expo- America A region Western Pacific B1 region sure reduction of about 70 to 85 percent. The greatest exposure Engineering control 105.89 109.35 reduction of 95 percent was achieved at the highest cost, with Comfort mask 111.04 117.19 training plus use of a full-face respirator, but an 80 percent Dust mask 191.38 173.90 reduction was achieved at half this cost when training was Half-face respirator 299.82 272.45 combined with a dust mask. Full-face respirator 304.87 265.74 Table 60.3 shows that engineering controls in both industrial and developing regions are the most cost-effective Source: Lahiri and others 2005. 1138 | Disease Control Priorities in Developing Countries | Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut interventions, with expenditures of between US$105 and Injury rate per 100 full-time workers US$109 per healthy year saved in the two regions. Although 20 exposure reductions with respect to each intervention type are 18 identical in both regions and the cost of interventions is some- 16 what higher in the America A region, it might seem perplexing 14 that the cost per unit of health gain is relatively lower in the 12 America A region than in the Western Pacific B1 region. The 10 reason for this result is that effectiveness (the denominator of 8 the average cost-effectiveness ratio) is represented by health 6 outcomes that are higher for this region because life expectancy Agricultural production 4 Construction in the America A region is higher than in the Western Pacific B1 2 Nursing and personal care region. Therefore, lives saved through interventions in indus- 0 trial regions contribute more toward the healthy years gener- 1980 1981 1982 1983 1984 1985 19861987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 ated by the model. The study concluded that engineering controls are the most Source: BLS 2002. cost-effective interventions in both regions and should be con- Figure 60.4 Incidence Rates of Nonfatal Occupational Injuries (Total sidered as the first choice in cases in which resources are scarce. Recordable Cases) by Selected Industry Sectors, 1980­2000 The results underestimate the health gains because other silica-related diseases such as tuberculosis and cancer are not considered. In the developing world, a patchwork of some of these approaches has brought success: in Vietnam by reducing silico- sis through industrial hygiene practices of wetting the process IMPLEMENTATION and through surveillance (effective but much less so than pro- tective equipment or, better yet, substitution of a safer prod- The health of a country's workforce, even more than the health uct), and in Nicaragua by efforts to reduce pesticide poisoning of the country's overall population, is critical to its economic through worker and health professional education and report- and national security. No country has become a successful ing systems. Even in highly developed countries, the continued economic power without sustained attention to the health of need for a responsive and broadly based framework, with gov- workers, who create the successful economy. Responsibility for ernment involvement at the center, is evident. the safety and health of workers lies with the government, the An example recently identified in the United States is the employers, and the workers themselves. However, it is the gov- comparison of injury indicators for workers across three ernmental framework, whether at a national or local level or prominent sectors: construction, agriculture, and health care both, that is the linchpin on which other efforts rest. (figure 60.4). Two decades of data demonstrate improving nonfatal injury rates for construction and agricultural workers (falling steadily from about 16 and 12 per 100 full-time work- Institutions and Programs ers, respectively). This rate is in contrast to injury rates for The potential to continually improve work-related health sta- health care workers, which have risen by about one-third from tus, as measured by morbidity and mortality data across multi- a starting point of about 10 per 100 workers (BLS 2002). Why ple economic sectors and across many countries, has been the difference? Although the reason is not fully delineated, compellingly demonstrated. Not surprisingly, because these major efforts (including government regulation, research conditions are inherently preventable, what may be the lowest investment and policy setting, education, and best-practice achievable level of risk is debated in industrial countries. In the efforts)--plus, in the case of construction workers, active trade United States, for example, occupational injury fatality rates union involvement--in the two industrial sectors clearly have been steadily declining, now approaching 3.8 per 100,000 were under way in this time period. In contrast, in the health workers, down from 7.5 just 20 years earlier (NIOSH 2000). care worker sector, no such national program existed in the Key elements in improving worker health and safety, regard- period, and industry pressures resulted in a number of poten- less of the level of development, include regulatory and tial workforce problems, such as increasing work demands on a enforcement framework; worker, employer, and health profes- stressed and ill-prepared workforce. sional education; surveillance and reporting systems; and dis- semination and implementation of best practices. Often these Lessons Learned elements overlap in multifaceted approaches to addressing As countries undergo rapid economic development, industrial- complex and disparate work settings. ization, and the effects of globalization, leaders can examine Occupational Health | 1139 available occupational health system models as they develop ministers of all member nations. Both ILO and WHO regional their national occupational health systems. The Republic of offices provide technical assistance and training. An unfortu- Korea has experienced major economic and societal changes nate gap exists, however, because often the national ministries since it emerged from the Korean War in the 1950s with a need of labor relate to the ILO and the ministries of health to the to establish occupational safety and health programs without WHO. Because working people are influenced by the national any historical experience. The system directed by the Industrial ministries of both labor and health, the international organiza- Safety and Health Act of Korea is modeled after the specialist- tions have been advised to correct the situation (WHO and ILO based system of Japan, which relies on medical screening and 1995). A promising WHO/ILO Joint Effort on Occupational specialists outside the workplace to deliver health and safety Safety and Health in Africa has embraced partners across min- services (Paek and Hisanaga 2002). Paek and Hisanaga note istries and broadly within and outside Africa to work together that national traditions and culture strongly influence the to assist countries, workers, and employers in the formal and choice of system when developing countries examine models informal sectors (WHO and ILO 2002b). from which to choose--for example, from the code-based sys- Several coalitions of organizations have successfully assisted tem of the United States, the performance-based system of the developing countries to increase professional capacity and to United Kingdom, and the management system of the Nordic improve worker health and safety. A regional illustration is the countries. Association of Southeast Asian Nations Occupational Safety and Government involvement is necessary but not sufficient, Health Network, which was established to promote regional regardless of the level of development. Because national and cooperation in training and research as well as harmonization of local legislation and policies create the framework within standards in safety and health. The primary international coali- which a society functions, international influence, assistance, tion is the WHO Global Network of Collaborating Centers in and requirements play a key role in encouraging developing Occupational Health, described previously. The U.S. Fogarty and industrializing nations to create appropriate laws and poli- International Training and Research Program in Occupational cies to support healthy workers. International trade, develop- and Environmental Health is another stable and experienced ment, and funding organizations have immense power, which network, consisting of U.S. universities working with universi- is not fully exercised to date, to influence countries on working ties in more than 30 developing nations to increase professional conditions. The large multinational trade agreements have also capacity. Some of the institutions are also WHO Collaborating failed to ensure that worker health and safety is adequately Centers, and others are partners in Africa. The relationships addressed, and multinational corporations have generally not have provided opportunities for synergy and streamlining of transferred the safety and health systems of the industrial training and technical assistance (WHO and ILO 2002b). world to developing nations. The consequences are grave, as Globalization has brought work-related hazards to develop- seen in the deaths of 2 million workers each year from occupa- ing countries lacking the infrastructure and professional capac- tional injuries and illness. Great power lies in these institutions ity to handle them adequately. It is incumbent on the national and trade agreements to produce direct changes in the health of and international bodies responsible for globalization to assist workers globally. the recipient nations. Organizations with proven track records The ILO provides strong guidance internationally for indus- in occupational health could play key roles if international and trial and developing nations through its tripartite conventions national laws provide the appropriate context and funding. and recommendations. Each year the ministers of labor of all member nations, employer representatives, and worker repre- sentatives agree on policies in conventions, which the member Globalization states are asked to ratify. The models provided by the ILO are Increased globalization has caused important changes for many regularly used by some countries as they create their national developing countries. Dias, Mendes, and Schwartz (2002) iden- systems. Brazil, for example, has been a member country of the tify the series of developmental stages through which a country ILO since 1919 and has ratified 75 ILO conventions. Set in place passes: underdevelopment and poverty, industrial revolution by legislation in the 1970s, the Brazilian model for occupational and accelerated economic development, internal adjustments health services followed the ILO's Recommendation 112 (1959) to strengthen national competitive power to enter globalized as a paradigm (Dias, Mendes, and Schwartz 2002). Although markets, adjustments to foreign policy to integrate globalized few countries ratify most conventions, more than 150 countries markets, and aims toward long-term sustainable human devel- have ratified Convention 182, which was passed in 1999 and opment. These authors also analyze the positive and negative addresses the worst forms of child labor. A current critical need effects on occupational health conditions in countries such as is assistance to developing nations to meet the obligations Brazil, where crises and opportunities are experienced simulta- required by ratification. WHO sets international policies on neously. On the risk side, for example, free trade agreements health, including worker health, at its annual meeting of health have intensified rapid industrialization and the export of 1140 | Disease Control Priorities in Developing Countries | Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut industry and materials--many hazardous--to regions with Nonetheless, globalization does offer potential solutions. poor occupational infrastructure. These effects are likely One is the link provided by international lending agencies such increasing risk and rates of injury and occupational disease. as the World Bank and International Monetary Fund of health Globalization has also engendered major occupational and safety considerations to development loans. As a condi- health and safety development projects, most notably spon- tion of receiving the development funds, control of health sored by Scandinavian governments (Partanen and others and safety conditions could be mandated and enforced. A 1999). These initiatives have infused developing countries second opportunity is voluntary initiatives, as were recently with expertise, training programs, and equipment and have developed in the Apparel Industry Partnership, wherein a con- provided much-needed (external) economic incentives for sortium of European and U.S. garment manufacturers agreed adoption of change on national, regional, and local levels. The to control labor and safety practices in their facilities in devel- major concern is sustainability, because the incentives are oping countries by joint consent. external. Globalization has resulted in a rapid increase in the number Implications for Health System Development of multinational companies operating outside industrial Workers' health and safety in most of the developing world countries. Indeed, this outcome was the underlying economic may fall under the jurisdictions of both the ministry of labor intent of recent free trade agreements such as the North and the ministry of health, with little collaboration and coordi- American Free Trade Agreement (NAFTA). In general, these nation between the two. The ministry of labor most commonly companies bring with them a highly developed infrastructure is the governmental focus of any regulatory and enforcement in occupational health practice from their base countries. efforts, even though without requisite expertise and access to Unfortunately, although these model companies undoubtedly follow-up care, it may mandate services falling squarely within upgrade the availability of high-quality services and training, the traditional health system (for example, pulmonary func- enhance workers' awareness, and create pressure on other tion testing as ongoing screening for individual response to industries in the region to conform, the pressure likely goes in exposure to pulmonary toxicants). Whatever the country- both directions. The competitive advantage of lowered invest- specific organizational and structural constraints, the following ment in health and safety, as long as the labor markets are plen- set of principles can be applied in providing health services to tiful and the direct costs to employers of illness and injury low, workers: results in strong pressure to minimize--or at least reduce--the intensity and quality of services. · Coordination between occupational health services and over- Even at their best, multinationals may inadvertently create all health services. Occupational health services, consisting an occupational health and safety "caste system." Many provide of efforts to prevent work-related disease and disability as extremely high levels of care and service for their international well as to recognize and treat them once they occur, must managers and for technical support staff members, while offer- be coordinated with overall health services. The separation ing local resources to indigenous workers. More broadly, occu- between work-related and other health conditions, often pational clinics, industrial hygiene services, and the like are driven by regulatory and liability concerns, insurance, and often developed but available for the exclusive use of the multi- other external constructs (workers' compensation system or nationals, creating communities inside which modern occupa- the disability system, for example), is not only clinically tional health exists but outside which nothing changes except challenging but inefficient in optimizing individual health the incorrect perception that progress has occurred. Often the status. Although many work-related injuries and a few spe- reverse of progress has actually occurred because the limited cific illnesses, such as asbestosis, can be readily pinpointed as numbers of trained physicians, occupational health nurses, stemming from work alone, most health problems result industrial hygienists, and safety professionals are siphoned off from multiple causes. This fact is as true in the developing to the higher-paying, more prestigious positions. world as in the industrial one--whether it be cumulative Free trade zones, established by treaties such as NAFTA, cre- pesticide exposure from work and community sources, the ate special considerations. Although the agreements offer the interaction of poverty and poor health status with chemical potential to incorporate strong industrial world rules regarding work exposures, or the cumulative psychosocial stressors of labor, environment, and health in underdeveloped zones, the life both inside and outside work. A holistic approach to the host countries often resist such changes, perceiving these rules individual, recognizing the multisectoral, multiple determi- as trade restrictions. The final language regarding health and nants of health on overall health status, should be the goal safety in NAFTA, for example, is significantly less stringent in the provision of health services. than rules in the United States. Moreover, some multinationals · Attribution of causality and access to health services. In parts resist even these rules, seeking broad economic relief as a foun- of the developing world, as in the industrial world, a phe- dation of moving across the border (Frumkin 1999). nomenon exists wherein some threshold of causality (such Occupational Health | 1141 as "more probably than not work related") is the trigger to control strategies in the developing world. This research need workers' compensation or other employer responsibility for not be undertaken solely--and sometimes not even in part-- taking care of the illness or injury. All too often in this set- in the countries of concern, but rather is likely to be aided by ting, the incentive for the employer to disclaim responsibil- the capacity building derived from partnership between aca- ity leaves the worker, whatever the cause or causes of the demic institutions and government agencies across countries of condition under question, falling through the cracks of different levels of development. occupational health care and traditional health care. Universal access to health care, unfortunately not available in the United States or in much of the developing world, can Public Health Systems Research mitigate this problem. Although health services research has emerged as an important · Health professional workforce expertise in occupational health. area of inquiry in the health field in the industrial world over As discussed in the section on capacity building, adequate the past few decades, scant attention has been paid to public expertise does not exist in the developing world to address health systems research (Institute of Medicine 2003). Given traditional and emerging occupational health problems. that occupational health sits at the interface between individual Occupational health services are multidisciplinary, including and population health, this area of inquiry is particularly nonmedical (industrial hygiene and engineering expertise, germane to research in the field. This research would examine for example) in addition to health (nursing and physician) the effectiveness of government systems working in coordina- expertise. Without being prescriptive, health systems need tion with other sectors (academic institutions, employers, to ensure the existence of an adequately prepared workforce unions, voluntary agencies) in promoting occupational health (whether through broadly based training for all health per- status. sonnel, training for occupational health specialists, or most likely some combination of the two) if they are to grapple even minimally successfully with reducing the human and Occupational Health Policy Research economic burden of work-related injuries and illnesses. Public policy to address improving occupational health in the developing world should rest on a sound scientific base (that is, Unifying the three principles identified above is the need to be evidence based) and should be coupled with an under- recognize that occupational health should be in the main- standing of the local and national frameworks for policy stream of both health education (at the professional, employer, (whether through legislative, regulatory, or other means). and individual level) and health care. An argument against this Adequate research has not been undertaken to evaluate policy approach is the perception that doing so will result in an unten- development and implementation in public health in general able burden on already underfunded health care. We suggest and occupational health specifically. As with the need for new that not so doing will create a greater burden, in both financial health systems research, this area of inquiry would undoubt- and human terms. edly benefit from partnerships among countries in the indus- trial world and in the developing and industrializing world. RESEARCH AGENDA Before 1996, there was no known national effort to identify and Intervention Effectiveness Research promote an occupational health research agenda. That changed Intervention effectiveness research, a cornerstone of the U.S. with NIOSH taking the lead to launch the National NORA initiative, is critical to advancing occupational health in Occupational Research Agenda (NORA) (U.S. DHHS 1996). the developing world. The absence of data in this chapter to Since then, a number of other countries in the industrial world demonstrate cost-effectiveness of occupational health meas- have launched similar efforts (for example, Italy, Sweden, and ures is indicative of the need for more such information to the United Kingdom). Although all these efforts are relevant to target what will always be a demand for limited resources. the developing world, the reality is that country-specific Recognizing the relative dearth of intervention effectiveness research, even at the risk of reinventing the wheel, is often research in countries with high research investments, this rec- needed to strengthen political will to effect policy. Moreover, ommendation is made cautiously for countries with fewer although traditional epidemiological etiologic research in resources. However, it is assumed all too often that an accepted occupational health is not a priority or even feasible for much intervention in a country with higher economic productivity of the developing world, research targeted at local conditions might not be viable in one with fewer resources. The research and institutions is often what is most needed. Six areas are agenda for the developing world in this arena needs to be tai- identified, with appropriate modifications for local conditions, lored to what is known and proven coupled with local and as ongoing research priorities to address injury and disease national conditions and needs. 1142 | Disease Control Priorities in Developing Countries | Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut Control Technology and Protective Equipment Research most vulnerable--children and the poor--are also dispropor- Investigation of control technology and protective equipment, tionately at risk. Compounding this tragedy is that many effec- another NORA priority, is critical for developing effective and tive and economically feasible interventions are available to feasible control strategies in the developing world. Much of the address these largely preventable health conditions. primary research in this category can be done in the industrial Despite relatively little systemic data on cost and cost- world, but along with investments in intervention effectiveness effectiveness, even this "tip of the iceberg" picture demonstrates research, new technologies may still need to be tested in real work-related conditions contributing significantly to overall situations in developing countries. Simplified approaches to mortality and morbidity and demonstrates the overall societal management of chemicals suitable to the local work settings benefit of their prevention and treatment. Externalization of have been developed in Indonesia and are being evaluated, and costs by employers--to the society as a whole--often obscures the International Program for Chemical Safety is helping other the actual overall benefit of a framework that relies on govern- countries modify, implement, and evaluate the U.K. system, ment regulation and enforcement, education, and best prac- which was originally designed for use by small enterprises in tices. Effectively addressing these problems takes active involve- the United Kingdom (ILO 2003). ment from national and local government, employers, and workers and their representatives. The challenges to reducing Disease and Injury Research the burden are heightened to the degree that public health and health care delivery systems isolate occupational health from Many questions of epidemiologic importance to improving the the mainstream of health and health care. health of all workers can best be answered in settings in the Despite structural and political barriers to overcoming this developing world. This situation is not unique to occupational high burden of disease and injury, evidence exists of enormous health, but in the occupational health arena, it is important to progress in the industrial world and of isolated progress in recognize that workers are often the first exposed and are parts of the developing world. Targeted future investments exposed to the highest levels of potential hazards (as compared in research and public health and health systems are critical with their community counterparts). So, too, are levels of expo- to ensuring that progress continues and is more equitably sure to many hazards far greater in the developing world than distributed. elsewhere, and undertaking studies becomes efficient and fea- sible in these settings that would prove difficult if not impossi- ble in settings where exposures are lower and larger numbers of study participants are needed to detect meaningful differences REFERENCES in risk. Not surprisingly, then, sentinel studies of health effects Amador, R., C. Maldonado, R. Venezia, and C. Rivera. 2003. "Return on of interest to the industrial world have been undertaken in Investment in Prevention via the CERRSO Tool Kit." Central American other countries--for example, studies in Latin America identi- Regional Occupational Safety and Health (CERSSO) Project, San Salvador, El Salvador. fying the potential for acute pesticide intoxication to cause APHA (American Public Health Association). 2003. "Recent Conference chronic neurological effects (Rosenstock and others 1991). on Protecting the Nation's WorkforceNation's Health." http://www. apha.org/journal/nation/tnhfullstories.htm. Asuzu, M. C. 1996. "The Development and State of Health and Safety in Surveillance Research the Workplace in West Africa: Perspectives from Nigeria." West African Surveillance is a critical component of all effective occupational Journal of Medicine 15 (1): 36­44. health programs; thus, continuing research is needed into the Bedrikow, B., E. Algranti, J. T. Buschinelli, and L. C. Morrone. 1997. "Occupational Health in Brazil." International Archives of Occupational most effective ways to gather and interpret this information. and Environmental Health 70 (4): 215­21. Surveillance systems are often limited at best in many develop- Bhagia, L. J., T. Ramnath, and H. Saiyed. 2003. "Cost Benefit Analysis of ing countries, and evaluation research needs to be undertaken Engineering Control Devices in the Agate Industry." National Institute to determine the benefits of investing in gathering both gener- of Occupational Health, Ahmedebad, India. ic (absences from work, for example) and specific (blood lead BLS (Bureau of Labor Statistics). 2002. "Survey of Occupational Injuries and Illnesses." Washington, DC: U.S. Department of Labor, Bureau of levels, for example) information on which to target public Labor Statistics, Safety and Health Statistics Program. health action. Chen, M-S., and C-L. Huang. 1997. "Industrial Workers' Health and Environmental Pollution under the New International Division of Labor: The Taiwan Experience." American Journal of Public Health 87 (7): 1223­31. CONCLUSIONS Christiani, D., X. Tan, and X. Wang. 2002. "Occupational Health in China." Occupational Medicine 17: 355­70. The burden of occupational health problems is staggering in Collins, J., L. Wolf, J. Bell, and B. Evanoff. 2004. "An Evaluation of a `Best both human and economic costs, and workers in the develop- Practices' Musculoskeletal Injury Prevention Program in Nursing ing world bear this burden disproportionately. Moreover, the Homes." Injury Prevention 10: 206­11. Occupational Health | 1143 Contreras, R., and W. Dummer. 1997. "Occupational Medicine in Chile." Institute of Medicine. 2003. Who Will Keep the Public Healthy? Educating International Archives of Occupational and Environmental Health 69 Health Professionals for the 21st Century. Washington, DC: Institute of (5): 301­5. Medicine. Cullen, M. R., and R. Harari. 1995. "Occupational Health Research in Ives, J. H. 1985. The Export of Hazard: Transnational Corporations and Developing Countries: The Experience in Ecuador." International Environmental Control Issues. Boston: Routledge & Kegan Paul. Journal of Occupational and Environmental Health 1 (1): 39­46. Jeyaratnam, J. 1990. "The Transfer of Hazardous Industries." Journal of the Dababneh, A. J., N. G. Swanson, and R. L. Shell. 2001. "Impact of Added Society of Occupational Medicine 40 (4): 123­26. Rest Breaks on the Productivity and Well Being of Workers." Joubert, D. M. 2002. "Occupational Health Challenges and Success in Ergonomics 44 (2): 164­74. Developing Countries: A South African Perspective." International Dias, E. C., R. Mendes, and B. Schwartz. 2002. "Occupational Health in Journal of Occupational and Environmental Health 8 (2): 119­24. Brazil." Occupational Medicine 17: 523­27. Kawakami, T., and K. Kogi. 2001. "Action-Oriented Support for European Agency for Occupational Safety and Health at Work. 1998. Occupational Safety and Health Programs in Some Developing Annual Report. Bilbao, Spain: European Agency for Occupational Countries in Asia." International Journal of Occupational Safety and Safety and Health at Work. Ergonomics 7 (4): 421­34. Ezzati, M., A. D. Lopez, A. Rodgers, and C. J. L. Murray, eds. 2004. Kjellstrom, T., and L. Rosenstock. 1990. "The Role of Occupational and Comparative Quantification of Health Risks: Global and Regional Environmental Hazards in the Adult Health Transition." World Health Burden of Disease Attributable to Selected Major Risk Factors. Geneva: Statistics Quarterly 43: 188­96. World Health Organization. Krungkraiwong, S. 2000. "Occupational Safety and Health Improvement Ezzati, M., A. Lopez, A. Rodgers, S. Vander Hoorn, C. Murray, and the and Productivity in Small and Medium-Sized Enterprises Program in Comparative Risk Assessment Collaborating Group. 2002. "Selected Thailand, Case Studies in Safety and Productivity." National Safety Major Risk Factors and Global and Regional Burden of Disease." Council, Itasca, IL. Lancet 360 (9343): 1342­43. http://image.thelancet.com/extras/ Lahiri, S., C. Levenstein, D. Imel Nelson, and B. J. Rosenberg. 2005. "The 02art9066web.pdf. Cost Effectiveness of Occupational Health Interventions: Prevention of Fingerhut, M., and E. Kortum-Margot. 2002. "Network of WHO Silicosis." Amer J Ind Med 48 (6): 503­14. Collaborating Centres in Occupational Health, Communication and Mbakaya, C. F., H. A. Onyoyo, S. A. Lwaki, and O. J. Omondi.1999. "A Information Dissemination." Asian-Pacific Newsletter on Occupational Survey of Management Perspectives of the State of Workplace Health Health and Safety 9 (2): 28­30. and Safety Practices in Kenya." Accident Analysis Prevention 31 (4): Frumkin, H. 1999. "Across the Water and Down the Ladder: Occupational 305­12. Health in the Global Economy." Occupational Medicine 14 (3): 637­63. Murray, C., D. B. Evans, A. Acharya, and R. M. P. M. Baltussen. 2000. Galinsky, T. L., N. G. Swanson, S. L. Sauter, J. J. Hurrell, and L. M. Schleifer. "Development of WHO Guidelines on Generalized Cost-Effectiveness 2000. "A Field Study of Supplementary Rest Breaks for Data-Entry Analysis." Health Economics 9 (3): 235­51. Operators." Ergonomics 43 (5): 622­38. Nelson, D. I., M. Concha-Barrientos, T. Driscoll, K. Steenland, M. Galinsky, T. L., N. G. Swanson, S. L. Sauter, J. J. Hurrell, L. M. Schleifer, Fingerhut, L. Punnett, A. Prüss-Üstün, J. Leigh, and C. Corvalan. 2005. J. Martin, and others. 1999. "Three Studies of Rest Break Interventions "The Global Burden of Selected Occupational Diseases and Injury for IRS Data Transcribers." Abstract of paper prepared for "Work, Risks: Methodology and Summary." Amer J Ind Med 48 (6): 400­18. Stress and Health 99: Organization of Work in a Global Economy," a NIH (National Institutes of Health). 2003. "Fogarty International Training meeting of the American Psychological Association, Washington, DC, and Research in Occupational and Environmental Health (ITREOH) March. Program." Bethesda, MD, NIH. http://www.fic.nih.gov/. Giuffrida, A., R. F. Iunes, and W. D. Savedoff. 2002. "Occupational Risks in NIOSH (National Institute for Occupational Safety and Health). 2000. Latin America and the Caribbean: Economic and Health Dimensions." Worker Health Chartbook, 2000. Publication 2000-127. Washington, Health Policy and Planning 17 (3): 235­46. DC: U.S. Department of Health and Human Services. Harley, R., and V. Vallyathan. 1996. "History of Silicosis." In Silica and Paek, D., and N. Hisanaga. 2002. "Occupational Health in South Korea." Silica-induced Lung Disease, ed. V. Castranova, V. Vallyathan, and Occupational Medicine 17 (3): 39­408. W. Wallace. Boca Raton, FL: CRC Press. PAHO (Pan American Health Organization). 2002. "The Environment Hecker, S. 1991. Labor in a Global Economy. Eugene, OR: University of and Public Health." In Health in the Americas. Vol. 2. Washington, DC: Oregon Books. PAHO and WHO. ILO (International Labour Organization). 1995. Economically Active Partanen, T. J., C. Hogstedt, R. Ahasan, A. Aragon, M. Arroyave, Population, by Industry and by Occupation: Year Book of Labour J. Jeyaratnam, and others. 1999. "Collaboration between Developing Statistics. 54th ed. Geneva: ILO. and Developed Countries and between Developing Countries in Occupational Health Research and Surveillance." Scandinavian Journal ------. 1996. Year Book of Labour Statistics. 55th ed. Geneva: ILO. of Work and Environmental Health 25 (3): 296­300. ------. 2000. Rampal, K. G., T. C. Aw, and S. B. Jefferelli. 2002. "Occupational Health in ------. 2002a. Decent Work and the Informal Economy. Report VI of Malaysia." Occupational Medicine 17 (3): 409­25. the International Labour Conference, 90th Session, Geneva: ILO. Rosenstock, L., M. Kiefer, W. E. Daniell, R. McConnell, K. Claypoole, http://www.ilo.org/public/english/employment/infeco/ilc2002.htm. and the Pesticide Health Effects Study Group. 1991. "Chronic Central ------. 2003. "ILO Chemical Control Banding Toolkit." ILO, Geneva. Nervous System Effects of Acute Organophosphate Pesticide www.ilo.org/public/english/protection/safework/ctrl_banding/index. Intoxication." Lancet 338 (8761): 223­27. htm. Smith-Jackson, T. L., and A. Essuman-Johnson. 2002. "Cultural ILO and WHO. 2003. "SafeWork: Thirteenth Session of the Joint Ergonomics in Ghana, West Africa: A Descriptive Survey of Industry Committee on Occupational Health." ILO, Geneva. www.ilo.org/ and Trade Workers' Interpretation of Safety Symbols." International public/english/protection/safework/health/session13/. Journal of Occupational Safety and Ergonomics 8 (1): 37­50. 1144 | Disease Control Priorities in Developing Countries | Linda Rosenstock, Mark Cullen, and Marilyn Fingerhut Takala, J. 2002. "Introductory Report: Decent Work--Safe Work." Paper Wang, J. D., T. J. Cheng, and Y. L. Guo. 2002. "Occupational Health in presented at the 16th World Congress on Safety and Health, Vienna, Taiwan." Occupational Medicine 17: 427­35. May 27. WHO (World Health Organization). 2001. "Access to Biomedical Trapido, A. S. M., N. P. Mqoqi, C. M. Macheke, B. G. Williams, J. C. A. Journals." WHO, Geneva. http://www.who.int/inf-pr-2001-32.html. Davies, and C. Panter. 1996. "Occupational Lung Disease in Ex- ------. 2002. The World Health Report 2002--Reducing Risks, Promoting Mineworkers--Sound a Further Alarm" (letter). South African Medical Healthy Life. Geneva: WHO. Journal 86 (4): 559. ------. 2003. WHO Compendium of Activities of the Network of U.K. HSE (Health and Safety Executive). 2002. "COSHH Essentials--Easy Collaborating Centers in Occupational Health. Geneva: WHO. Steps to Control Chemicals." London, HSE. http://www.coshh- WHO and ILO (World Health Organization and International Labour essentials.org.uk. Organization). 1995. Report of the 12th Meeting of the WHO/ILO Joint UN (United Nations). 2000. International Standard Industrial Classi- Advisory Board on Occupational Safety and Health. Geneva: WHO. fication of All Economic Activities (ISIC). 3rd Revision. St/ESA/ ------. 2002a. "Long-Distance Occupational Hygiene Mentoring SER.M/4/Rev3. New York: United Nations. Program." WHO, Geneva. http://www.sheafrica.info/en/About/ U.S. DHHS (Department of Health and Human Services). 1996a. who_cc.htm. "National Occupational Research Agenda (NORA)." DHHS (NIOSH) ------. 2002b. "The WHO/ILO Joint Effort on Occupational Safety and Publication 96-115. Washington, DC, U.S. DHHS. Health in Africa." WHO, Geneva. http://www.sheafrica.info/en/ ------. 2003. "National Occupational Research Agenda (NORA) About.htm. Update." DHHS (NIOSH) Publication 2003-148. Washington, DC, World Bank. 2003. World Development Indicators. Washington, DC: World U.S. DHHS. Bank. Occupational Health | 1145 Chapter 61 Natural Disaster Mitigation and Relief Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio Sudden-onset natural and technological disasters impose a tims. What would be a minor incident in a large country may substantial health burden, either directly on the population or constitute a major disaster in a small isolated island state. Not indirectly on the capacity of the health services to address pri- only are "quantitative definitions of disasters unworkably sim- mary health care needs. The relationship between communica- plistic" as noted by Alexander (1997, 289), but when based on ble diseases and disasters merits special attention. This chapter the economic toll or the number of deaths, they are also mis- does not address epidemics of emerging or reemerging dis- leading with regard to the immediate health needs of the eases, chronic degradation of the environment, progressive cli- survivors or their long-term impact on the affected country. matic change, or health problems associated with famine and temporary settlements. Classification of Disasters In line with the definition of health adopted in the constitu- In the early 1970s, a series of well-publicized disasters (the civil tion of the World Health Organization (WHO), the chapter war and resulting famine in Biafra, the cyclone in Bangladesh, treats disasters as a health condition or risk, which, as any other and the earthquake in Peru) triggered the scientific interest of "disease," should be the subject of epidemiological analysis, the international public health community. systematic control, and prevention, rather than merely as an Disasters can be classified as natural disasters, technological emergency medicine or humanitarian matter. The chapter disasters, or complex emergencies. The latter include civil wars stresses the interdependency between long-term sustainable and conflicts. These classifications are arbitrary and refer to the development and catastrophic events, leading to the conclusion immediate trigger--a natural phenomenon or hazard (biolog- that neither can be addressed in isolation. ical, geological, or climatic); a technologically originated prob- lem; or a conflict. In reality, all disasters are complex events DISASTERS AS A PUBLIC HEALTH CONDITION stemming from the interaction of external phenomena and the vulnerability of man and society. According to the International Federation of Red Cross and The human responsibility in so-called natural disasters is Red Crescent Societies, internationally reported disasters in well acknowledged. The term natural disaster remains com- 2002 affected 608 million people worldwide and killed monly used and should not be understood as denying a major 24,532--well below the preceding decade's annual average human responsibility for the consequences. mortality of 62,000 (IFRC 2003). Many more were affected by myriad local disasters that escaped international notice. Disaster has multiple and changing definitions. The essential Disaster Terminology common element of those definitions is that disasters are The following definitions are adapted from those proposed by unusual public health events that overwhelm the coping capac- the Secretariat of the International Strategy for Disaster ity of the affected community. This concept precludes the Reduction (ISDR), a United Nations (UN) body established to universal adoption of a threshold number of casualties or vic- sustain the efforts of the International Decade for Natural 1147 Disaster Reduction (UN/ISDR 2004) and the WHO World US$ billions Percentage of GDP Health Report 2002 (WHO 2002): 700 16 14 · Hazards are potentially damaging physical events, which 600 may cause loss of life, injury, or property damage. Each haz- 12 500 ard is characterized by its location, intensity, frequency, and 10 probability. 400 · Vulnerability is a set of conditions resulting from physical, 8 300 social, economic, and environmental factors that increase 6 the susceptibility of a community to the effects of hazards. 200 4 A strong coping capacity--that is, the combination of all the 100 strengths and resources available within a community--will 2 reduce its vulnerability. 0 0 · Risk is the probability of harmful consequences (health bur- Richest Nations Poorest Nations den) or economic losses resulting from the interactions Economic losses Losses as percentage of GDP between natural or human-induced hazards and vulnerable or capable conditions. In a simplified manner, risk is Source: UN/ISDR, 2004. expressed by the following function: Figure 61.1 Disaster Losses, Total and as Share of Gross Domestic Product, in the Richest and Poorest Nations, 1985­99 Risk (Hazards Vulnerability) A public health approach to disaster risk management will Annual growth of GDP in Ecuador compared with preceding year (percent) aim to decrease the vulnerability by adopting prevention and 12 mitigation measures to reduce the physical impact and to 10 increase the coping capacity and preparedness of the health 8 sector and community, in addition to providing traditional North-East 6 emergency care (response) once the disaster has occurred. phenomenon phenomenon 4 Rockslides Niño Niño Distribution and Risk Factors 2 El Earthquake El 0 Health and relative economic losses of natural disasters dispro- 2.0 portionately affect developing countries (Alexander 1997; 4.0 UN/ISDR 2004). More than 90 percent of natural disaster­ related deaths occur in developing countries. Even though the 6.0 economic losses are far greater in industrial countries, the per- 8.0 centage of losses in relation to gross national product (GNP) in 198019811982198319841985 1986198719881989199019911992199319941995199619971998199920002001 developing countries far exceeds that percentage in industrial Years countries (figure 61.1).At an individual level,a sudden reduction Source: UN/ISDR, 2004. of US$5,000 from an annual income of US$50,000 is worrisome; Figure 61.2 Annual Growth of Gross Domestic Product and however, the ongoing loss of US$50 from a monthly income of Occurrence of Major Natural Disasters in Ecuador, 1980­2001 US$100 may be catastrophic. For this reason, statistics of economic damage and mortality alone are not true indicators of the effect of disasters on the Hydrometeorological hazards do not follow a well- health and development of people and communities. established distribution. Although the areas subject to seasonal Disaster impact statistics show a global trend: more disasters flood, drought, or tropical storms (cyclones, hurricanes, or occur, but fewer people die; larger populations are affected, and typhoons) are well known locally, global warming may possibly economic losses are increasing (IFRC 2000). redraw the map of climatic disasters. As the National Research Council (1999, 34­35) notes, "This change is far from uniform. Geographic Distribution of Risk. Natural disasters do not A pattern of response `modes' appears to be involved, in which occur at random. Geological hazards (earthquakes and volcanic warming is concentrated in northern Asia . . . while large eruptions) occur only along the fault lines between two tectonic regions of the northern Pacific and North Atlantic Oceans and plates on land or on the ocean floor. However, the local popula- their neighboring shores have actually cooled." El Niño­related tion often does not recognize the implications (the risks), as fluctuations in relation to the gross domestic product (GDP) of shown in the December 2004 tsunami in the Indian Ocean. Ecuador are shown in figure 61.2. 1148 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio The risk of massive technological disasters, such as the cata- locations, particularly urban areas. Following Hurricane Mitch strophic release of chemicals in Bhopal, India (methyl iso- in Tegucigalpa, Honduras, families that were relocated from cyanate), in December 1984, is serious in countries with signif- flooded areas to safer (but inconveniently remote) ground were icant industry (WHO 1992, 1996). Very few countries are rapidly replaced by new illegal settlers. In 2003, families killed immune to public health risks from hazardous chemical sub- by a landslide in Guatemala had been warned about their vul- stances (from insecticides to industrial by-products) or dis- nerability but were unable to afford resettlement in safer (and carded radioactive material from therapeutic or diagnostic use. more costly) areas. Subsidies alone may not have prevented this Technological hazards increase rapidly with the unregulated effect, given the overarching issue of land ownership by a few in industrialization of developing countries and the globalization Central America. of the chemical industry, suggesting that chemical emergencies may become a major source of disasters in the 21st century. Short-Term Health Burden Factors Affecting Vulnerability. Vulnerability to all types of Losses fall under three categories, which may have both direct disasters--and to poverty--is linked to demographic growth, and indirect components: rapid urbanization, settlement in unsafe areas, environmental degradation, climate change, and unplanned development. · lives and disabilities (both direct damage and an indirect consequence) Age The importance of age as a factor of vulnerability can be · direct losses in infrastructure and supplies (direct impact) significant in situations where physical fitness is necessary for · loss or disruption in the delivery of health care, both cura- survival. The higher fatality among children, elderly, or sick tive and preventive (indirect impact). adults following the 1970 tidal wave in Bangladesh (250,000 fatalities) and the 2004 tsunami in Asia (more than 180,000 The immediate health burden is directly dependent on the dead or missing) illustrates this point. nature of the hazard. National health budgets of developing countries are, in normal times, insufficient to meet the basic Gender Reports on immediate morbidity and mortality health needs of the population. In the aftermath of a major dis- according to gender are not as conclusive. An Inter-American aster, authorities need to meet extraordinary rehabilitation Development Bank paper indicated that 54 percent of the demands with resources that often have been drained by the 3,045 people who died as a result of Hurricane Mitch in emergency response (as distinct from the resources destroyed Nicaragua were male (IDB 1999). Stereotypes of gender vul- by the event). Beyond the immediate response, decision mak- nerability at the time of impact often do not apply. Depending ing in the allocation of resources among sectors is mostly influ- on the type of disaster, far more significant vulnerability factors enced by the magnitude of the economic losses rather than by than gender or age are the time of day of the impact (and, the health statistics (principally the disability-adjusted life year, therefore, the occupational activity of each group) and the or DALY, losses) or social costs. structural vulnerability of housing, factories, and public build- ings, including the location of the victims within the buildings. Earthquakes. As noted by Buist and Bernstein (1986), in the Following disasters, increased vulnerability of women is com- past five centuries, earthquakes caused more than 5 million monly noted in temporary settlements, where violence and deaths--20 times the number caused by volcanic eruptions. In sexual abuse are common. Specialized health care also may not a matter of seconds or minutes, a large number of injuries be available (Armenian and others 1997). (most of which are not life-threatening) require immediate medical care from health facilities, which are often unprepared, Poverty Economic vulnerability might play a much greater damaged, or totally destroyed, as was the case in the earthquake role than age and gender. What has been noted regarding the in Bam, Iran, in 2003. In the aftermath of that earthquake, greater vulnerability of poor countries also holds true at the which resulted in 26,271 deaths, the entire health infrastructure community and family levels. Disasters predominantly affect of the city was destroyed. All traumas were evacuated by air to the poor. Poverty increases vulnerability because of the the 13 Iranian provinces long before the arrival of the first for- unequal opportunity for healthy and safe environments, poor eign mobile hospitals. Table 61.1 illustrates the accelerated pace education and risk awareness, and limited coping capacity. A with which priorities evolve and overlap in the first week fol- notable exception was the 2004 tsunami in Banda Aceh, lowing an earthquake. Indonesia, where the middle- and upper-class neighborhood After a few weeks, national political solidarity and external close to the shore was particularly affected. assistance wane, and the local budgetary resources are drained. A major example is the settlement of a large number of eco- At the same time, health authorities face the overwhelming task nomically disadvantaged populations in highly vulnerable of providing services to a displaced population, rehabilitating Natural Disaster Mitigation and Relief | 1149 Table 61.1 Health Priorities Following Earthquakes Priority Time period Comments Search and rescue 0 to 48 hours Returns are rapidly diminishing. Most effective work is done by local teams. Trauma care 0 to 48 hours: initial lifesaving carea External assistance generally arrives too late for initial care. Traumas may 48 hours to 6 months: secondary care include burns and crush syndrome, especially in urban areas. Paraplegics and amputees require long-term care. Routine medical emergencies and Resumes as soon as the need for acute Emergencies include earthquake-related cardiovascular emergencies and primary health care lifesaving care subsides (within 24 hours) premature births. Attention to the dead Varies. Not a public health issue but a Priorities are identification and ritual burial. social and political one Disease surveillance Urgent--within 48 hours, unsubstantiated Surveillance is a sensitive public information and education issue. A simple, rumors of impending epidemics will be syndrome-based system is needed that will involve humanitarian circulating organizations. Provision of safe water A predominant issue within 48 hours The challenge is to provide a sufficient quantity of reasonably safe water. Temporary shelter 48 hours to several months Sanitation and provision of health services is a main issue. Accommodating families near their residence is preferable to setting up camps. Provision of food 3 days to 6 weeks Food provision is a social or economic issue. Food stocks and agricultural output are not affected by earthquakes. Psychosocial care 7 days to 6 months Mental health assistance is best provided by local personnel, if available. Source: de Ville de Goyet 2001. a. Following the earthquake in Mexico City in 1985 (10,000 deaths), bed occupancy rates did not exceed 95 percent despite the loss of 5,829 hospital beds. health facilities, restoring normal services, strengthening com- plants, and reservoirs. Studies by Bernstein, Baxter, and Buist municable disease surveillance and control, and attending to (1986) following the 1980 eruption of Mount St. Helens the long-term consequences, such as permanent disabilities, (United States) reviewed the transient, acute irritant effects of mental health problems, and possibly long-term increases volcanic ash and gases on the mucous membranes of the eyes in rates of heart disease and chronic disease morbidity and upper respiratory tract as well as the exacerbation of (Armenian, Melkonian, and Hovanesian 1998). chronic lung diseases with heavy ash fall. Concentrations of volcanic gases are rapidly diluted to nonlethal levels, which lead Tsunamis. Earthquakes on the ocean floor may cause cata- to inconvenience but negligible morbidity for the general pub- strophic tidal waves (tsunamis) on faraway shores. Waves lic. Lava flows present little health risk because of their very caused by the seismic event crest at less than a meter in open slow speed of progression. Mortality caused by ballistic projec- seas, but they are travel several hundred kilometers per hour, so tiles from a volcanic eruption is minimal. when they reach shallow waters, they can be 10 meters high. Attention to these public concerns may distract the author- Damage on the coast can be extensive. Usually, the number of ities from preparing for the greatest factors of mortality: the survivors presenting severe injuries is small in proportion to pyroclastic flows (Mount Pelé in Martinique, in 1902, with the number of deaths. 29,000 deaths) and lahars. Lahars are mud flows or mud and Volcanic Eruptions. Volcanoes persist as a serious public ash flows caused by the rapid melting of a volcano's snowcap, health concern, though they are often overlooked by authori- as in Colombia in 1985 (23,000 deaths), or caused by heavy ties and communities lulled by long periods of inactivity. rains on unstable accumulations of ash, as in the Philippines in Eruptions are preceded by a period of volcanic activity, which 1991. Historically, pyroclastic explosions or lahars have caused provides an opportunity for scientific monitoring, warning, about 90 percent of the casualties from volcanic eruptions. and timely evacuation. Potential contamination of water supplies by minerals from Some issues, such as ash fall, lethal gases, lava flow, and pro- ash; displacement of large populations for an undetermined jectiles, although of concern to the public, are of minimal period of time (over five years in Montserrat, a small island in health significance: Ash fall causes a significant burden on the Caribbean); accompanying sanitation problems; and men- medical services but is unlikely to result in excess mortality tal health needs are of great public health significance (PAHO or significant permanent problems. However, ash fall affects 2002a). Among the long-term problems, the risk of developing transportation, communications, water sources, treatment silicate pneumoconiosis requires further investigation.1 1150 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio Climatic Disasters. Many communities and health services In the short term, an increased number of hospital visits and have learned to live with seasonal floods of moderate intensity. admissions from common diarrheal diseases, acute respiratory Periodically, the magnitude of the phenomenon exceeds the infections, dermatitis, and other causes should be expected fol- local coping capacity and overwhelms the resources of the lowing most disasters (Howard, Brillman, and Burkle 1996; health systems. The health burden associated with seasonal Malilay and others 1996). This increase may reflect duplicate floods is well known locally: increased incidence of diarrheal reporting (diarrhea cases were reported through both the diseases, respiratory infections, dermatitis, and snake bites. The emergency and the routine surveillance systems in Maldives actual risk of compromised water supplies depends on the level after the 2004 tsunami), a temporary surge in surveillance, and of contamination of the community's water supply before the medical attention available to an otherwise underserved popu- disaster, compared with contamination after the flooding. lation rather than representing a genuine change in the epi- Saline contamination is a long-term issue following sea surges demiological situation. and tsunamis. Prolonged flooding endangers local agriculture In the medium term, heavy rainfalls may affect the trans- and occasionally requires food assistance on a large scale. The mission of vectorborne diseases. Following an initial reduction primary factors of morbidity remain overcrowded living con- as mosquito-breeding sites wash away, residual waters may ditions and poor water and sanitation in temporary settle- contribute to an explosive rise in the vector reservoir. When ments and other areas where water and sanitation services have associated with a breakdown of normal control programs, this deteriorated or are suspended. rise in the vector reservoir may lead to epidemic recrudescence Mortality and morbidity caused by tropical storms (hurri- of malaria or dengue. Retrospective studies (Bouma and Dye canes in the Atlantic Ocean and typhoons in the Pacific Ocean) 1997; PAHO 1998; UN/ISDR 2004, 156) all confirm a direct result from, in increasing order of importance, high winds, but delayed relationship between the intensity of rainfall heavy rainfall, and storm surge. When Hurricanes Mitch and (regardless of the existence of flooding) caused by the El Niño George hit the Caribbean in 1998, traumatic injuries (lacera- phenomenon and the incidence of malaria. Flooding has con- tions or electrocution) caused by high winds of up to 150 miles tributed to local outbreaks of leptospirosis (in Brazil and per hour were relatively few; deaths from extensive rainfall Jamaica, for example; PAHO 1982) and hepatitis A in Latin (leading to flash floods and landslides) constituted the bulk of America and Africa (WHO 1994). the more than 13,000 fatalities (PAHO 1999). In the In summary, what can be expected and prevented is a local Bangladesh delta, storm surges up to 6 meters traveled unim- surge in problems that the health services are normally used to peded over hundreds of kilometers and claimed between handling. 250,000 and 500,000 lives in 1970 and up to 140,000 lives dur- ing five cyclones in the 1990s--primarily during one storm in 1991. Another cost is the need for specialized psychosocial Long-Term Impact and Economic Valuation assistance to large numbers of the population who survive the In addition to the delayed impact on transmission and control sustained violence of nature. of endemic diseases and the burden of disabilities (paraplegia, Cumulative mortality caused by small, undocumented amputation, burns, or chronic or delayed effects of chemical or mudslides and rockslides from water-saturated, unstable slopes radiological exposure), the health sector bears a significant probably approach the toll from well-known landslides share of the economic burden. Disasters must be seen in a sys- (earthquakes in Peru in 1970 and in El Salvador in 2001, and temic (that is, intersectoral) manner: what affects the economy the rains in Caracas, Venezuela, in 1999). Morbidity problems will affect the health sector--and vice versa. After the emo- are often minimal, as survivors in the path of the landslide tional response of the first few days, decision makers in a crisis are few. react primarily to political and economic realities, not to health indicators. Economic valuation of the social burden--that is, Impact on Communicable Diseases placing a monetary value on the cost--becomes a critical tool Disasters related to natural events may affect the transmission as the various sectors compete for scarce resources. The health of preexisting infectious diseases. However, the imminent risk sector, in particular, must learn how to use this tool in spite of of large outbreaks in the aftermath of natural disasters is over- being absorbed by its immediate relief responsibilities. stated. Among the factors erroneously mentioned is the pres- ence of corpses of victims, many buried beneath rubble. Dead Valuation of Disasters. The Economic Commission for Latin bodies from a predominantly healthy population do not pose a America and the Caribbean (ECLAC) has developed over the risk of increased incidence of diseases (Morgan 2004). decades a methodology for the valuation of disasters (ECLAC Catastrophic incidence of infectious diseases seems to be con- 2003). This tool, intended for reconstruction, has also proved fined to famine and conflicts that have resulted in the total fail- its usefulness by developing historical records of major events, ure of the health system. particularly of the health burden expressed in economic terms. Natural Disaster Mitigation and Relief | 1151 Valuation is made using all possible sources of information, nents of the cost of illness--the cost of treatment and the cost from georeferenced satellite mapping and remote sensing to of lost opportunities (lost income and employment, loss of more conventional statistical data, direct observation, and sur- time and productivity)--are sharply increased. The social bur- veys, with a reliance on information gathered immediately after den is heavier on the poorest, who are unable to adjust their the event. Economic valuation rests on the basic concepts of willingness to pay to absorb the additional expenses of alterna- direct damage and indirect losses. tive (private) providers of care. Direct damage is defined as the material losses that occur The same approach applies to the economic valuation of as an immediate consequence of a disaster.2 Direct damage is lives lost. Kirigia and others (2004) found a statistically signifi- measured first in physical terms. The physical loss includes cant impact of disaster-related mortality on the GDP of African assets, capital, and material things that can be counted: hospi- countries. One single disaster death reduced the GDP per tal beds lost, equipment and medicines destroyed, damaged or capita by US$0.01828. Lost lives are given a higher economic affected health service installations (number and type of instal- value in places where productivity is high. lations, stocks of medicines, laboratory facilities, operating Because economic valuation uses standard sectored proce- rooms, and so on), and pipes and water plants destroyed. dures that allow comparability of results, it can be used in the The physical plant then is valued both in terms of dis- decision-making process and for policy formulation since it counted present value and estimated replacement cost. Recon- identifies sectors, geographical areas, and vulnerable groups structing facilities with the same vulnerability and level of that are more severely affected economically. Over the years, a service as before would be unacceptable; the affected health number of conceptual improvements have been made to allow infrastructure must be replaced by more resilient and efficient for the measurement of aspects not included in national installations to ensure better and sustainable service. This need accounting systems--to bring attention to environmental is most evident in developing countries where impacts tend to losses as a cross-cutting issue; to highlight the contribution of be concentrated in those most at risk (the poor, marginalized, specific groups, namely women, as agents for change; and to and less resilient sectors of the population). focus on the better management of both the emergency and the Indirect effects refer to production of goods and services that reconstruction processes. It is also a valuable tool for prepared- will not occur as an outcome of the disaster, reduced income ness and mitigation of future damage. associated with those activities not occurring, and increased Table 61.2 summarizes the valuations made by ECLAC over costs to provide those goods and services. the years for Latin America and the Caribbean in terms of deaths, In the case of health services, indirect effects encompass affected populations, and economic losses (2003 values). Of both the income losses associated with the diminished supply interest are the decrease in the number of deaths and the increase of health care services and the increased costs of providing the in total damage (in particular, indirect damage) over time. services following the disaster. Indirect effects are valued at The distribution of direct and indirect damage in the health the current market value of goods or services not produced sector also varies. According to ECLAC (2003), direct damage and the costs associated with the necessary provision of servic- between 1998 and 2003 in Latin America ranged from 44.6 per- es under emergency, disaster-related conditions. Both compo- cent to 77.2 percent of total damage. Table 61.2 Impact of Disasters in Latin America and the Caribbean Population Damage (2003 US$ millions) Date Deaths Affected Total Direct Indirect 1972­80 38,042 4,229,260 9,376 5,420 3,956 1981­90 33,638 5,442,500 19,603 13,916 5,687 1991­2000 11,086 2,318,508 20,902 10,401 10,501 2001­2002 120 4,828,470 4,498 2,270 2,228 Total of major events 1972­2002 82,886 16,818,738 54,379 32,007 22,372 Overall estimate including small disastersa 103,608 21,023,422 67,974 40,009 27,965 Average per year 3,454 700,781 2,266 1,334 932 Source: ECLAC 2003. a. The full image should include the recurrent small disasters that do not make the headlines but have a cumulative negative effect. Such disasters can be more pervasive and damaging to the develop- ment process because their economic, social, psychological, and political effects are hardly perceived. An estimate of the average losses of small disasters would be at least 25 percent greater than those of large disasters. 1152 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio Specific Damage to the Health Infrastructure. Damage to The developmental burden is significant. In the past 30 years housing, schools, channels of communication, industry, and so in Latin America and the Caribbean alone, an estimated 400 on contributes to the health burden. However, the following urban water supply systems and 1,300 rural systems (in addition analysis focuses on the health infrastructure (understood as to 25,000 wells and 120,000 latrines) were severely damaged, at health care facilities, including hospitals, health centers, labora- an estimated cost of almost US$1 billion--a major setback to tories, and blood banks) and the drinking water and sanitation efforts to expand coverage and improve those services. In severe infrastructure. flooding, the sudden interruption of these basic services coin- cides with the direct effect on the transmission of waterborne or Damage to Hospitals and Health Installations Most data and vectorborne diseases. In the case of earthquakes, the number of examples presented here come from Latin America and the people who are adversely affected by water shortage may far Caribbean because of the disaster reduction programs in the exceed those injured or suffering direct material loss. health sectors of those regions. In the past two decades, damage As in the case of health care facilities, the rehabilitation of to approximately 260 hospitals and 2,600 health centers resulted public water systems is slow, particularly for community- in interruption of services at a direct cost of US$1.2 billion. In owned or community-operated rural systems, which may not the 1985 earthquake in central Mexico, 5,829 beds were be repaired for decades. The foregoing demonstrates the need destroyed or evacuated (PAHO 1985), at a direct cost of for water authorities to harmonize their short-term objectives, US$550 million (ECLAC 1998). Hurricane Gilbert (1988) which are oriented almost exclusively to increasing the cover- damaged 24 of the 26 hospitals on Jamaica, and the El Salvador age of these services, with the long-term objective of reducing earthquake (2001) resulted in the loss of 2,000 beds--40 per- vulnerability to extreme natural hazards. cent of the country's hospital capacity (PAHO 2002b). The health burden is not limited to the loss of medical care. The control of communicable diseases and other public health pro- INTERVENTIONS: FROM RESPONSE grams suffer from loss of laboratory support and diagnostic TO PREVENTION capabilities of hospitals. Further research on the actual impact The immediate lifesaving response time is much shorter than of these losses, in terms of DALYs, is essential. humanitarian organizations recognize. In a matter of weeks, if A common misperception is that damage to critical health not days, the concerns of both the population and authorities facilities is promptly repaired. Experience shows that damaged shift from search and rescue and trauma care to the rehabilita- health infrastructure recovers at a slower pace than infrastruc- tion of infrastructure (temporary restoration of basic services ture in other service sectors, such as trade, roads, bridges, and reconstruction). In Banda Aceh, Indonesia, after the telecommunications, and even housing. For example, as a result December 2004 tsunami, victims were eager to return to nor- of the earthquake that affected El Salvador in 1986, renovation malcy while external medical relief workers were still arriving in of the general hospital, the most sophisticated referral hospital large numbers. in the capital, was completed 15 years after the earthquake. The only national pediatric facility was fully rehabilitated and strengthened six years after the earthquake. Two years after the Response and Rehabilitation earthquake of 2001 in El Salvador, several key hospitals still Immediate emergency response is provided under a highly polit- remained vacated or services were transferred to unsuitable ical and emotional climate. The public demands visible, albeit temporary facilities. The factors are many: low priority assigned perhaps unnecessary, measures at the expense of proven low-key to a nonproductive sector, the sector's inexperience in develop- approaches. The international community, eager to demon- ing comprehensive proposals for funding, conflicting attempts strate its solidarity or to exercise its"right of humanitarian inter- to use the reconstruction process to influence the ongoing vention,"undertakes its own relief effort on the basis of the belief reform and decentralization processes, the novelty of the engi- that local health services are unwilling or unable to respond. neering and design issues for safe hospital construction, the Donations of useless medical supplies and medicines and the complicated negotiation process for loans, and the administra- belated arrival of medical or fact-finding teams add to the stress tive inexperience of the health sector in executing large invest- of local staff members who may be personally affected by the dis- ment projects. Indeed, few large health installations have been aster. The cultural disregard of the humanitarian community to built directly by developing countries in the past decades. cost-effective approaches in times of disaster and the tendency to base decisions on perceptions and myths rather than on facts Damage to Water and Sewage Systems The primary goal of and lessons learned in past disasters contribute to making disas- water and sewage systems is to safeguard the public health of ter relief one of the least cost-effective health activities. the population. For that reason, these systems are considered The responsibilities of the national or local health authori- part of the health infrastructure. ties are significant. Natural Disaster Mitigation and Relief | 1153 Assessment of the Health Situation. A country's ministry of Environmental Health. Typical interventions in the aftermath health is expected to assess the health situation. To influence the of disasters include strengthening the monitoring and surveil- course of humanitarian response, this assessment must be rapid lance of water quality, vector control, excreta disposal, solid and, therefore, simple; transparent in collaboration with the waste management, health education, and food safety. main actors--nongovernmental organizations (NGOs) and A first priority is water supply. It is often preferable to have donors; and technically credible. The input of WHO, as the lead a large quantity of reasonably potable water than a smaller agency in health matters, is most valuable. Confusion should be amount of high-quality water (UNHCR 1998). Massive distri- avoided between assessing emergency needs and inventorying bution of water purification disinfectants can be effective if the or valuating the damage. In the first hours or days, relief actors public is already familiar with their use and not confused by base their decision making on the ministry of health's assess- the availability of many different brands and concentrations ment of what is required and, more importantly, what is not of donated chemicals. required for emergency response. Later, the international Health education and hygiene promotion efforts target pop- community will request detailed data, such as the number of ulations in shelters, temporary camps, collective kitchens, or persons affected, buildings damaged, and monetary valuation. prepared food distribution centers. The cost-effectiveness of the external relief effort could Mass Casualties Treatment. Following natural disasters, often be increased by shifting resources from the overattended hospital capacity may be considerably reduced by actual dam- medical response to the improvement of environmental health age to the facility or, in the case of a seismic event, an often in temporary settlements. unnecessary--but hard to reverse--evacuation. Triage of patients is required in order to first treat those likely to benefit Transparent Management of Donations and Supplies. If most, rather than the terminally injured or those whose care can donations and supplies are managed transparently during the be delayed. Lifesaving primary care takes place in the first six emergency, the flow of assistance to the intended beneficiaries hours (the golden rule of emergency medicine), making most of will be improved. Unsolicited and often inappropriate medical the foreign field hospitals irrelevant for intensive acute care of donations compete with valuable relief supplies for scarce traumas (WHO and PAHO 2003). Effectiveness of immediate logistical resources. Good governance is critical, and effective care will depend on local preparedness before the disaster, not logistics cannot be improvised following a disaster. A human- on faraway resources. itarian supply management system developed by PAHO and WHO successfully helped developing countries improve Strengthened Surveillance, Prevention, and Control of transparency and accountability in managing humanitarian Communicable Diseases. Because the surveillance,prevention, supplies and donations (de Ville de Goyet, Acosta, and others and control of communicable diseases are strengthened, the 1996). anticipated massive outbreaks generally do not actually occur. Traditional surveillance systems that are based on periodic Coordination of the Humanitarian Health Effort. Coordina- notification of diseases by the health services are inadequate in tion of the humanitarian health effort is essential to maximize a crisis situation. Early warning requires flexible and simple the benefit of the response effort and ensure its compatibility syndrome-based monitoring in temporary settlements and with the public health development priorities of the affected health centers, with information collected not only by the offi- country. Effective coordination in the health sector must do the cial health services but also by the medical humanitarian following: organizations. Systems that do not include consultation with NGOs are unlikely to succeed. Disease control programs in place before the disaster are the · Be comprehensive and include all external health actors. fruit of local experience and external technical advice. In a dis- · Be based on mutual respect rather than regulatory authority aster situation, there is generally no need to resort to new and alone. Dialogue and consultation are more effective than expensive control measures. The key is to quickly resume, enforcement. strengthen, and better monitor the routine control programs. · Benefit all parties, starting with the victims. It should aim to No public health concerns justify the hurried disposal of support and facilitate the activities of other partners. corpses through mass burial or unceremonious incineration. · Be evidence-based and transparent. Information is made to This practice is socially and culturally damaging. In addition, be shared and used, not jealously guarded. improvised mass immunization campaigns, especially by external relief groups, should be discouraged in favor of oppor- Coordination cannot be improvised in the aftermath of a tunistically strengthening national routine immunization disaster. Preparedness before the occurrence of the hazard is coverage, especially in temporary settlements. essential. 1154 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio Emergency Preparedness of the Health Sector Prevention and Mitigation Effective response by national health authorities cannot be The slogan "prevention is better than cure" was invented by the impromptu. Ministries of health that neglected to invest in health sector. However, this sector has been slow to adopt the capacity building before emergencies have generally experienced concept of preventing deaths and injuries from disasters serious difficulties in exercising their technical and political through the mitigation (that is, reduction) of damage to its leadership in the immediate aftermath of a disaster. Disaster own facilities. As is unfortunately often the case, political action preparedness is primarily a matter of building institutional is often triggered only by a major disaster, such as the collapse capacity and human resources, not one of investing heavily in of Hospital Juarez in Mexico in the earthquake of 1985; in that advanced technology and equipment. disaster 561 patients and employees died, (Poncelet 1997). Building local coping capacity is one of the most cost- Evaluating the damage (the past vulnerability) helps establish effective ways to improve the quality of the national response mitigation criteria for the future. and the external interventions. The level of protection required for each health installation Disaster preparedness is not merely having a disaster plan must be negotiated--from life protection, which prevents an written by experts. It must involve the following: immediate structural collapse to permit the evacuation of peo- ple; to investment protection, which minimized the economic · Identifying vulnerability to natural or other hazards. The losses; to operational protection, which guarantees the sustain- health sector should seek information and collaborate with ability of services under any extreme circumstances. Though other sectors and institutions (civil protection, meteorology, a commercial or office building may be structurally designed environment, geology) that have the primary responsibility only to prevent loss of lives, key hospitals must remain opera- for collecting and analyzing this information. tional during the times they are most needed. · Building simple and realistic health scenarios of a possible Local engineering and architectural experts play a key role in and probable occurrence. It is challenging enough to pre- developing the knowledge, technical abilities, and cost- pare for a moderate-size disaster; building and sustaining a effectiveness analysis to establish mitigation priorities. Technical culture of fear based on unrealistic worst-case scenarios may mitigation guidelines prepared at a global level (PAHO, WHO, serve the corporate interests of the disaster community but World Bank, and ProVention Consortium 2004) need to be not the interests of the public at large. adapted to local culture, conditions, and resources. · Initiating a participative process among the main actors to Reducing the physical vulnerability of infrastructure develop a basic plan that outlines the responsibilities of can take place on three different occasions (UN/ISDR 2004, each actor in the health sector (key departments of the 324): ministry of health, medical corps of the armed forces, pri- vate sector, NGOs, UN agencies, and donors). What matters · When reconstructing the infrastructure destroyed by a disaster. is the process of identifying possible overlaps or gaps and At that time, risk awareness is high, political will is present, building a consensus--not the paper plan itself. Disasters and resources are available. often present problems that are unforeseen in the most · When planning new infrastructure. Reducing vulnerability detailed plans. is most cost-effective and politically acceptable when it is · Maintaining a close collaboration with these main actors. A included at the earliest planning and negotiation stage, good coordinator is one who appreciates and adapts to the whether it involves a 1 to 2 percent additional cost for wind strengths and weaknesses of other institutions. Stability is resistance or a 4 to 6 percent additional cost for earthquake essential. Changes of key emergency staff members during a resilience. Full resistance to any damage is prohibitively disaster situation or when a new administration or minister expensive. take over have occasionally complicated the tasks. · Strengthening of existing facilities (retrofitting). This most · Sensitizing and training the first health responders and expensive measure has been adopted by several developing managers to face the special challenges of responding to countries (Chile, Colombia, Costa Rica, Mexico, Peru, and disasters. Participation of external actors (UN agencies, others) to protect their most critical health facilities. In the donors, or NGOs) in designing and implementing the train- earthquake in Colombia in 1999, partial retrofitting of the ing is critical. The incorporation of disaster management in main hospital is credited for saving the installation. Costs the academic curriculum of medical, nursing, and public vary greatly (see table 61.3). health schools should complement the on-the-job training programs of the ministry of health, UN agencies, and Mitigation of Damage to Hospitals. Mitigation does not NGOs. Well-designed disaster management training pro- pretend to eliminate all possible damage from hazards but aims grams will improve the management of daily medical emer- to ensure the continuing operation of the health facility at a gencies and accidents as well. level previously defined by the health authority. Hospitals Natural Disaster Mitigation and Relief | 1155 Table 61.3 Retrofitting of Hospitals in Costa Rica Duration of Cost of Percentage of total Hospital Number of beds retrofitting (months) retrofitting (US$) value of the hospital Hospital Mexico 600 31 2,350,000 7.8 Children's Hospital 375 25 1,100,000 4.2 Hospital Monseñor Sanabria 289 34 1,270,000 7.5 Source: PAHO and WHO 2000. should be subject to stricter norms than other less critical The health sector should, therefore, coordinate with the facilities that are designed to prevent only total collapse and institutions in charge of constructing, operating, and main- loss of life. taining water and sanitation services, both urban and rural, to Hospital mitigation interventions fall into three categories: promote reduction of the vulnerability of existing systems. The health sector should also ensure that health aspects and mitiga- · Functional mitigation to ensure that the necessary support- tion of damage be included in the regulatory framework and ing infrastructure services permit continuing operation: operating procedures of water and sanitation services. water, electricity, road access, communications, and so Protecting the water supply is feasible in developing forth. Improving routine maintenance will facilitate opera- countries. The Costa Rican Institute of Aqueducts and tions under normal circumstances and in the event of Sewage Systems reduced the vulnerability of one of the main extreme hazards. aqueducts of the country, the Orosi Aqueduct. Over 10 years, · Nonstructural mitigation to reduce losses and health injuries Costa Rica invested almost US$1.5 million in studies and from falling or moving objects. Measures include, for reinforcements, an amount equivalent to 2.3 percent of the total instance, proper anchoring of equipment for earthquakes or cost of the aqueduct. This investment would prevent a loss of strong winds or the location of only noncritical services on nearly US$7.3 million in direct damages alone (FEMICA 2003). flood-prone floors. · Structural mitigation to ensure the safety of the structure itself (columns, beams, load-bearing walls). INTERVENTION COST, COST-EFFECTIVENESS, Given the high economic, health, and political costs repre- AND ECONOMIC BENEFITS sented by the avoidable loss of critical health facilities, health The highly emotional and sensationalized climate of disaster authorities and funding agencies should require that, in all new response has long prevented the adoption of a cost- health infrastructure projects, natural hazards be a decisive effectiveness approach in decision making. When survival of factor for selecting the facility's location and for formulating both people and political institutions is threatened, perceptions the specifications at the earliest stage of the process. and visibility tend to prevail over facts and analysis, resulting in Mitigation of Damage to Water Systems. Unlike hospitals, a lack of evidence-based studies on costs and benefits. water supply systems are geographically extensive and thus are The willingness to spend hundreds of thousand of dollars exposed to different types of hazards. The search for technical per victim rescued from a collapsed building in a foreign coun- solutions is more complex, given the diversity of the water try is a credit to the solidarity of the international community, system's components. Finally, in many countries, the health but it also presents an ethical issue when, once the attention has authorities have no jurisdiction over the construction or oper- shifted away, modest funding is unavailable for the mid-term ation of those services owned or administered by many local or survival of tens of thousands of victims. municipal agencies. Even a short disruption of water services may have serious and direct implications for the health of individuals, the oper- Cost-Effectiveness of Selected Humanitarian Interventions ation of health services, and the community at large through its Emergency health interventions are more costly and less effec- impact on business. A probabilistic model studied the disrup- tive than time-tested health activities. Improvisation and rush tive potential of a water outage in the event of an earthquake inevitably come with a high price. The preferential use of in Los Angeles county in the United States. As noted by the expatriate health professionals; the emergency procurement authors, "water outage is more likely to be disruptive for busi- and airlifting of food, water, and supplies that often are avail- nesses in some industries, such as health services, than for able locally or that remain in storage for long periods of time; others" (Chang and Chamberlin 2004, 89). and the tendency to adopt dramatic measures contribute to 1156 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio making disaster relief one of the least cost-effective health from the international community. Guidelines for the use of activities. foreign field hospitals are available from WHO and PAHO (2003). Search and Rescue. Few developing countries have established the technical capacity to search for and attend to victims In-Kind Donations. Unsolicited donations of inappropriate trapped in confined spaces in the event of the collapse of mul- medical supplies not only are of limited use, but often cause tistory buildings. Industrial nations routinely dispatch search serious logistic, economic, and political problems in the recip- and rescue (SAR) teams. Costs are high and effectiveness is ient country. Warehousing those supplies and, in many reduced by delayed arrival and quickly diminishing returns. instances, building facilities (incinerators, for example) for the Following the 1988 earthquake in Armenia, in the former safe disposal of pharmaceutical donations diverts humani- Soviet Union, the U.S. SAR team extracted alive only two vic- tarian funds from more effective uses. Recipient countries col- tims at a cost of over US$500,000. In Turkey in 1999, 98 per- lectively share part of the responsibility by not clearly indicat- cent of the 50,000 people pulled alive from the rubble were sal- ing what they do not want to receive and by not speaking out vaged by relatives and neighbors. In Bam in 2003, the absence once inappropriate items arrived. of high-rise and reinforced concrete buildings ruled out the need for specialized teams. Nevertheless, according to UN sta- Disease Prevention and Control. Postdisaster interventions tistics, at least US$2.8 million was spent on SAR teams. An in surveillance and control of communicable diseases should alternative solution consists of investing these resources in focus on strengthening existing programs. Benefits will outlive building the capacity of local or regional SAR teams--the only the crisis. Improvised mass immunizations (instead of ones able to be effective within hours--and training local improved sanitation and public awareness) and vector control hospitals to dispatch their emergency medical services to the by aerial spraying or fogging (instead of breeding-site reduc- disaster site. tion or waste disposal) are just two examples of wasteful managerial decisions. Field Hospitals. The limited lifesaving usefulness of foreign field hospitals has been discussed. Again, the lessons learned Shelters. Tent cities should be a last resort. Family-size tents from the Bam earthquake are clear. The international commu- may be expensive and do not last long. Establishing large set- nity spent an estimated US$10.5 million to dispatch approxi- tlements is easy, but such settlements are difficult to sustain and mately 10 mobile hospitals,3 which arrived from two to five nearly impossible to terminate. They come with their own san- days after the impact, long after the last casualty had been itation problems and social shortcomings (lack of privacy, loss evacuated to other Iranian provinces. This delay alone, hard to of family identity, and loss of empowerment). Distributing reduce further, rules out any significant contribution to imme- construction material (or, preferably, cash subsidies) is more diate trauma care and led the hospitals to compete for routine cost-effective and tailored to the needs and priorities of end outpatient care with the teams of Iranian volunteers from users. across the country. A few of the mobile hospitals, better pre- pared to meet nontrauma needs and to stay much longer than Cash Assistance. Developed societies long ago abandoned the the usual two to three weeks, have been invaluable. No data are distribution of in-kind relief goods and services to their available on the number of lives actually saved by mobile hos- nationals in favor of direct financial assistance in the form of pitals (that is, lives that would not have been saved by local subsidies, grants, or tax relief. The individual is free to deter- means). Less understood are the negative effects of such hos- mine actual priorities and to seek the most cost-effective source pitals on local health services, which are often marginalized of services (shelter, medical, food, or other). It is therefore sur- and discredited for their lack of technology and sophistication prising that external assistance from these same countries but which must cope once the external facility leaves. remains focused on the costly delivery of predetermined serv- The cost of mobilizing a mobile hospital for a few weeks ices or commodities. often exceeds US$1 million, funds that would be more pro- The most immediate lifesaving needs can be addressed only ductive in the construction and equipping of a simple but locally with existing resources and capacity. No cash contribu- sturdy temporary facility. Such an approach was adopted by tion will meet those immediate needs. Beyond the acute phase, the U.S. Army Southern Command in Wiwili, Nicaragua, in in many countries with market economies, most other services the aftermath of Hurricane Mitch. In the case of Bam, Iran, the and goods are easily procured by those with financial means, cost of rebuilding the entire primary and secondary health suggesting that income availability is often the single limiting care facilities and teaching institutions was estimated by the factor in rehabilitation. government of Iran to be US$10.75 million, an amount very Undoubtedly, this approach would affect considerably the similar to that expended for the dispatch of field hospitals type (and number) of humanitarian actors by transferring Natural Disaster Mitigation and Relief | 1157 power and decision making to the local beneficiaries and From a ministry of health point of view, competition for relying on local economic forces for delivery to the end user. It disaster resources is with other sectors or humanitarian organ- may also bring its own set of problems (and abuses), though izations, not within the sector (as it would be, for instance, with perhaps that is a small cost, considering the economic and malaria or tuberculosis control projects). social benefits of the most interested party--the victim--being in charge. Funding for Preparedness. "By strengthening our public health planning for natural disasters and disease outbreaks, we will be in a better position to care for our populations, regard- Cost-Effectiveness of Prevention and Mitigation less of the type of hazard that confronts our health depart- The social benefits of making hospitals and water systems more ments" (Rottman 2003, 1). This message, addressed to the resilient to the effects of natural hazards are recognized but too public health community in the United States, is even more rarely applied. On the economic side, mitigation also increases pertinent for developing countries. Most humanitarian offices the investment capacity in the health sector by preventing in more developed countries allocate a modest but increasing losses and the need for reconstruction (PAHO and UN/ISDR proportion of their funds for predisaster capacity building. The 1996; Bitrán 1996). capacity of the ministries of health to secure directly nonreim- The most compelling case for the cost-effectiveness of bursable funding depends on the following: mitigation can be made during the planning phase for new installations, when costs of additional structural safety are · The existence of an established disaster program within the minimal. Although the social benefits of prevention and risk ministry, demonstrating a long-term commitment to health management are more evident in the health sector than in oth- disaster preparedness. ers, further studies are needed to provide decision makers with · An ongoing dialogue with local representatives of donors quantified parameters of the economic benefits brought about and their prior involvement in disaster-related activities or by investment in risk management and disaster reduction. meetings of the health sector. PAHO and UN/ISDR (1996) studies indicate that such · A realistic projection of concrete activities, taking into con- increased investment fluctuates between 4 and 8 percent of a sideration the efforts of others, especially NGOs. One- or hospital's local construction cost.When the value of services lost two-year training or capacity-building projects are more is added to the infrastructure loss, the additional investment is likely to be supported than those of longer duration that reduced to between 2 and 4 percent of direct and indirect losses have recurrent costs or involve the purchase of equipment observed. Even though this is a gross estimate that requires fur- (radios, vehicles). ther research in other regions and types of health facilities, the · The technical endorsement and support of WHO and other figure is ratified by the estimated cost of reinforcement, which UN agencies. fluctuates but averages between US$2,000 and US$5,000 per bed, compared with the average cost of a new hospital bed of A multisectoral preparedness component is also increas- between US$100,000 and US$150,000 (at 1996 prices). ingly included in loans negotiated in the aftermath of disasters. Prevention of chemical and radiation accidents can be a Intended to strengthen the capacity of the civil protection highly cost-effective expense that is normally absorbed by the agency, the funding is no substitute for local political commit- respective industries. Respect for existing norms in the use of ment to assume recurrent expenses, the only guarantee of radiotherapy and diagnostic equipment and, once such equip- sustainability. ment is decommissioned, its proper disposal reduces DALYs from accidents at a modest cost. Resources for Emergency Response. The amount of external resources available for response, financial or material, is influ- enced by the type of hazard, geopolitical considerations, and Mobilization of Resources the number of deaths (rather than that of survivors in need of Funding for preparedness and response programs follows rules assistance). Funding is channeled mostly through humanitar- and procedures that are distinct from those applicable to devel- ian NGOs, the Red Cross system, or multilateral organizations, opment projects. Most donors maintain a specific office or rather than through national governments. Consequently, the department for humanitarian affairs with a separate budget priority of the health authorities, rather than to seek direct con- line. Procedures are also streamlined for quick response to tributions to the ministry, should be to ensure that health needs unexpected situations. Processing a request takes a matter of are properly identified and adequately covered by those agen- days in emergencies and takes months for preparedness or mit- cies benefiting from the donations. Ministries of health often igation projects, but it can take years in typical development can obtain indirect financial support for their own activities projects negotiated with donors or financial institutions. through UN projects. 1158 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio Concentrating on several key factors will improve the flow within their ministries of health. Some lessons can be learned of external resources toward health priorities: from this process: · Issuing a rapid and reliable assessment of what is needed and · The occurrence of a major disaster in the country or its what is not needed for the emergency response, rather than neighbor is the initial catalyst for health authorities to rec- waiting for a detailed assessment of the physical damage. ognize that disasters represent public health risks that must · Focusing on tomorrow's emergency health problems. be addressed in an institutionalized manner. External response is unable to address today's short-lived · Access to and support from the political level has deter- problems. mined the success or failure in coordinating the external and · Keeping a long-term view. Funding for emergency response domestic health response. is limited to a few months, whereas the health problems · A multihazard program covering the entire health sector is caused by the disaster will stay much longer. Projects should most effective. Assigning responsibility for coordination offer sustained benefits beyond their conclusion. and management among different technical departments · Recognizing shortcomings in governance when in contact according to the type of hazard (chemical or natural, for with the many bilateral fact-finding or assessment missions instance) does not work. coming to the disaster site. · A risk management program should cut across departments (medical care, epidemiology, water supply, sanitation, nutri- Funding for Reconstruction. Funding for reconstruction is tion, and so forth) of the ministry of health and become sec- multisectoral and is often coordinated by an international tor wide. financing institution (global or regional), together with a con- · The synergy between normal development, preparedness, sortium of large donor countries. The health sector will com- and disaster response activities should be recognized. pete with other social priorities and the "productive" sectors in Poor development practices increase vulnerability, whereas an arena where the health burden (measured in DALYs) does preparedness improves the attention to daily health not carry the same weight as economic factors. Success will challenges. Programs narrowly focused on operational depend on an exhaustive monetary valuation of the health response have generally failed. damage, rapid formulation of projects, political support from the country's highest authorities, and technical support and In Asia, the Asian Disaster Preparedness Center also has endorsement of specialized UN agencies and larger NGOs. documented some interesting experiences (http://www.adpc. Funding for Mitigation of Damage. Protecting the national ait.ac.th/). capital investment of the health sector is primarily the respon- sibility of the country at risk. Development agencies or finan- THE RESEARCH AND DEVELOPMENT AGENDA cial institutions may contribute only marginally to the actual cost of retrofitting installations or improving the design of new Disasters in any one country are relatively infrequent. In addition facilities. to being a dangerous temptation for the authorities to postpone Modest funding for pilot or demonstration prevention pro- preventive actions, this infrequency is an impediment for grams may be available from both the humanitarian and the research and institutional memory.On one hand,the humanitar- development sources of donor countries. Humanitarian offices ian culture tends to raise ethical questions on the role of observers may support promotion of the concept, development of guide- at a time when action at all costs is expected. On the other hand, lines or studies on vulnerabilities, and training. few health academicians wish to embark on projects when con- The health sector will benefit from close contacts with trol groups and time for advance planning are unavailable. financial institutions, the ministry of foreign affairs, and other Particularly encouraging are the increased numbers of pub- national ministries. Negotiations to ensure that new installa- lications and guidelines by UN organizations and NGOs and tions are able to withstand disasters must be initiated at the the trend toward organizing workshops on lessons learned a earliest opportunity, and the corresponding additional costs few months after a major disaster. These meetings of national should be considered in the earliest stages of the project. experts and officials together with representatives from exter- nal actors are invaluable for identifying and sharing opera- tional or institutional successes and failures for the collective IMPLEMENTATION OF CONTROL STRATEGIES: benefit of other countries at risk. LESSONS OF EXPERIENCE AND CHALLENGES FACED Epidemiological Research All countries in Latin America and the Caribbean have estab- Most of the DALYs attributable to disasters occur immediately lished programs and structures for disaster risk management at the time of the disaster. Epidemiological research should, Natural Disaster Mitigation and Relief | 1159 therefore, complement engineering studies to design better methodology and developing quantitative indicators to esti- facilities and preparedness measures. After the initial disaster, mate those indirect costs should be a research priority. basic questions need to be answered: How many secondary deaths and disabilities can actually be prevented by improving search and rescue and trauma care? How critical is the time fac- CONCLUSIONS tor in reducing DALY losses and assessing the effectiveness of foreign SAR and field hospitals teams? How can researchers Natural hazards are not likely to decrease in the foreseeable objectively assess the risk of outbreak following disasters? In future. Though geological events may occur independently of particular, how can they better differentiate between cases any human control, available data suggest that mankind plays a attributable to increased transmission and those resulting from role in global climate. Technological hazards may also increase improved surveillance and medical attention provided to the rapidly as a result of the unregulated development of industries victims? What is needed are data to put to rest unquestioned in most countries and possibly the use of weapons-grade haz- assumptions and clichés. The alternative is to continue to divert ardous substances against civilian populations. An increase in scarce resources away from routine disease control programs the number of hazards should not mean that the resulting and toward costly measures of doubtful effectiveness. health burden will also increase. A sustained effort is needed to minimize risk, both by reducing vulnerability through preven- tion and mitigation and by increasing capacity through pre- Strategic Research paredness measures. Research is required that will compare the effectiveness of pre- paredness and response strategies and approaches: A Strategic Approach · With respect to preparedness, how should researchers assess The prime objective of a developing country is to develop. the effectiveness of training and coordination versus that of Emergencies and disasters have proven to be major obstacles investing in hardware and stockpiles? For instance, will the and setbacks in the path toward sustainable development. accreditation of hospitals based on their safety and readiness Conversely, the shortcomings in development programs and improve their disaster performance? institutions reduce the effectiveness of the health response in · With respect to mitigation, how should limited funding for times of crisis. Development and disaster risk management retrofitting health facilities be allocated? Is nonstructural cannot be addressed separately. Reducing risk is not a luxury mitigation a workable alternative in the absence of struc- reserved for more developed societies; it is a necessity in coun- tural measures? tries with fragile economies and health systems. It is clearly a · With respect to response, what is the effect of international public health priority. assistance in terms of reductions in DALY losses that could Disasters, as any other public health problem, need to be not be achieved locally? Is it contributing to strengthening the addressed on a long-term and institutionalized basis through capacity of the developing countries? What type of humani- the establishment in the ministry of health of a program or tarian assistance has proven to be development friendly? department for prevention, mitigation, preparedness, and · Finally, how should researchers measure the effectiveness response for all types of disasters. Trends in Latin America sug- of preparedness or mitigation given the unpredictability of gest that such an approach in the context of sustainable devel- disasters? opment contributes to narrowing the gap in disaster-related deaths and disabilities (as measured by DALYs) between indus- trial and developing countries. Economic Research Disaster risk reduction is not merely a health issue. The eco- Humanitarian response is resistant to concepts of cost- nomic and political dimensions should not, however, be effectiveness. Economists should contribute to the comparative allowed to overshadow the fundamental fact that disasters are, study of the immediate and long-term effects of external inter- above all, human tragedies incompatible with the definition of ventions versus less costly alternatives such as relying on local health adopted by the WHO constitution. On one hand, the resources and building local capacity. A cost-benefit analysis of health sector should adapt and use the methodology of eco- international medical interventions prior to and during a nomic valuation of disaster impact as developed by ECLAC; on disaster situation is also overdue. the other hand, the financial world should also learn to give Economic assessment of the damage to the health sector equal consideration to the health burden (DALYs) in its deci- remains focused on physical losses and fails to sufficiently sion making for development or reconstruction. For this to take consider the broader burden on a society caused by the loss of place, health and humanitarian actors need to dramatically health services over a sustained period. Refining the existing improve the availability of data. 1160 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio Disaster risk reduction is not the exclusive domain of a few where a sustained effort over 25 years, with the support of experts or officials. It is the collective responsibility of all disci- PAHO, WHO, and donor countries, traced the way to the plines and programs in the health sector, as well as a remark- reduction of risks from extreme events. This chapter owes able tool or gateway for collaboration with other sectors. Alone, greatly to a large number of experts and professionals in the the ministry of health cannot reduce the health burden or play health sector of those countries. its coordinating role in the response. Disaster risk reduction is unlikely to produce immediate results. It requires sustained commitment over the years. NOTES 1. In a nonnatural phenomenon, such as the attacks in New York on Learning from Errors September 11, 2001, a similar risk has been detected and is perceived as a Learning from past disasters is difficult. At a national level, the remnant potential long-term health risk similar to the effect of air con- tamination from ash from volcanoes. relatively long periods between major disasters result in few 2. Evidently these direct losses are not easy to determine in long- decision makers having prior disaster management experience. developing events (such as the ones associated with slow processes or At an international level, the frequent turnover of relief work- climatic variability), because over time there will be overlapping damage, in contrast to the damage that occurs in sudden events such as hurricanes ers ensures that many of the actors are relatively inexperienced or earthquakes. and susceptible to adopting myths and clichés, which are rarely 3. Data came from reports of the UN Office of Coordinator for challenged by the media and the academic world. It is time for Humanitarian Affairs (http://www.reliefweb.org), supplemented by an international initiative to identify the best practices, and it is authors' estimated costs for donors that did not report actual costs. time for affected countries and scientists to point out the inad- equacies of responses. Humanitarian health interventions, as any other health REFERENCES intervention, should be subject to cost-benefit reviews that Alexander, D. 1997. "The Study of Natural Disasters 1977­1997: Some compare their benefits in terms of DALY loss reduction to Reflections on a Changing Field of Knowledge." Disasters 21 (4): other alternatives, including a possible shift of international 284­304. emphasis from immediate medical response to preparedness or Armenian, H. K., A. K. Melkonian, and A. P. Hovanesian. 1998. "Long- Term Mortality and Morbidity Related to Degree of Damage Following rehabilitation projects. the Earthquake in Armenia." American Journal of Epidemiology 148 Local health services are best situated to address the health (11): 1077­84. consequences of disasters. They should be better prepared to Armenian, H. K., A. Melkonian, E. Noji, and A. P. Hovanesian. 1997. do so. A formalized mechanism to transmit and share those les- "Deaths and Injuries Due to the Earthquake in Armenia: A Cohort sons learned from past errors and to build the response capac- Approach." International Journal of Epidemiology 26 (4): 806­13. ity is required in the health sector. Bernstein, R. S., P. J. Baxter, and A. S. Buist. 1986. "Introduction to the Epidemiological Aspects of Explosive Volcanism." American Journal of Finally, the greatest potential for saving lives is in reducing Public Health 76 (Suppl.): 3­9. the risks and the vulnerability through better infrastructure, Bitrán, D. 1996. "Impacto económico de los desastres naturales en la land-use management, public awareness, and training. infraestructura de salud." Paper presented to the PAHO International The challenge in risk reduction is to sustain public support Conference on Disaster Mitigation in Health Facilities, Mexico City, January. Document LC/MEX/L.2911. ECLAC, Mexico City. and political will in periods of calm. International organiza- Bouma, M. J., and C. Dye. 1997."Cycles of Malaria Associated with El Niño tions--WHO in particular--have a unique and critical role to in Venezuela." Journal of the American Medical Association 278 (21): play as advocates for a long-term approach to disaster risk 1772­74. management in the context of sustainable development. Buist, A. S., and R. S. Bernstein. 1986. "Health Effects of Volcanoes: An Approach to Evaluating the Health Effects of an Environmental Hazard." American Journal of Public Health 76 (Suppl.):1­2. ACKNOWLEDGMENTS Chang, S. E., and C. Chamberlin. 2004. "Assessing the Role of Lifeline Systems in Community Disaster Resilience." In Research Progress and Accomplishments 2003­2004, Multidisciplinary Center for Earthquake The field of disaster epidemiology, a concept first introduced Engineering Research. Buffalo, NY: 87­94. in the early 1970s by M. F. Lechat of the University of Louvain de Ville de Goyet, C. 2001. "Earthquakes in El Salvador." Revista in Belgium, is now calling on many disciplines and fields of Panamericana de Salud Publica 9 (2): 107­13. knowledge. The authors express their gratitude to Caridad de Ville de Goyet, C., E. Acosta, P. Sabbat, and E. Pluut. 1996. "SUMA: A Borras for her contribution on radiological disasters and to Management Tool for Post-Disaster Relief Supplies." World Health Jean-Luc Poncelet, Karl Western, Guy Arcuri, Steve Devriendt, Statistics Quarterly 49 (3­4): 189­94. and Roberto Jovel for their advice, comments, and suggestions. ECLAC (Economic Commission for Latin America and the Caribbean). 1998. Análisis Costo-Efectividad en la Mitigación de Daños de Desastres This chapter relies heavily on the successes and failures of Naturales sobre Infraestructura Social. Document LC/MEX/R.643. the health sectors in Latin America and the Caribbean, a region Mexico City: ECLAC. Natural Disaster Mitigation and Relief | 1161 ------. 2003. Handbook for Estimating the Socio-economic and Environ- ------. 2002b "Terremotos en El Salvador 2001." In Crónicas de Desastres mental Effects of Disaster. LC/MEX/G.5. Mexico City: ECLAC. No. 11. Washington, DC: PAHO. FEMICA (Federación de Municipios del Itsmo Centro Americano). 2003. PAHO and UN/ISDR (Pan American Health Organization and United "Acueducto Orosi: Una experiencia regional sobre implementación de Nations/International Strategy for Disaster Reduction). 1996. Lecciones medidas de prevención y mitigación," Paper presented at the 10th aprendidas en América Latina de mitigación de desastres en instalaciones meeting of the Network for Decentralization and Municipal de salud: Aspectos de costo-efectividad. Washington, DC: PAHO. Development on the topic Local Risk Management: A Challenge for PAHO and WHO (Pan American Health Organization and World Health the Development of Municipalities in Central America. Antigua, Organization). 2000. "Mitigación de desastres en instalaciones de Guatemala, October 21­24. salud: Aspectos estructurales." In Mitigación de desastres en instala- Howard, M. J., J. C. Brillman, and F. M. Burkle. 1996. "Infectious Disease ciones de salud--Material técnico y de capacitacion (CD-ROM). Emergencies in Disasters." Emergency Medicine Clinics in North Washington, DC: PAHO and WHO. America 14 (2): 413­28. PAHO (Pan American Health Organization), WHO (World Health IDB (Inter-American Development Bank). 1999. "Hurricane Mitch: Organization), World Bank, and ProVention Consortium. 2004. Women's Needs and Contributions." Document SOC-115, IDB, Guidelines for Vulnerability Reduction in the Design of New Health Washington, DC. Facilities. Washington, DC: PAHO and World Bank. IFRC (International Federation of Red Cross and Red Crescent Societies). Poncelet, J.-L. 1997. "Earthquakes in Latin America: The Role of Cities in 2000.World Disasters Report 2000: Focus on Public Health.Geneva: IFRC. Disaster Management." In Earthquakes and People's Health: Proceeding ------. 2003. World Disasters Report 2003: Focus on Ethics in Aid. Geneva: of a WHO Symposium, Kobe, Japan, January 27­30. Geneva: WHO. IFRC. http://www.helid.desastres.net. Rottman, S. 2003. "Director's Message." UCLA Center for Public Health and Kirigia, J. M., L. G. Sambo, W. Aldis, and G. Mwabu. 2004. "Impact of Disasters Newsletter 9 (Fall): 1. Disaster-Related Mortality on Gross Domestic Product in the WHO African Region." BMC Emergency Medicine 4 (1): 1. http://www. UNHCR (United Nations High Commissioner for Refugees). 1998. biomedicalcentral.com/1471-227x/4/1. Handbook for Emergency Situations. Geneva: UNHCR. Malilay, J., M. G. Real, A. Ramirez Vanegas, E. Noji, and T. Sinks. 1996. UN/ISDR (United Nations/International Strategy for Disaster Reduction). "Public Health Surveillance after a Volcanic Eruption: Lessons from 2004. Living with Risk: A Global Review of Disaster Reduction Initiatives. Cerro Negro, Nicaragua, 1992." Bulletin of the Pan American Health Vol 1. Geneva: UN/ISDR. Organization 30 (3): 218­25. WHO (World Health Organization). 1992. Assessing the Health Morgan, O. 2004. "Infectious Disease Risks from Dead Bodies Following Consequences of Major Chemical Incidents--Epidemiological Natural Disasters." Revista Panamericana de Salud Publica 15 (5): Approaches. Geneva: WHO. 307­12. ------. 1994. Health Laboratory Facilities in Emergencies and Disaster National Research Council. 1999. From Monsoons to Microbes: Situations. Geneva: WHO. http://www.helid.desastres.net. Understanding the Ocean's Role in Human Health. Washington, DC: ------. 1996. Health Consequences of the Chernobyl Accident: Results of National Academy Press. the IPHECA Pilot Projects and Related National Programmes. Geneva: PAHO (Pan American Health Organization). 1982. Disaster Reports WHO. Number 2: Jamaica, St. Vincent, and Dominica. Washington, DC: PAHO. ------. 2002. Reducing Risks, Promoting Healthy Life: World Health Report ------. 1985. "Earthquake in Mexico." In Disaster Chronicles No. 3. 2002. Geneva: WHO. Washington, DC: PAHO. WHO and PAHO (World Health Organization and Pan American Health ------. 1998. "El Niño and Its Impact on Health." Official document Organization). 2003. Guidelines for the Use of Foreign Field Hospitals CE/122/10. PAHO, Washington, DC. Following Sudden-Impact Natural Disasters. Washington, DC: WHO and PAHO. http://www.paho.org/English/DD/PED/FieldH.htm. ------. 1999. "Huracanes Georges y Mitch, 1998." In Cronicas de Desastres No. 7. Washington, DC: PAHO. ------. 2002a. "Protección de la salud mental en situaciones de desastres y emergencias." In Serie de Manuales y Guías sobre Desastres No. 1. Washington, DC: PAHO. 1162 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio Chapter 62 Control and Eradication Mark Miller, Scott Barrett, and D. A. Henderson The Controversy: Control or Eradication? We cannot refrain altogether from examining the roots of this controversy if only because the extreme views for and against eradication have exerted and are still exerting a . . . highly detrimental influence on public health practice. --P. Yekutiel, Eradication of Infectious Diseases: A Critical Study Eradication of an infectious disease is an extraordinary goal. Its endemic focus was serving to reinfect areas otherwise free of possibility became apparent as soon as Edward Jenner demon- yellow fever. In the 1930s, F. L. Soper set out to eradicate the strated an ability to provide immunity to smallpox. Writing in Aedes aegypti vector from the Americas. By 1961, Soper reported 1801, Jenner observed that, through broad application of vac- that he had largely succeeded except for the United States, where cination, "it now becomes too manifest to admit of controversy the program received little support. By the 1980s, Aedes aegypti that the annihilation of the Small Pox, the most dreadful had become reestablished in Central and South America. scourge of the human species, must be the result of this prac- In 1953, Brock Chisholm, the first director-general of the tice" (Jenner 1801). Louis Pasteur claimed that it was "within World Health Organization (WHO), tried to persuade the the power of man to eradicate infection from the earth" (Dubos World Health Assembly (WHA) to undertake smallpox eradi- and Dubos 1953). And yet, by and large, public health has pro- cation, but a number of countries objected on the grounds that ceeded with more modest goals of local and regional disease eradication was not technically feasible. Instead, in 1955, control. Notable successes have occurred. Indeed, some dis- under the leadership of his successor, Marcolino Candau, eases now thought of as "tropical" were previously endemic in WHO began a global effort to eradicate malaria primarily by temperate climates. Systematic application of hygiene, sanita- means of household spraying of DDT. The relatively sophisti- tion, environmental modification, vector control, and vaccines cated science of malaria control was abandoned in favor of this have led, in many countries, to the interruption of transmission simplistic technology (Jeffrey 1976). Despite an expenditure of of microbes causing such diseases as cholera, malaria, and more than US$2 billion, the effort failed. yellow fever. Even while the malaria eradication effort was under way, the Intensive efforts to eliminate breeding sites of the yellow Soviet Union, in 1958, proposed to the WHA that smallpox be fever mosquito vector, Aedes aegypti, interrupted transmission eradicated. A resolution to this effect was offered in 1959 and of this disease in Havana in 1901 and throughout Cuba soon passed unanimously. However, the resolution provided little thereafter. Subsequently, yellow fever and malaria were able to international funding or support. Over the next seven years, be controlled in Panama, thus permitting construction of the disease transmission was interrupted in some 30 countries in Panama Canal. In 1915, the Rockefeller Foundation launched Africa, Asia, and South America, but endemic smallpox per- an effort to eradicate the disease worldwide. Transmission sisted in the Indian subcontinent, Indonesia, most of Sub- appeared to have ceased in the Americas by 1928, but then cases Saharan Africa, and Brazil. WHO launched an intensified effort reappeared, and by 1932, it became clear that a nonhuman in 1967 to eradicate the disease within a decade. This new 1163 resolution included an annual budget of US$2.4 million, to be eradicating that disease. In the Americas, spurred on by the paid according to the WHO scale of assessments. The resolu- success of regional cessation of transmission of wild poliovirus, tion passed by the narrowest of margins, but a reinvigorated eventual consensus was reached to intensify measles control effort was soon under way and paved the way for a historic efforts, primarily through surveillance and periodic pulse public health achievement (Henderson 1988). Following an application of measles vaccine in national campaigns. As a con- extraordinary worldwide effort, the last case of smallpox was sequence, transmission of measles virus was temporarily inter- isolated in October 1977, and the disease was certified as being rupted in the Americas on several occasions but reestablished eradicated in 1979, 170 years after Edward Jenner first dreamed again by importations (CDC 1998a). Although the U.S. Centers of that possibility. Understanding how and why smallpox for Disease Control and Prevention (CDC) and WHO have eradication succeeded is essential to the study of control and advocated extending measles "elimination" through vaccina- eradication. tion campaigns and second-dose opportunities to other The smallpox success was inspirational, even though the regions (Biellik and others 2002; CDC 1998a, 1998b, 1999a, leaders of WHO's smallpox eradication effort cautioned that, 1999b, 2003d, 2004b, 2004d, 2004f), the intensive control among all the diseases that might be considered candidates for efforts required to break transmission of this highly infectious eradication, smallpox was unique (Fenner and others 1988) agent make global eradication unlikely at this time. and that they foresaw no other disease as a candidate for eradi- cation (Henderson 1982). At a meeting convened by the DEFINITIONS Fogarty International Center of the National Institutes of Health in 1980, scientists, public health officials, and policy Yekutiel (1980, 5­8) provides an excellent treatise on the con- makers discussed the merits of eradicating other diseases, with cept of eradication, which includes a summary of the multiple schistosomiasis, dracunculiasis, poliomyelitis, and measles definitions that have been formulated (Andrews and Langmuir identified as possible candidates (Henderson 1998a). However, 1963; Cockburn 1961, 1963; Payne 1963a, 1963b; Spînu and no consensus was reached at that time on moving forward with Biberi-Moroianu 1969). A conference devoted to eradication any of those diseases. held in Dahlem, Germany, in 1997 (Dowdle and Hopkins Poliomyelitis became the next principal target when mass 1998) set out to provide precise definitions for control, elimina- vaccination campaigns, proposed by Albert Sabin (1991), tion, eradication, and extinction in a biological, economic, and proved remarkably successfully in Cuba and Brazil. In 1985, political context (Dowdle 1998, 1999; Ottesen and others an American Health Organization coordinated campaign was 1998); however, a number of eminent public health officials launched to interrupt poliovirus transmission in the Americas (Cochi and others 1998; de Quadros 2001; Goodman and by 1991, and this effort succeeded. Some believed that global others 1998b; Henderson 1998b; Salisbury 1998) challenged eradication might be possible, although others were concerned these definitions at two subsequent meetings at the CDC that the far less developed infrastructure of health, transporta- (Goodman and others 1998a, 1998b) and the U.S. Institute of tion, and communications services in many parts of Asia and Medicine (Knobler, Lederberg, and Pray 2001). Africa would make it an unachievable task. In 1988, the WHA Unfortunately, broadly accepted, standard definitions for adopted a resolution to eradicate polio, but at that time, a key concepts pertaining to disease control and eradication do longer-term strategy for ending polio vaccination was neither not exist in the literature. Making matters more confusing, formulated nor agreed on by the public health and scientific certain of the concepts have been given names that are part of community. our everyday language and so are easily misinterpreted by non- The WHA has adopted only one other resolution to eradi- specialists as meaning something different from the meanings cate a disease--guinea worm, or dracunculiasis. The eradica- understood by those who are preoccupied with eradication tion of this disease can be achieved by applying simple tech- programs. Most unfortunate is the all too casual use of the nologies for providing water that is free of the vector copepod words elimination and eradication to promote programs that and parasite and for treatment of patients with the disease. This cannot reasonably be expected to achieve the promise implicit eradication program has made steady progress but has been in these words. Moreover, the two words themselves are com- hampered in part by civil and political unrest and lack of pro- monly used interchangeably. gram priority because of low mortality and low incidence in some remaining endemic areas. However, given the environ- Control mental restriction of the parasite to rural tropical areas and its Two concepts are central to this chapter: control and eradica- relatively low transmissibility, eventual global eradication tion. By control, we mean a public policy intervention that seems within reach. restricts the circulation of an infectious agent beyond the level One other case--that of measles--is worth noting. A num- that would result from spontaneous, individual behaviors to ber of public health authorities have raised the possibility of protect against infection (Barrett 2004). 1164 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson Although control is a range rather then a level, a particular tion. The 1993 edition of the New Shorter Oxford English level of control may be an aim of policy. Because every choice Dictionary, for example, defines eliminate as to "remove, get rid entails consequences, choice of the "optimal" level of control of, do away with, cause to exist no longer." This same diction- requires economic analysis. Optimal here is defined in relation ary defines eradicate as "pull up or out by the roots, uproot, to the model that gives rise to the result. Control is local and so remove or destroy completely, extirpate, get rid of." This ambi- needs to be looked at from the local perspective. Because one guity invites misunderstanding among those not intimately country's (or region's) control may affect other countries involved in an eradication effort. For purposes of clarity, we sel- (regions), a global perspective exists as well. The level of con- dom use the term elimination in this chapter and then only to trol that is optimal for one country (region) may not be opti- signify control measures sufficient to interrupt microbe trans- mal from the perspective of the world as a whole. Thus, a need mission in a specified area. exists to distinguish between, say, a locally optimal level of con- trol and one that is globally optimal. Extinction Finally, control requires ongoing intervention. Sustaining a Finally, the literature sometimes refers to extinction as a possi- given level of control requires an annual expenditure. ble policy goal. In the context of infectious disease control, the concept is problematic for two reasons. First, proving that an Eradication organism has become extinct is impossible. To do so would Eradication differs from control in that it is global. The term require demonstrating not only that the organism no longer denotes the certified total absence of human cases, the absence exists in nature but also that it no longer exists in any con- of a reservoir for the organism in nature, and absolute contain- trolled environment--a practical impossibility. Second, de ment of any infectious source. Eradication permits control novo synthesis of viral agents from published genomes interventions to stop or at least to be curtailed significantly. (Cello, Paul, and Wimmer 2002) now put the concept in peril, Finally, eradication is binary. Control levels can vary, but a although much research remains to be done in this area. disease is either certified as eradicated or not. Extinction, in the context of infectious diseases, may no longer Every disease can be controlled, even if only by using simple be irreversible. measures, such as quarantine. The ultimate achievement of Clearly, policy making will be improved by stating the goal control is eradication. But not every disease that can be con- of any particular intervention in precise language. trolled can be eradicated. Very few diseases, in fact, are poten- tial candidates for eradication. The criteria for the feasibility for FRAMEWORKS FOR ERADICATION eradication as a preference over control are discussed in the section titled "Economic Considerations." Numerous issues need to be considered in planning expanded control measures that lead, possibly, to regional cessation of Elimination transmission or global eradication of disease. These complex issues will be further examined in the chapter. Control and eradication are the essential concepts, but two other terms bear mention. The first is elimination. Some who are concerned with eradication programs have explicitly Scientific Considerations defined this term to denote the cessation of transmission of an Scientific considerations include the nature of potential reser- organism throughout a country or region. In contrast, eradica- voirs for disease-causing microbes or their vectors, technolo- tion is defined as a global achievement. Like control, elimina- gies available for interrupting disease transmission, changes in tion is location-specific and would require ongoing interven- host capabilities to deter infections and disease, and satis- tions to be sustained in order to prevent reemergence of the factory containment of organisms in laboratories. disease from microbe importations. Two problems exist with the term elimination. First, it has Geographic and Environmental Controls. The limit of been used to describe different phenomena, not just that endemicity for microbes and their associated diseases is deter- described in the definition given above. For example, some mined in part by their ability to exist in nature outside the public health officials have promoted programs aimed at "elim- human host. Both geographic and temporal variations deter- inating a disease as a public health threat," which is interpreted mine the ecological niche of microbes, resulting in variable to mean reducing incidence to an "acceptable" level but not annual incidence rates throughout the world. This niche limi- necessarily to zero. This usage is very different from the one tation is further extended to intermediary vectors and hosts in outlined above and is almost certain to be misunderstood. complex biological systems. Natural environmental barriers Second, the definition of the word elimination in common use, also may isolate the habitats of helminths. Infectious agents that as applied to disease control, is indistinguishable from eradica- are not limited to an environmentally restricted intermediary Control and Eradication | 1165 host or those that have longer latent periods, thereby allowing and their intermediary and human hosts. The reproductive translocation, may have a global pattern of distribution. rate, R, is influenced by many local factors, including popula- Examples include the highly transmissible viral agents such as tion density (of vectors, intermediary hosts, and humans) and measles, rubella, influenza, and varicella. Although these agents other environmentally determined conditions, all highly vari- are not geographically constrained, their transmission patterns able throughout the world. For a disease to be controlled to are directly and indirectly influenced by seasonal environmen- stop transmission, the intervention-altered reproductive rate tal factors and population-based immunity. must be maintained below 1.0. At the same time, all reservoirs of the responsible microbe must be controlled. Potential Reservoirs. A microbe and associated disease can Three main components of possible eradication programs not be eradicated if the microbe is capable of persisting and are multiplying in a reservoir. Microbes that thrive in nonhuman species may reemerge if control efforts cease, thus leaving · surveillance, including environmental sampling where human populations susceptible. Similarly, if the infectiousness appropriate and clinical testing of a human is long lived or could lead to potential recrudes- · interventions, including vaccination and chemotherapy or cence, surveillance efforts would have to continue as long as the chemoprophylaxis or both last individual remained potentially capable of transmitting · environmental controls and certification of eradication. infection, as would be the case with tuberculosis or hepatitis B infection. Each of these components must be undertaken at local, community, national, regional, and global levels. Eradication Transmissibility. The inherent rate of a microbe's ability to differs from control in that it is expected to be permanent. cause secondary infections is defined by an organism's repro- Success depends on having adequate surveillance to identify ductive rate in a fully susceptible (R0) and partially susceptible potentially infectious persons and on stopping transmission (R) population. The reproductive rate of organisms that infect before infection of a new cohort of susceptible persons arises individuals only once because of durable immunity is inversely as a result of births, migration, or the waning effectiveness of proportional to the average age of infection in an endemic prophylactic measures. area. Agents that cause childhood infections, such as viral res- piratory agents, are far more transmissible than helminths and Disease Surveillance. Effective surveillance requires a sensitive subsequently require more intensive control efforts to interrupt system to detect the presence of microbes within the environ- transmission. ment, intermediary hosts, and clinical cases. Surveillance and response systems need to be more efficient than the rate of trans- Natural Resistance to Reinfection. Many natural infections mission of the targeted agent.As eradication progresses, the sen- induce long-lived immunity to reinfection. Although the most sitivity of detection systems must be steadily enhanced to detect commonly used vaccines have been available for fewer than all existing foci. Nonclinical or latent infections pose formidable 50 years--less than the lifetime of an individual--they, too, are barriers to eradication efforts. Operationally, the need for near- assumed to offer long-lasting immunity. Because eradication perfect sensitivity comes at the expense of lower specificity. depends on reducing susceptible populations in potentially Thousands of skin lesions from suspected smallpox patients endemic areas, long-lived protection through immunization or were tested in reference laboratories during confirmation of natural disease is important to successful programs. smallpox eradication, and tens of thousands of stool specimens are being examined for poliovirus. Highly sensitive systems used Laboratory Containment. Laboratory specimens containing to detect measles cases in the Americas began to identify a greater the organism targeted for eradication could serve as reservoirs. proportion of rubella and parvovirus infections because of the Considerable effort may be necessary to ensure their maximum nonspecific surveillance of rash illness. Such findings are impor- security. That these microbes may be inconspicuous in speci- tant because the identification of other diseases that mimic the mens collected for other purposes poses special challenges. targeted disease can lead to a misdirection of resources. This situation is especially true for the poliomyelitis virus, However, the ability to detect such similar clinical cases can serve which may be found in many stool specimens collected for as a proxy measure for the adequacy of surveillance. For exam- studies completely unrelated to current poliomyelitis eradica- ple, identification of a minimum incidence of cases of acute flac- tion efforts. cid paralysis that is not related to polio has served as an indicator of adequate efforts of case finding for polio. Operational Considerations Optimization of control requires a fundamental appreciation Interventions. Interventions to block transmission of the tar- of the biological systems that govern the ecology of microbes geted infectious agent should be easy to deploy and adaptable 1166 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson to diverse conditions, given the global goal of eradication. Cost herd immunity is conferred on all susceptible persons. When a considerations and local acceptance of the required sacrifices population is immunized to this level, a disease ceases to be (both short and long term) are crucial for success. endemic, and imported infections cannot spark an epidemic. Interventions may be designed for environmental control of This level is determined by the epidemiology of a disease, microbes, isolation (quarantine) of clinically infectious indi- but whether it pays to vaccinate to this level depends on the viduals to limit their contacts with susceptible persons, treat- economics, and the economics depend in turn on the social ment of clinical cases to limit the duration of infectiousness, or costs and not only the social benefits of vaccination. These reduction in the infected pool of individuals through immuno- costs consist of the direct costs of producing, distributing, and or chemoprophylaxis. administering a vaccine. The economics depend also on the costs borne by the individuals who are vaccinated, such as those Certification. The last tool for eradication is a certification incurred by individuals who experience vaccine complications. process whereby independent, respected parties certify the The proportionate costs of reaching people who live in remote absence of disease transmission or the existence of any specific areas and those who are at special risk, such as migrants and the microbe in an uncontrolled reservoir, including laboratories homeless, increase as the fraction of the population vaccinated (Breman and Arita 1980). Although certification can be increases. implemented on a regional basis, it must ultimately be done The economics of varying levels of disease control depend globally. Certification is one of the greatest challenges in any on the relationship between the marginal social benefits and eradication effort, given the exceedingly great difficulty of ver- the marginal social costs of vaccination. As vaccination levels ifying a negative finding in a reasonably short period of time. increase, the marginal social benefits of vaccination fall, where- When certification is completed, curtailment of control meas- as the marginal social costs rise. Social welfare is maximized ures should be possible. where these two relations intersect, which might be called the Strengthening control efforts sufficiently to achieve eradica- "optimal" level of vaccination--a level that may or may not tion is a difficult and expensive task. It requires that scaling up achieve cessation of transmission or eradication. of such efforts occur over a wide area--at the community, national, regional, and global levels. Its efficacy depends heavily Short-Term versus Long-Term Net Benefits. Control pro- on the adequacy of local financial and human resources, as well grams require ongoing intervention. Sustaining a given level of as on a broad range of logistical factors. protection requires that, over time, a certain proportion of new susceptible persons be vaccinated. Eradication differs from control in being permanent. The economics of eradication Economic Considerations must therefore take account of long-term benefits as well as Control and eradication programs have many economic short-term costs. dimensions: private versus social net benefits, short-term ver- The long-term benefits of eradication consist of avoided sus long-term net benefits, and local versus international net future infections and vaccination costs--a dividend. To calcu- benefits. Such interventions also have implications for existing late this benefit, one projects future infection and vaccination public health programs. levels in the absence of eradication, attaches values to these, and discounts them. If this sum exceeds the costs of eradica- Private versus Social Net Benefits. Individuals have private tion, then eradication enhances social well-being, and it there- incentives to protect themselves from disease--by means of fore should be undertaken. vaccination, for example. But when individuals protect In deciding on the benefits of eradication, the cost of future themselves--when they elect to be vaccinated--they offer a infections and vaccination should ideally be compared with the measure of protection to others by helping limit the spread of best alternative to eradication: the level of optimal control infection. In brief, the social benefit of vaccination is greater (Barrett and Hoel 2003). than the private benefit alone. As more people become vacci- The costs of eradication must also take into account ongo- nated, the marginal private and social benefit of vaccination-- ing surveillance requirements, laboratory containment, and that is, the benefit of vaccinating an additional susceptible perhaps the maintenance of stockpiles of vaccine in the chance person--declines. The marginal private benefit is likely to fall event of disease reemergence. From an economic perspective, because, as more people are vaccinated, the probability of a sus- attractive candidates for eradication are those diseases that ceptible person becoming infected falls. The marginal social some countries have themselves targeted for interruption of benefit is likely to fall for the same reason and for one other: as transmission nationally or regionally. more people become protected, the total number of susceptible persons falls. The marginal social benefit of vaccination falls Local versus International Net Benefits. Control differs from sharply at the critical level of immunization--the level at which eradication in another important way. Control refers to Control and Eradication | 1167 location-specific interventions. Eradication, by contrast, is eradicated. In these two cases, no need exists for an interna- global. In economic terms, eradication is a global public good. tional policy. No country can be excluded from the benefit of eradication, · Third, each country may have an incentive to eliminate a and no country's consumption of that benefit diminishes the disease only if all other countries have eliminated it. In this amounts available to other countries. Control, by contrast, case, achieving global eradication requires coordination. supplies only a local public good. Here a role exists for international policy, but all that is Eradication requires a global effort. A disease can be eradi- required is for each country to be assured that all others will cated only if microbe transmission ceases everywhere. This eliminate the disease. spatial dimension to eradication is of fundamental importance · Finally, and noting that the "last" country to eliminate a dis- because no world government can implement an eradication ease would get just a fraction of the global dividend from policy; the WHA can declare its support for eradication, but eradication, under some circumstances no incentive may WHO does not have the power to enforce the execution of a exists for this country to eliminate the disease--even if all national program in support of that goal. The outcome experi- other countries have done so and even if the entire world enced by any country depends not only on whether the coun- would be better off if it did. This case is the most worrisome, try itself eliminates the disease within its borders but also on because implementation of the efficient outcome would whether all other countries do so. Indeed, eradication is a likely require enforcement. weakest-link public good. Whether eradication is achieved depends on the level of All this hypothesizing assumes that countries are symmet- control adopted by the country that undertakes the least ric, and of course they are not. Some countries gain less from control. In practical terms, any country in which disease is control and would gain less from eradication than others. Some endemic can prevent eradication from being achieved. In 2004, are unable to implement an elimination program, even if they the global polio eradication initiative, after investing more than would very much like it to succeed. In these situations, achiev- US$3 billion and involving some 20 million volunteers over a ing an eradication goal will require international financing and period of 16 years, was placed at risk of failure by the actions of technical assistance, with the countries that benefit most from one local administration. In the Kano state of Nigeria, local eradication compensating the other countries for the costs of leaders claimed that the polio vaccine was tainted with the eradicating the disease. National and international reproach AIDS virus and sterility drugs and declined to participate in a are often expressed if a country lags in its eradication efforts. national immunization day program. The European Union International financing has been a key element in all eradica- then declined to pay for the national program in Nigeria, tion programs. believing the money would be wasted (Roberts 2004). One We have thus far looked at eradication from the perspective consequence was the subsequent spread of polio to nine for- of only the self-interests of states. But eradication also has merly polio-free countries. Concerted efforts by WHO later implications for development. In particular, eradication has persuaded local leaders in Nigeria to rejoin global efforts, but two advantages over control programs. The first is that the rich special vaccination programs had to be launched over a popu- countries may gain directly if the goal is achieved, giving them lation area of more than 300 million persons. This situation a vested interest in ensuring that the goal is achieved. The sec- dramatically illustrated the vulnerabilities inherent in a ond is that eradication is permanent, making an investment in weakest-link public good. eradication financially sustainable. This second advantage is What are the incentives for states to participate in an eradi- important because financial sustainability has proved to be a cation effort? To begin, assume that countries are symmetric, key problem for disease control programs in developing coun- meaning that all countries have the same benefits and costs of tries (Kremer and Miguel 2004). control. Assume as well that eradication is feasible. Four possi- ble situations then exist (Barrett 2003): Vertical versus Horizontal Programs. Control and eradica- tion programs cannot be viewed in isolation. All programs have · First, the global net benefit of eradication may be negative-- implications for the delivery of comprehensive primary care the cumulative programmatic costs outweigh the net pres- services. An important question is whether targeted, or so- ent value of the cumulative benefits. In this case, elimination called vertical, programs draw critical resources away from would also yield a negative net benefit to every country, and other health care programs or whether they serve instead to so no country would eliminate the disease. augment competence and capacity. The evidence is mixed. · Second, the global net benefit of eradication may be so large Evidence suggests that disease-specific systems can serve that each country would choose to eliminate the disease to expand polyvalent services (Aylward and others 1998). even if others did not. In this case, all countries would Smallpox eradication, for example, gave many national govern- eliminate the disease, and the disease would therefore be ments the confidence to introduce the Expanded Program on 1168 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson Immunization, with the ability to deliver vaccines and prorating estimates of the benefits of eradication for India and micronutrients in routine schedules and through national cam- the United States to all developing and industrial countries, paigns. However, other evidence suggests that some vaccination respectively, suggest that developing countries benefited programs have adversely affected primary health services more from smallpox eradication than industrial countries. (Steinglass 2001; Taylor, Cutts, and Taylor 1997) and may have Qualitatively, a consistent picture emerges: smallpox eradica- even increased costs. Implementation of international initia- tion was not only an extraordinary investment for the world; it tives can also expose conflicts of priorities. The polio eradica- was also an investment that benefited every country, rich and tion initiative, for example, has successfully vaccinated children poor alike. in the poorest of countries against this disease, but in some When the eradication effort began, smallpox was no longer of these countries it has failed to timely include the co- endemic in most industrial countries. Nonetheless, these coun- administration of measles and other common childhood vac- tries needed to maintain populationwide immunity under the cines, which would have had a much greater effect on child threat of possible imported cases from endemic countries. mortality. They would gain from eradication not only through the cessa- tion of vaccination and its associated costs but also by being DISEASE-SPECIFIC CASE STUDIES able to decrease the number of quarantine inspectors at ports of entry and by averting costs of care related to the adverse In this section, we apply the reasoning developed previously to events from this live vaccine. provide an empirical analysis of the three most recent eradica- The still-endemic countries would also save vaccination tion programs--smallpox and the two ongoing programs, costs, although most were vaccinating only a comparatively poliomyelitis and dracunculiasis. small proportion of their populations. The greater benefit to them was the avoided cost of disease, including the extraordi- nary death toll. A number of developing countries had Smallpox accorded smallpox prevention a high priority, as was evidenced As noted before, smallpox eradication was achieved in October by the number that succeeded in interrupting transmission 1977, 11 years after the intensified program began. Following without international assistance. This list includes China, implementation of a rigorous certification procedure, the which was not a member of WHO at the time the eradication WHA declared smallpox eradicated in 1980. effort commenced. Fenner and others (1988) have estimated the annual bene- Indeed, and as shown in table 62.1, the still-endemic coun- fits of smallpox eradication to developing and industrial coun- tries contributed an estimated two-thirds of the US$298 mil- tries (see table 62.1). These aggregate estimates, obtained by lion cost of eradication. International sources funded the balance. If the latter cost is interpreted as the incremental cost of achieving eradication, the benefit-cost ratio of global small- pox eradication was over 450:1, a singularly high figure. Even Table 62.1 Benefits and Costs of Smallpox Eradication including the expenditure by endemic countries, the benefit of (Millions of U.S. dollars) eradication exceeded the cost by an unusually large amount. Brilliant (1985) calculated the annual costs of the smallpox Annual amount eradication campaign for India to be about US$17 million per Beneficiary year, including indirect costs (lost productivity caused by India 722 adverse reactions to vaccination) and opportunity costs (health United States 150 workers being diverted from other programs). These costs were All developing countries 1,070 only a fraction of the annual benefits of eradication to India, All industrial countries 350 which, by Brilliant's calculations, were US$150 million. The Total annual benefit 1,420 benefit estimates by Fenner and others (1988) are much larger, Expenditure and those of Ramaiah (1976) are smaller, but all three studies Total international, on eradication 98 draw the same (qualitative) conclusion: smallpox eradication was a good investment for India. Basu, Jezek, and Ward (1979, Total national, by endemic countries 200 312) present estimates identical to those in Ramaiah (1976), Combined total, on eradication 298 but without giving attribution. Benefit-cost ratio Originally, India had decided to undertake a smallpox pro- International expenditure 483:1 gram just one month after the WHA voted to eradicate the dis- Combined total expenditure 159:1 ease globally in 1959. The attempt failed, however, largely for Source: Adapted from Fenner and others 1988. administrative reasons (Basu, Jezek, and Ward 1979; Brilliant Control and Eradication | 1169 1985; Fenner and others 1988). Essentially, India had an the eradication of the disease. A huge organizational effort, but economic incentive to control smallpox on its own (Brilliant only a relatively small incremental cost, was needed to achieve 1985, 33) but lacked organizational capacity and an effective eradication. The specter of global terrorism has recently caused strategy for achieving this goal. Note, however, that India had some countries to prepare themselves for a possible smallpox other health priorities, including family planning. According to attack by stockpiling vaccine. Although such actions reduce the Brilliant (1985, 33), "for India's health planners, occupied then benefits of eradication, the economics remain favorable. by emergencies and competing political demands on scarce Smallpox, however, was a special case. Many attributes of resources, the long-term benefits from disease eradication were the disease and the vaccine favored eradication. The vaccine not a great motivation. Health planners are sensitive to imme- was heat stable and required only a single dose to protect a per- diate political realities, and the benefits of smallpox eradication son for a period of at least 5 to 10 years. Vaccination was easily would be realized only at some future time when the $3 million performed and protected immediately on application. Every annual expenditures for smallpox could be applied to other individual who became infected exhibited a typical, easily rec- health problems. In the meantime, however, the cost of putting ognized rash, thus permitting accurate surveillance without so many scarce resources into one program rather than into recourse to laboratory diagnosis. The disease spread slowly so many health needs was high." that transmission could readily be stopped by isolating the Table 62.2 provides estimates of the benefits of smallpox patient and vaccinating contacts within the area. eradication to the United States. The total benefit of eradica- tion to the United States is about the same order of magnitude as India's, but the breakdown is different. Whereas India Poliomyelitis benefited mainly from avoided infections, the United States The polio eradication program, launched by the WHA in 1988, benefited mainly from avoided vaccinations. By the time the has made substantial progress (CDC 2003a, 2003b, 2003c, eradication program was launched, the United States had 2004a, 2004c, 2004e, 2004g, 2005). The incidence of paralytic already interrupted smallpox transmission, but vaccination poliomyelitis in children fell by more than 99 percent, from an was costly, both in economic and human health terms (a small estimated 1,000 cases per day worldwide in 1988 to fewer than number of people died every year from infections arising from 4 cases per day in 2003. The number of poliomyelitis-endemic the live vaccine). Defending the nation from imported countries also fell, from 125 in 1988 to just 6 by 2003 infections imposed additional costs. (Afghanistan, the Arab Republic of Egypt, India, Niger, Nigeria, In health terms, smallpox eradication saved millions of lives; and Pakistan). This laudable reduction was the result of in economic terms, it yielded a benefit many times greater than repetitive vaccination campaigns with easily administered oral the cost. Identifying another investment that has yielded com- polio vaccine to whole regions, to nations, and to large sub- parable returns and has benefited every country is difficult. populations. One reason that the economics of smallpox eradication were so During 2004, however, polio immunization activities in favorable is that all countries had strong incentives to join in northern Nigeria were halted for an extended period for fear of tainted vaccines, and this permitted the development of epi- demics extending throughout the country. The disease spread as well to 10 other African countries and to Saudi Arabia, Table 62.2 Benefits of Smallpox Eradication to the United Yemen, and Indonesia. Transmission has again been reestab- States, 1968 (Millions of U.S. dollars) lished in several African countries (Burkina Faso, Central African Republic, Chad, Côte d'Ivoire, and Sudan). Heroic Amount efforts are being made to control these outbreaks by large-scale Direct costs for medical services immunization, but in countries such as these, where health Vaccination 92.8 services are stressed and the health, communication, and trans- portation infrastructures are weak, disease transmission is dif- Treatment of complications 0.7 ficult to interrupt. Meanwhile, other countries throughout the Indirect costs, loss of productivity world that appear to be polio free are continuing their vaccina- Work losses attributable to vaccination and reactions 41.7 tion programs but finding it increasingly difficult to maintain Permanent disability attributable to complications 0.4 a momentum of interest, effort, and financing. Premature death 0.1 The difficulties of maintaining credible surveillance systems Cost of international traffic surveillance and delays in 14.5 throughout the developing countries were vividly demon- clearance of vessels strated by the discovery of polio in Sudan in May 2004, more Total 150.2 than three years after the last case had been reported (CDC Source: Sencer and Axnick 1973; see also Fenner and others 1988, table 31.2. 2005). In the interim, specimens from 75 to 90 percent of such 1170 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson cases were processed in the laboratory, and measures of derived pathogenic viruses (Kew and others 2002). Such an surveillance for acute flaccid paralysis cases were reported to outbreak could occur with disastrous speed because the polio have been entirely satisfactory. At first, the Sudanese cases were virus is far more contagious than that of smallpox. In develop- considered to have resulted from importations from Nigeria, ing countries, virtually all cases of polio occurred among those and, indeed, some cases were. However, from more detailed under five years of age, older persons having been protected by laboratory studies, it was determined that type 1 wild virus the natural immunity of earlier infection. Within five years had been circulating undetected for more than three years and after vaccination ceased, therefore, the population immunity type 3 virus for nearly five years. level in the developing countries would be no better than it was Clearly, stopping the continuing transmission of wild before vaccination was introduced. With this is mind, it seems poliovirus is itself a formidable challenge, the success of which questionable as to whether all health ministers could be per- is by no means certain. A problematic discovery since the glob- suaded to call for a country-wide cessation of poliomyelitis al eradication program began was the finding that individuals vaccination itself, given the uncertainties of virus detection in with particular immunologic disorders shed polio vaccine so many remote and inaccessible areas of the world. virus for many months to years, thus serving as a reservoir for By definition, eradication implies certifying cessation of this virus. The virus, in turn, can revert to a neurovirulent virus transmission and the absence of reservoirs so that control form, which is capable of causing outbreaks of disease interventions can cease. As noted earlier in this chapter, it is (Bellmunt and others 1999). Such individuals may be wholly only for this reason that eradication yields a dividend. without symptoms and impossible to identify except through Although the interruption of wild poliomyelitis virus trans- fecal cultures. Moreover, no treatment is known to stop them mission is theoretically feasible, the obstacles to achieving and from shedding virus. They pose an all but insurmountable maintaining this goal are formidable. At this time, it is difficult challenge to the current poliomyelitis eradication effort. to foresee a future that does not envisage a continuing vaccina- The program is further hampered by the tool that has pro- tion program, perhaps with IPV use in countries that can vided so much success--oral poliovirus vaccine (OPV). In afford the substantial additional costs entailed and with OPV resource-poor environments, poliomyelitis is best controlled use in all other countries. with the inexpensive, live, and easily administered oral vaccine. The polio eradication initiative, like that for smallpox, has The live vaccine is excreted and can infect other susceptible had to rely primarily on voluntary donations provided both to contacts. The ability of OPV to immunize others indirectly WHO and bilaterally. Playing an especially important role have makes it an ideal vaccine for achieving high levels of been the Rotary International Foundation and the Bill & population-based immunity, especially in lower socioeconomic Melinda Gates Foundation. From 1988 to 2004, more than populations that are the most difficult to reach. However, the US$3 billion was spent on the effort (WHO 2003). excreted virus occasionally reverts to a pathologic state, causing What are the economics of polio eradication? Bart, Foulds, not only cases but outbreaks of vaccine-associated paralytic and Patriarca (1996) developed the first global cost-benefit polio, which may not emerge until months or even years after analysis of polio eradication, beginning with the costs incurred the vaccine has been administered (Kew and others 2004). since 1986, the year that the Pan American Health Organiza- Unfortunately, the alternative inactivated polio vaccine (IPV) is tion launched a regional eradication effort, and extending to not immediately an option in many nations, not least because 2040. They assumed that eradication would be achieved in global manufacturing capacity could not begin to meet 2005, using OPV, and that vaccination would cease after eradi- demand. Other problems include the current cost differential cation had been certified. Benefits (like costs, discounted at between OPV and IPV, the increased difficulty of administer- 6 percent) reflect the avoided costs of acute care and avoided ing the vaccine by syringe and needle, and the need to achieve vaccination costs after certification. Their analysis showed that higher coverage rates with IPV because it does not spread from the initiative would break even by 2007 and yield a net benefit person to person as does OPV. to the world of more than US$13 billion by 2040--an encour- Tragically, if OPV use were discontinued, in the absence of aging result, but it was based on the assumption that all vacci- alternative immunity, polioviruses would likely circulate nation would stop abruptly in 2005. silently (Eichner and Dietz 1996) and reemerge. Preliminary Khan and Ehreth (2003) developed a similar analysis but results from a model presented by WHO indicate a greater than provided regional detail. They estimated the costs and medical 60 percent chance of an outbreak within two years of the pos- costs avoided of polio immunization and eradication over the sible global cessation of OPV (WHO 2004) because of contin- period 1970 to 2050, assuming that vaccination could cease uous circulation of undetected live vaccine viruses that can after 2010. As table 62.3 shows, Khan and Ehreth estimated that revert. Outbreaks have already been observed in several regions polio immunization and eradication would entail a negative where decreasing use of live vaccine has left pockets of suscep- net cost overall, with Europe and the Americas saving the most tible persons who eventually have been exposed to vaccine- and with other regions incurring a positive net cost. Compared Control and Eradication | 1171 Table 62.3 Net Costs of Polio Immunization and Eradication Table 62.4 Postpolio Eradication Costs (Millions of U.S. dollars) (Millions of U.S. dollars) Medical care Immuni- Cost/DALY Continue OPV Stop OPV Universal IPV WHO region cost savings zation costs Net costs saved Low-income countries 1,364 487 4,418 Africa 1,100 3,942 2,842 442 Middle-income countries 12,196 12,196 12,196 Americas 76,900 25,460 51,440 4,983 High-income countries 6,409 6,409 6,409 Eastern 1,930 3,512 1,582 426 Subtotal 19,969 19,092 23,023 Mediterranean Global response capacity 1,120 1,320 1,120 Europe 38,250 17,249 21,001 2,780 Total 21,089 20,412 24,143 Southeast Asia 1,270 6,519 5,249 1,041 Source: Sangrujee, Cáceres, and Cochi 2004. Western Pacific 8,670 10,327 1,657 356 Note: Costs are expressed in present value terms, calculated over the period 2005 to 2020, and World 128,120 67,009 61,111 1,457 discounted at 3 percent. Source: Khan and Ehreth 2003. Note: Cost savings, immunization costs, and net costs are present values for 2000 in millions of U.S. dollars, calculated for the period 1970­2050 and discounted at 5 percent. These estimates The respective cost to middle- and high-income countries is assume that immunization by OPV can cease after 2010. the same for all three scenarios, reflecting the assumption that the high-income countries will switch to IPV by 2005 and middle-income countries will do so between 2006 and 2008. with other health interventions, this cost to developing The scenarios differ only for the low-income countries. In the countries may still be comparatively cost-effective. However, first scenario, these countries are expected to continue routine Khan and Ehreth comment that the cost per disability-adjusted immunization using OPV; in the second, immunization ceases life year (DALY) saved is high for developing countries (see in 2011, followed by a system of surveillance and response. In table 62.3). As they explain (Khan and Ehreth 2003, 705), "This the third scenario, the low-income countries join the others in implies that without the financial support from developed switching to IPV between 2008 and 2010. Of these three sce- countries of the world many developing countries would not narios, the second comes closest to the 2005 post-eradication have opted for polio interventions for implementation. From strategy now advocated by the polio eradication program the developed countries' point of view, providing support leadership. for the polio program is not simply helping the poor and Unfortunately, this analysis is also deficient. First, interrup- the disadvantaged, it actually represents a good economic tion of transmission will not occur before 2006, and certifica- investment." tion will take an additional three years. Hence, analysis of Unfortunately, both of these cost-benefit studies have post-eradication costs should begin in 2009 at the earliest, with substantial limitations. First, both show that eradication is the costs of continuing immunization needing to be borne up economically attractive if one incorporates all costs and benefits until that time. Second, the analysis assumes a capacity to sup- from the inception of this program. Because eradication has not ply IPV that exceeds current estimates. It is not obvious that this yet been achieved, this approach mixes retrospective evaluation scenario is feasible or, if it were, if the costs of scaling up pro- and prospective analysis (historical expenditures and benefits duction are adequately reflected in the calculations. Third, and are sunk and so are irrelevant to the current situation). Second, most importantly, table 62.4 indicates that only low-income benefits and costs are calculated in both studies relative to a countries would benefit from polio eradication over this 15-year world without immunization. A better approach would be to time scale--and yet the table does not include any estimate of calculate the net benefits of eradication compared with the the risk these countries would bear of a possible outbreak. alternative of an optimal control program. The choice is not Although this analysis suggests that the discontinued use of between doing nothing and eradication. It is between an OPV promises the greatest return to eradication, this assumes optimal level of control and eradication. Finally, both studies that circulating vaccine-derived polioviruses could be con- assume that vaccination can cease in 2005 or 2010. As explained tained if and when they emerged. However, preparing for this previously, this possibility is highly unlikely. possibility would require a far more effective global surveil- A more recent analysis by Sangrujee, Cáceres, and Cochi lance system than now exists, maintenance of a laboratory (2004) calculates the costs for 15 years following the goal of infrastructure, and stockpiles of OPV. In addition, controlling certification of eradication in 2005 for three different scenarios: outbreaks with OPV without the risk of viruses reverting to continued use of OPV, OPV cessation with optional use of the virulence will be exceedingly difficult in the setting of an accel- killed or inactivated polio vaccine, and OPV cessation with erating proportion of immunologic-naive individuals. The use universal IPV. Table 62.4 shows their results. of OPV in this scenario could very well cause poliomyelitis to 1172 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson again become endemic. In any case, the estimated cost of any of This last omission is especially relevant to the study's analy- the strategies exceeds $20 billion. sis of the eradication program in Sudan. The study projected The economics of polio eradication are thus not as favorable that, by 1998, infections would cease everywhere except Sudan. as concluded by either Bart, Foulds, and Patriarca (1996) or (Plainly, this prediction was wrong, although Sudan is the Khan and Ehreth (2003). Both studies assume that vaccination largest problem for the program, mainly because of the ongo- can cease without IPV being used as a substitute anywhere, ing civil war, which has limited accessibility to endemic areas; both exclude the costs of maintaining a response capacity, and see Hopkins and others 2002.) It then calculates the net present neither accounts for the real threat of reemergence. Sangrujee, value of eliminating the disease there. The results are not Cáceres, and Cochi (2004) take account of two of these consid- promising. They show that eradication is attractive only if the erations, but their analysis calculates only the costs for 15 years, disease can be eliminated in Sudan within five years. However, ignoring both the risk of reemergence and the benefits of this analysis ignores the dividend that eradication would earn eradication. Hence, each study provides only a partial glimpse Sudan. It also disregards the most important feature of of the economics of polio eradication and does not adequately eradication--that if the disease were certified to have been address the fundamental difficulty (inability) of stopping eliminated from its last stronghold, it would yield a benefit to vaccination and maintaining eradication. all potentially vulnerable countries. Thus, the economics of In conclusion, although the economics of polio eradication eliminating dracunculiasis from Sudan, if that is where the may have been thought to be favorable by some (Aylward and disease makes its last stand, will be much more attractive than others 2003), they are far less favorable than were the econom- suggested by this analysis. ics of smallpox eradication, even assuming that polio vaccina- tion could cease. CONCLUSIONS Dracunculiasis Of the several attempts to eradicate diseases, all but one has Dracunculiasis, or guinea worm disease, is a nematode infec- failed. Even the exception, smallpox, barely succeeded despite tion, which is controlled not by vaccination but by education of the many factors favorable to eradication. Whether any eradi- the affected population, provision of nematode-free water cation effort will ultimately succeed or fail cannot be known through wells or filtration, and treatment of cases. It is not a with certainty at the time it is launched. Eradication entails global disease but found only in the rural areas of a few very risk. Money spent on eradication may not ultimately pay a poor tropical countries. This last difference is especially impor- dividend. Health risks may also exist. If eradication fails and tant from an economics perspective. It means that interna- vaccination levels drop after the eradication goal is abandoned, tional financing of a guinea worm eradication program needs susceptible persons who were previously shielded from infec- to rely more heavily on development assistance rather than on tion may become infected at a later age, when the disease can the self-interest of donor countries. cause greater harm. The risk also exists that, even if eradication Thus far, the eradication program has been successful in succeeds, the disease may be reintroduced by accidental or reducing the number of cases of guinea worm 99 percent from deliberate release. the 1986 level (Carter Center 2004). The geographic range of The reasons for potential failure of an eradication effort are the disease has also been reduced from 20 to just 12 countries. many. A nonhuman host may not be discovered until the num- Although this achievement is important, eradication remains ber of infected humans drops to a very low level (as happened elusive many years beyond 1995, the year that the WHA set for with yellow fever). The tools of eradication may be vulnerable eradication in 1991 (Cairncross, Muller, and Zagaria 2002, 232). to resistance (insecticides and drugs in the case of malaria). Only one cost-benefit study of the guinea worm eradication Political problems and civil strife may prevent an eradication program has been published (Kim, Tandon, and Ruiz-Tiben program from being executed in critical areas where the disease 1997), and it is unfortunately flawed in a number of respects. makes its last stand (a problem today for guinea worm). First, as indicated previously, eradication costs should be com- Termination of vaccination may leave populations vulnerable pared with those associated with an alternative optimal control to microbe reintroduction from an unforeseen reservoir or program. Second, the cost-benefit analysis applies to the period vaccine strain reversion (a risk now facing the poliomyelitis ini- 1987 to 1998 and thus is backward looking. The analysis can tiative). Another potential reason for failure is the inability reveal whether the money spent previously yielded a benefit in to raise the financial resources needed to complete programs excess of the cost (it did), but it cannot reveal whether eradica- that extend beyond expected targets. All eradication programs tion was worth pursuing at the time that this study was have experienced serious financial stringencies during the undertaken. Finally, it takes no account of the investment course of their execution. decision of eradication--the main reason for pursuing the Most eradication programs to date have been launched eradication goal in the first place. as visionary, far-reaching efforts but with vastly incomplete Control and Eradication | 1173 information. Basic epidemiological information and knowledge REFERENCES of the effectiveness and operational constraints of interventions Anderson, R. M., and R. May. 1991. Infectious Diseases of Humans: and costs in different settings are often inadequate, and the Dynamics and Control. Oxford, U.K.: Oxford University Press. required monitoring, evaluation, training, and research compo- Andrews, J. M., and A. D. Langmuir. 1963. "The Philosophy of Disease nents of the program may be absent. If a program's administra- Eradication." American Journal of Public Health 53: 1­6. tors lack a careful, probing analysis of the epidemiology of the Aylward, R. B., A. Acharya, S. England, M. Agocs, and J. Linkins. 2003. various candidate diseases or of the technologies available, and "Polio Eradication." In Global Public Goods for Health: Health Economic if their comprehension of the potential costs and who would and Public Health Perspectives, eds. R. Smith, R. Beaglehole, D. Woodward, and N. Drager, 33­53. Oxford, U.K.: Oxford University bear them is limited, a program is likely to founder, causing a Press. dispirited staff, confused beneficiaries, and donor fatigue and Aylward, R. B., J. M. Olive, H. F. Hull, C. A. De Quadros, and B. Melgaard. ambivalence. It is crucial that the eradication methodologies 1998. "Ensuring Common Principles Lead to Mutual Benefits: Disease and assumptions in those regions of the world that would be Eradication Initiatives and General Health Services." In The Eradication of Infectious Diseases, eds. W. R. Dowdle and D. R. Hopkins. most likely to pose the most significant problems be tested and New York: John Wiley and Sons. addressed before launching an eradication program and that Barrett, S. 2003. "Global Disease Eradication." Journal of the European evaluation and research continue during the program. Economic Association 1: 591­600. Proposals for disease eradication have seldom been brought ------. 2004. "Eradication vs. Control: The Economics of Global to the WHA with specific plans, costs, and uncertainties fully Infectious Disease Policy." Bulletin of the World Health Organization 82 laid out. Nor have the expected sources of fiscal support and (9): 683­88. needed country support been addressed with specific commit- Barrett, S., and M. Hoel. 2003. "Optimal Disease Eradication." Disease Control Priorities Project Working Paper 22. ments requested of the members. The WHA has only a limited deliberative capacity, and too much cannot be expected of its Bart, K. J., J. Foulds, and P. Patriarca. 1996. "Global Eradication of Poliomyelitis: Benefit-Cost Analysis." Bulletin of the World Health members in session. However, designated special committees of Organization 74 (1): 35­45. the WHA can and should be appointed, consisting of both Basu, R. N., Z. Jezek, and N. A. Ward. 1979. The Eradication of Smallpox visionary eradicationists and field-experienced public health from India. New Delhi: World Health Organization. and social science personnel. The WHA should take up the Bellmunt, A., G. May, R. Zell, P. Pring-Akerblom, W. Verhagen, and question of eradication only after the subject has been thor- A. Heim. 1999. "Evolution of Poliovirus Type I during 5.5 Years of Prolonged Enteral Replication in an Immunodeficient Patient." oughly vetted and sufficiently large-scale pilot programs in the Virology 265: 178­84. most problematic areas have clarified that an adequate under- Biellik, R., S. Madema, A. Taole, A. Kutsulukuta, E. Allies, R. Eggers, and standing of the epidemiology exists and that the appropriate others. 2002. "First 5 Years of Measles Elimination in Southern Africa: technologies are available. 1996­2000." Lancet 359 (9317): 1564­68. In the past, members have not voted for a specific program Breman, J. G., and I. Arita. 1980. "The Confirmation and Maintenance of for which all the uncertainties have been laid out and the ben- Smallpox Eradication." New England Journal of Medicine 303 (22): 1263­73. efits and costs associated with different outcomes have also Brilliant, L. B. 1985. The Management of Smallpox Eradication in India. been calculated. Nor, with one exception, have they voted for a Ann Arbor: University of Michigan Press. resolution imposing responsibilities, including financing obli- Cairncross, S., R. Muller, and N. Zagaria. 2002. "Dracunculiasis (Guinea gations, on individual states. The next time a proposal to erad- Worm Disease) and the Eradication Initiative." Clinical Microbiology icate a disease is presented to the WHA, it should be compre- Reviews 15 (2): 223­46. hensive. It should demonstrate why the effort is worth taking, Carter Center. 2004. "Guinea Worm Eradication: Review, June 1, 2003." even if the final outcome is uncertain; it should bind states, Carter Center, Atlanta. http://www.cartercenter.org/printdoc.asp? docID=1785andsubmenu=news. morally if not legally, to fulfill the pledges needed to see the CDC (U.S. Centers for Disease Control and Prevention). 1998a. "Progress program through to its completion; and it should prepare con- toward Elimination of Measles from the Americas." Morbidity and tingencies should the eradication effort fail. Mortality Weekly Report 47 (10): 189­93. ------. 1998b. "Progress toward Global Measles Control and Regional ACKNOWLEDGMENTS Elimination, 1990­1997." Morbidity and Mortality Weekly Report 47 (48): 104. We would like to acknowledge Walter Dowdle and Maria Teresa ------. 1999a. "Global Measles Control and Regional Elimination, 1998­1999." Morbidity and Mortality Weekly Report 48 (49): 1124. Valenzuela for their critical review of our manuscript; Joel Breman and Philip Musgrove for providing editorial guidance; ------. 1999b. "Progress toward Measles Elimination--Southern Africa, 1996­1998." Morbidity and Mortality Weekly Report 48 (27): 585. John Sentz for his numerous hours devoted to research and ------. 2003a. "Global Progress toward Certifying Polio Eradication and editorial assistance; and Cherice Holloway for her preparation Laboratory Containment of Wild Polioviruses--August 2002­August of various versions of this manuscript. 2003." Morbidity and Mortality Weekly Report 52 (47):1158. 1174 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson ------. 2003b. "Progress toward Poliomyelitis Eradication--Angola and Fenner, F., A. J. Hall, and W. R. Dowdle. 1998. "What Is Eradication?" In the Democratic Republic of Congo, January 2002­June 2003." The Eradication of Infectious Diseases, eds. W. R. Dowdle and D. R. Morbidity and Mortality Weekly Report 52 (34): 816. Hopkins, 3­17. New York: John Wiley and Sons. ------. 2003c. "Progress toward Poliomyelitis Eradication--Southern Fenner, F., D. A. Henderson, I. Arita, Z. Jezek, and I. D. Ladnyi. 1988. Africa, 2001--March 2003." Morbidity and Mortality Weekly Report 52 Smallpox and Its Eradication. Geneva: World Health Organization. (22): 521. Goodman, R. A., K. L. Foster, F. L. Trowbridge, and J. P. Figueroa, eds. ------. 2003d. "Update: Global Measles Control and Mortality 1998a. "Comments and Discussion Following Work Group Reports." Reduction--Worldwide, 1991­2001." Morbidity and Mortality Weekly In "Global Disease Elimination and Eradication as Public Health Report 52 (20): 471. Strategies," Bulletin of the World Health Organization 76 (Suppl. 2): 104­8. ------. 2004a. "Brief Report: Global Polio Eradication Initiative Strategic Plan, 2004." Morbidity and Mortality Weekly Report 53 (05): 107. ------. 1998b. "Global Disease Elimination and Eradication as Public Health Strategies: Proceedings of a Conference. Atlanta, Georgia, ------. 2004b. "Measles Mortality Reduction--West Africa, 1996­2002." February 23­25, 1998." Bulletin of the World Health Organization 76 Morbidity and Mortality Weekly Report 53 (02): 28. (Suppl.) 2: 5­162. ------. 2004c. "Progress toward Global Eradication of Poliomyelitis, Henderson, D. A. 1982. "The Deliberate Extinction of a Species." January 2003­April 2004." Morbidity and Mortality Weekly Report 53 Proceedings of the American Philosophical Society 126: 461­71. (24): 532. ------. 1988. "Development of the Global Smallpox Eradication ------. 2004d. "Progress toward Measles Elimination--Region of the Programme, 1958­1966." In Smallpox and Its Eradication, eds. Americas, 2002­2003." Morbidity and Mortality Weekly Report 53 (14): F. Fenner, D. A. Henderson, I. Arita, Z. Jezek, and I. D. Ladnyi, 365­419. 304. Geneva: World Health Organization. ------. 2004e. "Progress toward Poliomyelitis Eradication--Nigeria, ------. 1998a. "Eradication: Lessons from the Past." Bulletin of the World January 2003­March 2004." Morbidity and Mortality Weekly Report 53 Health Organization 76 (Suppl. 2): 17­21. (16): 343. ------. 1998b. "The Siren Song of Eradication." Journal of the Royal ------. 2004f. "Progress toward Sustainable Measles Mortality College of Physicians of London 32 (6): 580­84. Reduction--South-East Asia Region, 1999­2002." Morbidity and Hopkins, D. R., E. Ruiz-Tiben, N. Diallo, P. C. Withers Jr., and J. H. Mortality Weekly Report 53 (25): 559. Maguire. 2002. "Dracunculiasis Eradication: And Now, Sudan." ------. 2004g. "Wild Poliovirus Importations--West and Central Africa, American Journal of Tropical Medicine and Hygiene 67 (4): 415­22. January 2003­March 2004." Morbidity and Mortality Weekly Report 53 Jeffrey, G. M. 1976. "Malaria Control in the Twentieth Century." American (20): 433. Journal of Tropical Medicine and Hygiene 25: 361­71. ------. 2005. "Progress toward Poliomyelitis Eradication--Poliomyelitis Jenner, E. 1801. The Origin of Vaccine Inoculation. London: Shury. Quoted Outbreak in Sudan, 2004." Morbidity and Mortality Weekly Report 54 in Fenner, Hall, and Dowdle 1998. (4): 97­99. Kew, O. M., V. Morris-Glasgow, M. Landaverde, C. Burns, J. Shaw, Z. Garib, Cello, J., A. V. Paul, and E. Wimmer. 2002. "Chemical Synthesis of and others. 2002. "Outbreak of Poliomyelitis in Hispaniola Associated Poliovirus cDNA: Generation of Infectious Virus in the Absence of with Circulating Type 1 Vaccine-Derived Poliovirus." Science 296 Natural Template." Science 297 (5583): 1016­18. (5566): 356­59. Cochi, S., C. de Quadros, W. Dowdle, R. Goodman, P. Ndumbe, D. Kew, O. M., P. F. Wright, V. I. Agol, F. Delpeyroux, H. Shimizu, N. Salisbury, and others. 1998. "Post-Conference Small Group Report." In Nathanson, and M. A. Pallansch. 2004. "Circulating Vaccine-Derived Global Disease Elimination and Eradication as Public Health Strategies, Polioviruses: Current State of Knowledge." Bulletin of the World Health eds. R. A. Goodman, K. L. Foster, F. L. Trowbridge, and J. P. Figueroa. Organization 82 (1): 16­23. Bulletin of the World Health Organization 76 (Suppl. 2): 113. Khan, M., and J. Ehreth. 2003. "Costs and Benefits of Polio Eradication: A Cockburn, A. 1961. "Eradication of Infectious Diseases." Science 133: Long-Run Global Perspective." Vaccine 21: 702­5. 1050­58. Kim, A., A. Tandon, and E. Ruiz-Tiben. 1997. "Cost-Benefit Analysis of the ------. 1963. The Evolution and Eradication of Infectious Diseases. Global Dracunculiasis Eradication Campaign." Policy Research Baltimore: Johns Hopkins University Press. Working Paper 1835, World Bank, Washington, DC. de Quadros, C. 2001. "Introduction." In Considerations for Viral Disease Knobler, S., J. Lederberg, and L. A. Pray, eds. 2001. Considerations for Viral Eradication Lessons Learned and Future Strategies, eds. S. Knobler, Disease Eradication Lessons Learned and Future Strategies. Washington, J. Lederberg, and L. A. Pray, 22­32. Washington, DC: National DC: National Academy Press. Academy Press. Kremer, M., and E. Miguel. 2004. "The Illusion of Sustainability." NBER Dowdle, W. R. 1998. "The Principles of Disease Elimination and Working Paper W10324, National Bureau of Economic Research, Eradication." Bulletin of the World Health Organization 76 (Suppl. 2): Cambridge, MA. 22­25. New Shorter Oxford English Dictionary. 1993. ------. 1999. "The Principles of Disease Elimination and Eradication." Ottesen, E. A., W. R. Dowdle, F. Fenner, K. O. Habermehl, T. J. John, Morbidity and Mortality Weekly Report 48 (Suppl.): 23­27. M. A. Koch, and others. 1998."Group Report: How Is Eradication to Be Dowdle, W. R., and D. R. Hopkins, eds. 1998. The Eradication of Infectious Defined and What Are the Biological Criteria?" In The Eradication of Diseases. New York: John Wiley and Sons. Infectious Diseases, eds. W. R. Dowdle and D. R. Hopkins. New York: Dubos, R., and J. Dubos. 1953. The White Plague: Tuberculosis, Man and John Wiley and Sons. Society. London: Gollancz. Quoted in Fenner, Hall, and Dowdle 1998. Payne, A. M.-M. 1963a. "Basic Concepts of Eradication." American Review Eichner, M., and K. Dietz. 1996. "Eradication of Poliomyelitis: When Can of Respiratory Disease 88: 449­55. One Be Sure That Polio Virus Transmission Has Been Terminated? ------. 1963b. "Disease Eradication as an Economic Factor." American American Journal of Epidemiology 143 (8): 816­22. Journal of Public Health 53: 369­75. Control and Eradication | 1175 Ramaiah, T. J. 1976. "Cost-Benefit Analysis of the Intensified Campaign Spînu, I., and S. Biberi-Moroianu. 1969. "Theoretical and Practical against Smallpox in India." National Institute of Health Administration Problems Concerning the Eradication of Communicable Diseases." and Education Bulletin 9 (3): 169­203. Archives roumaines de pathologie experimentales et de microbiologie 28: Roberts, L. 2004. "Polio Endgame. Polio: The Final Assault?" Science 303 725­42. (5666): 1960­68. Steinglass, P. 2001. Thematic Evaluations in 2001 Eradication of Sabin, A. B. 1991. "Measles, Killer of Millions in Developing Countries: Poliomyelitis. Report by the director general, Programme Development Strategy for Rapid Elimination and Continuing Control." European Committee of the Executive Board Eight Meeting, December 13, 2001. Journal of Epidemiology 7 (1): 1­22. Document EBPDC8/3. Geneva: World Health Organization. Salisbury, D. 1998. "Report of the Work Group on Disease Taylor, C. E., F. Cutts, and M. E. Taylor. 1997. "Ethical Dilemmas in Elimination/Eradication and Sustainable Health Development." In Current Planning for Polio Eradication." American Journal of Public Global Disease Elimination and Eradication as Public Health Strategies, Health 87 (6): 922­25. eds. R. A. Goodman, K. L. Foster, F. L. Trowbridge, and J. P. Figueroa. WHO (World Health Organization). 2003. Global Polio Eradication Bulletin of the World Health Organization 76 (Suppl. 2): 72­79. Initiative: Estimated External Financial Resource Requirements Sangrujee, N., V. M. Cáceres, and S. L. Cochi. 2004. "Cost Analysis of 2004­2008. Geneva: WHO. Post-Polio Certification Immunization Policies." Bulletin of the World ------. 2004. Report of the Strategic Advisory Group of Experts (SAGE). Health Organization 82 (1): 9­15. Geneva: WHO. http://www.who.int/vaccines-documents/DocsPDF05/ Sencer, D. J., and N. W. Axnick. 1973. "Cost Benefit Analysis." In Sage_Report_2004.pdf. International Symposium on Vaccination against Communicable Yekutiel, P. 1980. Eradication of Infectious Diseases: A Critical Study. New Diseases, Monaco 1973, 22: 37­46. Symposia Series in Immuno- York: Karger. biological Standardization. Basel, Switzerland: Karger. 1176 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson Chapter 63 Integrated Management of the Sick Child Cesar G. Victora, Taghreed Adam, Jennifer Bryce, and David B. Evans NATURE, CAUSES, AND BURDEN In addition, comorbidity is common. Among children who OF CHILD MORTALITY die, a large proportion present with two or more diagnoses (Black, Morris, and Bryce 2003). Comorbidity is also highly Every year, over 10 million children under five years of age die. prevalent at the community level and among children seeking Most of those deaths are due to a small number of causes. In health care. Nutritional factors--including underweight, the mid 1990s, it was estimated that 70 percent of all global micronutrient deficiencies, and inadequate infant feeding prac- child deaths were due to five conditions: diarrhea, pneumonia, tices (see chapter 28)--play a major role in morbidity and malaria, measles, and malnutrition (Gove 1997; Tulloch 1999). mortality, and yet these are often overlooked by practitioners. The World Health Organization (WHO) has since conducted a Also, there are many missed opportunities for preventive inter- comprehensive review of under-five deaths using additional ventions during outpatient visits--for example, immuniza- data and improved methods (Bryce and others 2005), and now tions and promotion of insecticide-treated mosquito nets. estimates that six causes accounted for 73 percent of these deaths in 2000­2003: pneumonia (19 percent), diarrhea (18 percent), malaria (8 percent), neonatal pneumonia or sepsis POLICY SHIFT TO INTEGRATED MANAGEMENT (10 percent), preterm delivery (10 percent), and asphyxia at birth (8 percent). Undernutrition is an underlying cause in at Until the mid 1990s, actions aimed at improving child health least half of all under-five deaths. Few conditions, therefore, were organized as vertical programs, each addressing a specific account for a large proportion of all deaths. disease or providing a given intervention or set of interventions These deaths are not randomly distributed.They tend to occur (Claeson and Waldman 2000). Typical examples of these pro- in the poorest countries of the world, mostly in Sub-Saharan grams are the Expanded Program on Immunizations (EPI), Africa and South Asia (Black, Morris, and Bryce 2003), and with- Control of Diarrhoeal Diseases (CDD), acute respiratory infec- in any country they affect mostly the poorest families (Victora tion (ARI) programs, malaria control programs, and nutrition and others 2003). Fortunately, cost-effective interventions are programs that include growth monitoring, breastfeeding pro- available to prevent most of these deaths. Chapters 19, 21, 24­28, motion and support, and micronutrient supplementation. and 56 in this volume, as well as the next section in this chapter, The need for an integrated approach to improve child health describe these interventions in greater detail.Achieving universal became evident in the mid 1990s for a number of reasons. coverage with these interventions would likely prevent 60 percent From the perspective of epidemiology, a small number of of those deaths (Jones and others 2003). Yet coverage levels for diseases accounted for a high proportion of deaths, and those nearly all of these interventions remain below 50 percent (Bryce diseases were often present in the same children and had over- and others 2003), and children from the poorest families are least lapping clinical signs. Integrated management was expected to likely to be reached (Victora and others 2003). increase the probability that children would receive treatment 1177 for all major diseases and to decrease the possibility that chil- on a set of adapted algorithms (Gove 1997) that guide health dren would receive correct treatment for one disease and die workers through a process of assessing signs and symptoms, from another unrecognized illness. The important role played classifying the illness according to treatment needs, and provid- by nutrition across these major diseases also suggested that an ing appropriate treatment and education to the child's caregiver. integrated approach to case management was needed to ensure Figure 63.1 shows a general outline of the approach for children that health workers addressed children's nutritional needs age two months to five years (WHO and UNICEF 2001). Sick throughout the clinical encounter. children attending a first-level health facility are initially checked A second set of reasons for the policy shift to an integrated for danger signs and for the main symptoms of the key IMCI dis- approach was based on the need to promote managerial effi- eases: diarrhea, malaria, pneumonia, measles, and other severe ciency. The vertical approach required countries to appoint infections. Next, all children are assessed for malnutrition and managers at national, provincial, and district levels to run each anemia, and vaccination status is verified. Children under two program. It also led to separate training activities; for example, years of age, as well as older children presenting low weight for health workers might be required to leave their posts on a num- age, receive nutrition counseling. Other health problems related ber of occasions to be trained for the programs. Similar exam- by caretakers are then assessed, and children are classified ples of duplication of effort were often found in supervision according to a color code: pink (immediate referral), yellow and provision of essential drugs. There was a strong logical (management in the outpatient facility), or green (home man- basis for believing that integrating the management structure agement). Separate case-management algorithms are available of child health programs would lead to improved efficiency. for children under two months of age.IMCI health worker train- A third group of reasons for the shift to integrated case ing emphasizes the integration of curative care with preventive management related to the need to improve the quality of case measures, including nutrition and vaccinations. A special train- management provided by health workers. Vertical programs ing module addresses how to communicate effectively with trained health workers to manage one disease at a time, and mothers. The training course was originally designed to last decisions about how best to assess and treat those diseases, as 11 days, including a large amount of hands-on experience. well as how to promote nutrition and educate caretakers, were often left to individual health workers. An integrated set of guidelines for managing sick children ensured that health Improving Health Systems workers, including those with low levels of training, applied the The second component of IMCI is aimed at providing support best available knowledge of case management systematically for child health service delivery, including drug availability, and in correct sequence. effective supervision, referral services, and health information The realization that a few diseases were responsible for systems. Tools were developed for implementing specific most child deaths, that comorbidity was highly prevalent, that system-strengthening interventions, including a planning effective interventions were available, and that there were guide for national and district managers, an integrated health many missed opportunities for prevention led to the recogni- facility assessment tool, and a tool for improving referral level tion that an integrated approach was needed. Thus, WHO and care. In particular, several countries--beginning in Latin United Nations Children's Fund (UNICEF) launched the America through the Pan American Health Organization Integrated Management of Childhood Illness (IMCI) strategy (PAHO) and more recently in Africa with WHO's Regional in the mid 1990s (Tulloch 1999). Tanzania and Uganda began Office for Africa--made substantial efforts to improve the implementing IMCI in 1996. By 2003, more than 100 coun- management and availability of the specific drugs required for tries had adopted the strategy (http://www.who.int/child- IMCI (A. Bartlett, personal communication). adolescent-health). Improving Family and Community Practices INTERVENTIONS The third component, known as community IMCI, focuses A key aspect of IMCI was the integration of effective interven- on 12 key family practices relevant to child health and tions to improve child health and nutrition into a coordinated development (see http://www.who.int/child-adolescent- strategy. IMCI has three components, each of which was meant health/PREVENTION/12_key.htm). Community IMCI sup- to be adapted at the country level according to local epidemiol- ports the development and implementation of community- ogy, health system characteristics, and culture. and household-based messages and interventions to increase the proportions of children exposed to these practices. These Improving Health Worker Performance behaviors address breastfeeding, complementary feeding, The first component of IMCI includes health worker training micronutrients, personal hygiene, immunizations, insecticide- and the reinforcement of correct performance. Training is based treated nets, mental and social development, continued feeding 1178 | Disease Control Priorities in Developing Countries | Cesar G. Victora, Taghreed Adam, Jennifer Bryce, and others The Integrated Case Management Process Outpatient health facility Check for danger signs · Convulsions · Lethargy or unconsciousness · Inability to drink or breastfeed Assess main symptoms · Cough or difficulty breathing · Diarrhea · Fever · Ear problems Assess nutrition and immunization status and potential feeding problems Check for other problems Classify conditions and identify treatment actions according to color-coded treatment charts Pink Yellow Green Urgent referral Treatment at outpatient Home management health facility Outpatient health facility Outpatient health facility Home · Give prereferral treatments · Treat local infection Caretaker is counseled on how to · Advise parents · Give oral drugs · Give oral drugs · Refer child · Advise and teach caretaker · Treat local infections at home · Follow up · Continue feeding · Determine when to return immediately · Follow up Pink Urgent referral Referral facility · Provide emergency triage and treatment · Diagnose · Treat · Monitor and follow up Source: WHO and UNICEF 2001. Figure 63.1 Schematic Outline of IMCI Case Management for Children Age Two Months to Five Years Integrated Management of the Sick Child | 1179 and increased fluids during illness, home treatment of infec- tions of childhood interventions in different parts of the world tions, care-seeking practices, compliance with health worker (see http://www.who.int/evidence/cea). This research resulted recommendations, and prenatal care. in estimates of the cost-effectiveness of single or combined This chapter addresses issues related to the integrated deliv- interventions compared with doing nothing or with incremen- ery of these interventions, most of which are covered in greater tal intervention or current practice. Because large-scale trials detail in other chapters in this book. These include case man- on the effects of joint interventions have not yet been under- agement of ARI (chapter 25), diarrhea (chapter 19), malaria taken, the joint effects were modeled using the effectiveness of (chapter 21), and malnutrition (chapter 28); community inter- the individual interventions taken from systematic reviews. ventions to improve nutrition, including breastfeeding promo- The interventions included vitamin A and zinc fortification tion and complementary feeding (chapter 56); insecticide- and supplementation, oral rehydration therapy, case manage- impregnated bednets (chapter 21); anthelmintic treatment ment for pneumonia, and supplementary feeding and growth (chapter 24); vaccinations (chapter 20); and micronutrient monitoring. Costs and effects were estimated at various levels supplementation (chapter 28). of population coverage and in various combinations. The results showed that a childhood package consisting of vitamin A and zinc supplementation, oral rehydration therapy, INTERVENTION COST AND COST-EFFECTIVENESS and case management of pneumonia was cost-effective com- pared with doing nothing in most settings but that including One of the rationales for developing the IMCI strategy was the supplementary feeding and growth monitoring was not cost- belief that treating the sick child in an integrated manner, by effective. Implementation of this combination at 50 percent building on interventions that had already been shown to be coverage was estimated to cost, in 2000 prices, approximately cost-effective, would result in gains in efficiency. Two types of US$4.10 per child (US$0.60 per capita) in poor African coun- questions can be asked from an economic perspective. First, is tries such as Tanzania. The cost-effectiveness was US$38 per treating children on the basis of the IMCI strategy cost-effective? DALY averted. Costs increase faster than the increase in cover- Second, do the additional health benefits gained by switching age. In Tanzania, it was estimated to cost an additional from routine practice to IMCI justify the additional costs US$12.10 per child under age five (US$1.80 per capita) to reach (if any)? 95 percent coverage, with a resulting incremental cost- Only one publication has reported the cost-effectiveness of effectiveness ratio of US$60 per DALY averted. the IMCI strategy as a whole. Using a modeling exercise, the This study is important because it is one of the few that has World Development Report 1993 identified IMCI as being able specifically explored the cost-effectiveness of undertaking com- to avert 14 percent of the global burden of disease in children bined interventions in the same population, and the effect of under age five in resource-poor countries at a cost of only increasing coverage on costs. These absolutely critical questions US$1.60 per capita per year, with a cost-effectiveness of US$30 for policy makers considering different intervention strategies to US$100 per disability-adjusted life year (DALY) averted. No are also critical to IMCI. However, the WHO study did not ana- details of the methods used to derive those estimates are avail- lyze the same interventions included in the IMCI package, nor able (World Bank 1993). It is not clear if the costs are the addi- did it evaluate the effect of moving from current practice to tional costs of moving from current practice to IMCI or the costs IMCI-based care. of undertaking care for children under age five using IMCI, nor Some information on the effect of moving from current is it clear if the effectiveness is the additional effectiveness of practice is available from two studies in Kenya and Nigeria. changing current practice or the total effect of the package. Those studies compared the cost to the provider of traditional Detailed studies of the cost-effectiveness of some of the com- prescribing patterns with the costs of pharmaceuticals that ponents of IMCI are available. For example, oral rehydration would have resulted if the IMCI guidelines had been followed therapy for diarrhea, case management for pneumonia, and strictly. In Nigeria, the traditional prescribing method was five childhood vaccinations have been shown to be very cost-effective times more expensive: US$1.44 per child visit for pharmaceuti- when evaluated as separate interventions (see chapters 19, 25, cals compared with US$0.29, using 1996 estimates and 26). It is likely that the combination of different sets of (Wammanda, Ejembi, and Iorliam 2003). In Kenya, also in childhood interventions, as proposed by IMCI, would also be 1996, the traditional method was almost three times costlier per cost-effective, although this depends on the relationship between child visit if the low-cost combination of drugs was assumed costs and effects when the interventions are undertaken at the (US$0.44) and similar if the high-cost combination was same time in the same population. assumed (US$0.16) (Boulanger, Lee, and Odhacha 1999). In No published studies of the extent and nature of efficiency Bangladesh, it was estimated that strict adherence to the IMCI gains through integration were found. WHO has recently protocol could result in US$7 million in savings at the national explored some of these gains for slightly different combina- level simply from more rational use of drugs--almost 3 percent 1180 | Disease Control Priorities in Developing Countries | Cesar G. Victora, Taghreed Adam, Jennifer Bryce, and others of the total health budget of the government of Bangladesh training for health workers had been achieved, but there had (Khan, Ahmed, and Saha 2000; Khan, Saha, and Ahmed 2002). been no increase in the provision of under-five interventions at These estimates were based on models and assumptions, the community level, as opposed to the facility level. sometimes using evidence from separate systematic reviews Cost data were collected for the start-up period of imple- addressing the costs and effects of an intervention. It would be menting IMCI (from 1996 to 1997)--defined as the time from valuable if information on relative cost-effectiveness of differ- the national decision to implement IMCI to the time when ent combinations of interventions at variable levels of coverage IMCI started to be provided in health facilities--and for the could be derived from field studies rather than developed sole- maintenance of child health services in both types of districts. ly by modeling. Such an approach would allow the use of com- Costs were estimated from the societal perspective and were parable methods and counterfactuals across the evaluation sites collected from the national, district, hospital, health facility, to make the results more useful and generalizable to other and household levels. Costs at all these levels were summed to settings. In addition to answering questions related to the cost- obtain the total cost to the district of providing care for chil- effectiveness of IMCI, it would clarify gains that can be obtained dren under age five. So that comparison could be made across from delivering interventions at the same time as part of an inte- districts, cost estimates were standardized to a hypothetical dis- grated package rather than delivering them in a vertical manner. trict with a population of 50,000 children under age five. This This is one of the reasons the Multi-Country Evaluation of figure corresponds to a total population of around 300,000, IMCI Effectiveness, Cost, and Impact (MCE) was launched which is roughly the average district population for Tanzania. (Bryce and others 2004). Five countries are currently participat- Estimates of the additional cost to the district of implementing ing in in-depth studies--Bangladesh (in 20 catchment areas), IMCI were based on the difference in cost of under-five care Brazil (in 46 municipalities), Peru (in all 24 departments in the between the IMCI districts and the comparison districts, country), Tanzania (in 4 districts), and Uganda (in 10 districts). which, at the time of the study, had not yet implemented IMCI Seven other countries--Bolivia, Cambodia, Kazakhstan, the (Adam and others 2004b). Kyrgyz Republic, Morocco, Niger, and Zambia--were assessed For 1999, the cost per child of caring for children under age for the evaluation but could not be included, mostly because of five in IMCI districts was US$11.19, 44 percent lower than in insufficient implementation of IMCI. the comparison districts (US$16.09) (Adam and others 2004b). The overall objective of the MCE is to evaluate the actual The lower cost per child in IMCI districts was due to lower hos- changes associated with IMCI as it is implemented in different pitalization and administrative costs at the district level. There settings. All studies measured an identical set of indicators and, was no statistically significant difference in costs incurred to with minor exceptions, used identical data collection tools treat children at primary care facilities and at the household (Bryce and others 2004). The remainder of this section presents level (figure 63.2). the main findings from two MCE countries, Tanzania and Brazil, for which evaluation results are currently available. 1999 US$ 8 IMCI (US$11.19) 7 Comparison (US$16.09) MCE in Tanzania The MCE in Tanzania uses an observational design to compare 6 two districts where IMCI has been implemented since late 1997 5 (IMCI districts) with two districts where implementation began in 2002 (comparison districts). The four districts had 4 reasonably well-functioning health services, comparable levels 3 of per capita health expenditure, high utilization rates of gov- ernment health facilities, and high coverage of selected inter- 2 ventions (for example, EPI). Large numbers of governmental 1 and nongovernmental health actors were also active in the dis- tricts, many of which were involved in health worker training 0 and community activities, although their coverage was patchy. National District Hospital Primary Household facility The two IMCI districts had engaged in activities designed to Level strengthen district management skills; the districts also had authority for priority setting and control over their health Source: Adam and others 2004b. Note: Standard district with 50,000 children under age five. budgets. These activities of national health sector reform had not started in the comparison districts at the time of the study. Figure 63.2 Cost Components of Under-Five Care per Child in a In the comparison districts, a high level of coverage of IMCI Standard District Integrated Management of the Sick Child | 1181 Hospital costs were 2.5 times higher in the comparison dis- variation in the other parameters, only to the assumption tricts, not because of differences in the cost per under-five about rates of hospitalization. Therefore, if one assumes that admission, but because more children under age five were hos- hospital admission rates were not related to IMCI, there is no pitalized in those districts relative to IMCI districts (6 percent difference in the cost of under-five care in the two types of dis- in IMCI districts compared with 15 percent in comparison dis- tricts. Otherwise, the costs in IMCI districts are lower than in tricts; t­test: p 0.001). There are two possible explanations: the comparison districts. (a) improved quality of care and drug availability for children In the IMCI districts, IMCI was implemented concurrently under age five at IMCI primary facilities reduced the need for with measures designed to strengthen district management, referral and subsequent admission to hospitals, or (b) factors such as evidence-based planning and expenditure mapping other than IMCI, such as differences in quality or geographical at district level (http://web.idrc.ca/en/ev-3170-201-1-DO_ access to the hospitals in the different settings, meant that chil- TOPIC.html). In fact, it has been argued that the decision to dren in non-IMCI districts were more likely to seek care at hos- implement IMCI in the study districts was a result of the intro- pitals. Given that IMCI training had only started one year duction of the evidence-based planning. It is not possible to before data collection of hospital admissions, the second possi- separate the effects of IMCI from district-strengthening meas- bility may have played a bigger role in this finding. Even if one ures. The findings of the MCE study in Tanzania, therefore, can takes the most conservative assumption--that all the difference be interpreted as the costs of IMCI in the presence of a strong was due to other factors--and excludes the hospital component health system with adequate managerial capacity. from the analysis, the total cost per under-five child in IMCI The US$11.20 cost per child of treating children under age districts was still lower than in comparison districts (6 percent). five using IMCI in Tanzania translates into a per capita cost of The other important difference in costs between both types US$1.70, compared with US$2.30 for routine care. This finding of districts was found in costs incurred at the district level, is similar to previous per capita estimates of the cost of IMCI which were 50 percent higher in the comparison districts. in resource-poor countries (World Bank 1993). In addition, the These costs were mainly linked to more frequent trips for drug Tanzania evaluation had similar findings with respect to sav- distribution and general purpose supervision in comparison ings from drug costs to those expected based on previous stud- districts than in IMCI districts. ies (Khan, Saha, and Ahmed 2002; Wammanda, Ejembi, and Similar costs of training were observed in both types of dis- Iorliam 2003; World Bank 1993). tricts during the study period. This finding was unexpected The effects of IMCI can be assessed in terms of changes in given the emphasis of IMCI on training, but a wide variety of intermediate outcomes, such as improved quality of care at training courses were performed in comparison districts for health facilities, or in terms of final outcomes, such as changes preventive, curative, and administrative issues during the study in under-five mortality or DALYs averted. In the Tanzania eval- period. These courses included training for immunization, for uation, a health facility survey was carried out in 2000 to com- use of insecticide-treated bednets, and for use of district Health pare the quality of case management and health systems sup- Management Information System forms. port in IMCI and comparison districts. The results indicate At the facility level, univariate comparison between IMCI that children in IMCI facilities received better care than chil- and comparison district health facilities showed a 16 percent dren in comparison districts. Their health problems were more difference in the average cost per under-five visit (including thoroughly assessed, they were more likely to be diagnosed and vaccination visits) at government health centers and dispen- treated correctly as determined through a gold-standard reex- saries (US$1.40 and US$1.60 in IMCI and comparison districts, amination, and the caretakers of the children were more likely respectively; t-test: p 0.5). The average number of visits per to receive appropriate counseling and reported higher levels of child per year was 30 percent higher in the IMCI districts (3.28) knowledge about how to care for their sick children (Tanzania compared with comparison districts (2.49). Taken together, the IMCI Multi-Country Evaluation Health Facility Survey Study lower cost per visit but higher number of visits per child per Group 2004). year in IMCI facilities resulted in similar overall costs per child Estimating the effectiveness of IMCI training in improving under age five for treatment in the two types of districts. health workers' performance required measuring the propor- Multivariate regression analysis, however, led to a different con- tion of children correctly managed in IMCI and comparison clusion. Taking into account differences in other determinants facilities. Correct management is defined as the correct drug across facilities, in particular the number of visits per facility, being provided in the correct formulation (amount, times per the cost per visit was at least 30 percent lower in IMCI facilities day, number of days) and the health worker explaining cor- (t-test: p 0.001). rectly to the caretaker how the drug should be administered at Sensitivity analysis showed the importance of hospitaliza- home. Not prescribing an antibiotic or antidiarrheal drug for a tion costs in interpreting total district costs--the difference child who did not need one was also considered to be correct between IMCI and comparison districts was not sensitive to performance. In Tanzania, 65 percent of children under age five 1182 | Disease Control Priorities in Developing Countries | Cesar G. Victora, Taghreed Adam, Jennifer Bryce, and others Box 63.1 Impact of IMCI on Mortality and Nutrition in Tanzania Tanzania is the only MCE site where the evaluation has per 1,000 child-years or approximately 120 deaths per been completed. Its design included a comparison of mor- 1,000 children between birth and the age of less than five tality in four districts--two with and two without IMCI-- years. The quality of health care provided in the IMCI dis- over the two-year period starting in mid 2000. tricts was substantially higher than in the control districts Demographic surveillance systems were used to compare (see box 63.2). Over the following two years, mortality lev- under-five mortality rates in areas of the IMCI and control els became 13 percent lower in IMCI districts than in the districts. Adjustments for age (zero to one and one to four comparison areas, corresponding to a rate difference of years) and rainfall were made using Poisson regression 3.8 fewer deaths per 1,000 children per year. Stunting rates models. During the IMCI phase-in period (July 1999 to also became significantly lower in the IMCI districts. June 2000), under-five mortality levels were almost identi- Contextual factors, such as mosquito net use, all favored cal in IMCI and comparison districts, at about 27 deaths the comparison districts. Source: Armstrong Schellenberg and others 2004. presenting to the surveyed IMCI facilities were correctly man- evaluation found that IMCI-trained providers spent 1 minute aged, compared with 16 percent in the comparison facilities. and 26 seconds longer per consultation with under-five chil- When the information on costs and effectiveness are taken dren than untrained providers did. The difference was much together, the cost per child correctly managed is six times less greater when patient load was low but decreased as the number in IMCI districts (US$4.02) than in the comparison districts of patients a provider saw per day increased. This finding sug- (US$25.70) (Bryce and others forthcoming). gests that the system's ability to absorb IMCI depends on cur- Some of the differences in costs might be due to factors rent capacity utilization. In terms of the assessment of quality other than IMCI, so these ratios have to be interpreted with of care in the surveyed facilities, IMCI-trained health workers care. What is clear, however, is that treating children using were shown to provide significantly better care than those who IMCI in Tanzania was no more costly--and probably less had not been trained (Amaral and others 2004; Gouws and costly--than treating children using routine care. At the same others 2004). (See "Lessons about Implementation Success time, it resulted in higher quality of care. To the extent that this and Failure" later in this chapter for additional results on qual- higher quality of care leads to better health outcomes, IMCI is ity of care.) cost-effective--it costs less (or at least no more) and results in These results are important for policy development relating better outcomes (see box 63.1). to child health. Where current caseloads are relatively low, providers spend additional time to provide better child health services, using IMCI as the basis of part of their current activi- MCE in Brazil ties. In this study, the mean number of consultations per Brazil is another MCE site where IMCI is being implemented provider per day was 34, and 95 percent of the providers had in the context of an ambitious family health program (FHP), caseloads of fewer than 50 patients per day. If this finding is which is supported by the Ministry of Health and the World representative of the rest of Brazil, it would be possible to Bank and based at first-level government facilities. IMCI introduce IMCI relatively easily throughout the country with- implementation started in 1996 and is moving ahead in the out encountering capacity constraints in terms of provider whole country, particularly in the northeast regions. IMCI time. In areas with high patient loads, however, it would be training is targeted at FHP team members. important to explore whether it is possible to maintain high The MCE in Brazil was carried out in four states, all in quality of care under IMCI in those areas and what the alter- northeast Brazil, the poorest area of the country. In total, 23 natives should be (Adam and others, forthcoming). municipalities with both FHP and IMCI were compared with matched municipalities with FHP but without IMCI. Early results from one component of the evaluation--the time and Summary of MCE Results motion study--provide useful insights. After controlling for The available results from the MCE so far show that costs possible confounding factors using regression analysis, the of child health care in Tanzania were comparable or lower in Integrated Management of the Sick Child | 1183 districts with IMCI than with routine case management. Proportion of children (percent) Quality of care was higher, and a 13 percent reduction in mor- 100 First source of care from an appropriate provider tality was also found in the IMCI districts in the study period. 90 Antibiotics for probable pneumonia This finding strongly suggests that IMCI is a cost-effective Antimalarials for fever 80 intervention compared with routine care, as it costs less and is 70 more effective in saving children's lives. In Brazil, the MCE study also showed improved quality of 60 care for children under age five after health workers were trained 50 in IMCI. The results also showed that staff time constraints were 40 unlikely to limit the application of IMCI, because, in settings 30 with low caseloads, health workers could be expected to use the available excess capacity to provide better care for children 20 under age five. Assuming that primary health facilities will 10 experience savings on drug costs similar to those observed in 0 Tanzania and previous modeling studies, one could also argue Most Very Poor Less Least poor poor poor poor that IMCI is a good value for money in the Brazilian setting. Socioeconomic status Source: Victora and others 2003, based on data taken from Schellenberg and Equity Issues others 2003. Some of the equity implications related to care seeking and Figure 63.3 Proportions of Children, by Socioeconomic Quintiles, treatment have also been evaluated in the four rural districts Who Were Brought to an Appropriate Provider and Who Received included in the IMCI evaluation in Tanzania (Schellenberg and Correct Care, in Rural Tanzania others 2003; Victora and others 2003). The MCE analysis found no association between sex and any indicator of morbidity, care vary markedly, even within an apparently homogeneous rural seeking, case management, or compliance with treatment or population. This observation agrees with data on mortality and follow-up instructions, suggesting that mothers and health nutritional status inequalities for Tanzania (Schellenberg and workers treat boys and girls similarly (Schellenberg and others others 2003). There is no evidence yet showing whether the 2003). This finding is in accordance with that of Gwatkin and introduction of IMCI has reduced or increased this type of others (2000). Similarly, there were no statistically significant inequality. associations between socioeconomic status and reported prevalence of fever, diarrhea, severe diarrhea, or pneumonia. However, hospital admissions were almost half as common in IMPLEMENTATION OF PROGRAMS: the lowest socioeconomic status quintile as in the highest LESSONS OF EXPERIENCE (t-test: p 0.0093), suggesting that referral care is more readily accessible to the wealthy. The IMCI strategy was thoroughly evaluated from its onset. Positive associations were observed between socioeconomic Only two years after the first health worker training course took status and care seeking from an appropriate provider (a quali- place, the MCE was launched (Bryce and others 2004). MCE fied health worker practicing allopathic medicine) for fever researchers visited 12 countries and carried out in-depth stud- without cough or diarrhea, for care seeking for episodes per- ies in five of those. The MCE and the more recent Analytic ceived as severe, and for using an appropriate provider as the Review of IMCI (DFID and others 2003), which included vis- first source of care. This finding is illustrated in figure 63.3, its to six countries, provide the background for this section. which puts households into five wealth categories. The poorest group was at least 25 percent less likely to have sought care than the least poor. Among the children who sought care, antibiotic Institutions and Programs use for probable pneumonia was less than half as common IMCI introduction was highly successful. As of December among the poorest than among the least poor. Also, children in 2002, WHO's global monitoring team reported that IMCI had the lowest socioeconomic group were half as likely to have been been introduced in 109 countries (see http://www.who.int/ given antimalarials as those in the highest category (t-test: child-adolescent-health/overview/child_health/map12_ p 0.0001). 02.jpg). Twelve countries were included in the introduction These findings suggest that, although the prevalence of dis- phase, in which the strategy was officially endorsed, a national ease (self-reported) does not differ by socioeconomic status, IMCI coordination group was appointed, and key ministry care seeking and the probability of receiving appropriate care of health staff members were trained in the IMCI clinical 1184 | Disease Control Priorities in Developing Countries | Cesar G. Victora, Taghreed Adam, Jennifer Bryce, and others guidelines. Another 50 countries were in the early implementa- gram is highly regarded, and trainees are pleased with its logi- tion phase, which included development of a national plan, cal, consistent approach to child health problems. Innovative selection of initial districts for implementation, adaptation of clinical skills, such as the use of palmar pallor to diagnose ane- the IMCI clinical guidelines and materials, training of course mia and the use of breathing rate for pneumonia, are often facilitators, and planning at the district level. Finally, 47 coun- praised. Nutritional counseling, an area in which most health tries were in the expansion phase, which included scaling up workers receive little formal training in school, is also greatly IMCI activities in districts already covered and expanding to appreciated. When asked about the limitations of IMCI, health cover additional districts (WHO and UNICEF 1999). workers often mention the increased time required for a con- The fact that a country has adopted the IMCI strategy, how- sultation and the difficulty of following the IMCI guidelines ever, does not mean that a high population coverage has been when there is a high patient load. reached. The best available estimates of IMCI coverage are pro- Several studies have shown that health workers trained in vided by the percentage of health workers who underwent IMCI do perform better than those not trained. Health facility IMCI training and are managing sick children. For example, surveys carried out in Tanzania (Schellenberg and others the Brazil MCE (Amaral and others 2004) has shown that 2003), Brazil (Amaral and others 2004), and Uganda show that IMCI is being implemented in all 27 states, but in some IMCI training substantially improves health worker perform- of those states only a few health professionals were trained. In ance in assessing and managing sick children, and in counsel- the three states selected for the evaluation because of report- ing their caretakers. Box 63.2 summarizes MCE findings on edly strong IMCI implementation, there was at least one IMCI- antibiotic prescribing patterns, a critical area for managing sick trained health worker in 239 out of 443 municipalities children (Gouws and others 2004). (54.0 percent), but only 23 municipalities (5.2 percent) had at Important constraints to IMCI implementation were also least 50 percent of health workers trained after three years. In identified through visits to 17 countries by the MCE and Peru, also a leading country in IMCI training, approximately 10 Analytic Review teams. Using the framework developed by the percent of all doctors and nurses providing child care were Commission on Macroeconomics and Health (Hanson and oth- trained after seven years of IMCI implementation (Huicho and ers 2001),the teams described shortcomings in three areas: com- others 2005). Therefore, levels of training coverage in most munity and household issues, health service delivery issues, and countries appear to be low. issues related to health sector policy and strategic management. Community and Household Issues. Coverage levels for effec- Lessons about Implementation Success and Failure tive interventions to improve child survival are remarkably low Both the MCE (Bryce and others 2004) and the Analytic in most developing countries. A review of the 42 countries that Review of IMCI that were carried out in 2002­3 (DFID and account for 90 percent of global child deaths showed that only others 2003) confirm that IMCI has been highly successful in two out of nine key interventions reached more than half of all motivating managers and health workers. The training pro- children (Bryce and others 2003). This finding agrees with Box 63.2 Improving the Use of Antimicrobials through IMCI Case-Management: Findings from the MCE Antimicrobial drugs, including antibiotics and antimalari- IMCI are significantly more likely than those receiving care als, are an essential child survival intervention. Prompt and from workers not yet trained in IMCI to receive correct pre- correct provision of drugs to children under age five who scriptions for antimicrobial drugs, to receive the first dose need them can save lives. Ensuring that these drugs are not of the drug before leaving the health facility, to have their prescribed unnecessarily and that those who receive them caregivers advised on how to administer the drug, and to complete the full course can slow the development of have caregivers who are able to describe correctly how to antimicrobial resistance. Analysis of data collected through give the drug at home as they leave the health facility. IMCI observation-based surveys at randomly selected first-level training is an effective intervention to improve the rational health facilities in Brazil, Tanzania, and Uganda shows that use of antimicrobial drugs for sick children visiting first- children receiving care from health workers trained in level health facilities in low- and middle-income countries. Source: Gouws and others 2004. Integrated Management of the Sick Child | 1185 those of the MCE, showing that the third component of current rate of training, several decades will be needed before IMCI--improving family and community practices--was full coverage is reached (Huicho and others 2005). Similar poorly implemented. At the global level, UNICEF was primarily results were observed in Uganda (J. Nsungwa-Sabiti, personal in charge of developing this component (see http://www. communication). childinfo.org/eddb/imci/practices.htm), and at the country Staff turnover is also a major problem. In Peru, between level, UNICEF often acted through nongovernmental organi- 1996 and 2001, 43 percent of IMCI-trained health workers had zations (NGOs). Coverage with these community-based pro- already been rotated since their training (Huicho and others grams tended to be patchy. In Peru (Huicho and others 2005) 2005). In Tanzania, where staffing patterns appear to be quite and Tanzania, the districts that were selected for implementa- stable in comparison with the situation in other countries, tion of the community component were not the same as those 23 percent of trained staff had moved within three years of prioritized for health worker training, which were chosen by initial training (C. Mbuyia, personal communication). the ministry of health with WHO support. This precluded any Problems with turnover were also observed in Bolivia, Brazil, possible synergy at the district level between improved quality and Niger. These health workers did not necessarily leave gov- of care in health facilities and community interventions, ernment employment, but high rotation means that IMCI may including those aimed at improving care seeking and compli- not be continually delivered to the same target population over ance with health workers' advice. time. All the countries that were visited, however, have a number Another relevant issue mentioned in several countries was of programs and projects that deliver child survival interven- that of low staff motivation, which was often associated with tions at the community level. Many of these interventions are low salary levels. In Uganda, the performance of health work- part of the key IMCI family practices, but they are being deliv- ers fell dramatically in 2001 after the government discontinued ered in an uncoordinated manner by national, international, cost-sharing schemes that were used to supplement drug sup- and nongovernmental organizations in limited geographical plies and health worker salaries at the facility level (Burnham areas. The low population coverage of these projects makes it and others 2004). In Cambodia and Tanzania, salary levels are unlikely that they will ever result in a substantial effect on a so low that health workers need other sources of income to larger scale. The notable exceptions are the EPI programs, maintain their families. Issues related to human resources are which, despite some recent evidence of falling coverage (Bryce addressed in greater detail in chapter 71. and others 2003), still reach the vast majority of children in Poor supervision was a major issue in all countries that developing countries. were visited. IMCI recommends regular supervisory visits that On the basis of the experience obtained in these countries, should include systematic observation and feedback on case it appears that those key family practices that are most likely to management. In Peru, the average number of supervisory vis- be synergistic with facility-based IMCI--improved care seek- its was 0.19 per facility per year (Huicho and others 2005). In ing, home management of disease, and compliance with health Bangladesh, a baseline (pre-IMCI) health facility survey con- worker advice--are among those least likely to be supported by ducted in 2000 found that none of the facilities in the study existing programs. Existing programs seem to favor biological area had received a supervisory visit, including observation of interventions such as vaccines, micronutrient supplementa- case management, within the previous six months (S. E. tion, and insecticide-treated materials. Arifeen, personal communication). Common reasons given by The present criticism of community IMCI should not be health workers for erratic supervision activities are shortages of extrapolated to community-level child health interventions in vehicles, fuel, and staff members. general, which can often be highly successful. These interven- Problems with referral were also common. The Urgent tions are covered in chapter 56. Referral category in figure 63.1 requires immediate referral to a hospital. In several countries (Bangladesh, Cambodia, Niger, Health Service Delivery Issues. Given the difficulties in Uganda, and Tanzania), it was reported that children in this implementing the community component, IMCI was largely category are often not taken to a hospital because of distance or restricted in nearly all countries to training health workers in lack of funds for travel and hospitalization-related expenses. the improved management of care for young children. Even For example, in a Tanzania survey, only 5 of 13 children there, some difficulties were apparent. who had been referred were actually taken to a hospital In countries such as Peru, Brazil, and Uganda, after an ini- (Schellenberg and others 2003). Also, in some countries hospi- tial sharp increase in the number of health workers who were tal staff members who had not been trained in IMCI were trained, budgetary and other restrictions led to a decrease in reluctant to admit children with danger signs identified the number of training courses being offered. In Peru, about through the IMCI algorithm. This situation highlights the need 10 percent of all eligible health workers in the public sector for reinforcing training of referral-level health workers using were trained after seven years of implementation. At the IMCI guidelines. 1186 | Disease Control Priorities in Developing Countries | Cesar G. Victora, Taghreed Adam, Jennifer Bryce, and others Another important limitation, observed in Bangladesh, action to prevent nurses from being trained in using antibiotics Cambodia, Niger, and Uganda, is the low use of public sector for life-threatening conditions. This obstruction succeeded health care for a variety of reasons (accessibility, official or despite an MCE health facility survey that showed that IMCI- under-the-table user fees, perceived poor quality, lack of drugs, trained nurses performed as well as doctors in managing sick and so on). For example, using Ministry of Health documents in children (Amaral and others 2004). In Morocco, IMCI-trained Niger, the authors estimated that the average annual number of nurses are also unable to prescribe antibiotics because of cen- attendances by children under age five was 0.5. In Bangladesh, tral regulations. only 8 percent of children who were ill were taken to a qualified A particular challenge came to light when the MCE team provider (S. E. Arifeen, personal communication). In the pres- visited Cambodia and Niger and the Analytic Review team ence of such low utilization rates, it is unlikely that health work- visited Mali. These countries have high levels of under-five er training can have an effect on mortality rates, unless simulta- mortality and thus the greatest need for IMCI. They also have neous community activities improve care-seeking practices. weak health systems and low utilization rates and are therefore Although equipment and vaccines that are needed for IMCI having difficulty implementing IMCI successfully. Just as for delivery were available in most countries visited, availability of individuals, the inverse care law--which suggests that those drug supplies varied from country to country. Shortages were who most need high quality care are the least likely to get it-- reported in Cambodia and Zambia, and other countries, such seems to also apply to countries (Hart 1971). as Peru and Tanzania, reported that essential IMCI drugs were However, there is a possibility that IMCI (or other mostly available. approaches to managing sick children) might help strengthen selected health system functions through specific approaches, Health Sector Policy and Strategic Management Issues. as the Analytic Review team observed in regard to drug and Several of the problems described in the preceding section are commodity availability, service management, and health directly related to health sector issues. In addition, issues related worker motivation through IMCI in the Arab Republic of to higher-level policy and management may also represent con- Egypt (DFID and others 2003). straints to successful IMCI implementation. In some countries, IMCI was not fully institutionalized at national or subnational levels. For example, a national coordi- Implications for Health System Development nator was not appointed or was appointed on a part-time basis. The first component of IMCI, which involves training of health In Peru, IMCI was implemented side by side with CDD and care workers, has been implemented in many developing coun- ARI programs, which it was expected to replace, and in several tries and has resulted in important improvements in the quali- districts the ARI coordinator's tasks were expanded to also ty of care delivered to children in first-level facilities in limited encompass IMCI. In several countries, IMCI activities did not geographic areas. The potential population-level effect of IMCI have a separate budget line, or they were not included in dis- case-management training has not been realized, however, for trict health plans, or neither. A report on the Analytic Review of three reasons: IMCI (DFID and others 2003, 39) states, "IMCI was generally introduced as a strategy, not as a program. If this was not a bar- · Sufficient resources were not available for full rier in the pilot phase, it seemed to generate problems for rapid implementation. scaling up. In five of the six Analytic Review countries, IMCI · Few health systems in low-income countries are capable of focal persons did not have the rank or the responsibility of providing the policy, personnel, and managerial support previous disease specific program managers within their needed to expand and sustain high levels of IMCI training Ministry of Health, and IMCI did not have a budget line and coverage. a strong management structure." The report also argued that · At the time of this writing, not one country had succeeded decentralization, as part of health sector reform, reduced man- in mounting a behavior-change program capable of agerial capacity at the central level and had, at least in the short improving care seeking, home management of illness, and term, a negative effect on IMCI implementation. nutrition-related practices to coverage levels that will result Conflict between IMCI guidelines and existing policies in population-level changes in service utilization or health and regulations was present in some countries, particularly status. the former Soviet republics of Kazakhstan and the Kyrgyz Republic, where policies for hospital admission--requiring, for One implication for health system development is that sup- example, that all children with diarrhea be hospitalized--were port should be continued and expanded for integrated case in conflict with IMCI guidelines. Another regrettable example management in first-level facilities as an essential component comes from Brazil, where both doctors and nurses were being of an effective child survival strategy. A second implication, trained in IMCI until medical associations threatened legal however, is that greatly expanded efforts must be directed Integrated Management of the Sick Child | 1187 simultaneously to the development of new and innovative and ARI programs. In fact, at least in theory, the efficiency gains approaches to strengthening health systems and to reaching represented by the integration promoted by IMCI should make families and communities with known and affordable child it easier for developing countries to implement as the key child survival interventions. health strategy, so a return to vertical programs is not the An important distinction can be made between interven- answer. tions and delivery strategies (Bryce and others 2003). The same Given these difficulties, however, there may be a tempta- intervention (vitamin A capsules, perhaps) can be delivered tion to bypass health services altogether in the poorest coun- through different strategies--for example, to children attend- tries by promoting the delivery of child health interventions ing health facilities, on National Immunization Days, or directly directly to families and households. There are successful at the household level through community networks. In spite examples of such community delivery schemes--for example, of its community component--which in most countries has projects dispensing antimalarials (Pagnoni and others 1997; not been operational anyway--IMCI, as implemented to mid Kidane and Morrow 2000) and antibiotics for pneumonia 2004, relies on health facilities as its key delivery strategy. (Sazawal and Black 1992). This approach may, in fact, be the The first component of the IMCI strategy--a focus on most viable short-term solution for countries with weak improving health worker skills--was innovative to the extent health services, but it should not be forgotten that most suc- that it provided clear technical guidelines and yet required cess stories represent small-scale pilot projects with strong country-level adaptation. Similar levels of technical clarity and managerial backup. In countries with weak systems, the man- country-level flexibility did not exist for the second and third agerial support for implementing and sustaining high-quality, components of the IMCI strategy, which focused on improving community-based interventions is also likely to be lacking, so health systems to support IMCI and improving family and it may be naive to assume that such programs will have the community practices. Within these two components, the IMCI effects that health services have failed to deliver. Also, just as strategy has been criticized for attempting to become a uniform first-level health facility care depends on referral services for global strategy, with guidelines for implementation that do backup, community delivery schemes will require operational not allow room for country-level modifications, especially the first-level health facilities to handle complications and treat- incremental approaches to implementation needed by weak ment failures. health systems (Bryce and others 2003). The first component of There is no substitute for strengthening health systems in IMCI can serve as a model for the types of development work the poorest countries. In the long run, strengthening these sys- that must now move forward in the health systems and family tems will be the key intervention for reducing child mortality practice areas; however, in these areas, key decisions about how as well as for promoting healthy growth and development. best to deliver interventions will need to be made at the country Delivery strategies that reach communities either directly or level and below. through other mechanisms are needed in the short term, but One example of progress is that WHO and its partners have the long-term goal of improving health services is paramount. now developed a process for assessing country-level opportu- nities and requirements for achieving population-level behavior change in relation to key family practices and for THE RESEARCH AND DEVELOPMENT AGENDA developing feasible and collaborative work plans for effectively implementing child health activities at the community level (A. This section starts by addressing health systems research issues Bartlett, personal communication). related to scaling up child health interventions effectively. Key As the MCE and Analytic Review have shown, IMCI issues on the research agenda are how to make the best possi- requires a functional health system with managerial capacity; ble use of integrated case management in settings where health an ability to train health workers and to keep them on the systems are weak and how to design alternative delivery strate- job; an efficient means of supplying drugs, vaccines, and equip- gies to improve child survival in such settings. ment; and the capacity to maintain regular supervisory activi- Specific issues relate to how to counteract the main con- ties. It also requires appropriate care-seeking practices, leading straints to scaling up IMCI that were described earlier in to a reasonable level of health services utilization by children "Implementation of Programs: Lessons of Experience." For under age five. In most countries, appropriate health services example, research is needed in the following areas: utilization is unlikely to be achieved without strong family and household-level interventions such as those promoted by com- · how to increase utilization in the public sector (and the role munity IMCI. of user fees in this respect) These problems, however, are not specific to IMCI; they · how to develop viable public sector­private sector affect every other delivery strategy that relies heavily on health partnerships facilities, including the predecessors of IMCI, namely the CDD · how to reduce staff turnover 1188 | Disease Control Priorities in Developing Countries | Cesar G. Victora, Taghreed Adam, Jennifer Bryce, and others · how to improve supervision and make it sustainable IMCI implementation has not yet materialized, and country · how to institutionalize IMCI case-management training reports of the barriers to achieving and maintaining high cov- and supervision at the district level. erage levels suggest that effects will not be seen unless IMCI in first-level facilities is buttressed by equally strong or stronger In addition, research is needed on how to strengthen health efforts to develop health system capacity and reach families and systems and improve family and community practices for child communities. In fact, progress in child survival in the late 1990s survival in ways that take into account and build on features of and early 21st century has been slower than in earlier decades the country context. These features include the epidemiologic (Bellagio Study Group on Child Survival 2003), and current profile and the characteristics of the health system. Country- trends suggest that the Millennium Development Goals are level assessments and planning and support for longer-term unlikely to be achieved for most countries unless major new implementation are required to achieve high and equitable investments are made very soon. coverage and population effect in different epidemiologic, The Tanzania results suggest that, in a setting where IMCI health system, and cultural settings (Bellagio Study Group on was implemented in conjunction with health system strength- Child Survival 2003). Innovative research is needed involving ening and where utilization of health facilities is high, an effect country-level investigators and program staff members, with on mortality and nutritional status is likely. However, experi- international partnerships if required, to develop and evaluate ence from other countries showed that reaching high and sus- different combinations of health facility­based and household- tained implementation was difficult. based delivery schemes. In particular, research should address Although IMCI has only partially lived up to initial expecta- how to go to scale with interventions that have been proven tions, it has many positive aspects that must be fostered. A effective in pilot studies. From the costing perspective, research return to isolated vertical programs for child survival will not should address the issue of how to estimate the cost of scaling solve the difficulties faced by scaling up IMCI effectively, and up using different scenarios of resource availability and integration should continue to be a key goal of child survival constraints. strategies in the future. In fact, much of the frustration Monitoring and evaluation are key components of the associated with reported underperformance of IMCI arises required research. Tools must be developed for use at the dis- from the fact that sufficient resources--financial, human, and trict and national levels, and capacity to use them must be organizational--were not planned for or available to support strengthened. its full implementation, either at national or at international levels. The meager training coverage levels observed in most of the MCE countries are clear evidence of insufficient imple- CONCLUSIONS: PROMISES AND PITFALLS mentation, and it is thus not surprising that coverage and effect were also less than expected. IMCI was introduced in the mid 1990s as an ambitious global Renewed and expanded efforts to reduce child mortality strategy that held many promises. Cost-effective vertical inter- should build on the proven effectiveness of IMCI case manage- ventions against the main causes of under-five mortality in the ment in first-level facilities, but they also should incorporate world were integrated into a single, facility-based health worker new knowledge. Country-specific planning is needed to reach training program. The program was accompanied by efforts to families and communities and to build on the existing health improve health systems support for child health care and system to achieve and sustain population-level coverage. to promote key family practices at the community level. Countries with weak health systems will require creative Integration was expected to further improve coverage levels approaches to intervention delivery in the short term at the and the cost-effectiveness of child survival interventions rela- same time that health systems are strengthened as a long-term tive to their delivery through separate vertical programs. strategy. The poorest strata of the population are also the need- IMCI case-management training was repeatedly shown to iest in terms of health care and the hardest to reach. The chal- improve the quality of care delivered in first-level health facili- lenge of improving equity is not unique to IMCI or to child ties, and the costing data reviewed above suggest that it can do so survival; it affects virtually every intervention and delivery at similar or lower costs than those of existing health services. strategy. Unless equity considerations become a key part of pol- IMCI, therefore, is able to deliver better child health care at no icy making and of monitoring outcomes, interventions may increase in costs. widen instead of narrow inequity gaps (Victora and others Nevertheless, community IMCI interventions only reached 2003). meager population coverage in the countries studied, and even A continuing challenge is how to raise and sustain the stand- health worker training was never effectively scaled up in most ing of child survival in the international agenda. The more than countries as a result of health system constraints. The major 10 million annual deaths of children under age five--more effect on child survival that was initially expected as a result of than 20 deaths per minute--represent twice the number of Integrated Management of the Sick Child | 1189 deaths attributable to AIDS, malaria, and tuberculosis com- Bryce J., E. Gouws, T. Adam, R. Black, J. A. Schellenberg, C. Victora, and bined (Black, Morris, and Bryce 2003). Putting child survival J-P. Habicht. Forthcoming. "Improving the Efficiency of Facility-Based Child Health Care: A Success Story from Tanzania." back on the public health and development agenda is an essen- Bryce, J., C. G.Victora, J-P. Habicht, J. P.Vaughan, and R. E. Black. 2004."The tial developmental step in the process of refining country-level Multi-Country Evaluation of the Integrated Management of Childhood and global child health strategies. Only through taking stock of Illness Strategy: Lessons for the Evaluation of Public Health the lessons learned in early IMCI implementation can a flexi- Interventions." American Journal of Public Health 94 (3): 406­15. ble, integrated program be developed that will improve child Burnham, G. M., G. W. Pariyo, E. Galiwango, and F. Wabwire-Mangen. 2004. "Discontinuation of Cost-Sharing in Uganda." Bulletin of the survival in particular and child health in general. World Health Organization 82 (3): 187­95. Claeson, M., and R. J. Waldman. 2000. "The Evolution of Child Health Programmes in Developing Countries: From Targeting Diseases to Targeting People." Bulletin of the World Health Organization 78 (10): ACKNOWLEDGMENTS 1234­45. DFID (U.K. Department for International Development), UNICEF A substantial portion of this chapter is based on early results (United Nations Children's Fund), World Bank, USAID (U.S. Agency from the Multi-Country Evaluation of IMCI Effectiveness, for International Development), and WHO (World Health Cost, and Impact. The MCE is funded by the Bill & Melinda Organization). 2003. The Analytic Review of Integrated Management Gates Foundation and is supported by technical assistance of Childhood Illness Strategy (Final Report). Geneva: World Health Organization. http://www.who.int/child-adolescent-health/ from the World Health Organization and a technical advisory New_Publications/IMCI/ISBN_92_4_159173_0.pdf. group representing global expertise in cost-effectiveness analy- Gouws, E., J. Bryce, J. P. Habicht, J. Amaral, G. Pariyo, J. A. Schellenberg, sis, measurement, research design, and child health. and O. Fontaine. 2004. "Improving Antimicrobial Use among Health Workers in First-Level Facilities: Results from the Multi-Country Evaluation of the Integrated Management of Childhood Illness Strategy." Bulletin of the World Health Organization 82 (7): 509­15. REFERENCES Gove, S. 1997. "Integrated Management of Childhood Illness by Out- patient Health Workers: Technical Basis and Overview." Bulletin of the Adam, T., D. G. Amorim, S. Edwards, J. Amaral, and D. B. Evans. World Health Organization 75 (Suppl. 1): 7­24. Forthcoming. "Capacity Constraints to the Adoption of New Gwatkin, D. R., S. Rustein, K. Johnson, R. P. Pande, and A. Wagstaff. 2000. Interventions: Consultation Time and the Integrated Management of "Socio-Economic Differences in Health, Nutrition, and Population in Childhood Illness in Brazil." Tanzania." World Bank, Washington, DC. http://www.worldbank.org/ Adam, T., D. Bishai, M. Khan, and D. Evans. 2004a. Methods for the Costing poverty/health/data/tanzania/tanzania.pdf. Component of the Multi-Country Evaluation of IMCI. Geneva: World Hanson, K., K. Ranson, V. Oliveira-Cruz, and A. Mills. 2001. "Constraints Health Organization. http://www.who.int/imci-mce. to Scaling up Health Interventions: A Conceptual Framework and Adam, T., F. Manzi, C. Kakundwa, J. Schellenberg, L. Mgalula, D. Empirical Analysis." Working Group 5 Paper 14, WHO Commission de Savigny, and others. 2004b. Analysis Report on the Costs of IMCI in on Macroeconomics and Health, Geneva. http://www.cmhealth.org/ Tanzania. Geneva: World Health Organization. http://www.who.int/ docs/wg5_paper14.pdf. imci-mce. Hart, J. T. 1971. "The Inverse Care Law." Lancet 1 (7696): 405­12. Amaral, J., E. Gouws, J. Bryce, A. J. M. Leite, A. L. A. Cunha, and C. G. Huicho, L., M. Dávila, M. Campos, C. Drasbeck, J. Bryce, and C. G. Victora. Victora. 2004. "Effect of Integrated Management of Childhood Illness 2005. "Scaling up IMCI to the National Level: Achievements and (IMCI) on Health Worker Performance in Northeast Brazil." Cadernos Challenges in Peru." Health Policy and Planning 20 (1): 14­24. de Saude Publica 2004. 20 (suppl. 2): 209­12. Jones, G., R. Steketee, R. E. Black, Z. A. Bhutta, S. S. Morris, and the Armstrong Schellenberg, J. R. M., T. Adam, H. Mshinda, H. Masanja, Bellagio Study Group on Child Survival. 2003. "How Many Child G. Kabadi, O. Mukasa, and others. 2004. "Effectiveness and Costs of Deaths Can We Prevent This Year?" Lancet 62 (9377): 65­71. Facility-Based Integrated Management of Childhood Illness (IMCI) in Tanzania." Lancet 364 (9445):1583­94. Khan, M. M., S. Ahmed, and K. K. Saha. 2000. "Implementing IMCI in a Developing Country: Estimating the Need for Additional Health Bellagio Study Group on Child Survival. 2003. "Knowledge into Action for Workers in Bangladesh." Human Resources for Health Development Child Survival." Lancet 362 (9380): 323­27. Journal 4: 73­82. Black, R. E., S. S. Morris, and J. Bryce. 2003. "Where and Why Are Khan, M. M., K. K. Saha, and S. Ahmed. 2002. "Adopting Integrated 10 Million Children Dying Every Year?" Lancet 361 (9376): 2226­34. Management of Childhood Illness Module at Local Level in Boulanger, L. L., L. A. Lee, and A. Odhacha. 1999. "Treatment in Kenyan Bangladesh: Implications for Recurrent Costs." Journal of Health and Rural Health Facilities: Projected Drug Costs Using the WHO- Population Nutrition 20 (1): 42­50. UNICEF Integrated Management of Childhood Illness (IMCI) Kidane, G., and R. H. Morrow. 2000. "Teaching Mothers to Provide Home Guidelines." Bulletin of the World Health Organization 77(10): Treatment of Malaria in Tigray, Ethiopia: A Randomised Trial." Lancet 852­58. 356 (9229): 550­55. Bryce, J., S. Arifeen, G. Pariyo, C. F. Lanata, D. Gwatkin, J. P. Habicht, and Mathers, C. D., C. J. L. Murray, and A. D. Lopez. 2006. "The Burden of the Multi-Country Evaluation of IMCI Study Group. 2003."Reducing Disease and Mortality by Condition: Data, Methods, and Results for Child Mortality: Can Public Health Deliver?" Lancet 362 (9378): the Year 2001." In Global Burden of Disease and Risk Factors, eds. Alan 159­64. D. Lopez, Colin D. Mathers, Majid Ezzati, Dean T. Jamison, and Bryce, J. C., Boschi-Pinto, K. Shibuya, R. E. Black, and the WHO Child Christopher J. L. Murray. New York: Oxford University Press. Health Epidemiology Reference Group. 2005. "WHO Estimates of the Pagnoni, F., N. Convelbo, J. Tiendrebeogo, S. Cousens, and F. Esposito. Causes of Death in Children." Lancet 365 (9465): 1147­52. 1997. "A Community-Based Programme to Provide Prompt and 1190 | Disease Control Priorities in Developing Countries | Cesar G. Victora, Taghreed Adam, Jennifer Bryce, and others Adequate Treatment of Presumptive Malaria in Children." Transactions Mortality: More of the Same Is Not Enough." Lancet 362 (9379): of the Royal Society of Tropical Medicine and Hygiene 91 (5): 512­17. 233­41. Sazawal, S., and R. E. Black. 1992. "Meta-Analysis of Intervention Trials on Wammanda, R. D., C. L. Ejembi, and T. Iorliam. 2003. "Drug Treatment Case-Management of Pneumonia in Community Settings." Lancet 340 Costs: Projected Impact of Using the Integrated Management of (8818): 528­33. Childhood Illnesses." Tropical Doctor 33 (2): 86­88. Schellenberg, J. A., C. G. Victora, A. Mushi, D. de Savigny, D. Schellenberg, WHO (World Health Organization) and UNICEF (United Nations H. Mshinda, and J. Bryce. 2003. "Inequities among the Very Poor: Children's Fund). 1999. IMCI Planning Guide: Gaining Experience with Health Care for Children in Rural Southern Tanzania." Lancet 361 the IMCI Strategy in a Country. WHO/CHS/CAH/99.1. Geneva: WHO. (9357): 561­66. http://www.who.int/child-adolescent-health/New_Publications/ Tanzania IMCI Multi-Country Evaluation Health Facility Survey Study IMCI/WHO_CHS_CAH_99.1.pdf. Group. 2004. "Health Care for Under-Fives in Rural Tanzania: Effect of ------. 2001. Model Chapter for Textbooks: Integrated Management of Integrated Management of Childhood Illness on Observed Quality of Childhood Illness. WHO/CAH/00.40. Geneva: WHO. Care." Health Policy and Planning 19 (1): 1­10. World Bank. 1993. World Development Report 1993: Investing in Health. Tulloch, J. 1999. "Integrated Approach to Child Health in Developing New York: Oxford University Press. Countries." Lancet 354 (Suppl. 2): SII16­20. Victora, C. G., A. Wagstaff, J. A. Schellenberg, D. Gwatkin, M. Claeson, and J. P. Habicht. 2003. "Applying an Equity Lens to Child Health and Integrated Management of the Sick Child | 1191 Chapter 64 General Primary Care Stephen Tollman, Jane Doherty, and Jo-Ann Mulligan Primary health care has always been a feature of health care THE SCOPE OF GENERAL PRIMARY systems and--from a modern perspective--involves four inter- CARE PRACTICE related aspects: a set of activities, a level of care, a strategy for organizing health care, and a philosophy that permeates health General primary care can be defined as the immediate--and care provision. At full stretch, then, the "primary health care often continuing--medical and health management of a child, approach" can affect the configuration and focus of the entire adult, or family when the patient first presents to the formal health system and extend to the development of communities health system. In low- and middle-income countries, such care (Vuori 1985). is often provided from publicly funded health posts and health It is not always easy to see how information on the cost- centers by nurses or other midlevel health workers, with med- effectiveness of individual interventions--the focus of part 2 ical doctors expected to play a support, training, and referral of this volume--contributes to the achievement of the wider role. features of primary health care. Indeed, Starfield (1998) points out that the importance of particular services or interventions is overrated, in part because of limited appreciation of the Comprehensive versus Selective Care "essential and unique functions" of primary care. These func- The 1978 Alma Ata declaration on primary health care (WHO tions are mutually reinforcing and include first-contact care; and UNICEF 1978) was informed by a number of well- continuity in care (in Starfield's words, "person-focused over described, small-scale health and development efforts in a time"); comprehensiveness of available services; and coordina- range of settings (Newell 1975; Tollman 1991). It focused tion with specialized services and other levels of care. The international health care efforts on low-cost, potentially high- functions point to the centrality of how primary care is impact interventions, both medical and social, at both the organized and delivered, something with which the cost- primary and community levels. In particular, the declaration effectiveness approach has hitherto not been greatly emphasized the importance of health promotion and commu- concerned. This chapter attempts to identify some of the com- nity development through, for example, the supply of water mon ground between the primary health care approach and and sanitation and the involvement of communities in decision the cost-effectiveness approach. We propose that resource making. constraints simultaneously require (a) the targeting of services Before long, the extent of resources and capacities needed to toward burdens of disease that are amenable to highly cost- implement such comprehensive activities led to the emergence effective interventions and (b) the general strengthening of of the concept of "selective primary health care" (Walsh and the health system, particularly at the primary care and district Warren 1979). This concept advocated a focus on a limited levels. number of priority conditions, such as children's health and 1193 particular tropical diseases. Selective interventions, often cen- of activities that is appropriate at the primary care level (see trally planned, and managed and operated by a dedicated staff, table 64.1). In many ways, this consensus existed before the pub- were intended as entry points into the health care system. Such lication of Investing in Health (World Bank 1993), but econom- an approach, when fully operationalized as a "vertical pro- ic evaluation has subjected conventional wisdom to empirical gram," proved especially useful in implementing control and validation and considerations of affordability; it has also scruti- eradication campaigns and in dealing with epidemics following nized the components of particular interventions (frequency of natural disasters (Unger, de Paepe, and Green 2003; World prenatal health care visits, for example). This evaluation has Bank 1994). served the purpose of providing a basis for consensus that is The selective approach attracted strong criticism for not more convincing to key constituencies, such as national treasur- acknowledging that primary care practice needs to take ies and donors. Consequently, several low- and middle-income account of the range of conditions that present, many of countries have delineated minimum or essential packages for which--by definition--were excluded from the selective their own national contexts. agenda (World Bank 1993). In addition, administrators The notion of such packages is not without controversy, responsible for vertical programs tended to have little contact however. Whereas some criticisms relate to methodological with local health officers and seldom coordinated well with issues--and hence to the advisability of generalizing the find- other vertical efforts, leading to duplication of training, super- ings of local studies and using them to establish priority serv- vision, and logistics management (Briggs, Capdegelle, and ices on a global scale--others relate to the disease-oriented Garner 2003). Other inefficiencies resulted from the need for basis of essential packages (Doherty and Govender 2004). The specialized personnel, which led to growing numbers of dedi- authors of essential packages are not necessarily proponents of cated staff members as well as excessive demand on service verticalization (in some instances, they have argued strongly users' time if they needed multiple services. Verticalization can for the cost-effectiveness that a horizontally well-integrated also lead to competition between programs, favoring of some service can achieve). However, the reality of activity by interna- conditions at the expense of others, weaknesses in continuity tional agencies, donors, and even governments means that, in of care, disruption of routine health services, and erosion of many instances, the implementation of narrowly defined pro- country-level delivery capacity (Coker, Atun, and McKee 2004). grams continues to be favored (with HAART--highly active Concerns about the appropriateness of individual vertical antiretroviral therapy for the treatment of HIV/AIDS--having programs were matched at that time by concerns about the the potential to follow suit). Continuing verticalization also continuing misallocation of resources toward expensive, cost- reflects the resource constraints faced by countries that are ineffective care, culminating in the publication of Investing in unable to mount a comprehensive set of services and suggests Health: World Development Report 1993 (World Bank 1993). that services in developing countries are chronically under- The report identified highly cost-effective interventions tar- funded (Jha and Mills 2002). geted at the major causes of the prevailing burden of disease in The disease-based construction of packages may have rein- low- and middle-income settings. These interventions were forced this tendency to verticalization, as may have incentives grouped into a "minimum package of health services," which, for program managers to monitor program activities in terms of it was argued, governments and donors should prioritize for their effect on specific diseases. This factor accounts partly for funding (World Bank 1993). The delivery vehicle for most of the suspicion that sometimes greets efforts at cost-effectiveness- these services was the primary care system. based planning, and it signals the need to continually promote The definition and costing of intervention packages has the effective integration of activities at the health facility level. evolved over time, and this evolution is reflected in the World Bank's Better Health in Africa (1994), the World Health Organization's World Health Report 2000 (WHO 2000), and General Primary Care at a Key Interface the report of Working Group 5 of the Commission on Although the primary care level constitutes the first point of Macroeconomics and Health (Jha and Mills 2002). Although patient or family contact, it is also a critical base for extending some services have been added and others have been described care to communities and vulnerable groups. These outreach more specifically (partly informed by new research), the essen- services may focus on individual preventive measures (such as tial components of the package have remained remarkably con- immunization, vitamin A, or oral rehydration therapy) or com- stant. Another World Health Organization (WHO) publication munitywide health-promoting efforts (such as education on (2002), World Health Report 2002, which focused solely on inter- child nutrition or adult diets and exercise; see chapter 56). ventions against risk factors, corroborated several of the inter- Increasingly, home-based care for chronic conditions, such as ventions in the package (Doherty and Govender 2004; see annex HIV and AIDS and poststroke rehabilitation, can be expected to 64.A). Thus, a broad consensus appears to exist on the nucleus feature in outreach services. These services depend substantially 1194 | Disease Control Priorities in Developing Countries | Stephen Tollman, Jane Doherty, and Jo-Ann Mulligan Table 64.1 Selected Sets of Interventions Used in the Cost that links, on the one hand, ambulatory care with hospital and Analysis of the Commission on Macroeconomics specialty services and, on the other, individual clinical care with and Health, 2002 community-wide or population-wide health, nutrition, and family planning programs. Acting as the fulcrum of a compre- Disease area Nature of interventions hensive care and support system, development of general Maternity-related Prenatal care primary care requires that local management teams plan serv- interventions Treatment of complications during pregnancy ices for their defined catchment communities (Jha and Mills Skilled birth attendance 2002)--recognizing that catchment populations can be more Emergency obstetric care difficult to define in urban (particularly high-density) settings. Postpartum care (including family planning) Delineating the community for which a local health system is Childhood disease­ Vaccinations (Bacillus Calmette- responsible--and thereby making explicit the population related interventions Guérin, oral polio vaccine, diphtheria- under the care of providers--makes it feasible to undertake (immunization) pertussis-tetanus, measles, hepatitis B, ongoing monitoring and evaluation of the performance of Haemophilus influenzae type B) local primary care services (quality of care and extent of cover- Childhood disease­ Treatment of various conditions (acute age, for example) and contributes to the assessment of health related interventions respiratory infections, diarrhea, causes of fever, systems more broadly. (treatment of childhood malnutrition, anemia), now increasingly Given the pivotal position of general primary care, distin- illnesses) combined as integrated management of childhood illness guishing sharply the cutoff between activities that occur at this level and those that occur elsewhere is difficult. The balance of Malaria prevention Insecticide-treated nets services provided at the primary level rather than other health Residual indoor spraying service levels is, in fact, a moving target, affected by an array of Malaria treatment Treatment for malaria factors. For example, diagnostic and technological innovation Tuberculosis treatment Directly observed short-course treatment for can influence substantially the level at which interventions are smear-positive patients delivered. In many middle-income countries, great potential Directly observed short-course treatment for exists to move elements of surgical, psychiatric, and medical smear-negative patients care upstream to the primary care level, provided that the nec- essary competencies, equipment, and technical or managerial HIV/AIDS prevention Youth-focused interventions support exist. However, the converse also holds: no matter how Interventions working with sex workers and clients appropriate in theory, if service delivery is ineffective, the downstream momentum to district and secondary hospital Condom social marketing and distribution levels (and into the private sector) is almost unstoppable, with Workplace interventions serious implications for the accessibility of care. In practice, Strengthening of blood transfusion systems therefore, factors such as geographic and financial inaccessibil- Voluntary counseling and testing ity, limited resources, poor capacity, and erratic drug supply Prevention of mother-to-child transmission and faulty equipment often mean that the services offered at Mass media campaigns the primary care level are disappointingly limited in their Treatment for sexually transmitted infections range, coverage, and effect. HIV/AIDS care Palliative care Clinical management of opportunistic illnesses "Fitting" of Interventions to the Health System Prevention of opportunistic illnesses Recognition is growing that focusing simply on selecting cost- Home-based care effective interventions as the basis for services development HIV/AIDS treatment Provision of highly active antiretroviral therapy is inappropriate. Paying close attention to the qualities of the Source: Jha and Mills 2002. delivery system that are required to support the introduction Note: Smoking cessation interventions, although considered a priority, were not included in this cost analysis because it was assumed that they would be financed by tobacco taxes. of these interventions and, in time, provide support to their scale-up is essential. The Multi-Country Evaluation of IMCI on community support and mechanisms for identifying, train- Study Group (Bryce and others 2003, 159) highlights the ing, and supporting village or community health workers. "importance of separating biological or behavioural interven- At the same time, primary care facilities mediate patient tions from the delivery systems required to put them in place, access to hospital care, particularly at the district level. General and the need to tailor delivery strategies to the stage of health- primary care, as a level of care, is thus located at a key interface system development." General Primary Care | 1195 Thus, seeking and evaluating the goodness of fit between HIV and AIDS, and sexually transmitted infections; malaria interventions with the potential to be highly cost-effective and management; management of hypertension, other cardiovas- the health system responsible for their effective delivery are cular risk factors, and--increasingly--stroke and cardio- critical. A major challenge in focusing on the primary care level vascular disease; and mental illness and substance abuse. In is to establish the most effective combinations or clusters of all cases, a systematic approach must be taken to establish interventions that can target multiple conditions and risk fac- explicit criteria, guidelines, or regulations regarding the appro- tors affecting key community groups (children, women, and priate treatment level (primary, secondary, and so on) for key older adults, for example) and that are appropriately adapted conditions and the interventions suitable for these conditions to local epidemiologic, economic, and sociocultural contexts. at different levels of severity, according to a country's stage Clustering interventions appears to be a pragmatic approach of health system development. Broadly categorizing countries that achieves a degree of comprehensiveness while at the same according to their income status and the general features of time acknowledging resource constraints. It also provides their health system is too blunt an approach to allow the opportunities to intensify training, improve the quality of care detailed mapping of interventions to the health system capa- provided, and assess community health impact (see chapter bility that is required, and it fails to recognize the wide dispar- 63). If clusters, such as integrated management of infant and ities in health system organization and effectiveness that exist childhood illness (IMCI) or reproductive health services, can within national boundaries. be fully integrated into health service planning, management, and operations, the health system will be provided with some focus while the shortcomings of vertical disease programs can be potentially avoided. Thus, clustering may allow for broader THE EFFECTIVENESS OF GENERAL PRIMARY CARE horizontal strengthening of local health system inputs. As stated earlier, persistent efforts to achieve such integration are Whereas it can be argued that highly cost-effective interven- essential, given that, as expressed by the Bellagio Study Group tions deserve to be implemented, no matter the level for on Child Survival (2003, 324), "in today's environment of dis- which they are designed, unique reasons exist for giving prior- ease-specific initiatives, cross-disease planning, implementa- ity to those based at the primary level. As implied earlier, these tion, and monitoring are hard to establish and maintain." reasons relate to the extent of the burden of disease that is However, clustering does not entirely address concerns that potentially avertable through primary-level care (the health interventions based solely on cost-effectiveness assessments, effect), the welfare benefits that accrue to households spared which seldom examine indirect and nonhealth benefits the experience of disease (the nonhealth effect), and the poten- (Doherty and Govender 2004), might deplete some of primary tial to provide widely accessible services (an equity effect based care's unique features, including responsiveness to the on degree of need). expressed needs of local communities. Indeed, Gilson (2003,) Unfortunately, although many small-scale projects and writes that "future analysis and policy development must assessments of single interventions have been able to measure recognise that health systems are complex socio-political insti- such effects (see part 2 of this volume), the empirical evidence tutions and not merely delivery points for bio-medical inter- with regard to large-scale and routine primary care programs-- ventions." Community expectations of the services provided at whether in industrial or low- and middle-income settings--is a health center or health post tend to be holistic and may well scant (Doherty and Govender 2004; Starfield 1998). The key depart from selected priority intervention categories. The problem is to demonstrate the causal link between provision of exclusion of services from clusters, or the exclusion of whole general primary care services and positive health outcomes-- clusters, can lead to inequity. Failure to manage this situation and especially to disentangle the influences of socioeconomic adequately can undermine clients' confidence in the public conditions. This difficulty is compounded by other factors, health system and affect provider behavior negatively. WHO ranging from the complexity of the study design required to (2000, 59) notes that providers "usually react to this . . . by convincingly evaluate routine programs (as opposed to field tri- cross-subsidizing the excluded activities through the budget als) to the difficulties faced by health ministries in ensuring that received to pay for the defined benefit package; or by charging monies targeted at primary care are translated into the delivery extra for the additional services." of quality health services. These considerations need to be taken into account when Thus, we are able to comment only in broad terms on the identifying clusters of interventions to be provided in local positive contributions of general primary care services, recog- contexts. Nonetheless, likely intervention clusters, with nizing that, although these contributions are potentially enor- potentially high goodness of fit at the primary care level, will mous, the gains made by such services over the past two generally include IMCI; maternal and reproductive health serv- decades have been mixed. Importantly, Almeida and others ices; clinic and community-based management of tuberculosis, (2001) caution against ascribing the failures of primary care to 1196 | Disease Control Priorities in Developing Countries | Stephen Tollman, Jane Doherty, and Jo-Ann Mulligan inherent weaknesses in the concept. In commenting on analy- (1993,) concluded that the results of surveys in Zaire and ses of the effectiveness of primary care, they point to "the Liberia "suggest that child survival programmes in Africa can cataclysmic effect on public health systems in less-developed reduce mortality substantially in populations living in different countries of the global economic recession of the 1980s and the environments at very different initial levels of child mortal- application of policies stressing privatisation and decreased ity. . . . In both countries, it appears that the programme public spending in that decade and the next; [this] resulted in reduced mortality under age 5 by about 20% or more." More rising poverty and under-funding of health services in many generally, many examples of successful health programs clearly less-developed countries, to the point of near-collapse in the depend on the existence of a strong primary health care system poorest countries" (Almeida and others 2001,). (see chapter 8). Given the paucity of evidence from developing countries, turning to the experience of high-income countries is useful, Health Effects although the configuration of primary care services in such set- Investing in Health reported that in countries with moderate to tings may be quite different. Following a detailed comparative high mortality only a few conditions accounted for the major- study of 11 industrial nations (which involved the method- ity of the burden of ill health (World Bank 1993). Thus, in ologically complex--and at times controversial--assigning of 1990, 55 percent of the global burden of disease was concen- primary care and health system scores by country and then trated in children under 15, and 75 percent of this burden was associating these scores with a range of health status indicators caused by 10 disease conditions or clusters (Bobadilla and oth- and total health care costs per capita), Starfield (1994,) con- ers 1994). Except for congenital malformations, all these causes cluded that "countries with a stronger orientation to primary could be aligned with highly cost-effective interventions, many care indeed are more likely to have better health levels and of which are classic components of general primary care lower costs." Shi (1994,) found that, in the United States, (labeled the "clinical services" component of the package). availability of primary care was "by far the most significant Indeed, almost all of the activities included in the "public variable related to better health status, correlating to lower health" component of the package also involve some element of overall mortality, lower death rates due to diseases of the heart individual service delivery in the primary care setting. and cancer, longer life expectancy, lower neonatal death rate, Together, it was estimated, these interventions could eliminate and less frequent low birth weight." Although working largely 21 to 28 percent of the burden of ill health in children. at the level of health output rather than outcome, Blumenthal, With respect to adults, the World Bank (1993) found the Mort, and Edwards (1995), in reviewing a number of studies in burden of disease to be less concentrated: here the 10 main the United States, found considerable evidence of the positive causes of disease and injury accounted for some 50 percent of effect of primary care services (see box 64.1). They argue that the burden.1 Most interventions against these problems were the literature does not adequately address the issue of whether found to be quite cost-effective, but their overall estimated primary care reduces the cost of providing care for underserved effect was moderate because they prevent or treat only part of populations, but they conclude that "a commitment to pri- the problems. Such interventions could thus eliminate 10 to mary care should be made for its potential to improve the 18 percent of the adult disease burden. satisfaction and health status of the American public, not for These figures give some sense of the potential effect of inter- its potential to save money" (Blumenthal, Mort, and Edwards ventions at the primary level when they are targeted at com- 1995,). mon, high-burden conditions in the population. Subsequent work by the World Bank estimated that the primary care level could potentially deal with up to 90 percent of health care Nonhealth Effects demands (World Bank 1994) and that only 10 percent of care Although most of the recent literature on primary care pack- needs require the services and skills typically associated with ages places value on primary care services because of their abil- hospitals. ity to reduce the burden of disease considerably and at low cost, Shifting from estimates to empirical evidence, we find that such services potentially bring other benefits to society. Among some studies have been able to demonstrate large-scale success the most striking may be the welfare benefits that accrue to in the sphere of child health. For example, using data from a households as a result of the prevention of illness. Severe national survey in Niger, Magnani and others (1996) showed disease can limit the ability of patients and caregivers to work, that children living in villages near health dispensaries were leading to the consumption of household assets in the 32 percent less likely to die than children without access to purchasing of care. Russell (2003) found that such costs modern primary care services (differential access resulted from amounted to just over 10 percent of household income in three the phased implementation of services, which produced a nat- developing countries, a proportion that can have a catastroph- ural quasi-experiment). Drawing on earlier work, Ewbank ic impact on the sustainability of poor households. Through General Primary Care | 1197 Box 64.1 Evidence of the Effectiveness of Primary Care Services The effectiveness of primary care services is illustrated as 3. The availability of primary care services is associated follows: with improvement in patients' self-perceived health 1. Community-based interventions improve access to ser- status. vices,reduce the use of emergency and outpatient depart- 4. The longitudinal care afforded by primary care services ments at hospitals, increase the use of noninstitutional is independently associated with improved patient ambulatory care, and reduce the use of hospital care satisfaction, reduced use of ancillary and laboratory (especially with respect to preventable hospitalizations). tests, improved patient compliance, shorter length of 2. Primary care is associated with improved control of stay, and improved recognition of patients' behavioral routine illnesses that have serious consequences if problems. untreated. Source: Blumenthal, Mort, and Edwards (1995). prevention and early treatment, geographically accessible and have an independent [of income levels] effect in improving financially affordable primary care services can reduce the neg- population health--in particular, the beneficial effects of ative economic consequences of ill health for households, primary care" (Shi and Starfield 2000,). The authors suggest reduce absenteeism, and enhance children's performance at that, at least within the particular settings studied, strengthen- school. ing the primary care aspects of health services could mitigate some of the adverse impacts that income inequality has on individuals' health status. Primary-level services are also potentially responsive to Serving of Equity Goals patients' nonhealth needs. These include a need for the range Primary care services have the advantage over hospital care of and quality of health services to meet community expectations tending to be more physically, financially, and culturally and a need for services to treat patients in a helpful and digni- accessible to local communities. Because of their staffing and fied manner. In addition, primary-level facilities can act as organization, they are less costly and more easily able to community resources (providing communal meeting places, provide comprehensive, integrated, personalized, and continu- for example), and primary care services can contribute sup- ous care (World Bank 1993). Because that part of the burden port to neighborhood sports and community development of disease that is addressed by primary care services dispro- activities. All in all, well-functioning primary-level services portionately affects the poor, primary care services are theoret- represent the face of the health system and have the potential to ically well placed to improve equity in health and health care. inspire trust in the system as a whole. Again, few data exist to demonstrate the equity effects of Another source of suspicion regarding the cost-effectiveness primary care delivered on a large scale in middle- and low- approach is the fear that efficiency concerns will override these income countries. This gap is compounded by the fact that positive features of primary care. Paalman and others (1998,) cost-effectiveness analyses seldom take into account the costs note that "the fact that the most efficient interventions . . . tend incurred by patients in seeking care (Doherty and Govender to specifically benefit the poor is more a result of coincidence 2004). than of principle." Indeed, the cost-effectiveness approach does However, in studies by Shi and Starfield (2000, 2001) exam- not intrinsically protect equity and could, for example, count ining income inequality and primary care in the United States, against the extension of services to populations living in a significant association between higher primary physician remote areas. Governments will, at times, need to make explicit supply and good health status was established, even in a context choices between serving equity goals and responding to of high income inequality: "The finding of a significant associ- efficiency concerns when determining service priorities. This ation between primary care and self-rated health contributes to tradeoff is easier to manage in wealthier countries, where the mounting evidence that specific aspects of health services resources are less constrained. 1198 | Disease Control Priorities in Developing Countries | Stephen Tollman, Jane Doherty, and Jo-Ann Mulligan Table 64.2 Cost-Effectiveness of the Health Interventions Subsequent packages have expanded the Investing in Health Included in the Investing in Health Minimum Package of list somewhat, yet the primary-level interventions put forward Health Services for Low- and Middle-Income Countries in 1993 remain among the most cost-effective available, (2002 US$, 2001 prices) especially when combined with population-based interven- Cost per DALY tions (Commission on Health Research for Development 1990; Jha and Mills 2002; WHO 2000). It is important to appreciate Low-income Middle-income that, because of the added costs of extending service delivery Interventions countries countries to people living in more rural and peripheral areas, achieving Public health universal coverage would probably raise marginal costs Expanded program of immunization, 15­22 32­38 considerably above the average figures normally quoted. including vaccine against Bobadilla and others (1994,) note that, in these instances, "the hepatitis B and vitamin A supplementation relative importance of cost-effectiveness versus equity will then determine whether to modify the package by leaving out some School health program 25­32 48­54 interventions, providing mobile services rather than fixed Tobacco and alcohol control program 44­70 57­70 facilities, concentrating on public health rather than clinical AIDS prevention program 4­6a 16­23a interventions for the high-cost population, or sacrificing some Other public health interventions -- -- efficiency in order to preserve equity." The need to redress (includes information, communication, and education on selected risk gender imbalances or respond to cultural preferences and other factors and health behaviors, plus factors, as well as the choice of interventions, also might affect vector control and disease costs. surveillance) Total 18 -- SCALING UP Clinical services Chemotherapy against tuberculosis 4­6 6­9 Adequate delivery of services (and health care more broadly) at Integrated management of the sick child 38­63 63­127 the primary care level is, we believe, fundamental to effective Family planning 25­38 127­190 functioning of health systems. However, for the most part, pri- Sexually transmitted disease treatment 1­4 13­19 mary care systems in low- and middle-income countries have Prenatal and delivery care 38­63 76­139 yet to receive the sustained attention and resources that their Limited care (includes treatment of 253­380 507­760 importance warrants. Early efforts at primary care expansion infection and minor trauma; for in the late 1970s and early 1980s were overtaken in many parts more complicated condition, of the developing world by economic crisis, sharp reductions includes diagnosis, advice, and in public spending, political instability, and emerging disease. pain relief, and treatment as Although essential packages based on cost-effectiveness criteria resources permit) have been criticized for their largely disease-oriented and verti- Total -- 168 cal approach, in most poor countries even these limited ver- Source: Bobadilla and others 1994; World Bank 1993. sions of general primary care remain incompletely applied and Note: -- not available, presumably because the authors were not able to aggregate data to country level. largely unaffordable in relation to current per capita health care a. Understates cost-effectiveness because the analysis examined the probability of transmission expenditure. At the same time, renewed awareness of the cen- to others in the first year only. trality of the primary level in responding to the consequences of the HIV/AIDS epidemic or to rapidly rising cardiovascular risk means that increasing demands will be placed on primary THE COST-EFFECTIVENESS OF PRIMARY care services. This section examines critical elements of any CARE INTERVENTIONS strategy to scale up primary care efforts; a prerequisite, how- ever, is an adequate understanding on the part of policy mak- Part 2 of this book details our best understanding of the cost- ers and planners of the position and role of primary care in the effectiveness of many of the individual interventions that have national health system (Travis and others 2004). been clustered into "essential packages." According to Investing Committing More Financial Resources in Health, most of these interventions are highly cost-effective, costing less than US$100 per disability-adjusted life year In the mid 1980s, Drummond and Mills (1987) found the best (DALY) averted (see table 64.2; World Bank 1993). estimate of the cost of effective primary health care (including General Primary Care | 1199 Table 64.3 Comparison of Per Capita Total Annual Health should be applied with great caution in poor communities Expenditure Required to Provide Minimum Packages (Gilson 1998). Yet the fact remains that an injection of addi- (2002 U.S.$, 2001 prices) tional resources is clearly one prerequisite for the successful scaling up of general primary care in the 21st century, backed Low-income Middle-income Report countries countries up by political commitment to the centrality of general pri- mary care (and primary health care more broadly) as a funda- Investing in Health 15 27 mental strategy for tackling the highest-burden diseases and (World Bank 1993) their causes. Better Health in Africa 16­20 --a (World Bank 1994) Commission on Macroeconomics Least Lower-middle Developing Human Resources and Health (Jha and Mills 2002) developed: 40 income: 39 Although increased financial resources are imperative, Other low Upper-middle Kurowski and others (2003) emphasize that "human resource income: 36 income: 331b availability is likely to determine the capacity to absorb addi- Source: Doherty and Govender 2004. tional financial resources and thus the pace of scaling up." a. Estimate not provided. These authors warn that human resource availability is likely to b. Higher figure because of range of services provided (beyond minimum package) and higher input costs; applies to a small subgroup of countries. be grossly insufficient to meet the scaling-up needs envisioned by the CMH. The skills and competencies necessary to deliver and sup- the recurrent and capital costs of basic and village-level health port effective primary care are in some respects similar to services but not of water and sanitation) to be 2 percent of the those required at other levels of the health system (see chap- annual per capita gross national product (GNP). This amount, ter 71), but certain competencies warrant special emphasis at they noted, is considerable, given that many governments in the primary care level (see box 64.2). Above all, if local developing countries do not spend even 2 percent of annual services are to meet community health needs, leaders at the per capita GNP on their entire health sector. primary care level will have to be freed from the constraints of More recently, the Commission on Macroeconomics and stifling, rule-bound bureaucracies and encouraged to develop Health (CMH) estimated that an additional US$40 billion to innovative and at times unorthodox responses to the demand- US$52 billion annual expenditure would be required by 2015 ing challenges they face. As expressed in the World Health to scale up 49 priority health interventions--not all at the Report 2000 (WHO 2000, 64), "a key challenge in health serv- primary care level--to reach high levels of coverage in 83 ice delivery is to balance the need for broad policy oversight deserving countries (that is, countries with a GNP per capita with sufficient flexibility so that managers and providers can below US$1,200, plus all countries of Sub-Saharan Africa) (Jha innovate and adapt polices to local needs and contexts in a and Mills 2002). Apart from the recurrent and capital costs of dynamic way." the interventions themselves, this estimate included manage- The creation of dynamic health teams at the primary level ment costs generated at levels above "close-to-client" services, is one of the greatest requirements for scaling up effective pri- expenditure to improve absorptive capacity, expenditure on mary care. The role of community health workers in such improvements in the quality of care, and 100 percent increases teams remains unresolved and bears further investigation. At in staff salaries to address the problems of staff recruitment and the same time, one of the most challenging constraints is to retention. The inclusion of these costs accounts largely for the overcome the loss of motivation and sense of resignation of greatly increased per capita estimates of the CMH package rel- the great body of primary care workers who work in under- ative to earlier estimates by the World Bank (see table 64.3), staffed settings; who lack consistent, quality support; and who and probably provides a better estimate of what is needed, have grown accustomed to a norm of inadequate service given the enormous challenges facing primary care service delivery (Narasimhan and others 2004). As Hongoro and delivery. Normand assert in chapter 71, the extent to which countries The CMH has placed great emphasis on donor funding of can improve access to good quality primary care will depend services to adequate levels (see chapter 12 for a more extensive in large part on a "better matches of skills to needs . . . and discussion of sources of financing). Other avenues of funding clearer understanding of how improved structures and incen- include reprioritizing government budgets or recovering costs tives will work." through health insurance schemes and user fees, although these all remain difficult options within low-income settings. In particular the experience of user fee schemes, which prolifer- Harnessing Private Sector Resources ated in the 1990s, suggests that such schemes have negative Private sector health care provision is widespread and growing. impacts on equity, especially at the primary care level, and It extends from local supply of drugs and equipment to fee- 1200 | Disease Control Priorities in Developing Countries | Stephen Tollman, Jane Doherty, and Jo-Ann Mulligan Box 64.2 Important Skills and Competencies at the Primary Care Level The primary care level warrants the following important · Responsiveness to community needs--whether or not skills and competencies: directly expressed--and heightened sensitivity to cul- · Ongoing monitoring of use of services with particular tural norms and local systems of leadership and reference to patient access and community coverage, authority and program adjustment as indicated · Facilitatory and responsive managerial styles by district · Continuous linking of general primary care services team leadership that support and encourage frontline with community outreach efforts in order to enhance health workers. program effects in the defined catchment community · Appropriate clinical skills for midlevel workers, espe- cially in the absence of doctors, and practiced judg- ment regarding timely referral to a hospital Source: Authors. for-service and insurance-based medical care to the many private partnership mechanisms, public sector contracts, and forms of traditional practice. Although general primary care in government regulation. These mechanisms are generally easier most African, Asian, and Latin American settings is a major fea- for not-for-profit providers to contemplate, because they have ture of publicly financed services that are provided by the pub- often been instrumental in bringing primary care to poor com- lic sector, private providers clearly play a significant role in munities. Some nongovernmental organizations (NGOs) are many low- and middle-income countries with respect to the able to offer services that can fill notable gaps, home-based care provision of primary care services (Berman and Rose 1996; to HIV and AIDS sufferers being but one example. Palmer and others 2003). Governments have thus viewed pri- The potential of for-profit providers to contribute to the vate sector providers as contributing additional human and care of the poor is less obvious, especially given the incentives related resources that can be deployed in the service of at least to overservice that are inherent in the fee-for-service reim- a portion of the population (usually those with means, includ- bursement system. Mills and others (2002) find that consumers ing employees with access to reasonable health insurance cov- of private sector primary care are often unable to assess the erage). By alleviating the workloads faced by public providers, technical quality of services, tending to place more weight on private sector providers have allowed the public services to aspects of perceived quality, such as interpersonal skills of focus more directly on poorer communities and patients with- providers and the comfort of the environment in which treat- out means. Out-of-pocket payments, health insurance, and ment occurs, than on technical competence. Mills and others donations (as opposed to government contracts) that fund pri- (2002,) argue that the effectiveness of private services is by and vate sector services thus result in additional financing for the large rather low: "poor treatment practices have been reported health sector. for diseases such as tuberculosis and sexually transmitted infec- Patients often prefer the private sector for a number of rea- tions, with implications not only for the individuals treated but sons. These reasons include geographic accessibility, conven- also for disease transmission and the development of drug ient opening hours, and more favorable staff attitudes, as well resistance." as perceived better quality in terms of shorter waiting times, Palmer and others (2003, 292) have reviewed a "new model greater privacy, higher standards of diagnosis, better (per- of private primary care provision" that has emerged in South ceived) treatment, and counseling (Doherty and Govender Africa. This innovation involves commercial companies pro- 2004). Although private providers are generally thought of in viding "standardized primary care services at relatively low relation to curative care, interest is growing in the role they cost" that are targeted at the low-income employed rather than could play in meeting public health objectives, especially with the very poor (Palmer and others 2003,). Regarding the grow- respect to the scaling up of primary care services (Palmer and ing popularity of these private clinics, the authors find that they others 2003). maintain excellent standards with respect to the quality of ser- It is important that the potential contribution of private sec- vices. The clinics also run at a cost per visit that is comparable tor resources be optimized through appropriate use of public- with public sector primary care clinics, demonstrating that the General Primary Care | 1201 acceptability of services to users and low-cost service delivery quality local information, coupled with local problem solving, are not incompatible objectives. planning, and ownership, was central to appropriate decision Palmer and others (2003, 295) suggest that the increasing making and consequent implementation. popularity of these (affordable) private clinics may provide an Because local data on intervention costs and coverage are opportunity to encourage employed but low-income workers generally not available to district planners and managers, local (who historically have used public sector health services at lit- cost-effectiveness analysis is difficult to incorporate into decen- tle or no charge) to make use of these clinic networks, which tralized priority setting. With TEHIP, priority setting was would enable the public sector to better tend to its "role as reg- driven more by the shares of the burden of disease that known ulator and providing services to the poorest." Potentially, this cost-effective interventions could address. New analytic tools redirection of care could remove some of the burden on the were devised that would help focus resource allocations on the public sector, and the task of regulation might be made easier major "intervention-addressable" disease burdens; targeted by the strong hierarchical control exercised within these clinic sets of cost-effective interventions were then applied--in place chains. To some extent, this shift has been the experience in Sri of embarking on a disease-by-disease or detailed cost- Lanka, where government services have been designed with the effectiveness approach (D. de Savigny, personal communica- explicit assumption that certain forms of care will be provided tion,). Available understanding on cost-effectiveness was used through the private sector (Rannan-Eliya 2001). to eliminate interventions known to be grossly cost-ineffective; However, Palmer and others (2003) point out that the model it was not used to prioritize or rank interventions generally has potential drawbacks. The comprehensiveness and continu- considered to be highly cost-effective. ity of services provided by these private clinics fall short of that TEHIP indicates that gross technical and allocative efficien- available in the public sector. Furthermore, how the behavior cies are relatively easy to address when incremental funding is of private clinics would change under a system of contractual available. As described by de Savigny and others (2002), the arrangements with the public sector is not clear. Whereas con- net effect of decentralized funding, together with a mutually tracting with the not-for-profit sector tends to accommodate reinforcing series of planning, management, and capacity- government objectives fairly easily (Gilson and others 1997), development inputs, was a proportional and absolute increase the experience of contracting with the for-profit sector has had in resources for more efficient delivery of prioritized, cost- mixed results. These and other concerns imply that, although effective interventions addressing the largest shares of the the for-profit sector is an important resource, arrangements for preventable local burden of disease; an increase in the use of the delivery of care through this sector should be developed government health services; and a decrease in mortality in with caution. It also bears mention that, where public sector infants, children under five, adolescents, and adults. This effect systems are weak, private sector services gain ground to the was achieved with relatively limited resources. extent of unbalancing the public-private mix, with potentially TEHIP and related experience make clear that delivery of serious consequences for costs and continuity in patient care effective primary care requires a greatly stepped-up capacity to and for coverage and equity more generally. provide an evidence base that is founded on current and evolv- ing local disease and risk factor burdens, the performance of local health services, client use of public as well as private and Setting Population Health and Clinical Care Priorities traditional services, and (where appropriate) the costs of pro- Along with securing additional resources for primary care viding care. Effective use of such information can profoundly delivery, country capacity to generate the information neces- enhance the ability of the health system to deliver on its core sary for setting and reviewing public health and clinical care service functions, target high-risk and vulnerable groups, assess priorities must be strengthened as a fundamental measure coverage in service provision, and gauge health effects. (Commission on Health Research for Development 1990). This Moreover, such information is vital to establishing the dimen- principle lies at the heart of influential pilot work--at times sions of the local disease burden that should be managed at the referred to as community-oriented primary care--that emerged primary care level (Kahn and others 1999). As cogently stated in the first half of the 20th century and now underpins the by the Bellagio Study Group on Child Survival (2003, 324),"the Tanzanian Essential Health Interventions Program (TEHIP), capacity of countries to obtain and use information to support 1997­2004. TEHIP, through a research and development arm child-health programmes will be a determining factor in tasked with devising practical tools for decentralized health reducing child mortality." planning, has tested "how and to what extent evidence can guide planning of the health sector at district level . . . [in order to] improve technical and allocative efficiency with regard to Developing a District Health System local choices for resource allocation and services offered" (de Drawing from theory and experience in other branches of the Savigny and others 2002,).2 A dynamic process of using high- public sector (Mills and others 1990)--and often as part of 1202 | Disease Control Priorities in Developing Countries | Stephen Tollman, Jane Doherty, and Jo-Ann Mulligan wider public sector developments--the health sector intro- specific interventions in various epidemiological, health duced decentralization widely in low- and middle-income system, and cultural contexts." From this perspective, a too- countries throughout the 1980s and 1990s (WHO 2000). narrow preoccupation with the cost-effectiveness of interven- Positive justification for this method of delivering health care tions cannot but have shortcomings: "whatever package of and primary care in particular lay,first,in its intended benefits-- services is delivered, the resulting effectiveness and equity will for patients and communities--through the provision of almost certainly depend on how the services are delivered, [in context-appropriate services of steadily improving quality. This other words] the strategy for organizing the care" (B. Starfield, service delivery was rightly seen as also conferring substantial personal communication,). financial benefit on households. Second, decentralization was Primary care is delivered through a system of facilities, equip- expected to lead to the strengthening of local responsibility and ment, and personnel; tackling inefficiencies in the system may accountability, with growing authority of district management have major positive benefits for quality of care, program cover- teams over local cost centers. Third, it was presumed that the age, and cost-effectiveness. In many settings a real opportunity more central management levels would invest in enhanced sup- exists to increase the efficiency of general primary care teams by port systems, including management support, further training, giving attention to working conditions, ensuring functional financial management and administration, laboratory services, equipment, and maintaining a stable drug supply. Meaningful and drug supply systems (World Bank 1994). step-ups in care, workable referral and communication systems, In developing settings, few health systems did not decen- gatekeeper functions where indicated, and effectively aligned tralize in some form or another over this period, and most management and support are all needed. Achieving such effi- based services development on a so-called district (or subdis- ciencies should result in many more patients being assessed and trict) health system model. Considerable effort was devoted to managed properly. Significant cost savings may accrue to the achieving a balance between primary care service delivery and health service (through patients being managed at the primary referral to the first-level (or district) hospital. Incentives as well care level rather than the first or specialist referral level) and to as penalties were invoked to encourage first use of primary care patients, families, and households (through care being delivered facilities. more rapidly and nearer to home). Notwithstanding the theoretical appeal of health system decentralization, numerous difficulties in implementation were encountered, with the consequence that the performance Demanding Services: Relationships with Local Communities of decentralized, primary care­oriented systems and national- Among poor and vulnerable communities, the need for care is level support to these systems have fallen way below expecta- demonstrably high, and the effectiveness of primary care serv- tions (Bellagio Study Group on Child Survival 2003). Although ices is likely to substantially influence demand on the public various factors can account for this outcome--and although sector. In relation to infectious as well as noncommunicable these factors will differ according to local and regional disease, outreach services have a major role to play in promot- circumstances--common difficulties include inadequate or ing positive health and health-seeking behaviors and in insufficient primary care skills and competencies, which result supporting community-level preventive and promotive efforts. in poor-quality care; breakdown in referral systems for emer- More generally, renewed efforts to enhance community rela- gencies and more complex cases (McCord and Chowdhury tionships with primary care workers and the health system as a 2003; Snow and others 1994); delegation of responsibilities whole--and to ensure that community voices actively and without the concomitant delegation of authority, especially in appropriately bear on local service development and decision relation to budgeting; authoritarian or strictly hierarchical making--can help bring clients and communities into con- managerial styles that are not conducive to local health services structive public health care partnerships. support and development; and weak or absent measures to develop workable cost-management systems appropriate to different service levels. These problems in achieving successful delivery need to be A RESEARCH AND DEVELOPMENT AGENDA addressed if decentralization is to achieve its intended benefits. FOR GENERAL PRIMARY CARE Again, greater appreciation of the role of decentralized systems in the broader health care architecture, the support needed to Throughout this chapter, we have emphasized the challenges ensure their effectiveness, and the time required to build the and constraints to the effective delivery of general primary care necessary capability are all necessary. As Bryce and others services. In this section, we single out a few areas that warrant (2003, 160) put it, "although research on interventions is plen- concerted research and development in the effort to establish a tiful, little is known about the characteristics of delivery strate- high-functioning primary care platform to support the imple- gies capable of achieving and maintaining high coverage for mentation of cost-effective interventions. General Primary Care | 1203 Evolving Health Transitions care--taking place in rural households, urban residences, or A critical consequence of fast-changing economic and social newer community-based facilities such as hospices. Although conditions is the rapid transition in health profiles in essen- home-based care of people living with AIDS is most promi- tially all low- and middle-income settings. This shift has nent, care and support for clients post-stroke or with other already led or is leading to the coexistence of persisting infec- forms of physical or mental disability are as important. tious disease; nutrition and reproductive health problems; Primary care outreach services, working with community emerging noncommunicable disease and related risk factors health initiatives, NGOs, and communities, are well placed to (such as hypertension, obesity, diabetes, stroke, and cardiovas- contribute expertise, training, and resources toward supporting cular disease); and a growing burden from accidents and inten- such efforts, which are growing rapidly in importance. tional injury. The challenge this transition poses to primary care systems is considerable. For the most part, these systems are oriented to maternal and child health and the management Research to Strengthen General Primary Care of acute illness. An accelerating health transition will require in the Public Sector extending the reach and capacity of widely established primary An abiding need exists for experimental and quasi-experimen- acute care systems (oriented to episodic care) to accommodate tal evaluation of innovation in general primary care services the need for effective systems of chronic and long-term care (whether delivered comprehensively or as clusters of interven- (including continuing, medium- to longer-term, patient or tions), providing greater insight into the enabling and client management and monitoring). constraining factors (which may be systemwide) and a more robust understanding of the effects, costs, and cost-effectiveness of modifications to these services in different settings. To Introduction of Antiretroviral Therapy for HIV and AIDS maximize the likelihood of success in efforts to scale up effec- Many countries, particularly those in southern Africa and East tive interventions or system innovations, such initiatives Africa, are moving to the rapid introduction and scale-up of should be carefully designed, implemented, and assessed in antiretroviral therapies for HIV and AIDS. Substantial and rap- partnership with senior health ministry officers (Berwick idly increasing financial investments are envisaged (indeed are 2004). Such evaluations are required to assess new forms of under way); a necessary accompaniment to such scale-up organizing primary care services (in particular, balancing per- should be improved drug supplies, strengthened laboratory sisting acute needs with the growing need for chronic and long- services, clinical training of primary care staff, and reassertion term care, or establishing the skills mix that is most effective in of the importance of health service relationships with local particular settings); similarly, they are necessary when assessing communities. Such measures--which can succeed only with delivery of interventions--such as the cost-effectiveness of sustained public sector commitment--have the potential to multi-disease intervention clusters in different epidemiological invigorate and motivate all facets of primary care delivery but, and social contexts, or the extent of uptake by vulnerable equally, could undermine existing services. The challenge is groups (such as children or the elderly) or marginal popula- how to realize the positive potential of antiretroviral therapies tions. Operational research efforts are needed in a range of in meeting the needs of HIV and AIDS sufferers and their com- spheres: to evaluate factors that facilitate or hinder effective munities, while ensuring a major contribution to strengthening performance by service providers or to develop easily managed primary care provision more generally. In other words, the monitoring systems to assist, for example, in assessments of challenge is to strengthen a particular service (HIV and AIDS intervention coverage. treatment and care) and primary care services simultane- ously--through building more effective health teams, improv- ing drug supplies, strengthening service monitoring and CONCLUSIONS evaluation, enhancing supervision and support systems, extending service coverage, and so forth. Such systemwide Clearly, a great many of the most cost-effective interventions strengthening can be expected to greatly improve the technical detailed in this volume depend on a high-functioning primary efficiency of key elements of the general primary care system. care system for their effective implementation. This system comprises the elements of the primary level of care (including facilities, equipment, drugs, personnel, and associated manage- Effective Support and Networking for Community ment support); their combination to form a competent deliv- and Home-Based Care ery capability; and the services that are thereby delivered. Along with the reorientation of primary care systems to sup- Because cost-effectiveness estimates are based on the presumed port chronic and long-term care are needs for "home-based" effective delivery of primary care services, it can be argued that 1204 | Disease Control Priorities in Developing Countries | Stephen Tollman, Jane Doherty, and Jo-Ann Mulligan implicit in the estimates have been overly optimistic assump- referral (district) hospital levels. This organizational tions regarding key constituents (staff, drugs, equipment, mon- and service unit is fundamental to effective health care itoring and evaluation, and so forth) of the primary care level provision, and failure to recognize the interrelationship and their functioning. Great efforts to render such systems as between component levels has had high health costs and effective as possible, subject to the constraints of particular resulted in great inefficiency. environments, are therefore justified. More generally, decisions on the best and most appropriate The health and development cost of weak or inadequate sites for delivery of interventions are not always straightfor- primary care systems to high-risk or vulnerable groups--and ward, will benefit from expert discussion, and will often be to communities more generally--is demonstrably high. context specific. Moreover, many interventions can be delivered However, effective general primary care that responds to the from multiple sites--although a hierarchy of preference will rapid health transitions under way in all socioeconomic con- usually exist, influenced by issues such as cost-effectiveness, texts offers the potential for major health and, hence, develop- ease of service provision, need for monitoring, access, and cov- ment gains that provide good value for money and enhance erage. Careful review of the extensive range of interventions equity. Critical make-or-break points include upscaled finan- presented in part 2 of this volume and their likely site of deliv- cial investments paralleled by major and sustained investment ery reveals the following: in human resources (principally the strengthening of local staff capacity, the building up of key skill sets--including support- · The interfaces between (a) community and primary care ive management--and the encouragement of innovation in levels and (b) primary care and district levels are critical services development); far greater attention to improving sites that profoundly affect the effectiveness of service delivery and service quality, monitoring service coverage, delivery. improving access by vulnerable groups and taking account of · No substitute exists for a well-functioning district health equity considerations in general; and establishment of a trust- system comprising community, primary care, and first- ing and constructive partnership with local communities. General Primary Care | 1205 Annex 64.A Comparison of Proposed Basic Packages of Interventions Alma Ata Commission on Declaration Investing in Better Health in World Health Macroeconomics and World Health (WHO and Health (World Africa (World Report 2000 Health Working Group 5 Report 2002 Intervention UNICEF 1978) Bank 1993) Bank 1994) (WHO 2000) (Jha and Mills 2002) (WHO 2002)a Maternity-related interventions Prenatal care Treatment of complications during pregnancy Skilled birth attendants Emergency obstetric care Postpartum care Family planning Nutrition: pregnant and lactating women Tetanus toxoid Childhood disease­related interventions (prevention) Bacillus Calmette-Guérin Polio Diphtheria-pertussis-tetanus Measles Hepatitis B Haemophilus influenzae type B Vitamin A supplementation Iodine supplementation Zinc supplementation Anthelmintic treatment School health program (incorporating micronutrient supplementation, school meals, anthelmintic treatment, health education) Childhood disease­related interventions (as part (as part (treatment) of IMCI) of IMCI) Acute respiratory infections Diarrhea Causes of fever Malnutrition (including nutrition and supplementary feeding) Anemia Feeding and breastfeeding counseling Malaria prevention b Insecticide-treated nets Residual indoor spraying Malaria treatment b 1206 | Disease Control Priorities in Developing Countries | Stephen Tollman, Jane Doherty, and Jo-Ann Mulligan Annex 64.A Continued Alma Ata Commission on Declaration Investing in Better Health in World Health Macroeconomics and World Health (WHO and Health (World Africa (World Report 2000 Health Working Group 5 Report 2002 Intervention UNICEF 1978) Bank 1993) Bank 1994) (WHO 2000) (Jha and Mills 2002) (WHO 2002)a Tuberculosis treatment b Directly observed treatment short course (DOTS) for smear-positive patients DOTS for smear-negative patients HIV/AIDS prevention (more limited than later packages?) Youth-focused interventions Interventions working with sex workers and clients Condom social marketing and distribution Workplace interventions Strengthening of blood transfusions systems Voluntary counseling and testing Prevention of mother-to-child transmission Mass media campaigns Treatment for sexually transmitted infections HIV/AIDS care Palliative care (see under limited care) Clinical management of opportunistic illnesses Prevention of opportunistic illnesses Home-based care HIV/AIDS HAART provision Tobacco control program (taxes, legal action, information, nicotine replacement) Alcohol control program Other public health interventions (includes (information, information, education, and communication education, and on selected risk factors and health communication) behaviors, plus vector control and disease surveillance) Limited care (includes treatment of infection and minor trauma; for more complicated conditions includes diagnosis, advice, and pain relief, and treatment as resources permit) Noncommunicable diseases and injuries (selected early screening and prevention) Populationwide interventions to reduce the risks of cardiovascular disease (salt- and cholesterol-lowering strategies) Water and sanitation (disinfection at point of use) Sources: Jha and Mills 2002; WHO 2000, 2002; WHO and UNICEF 1978; World Bank 1993, 1994. HAART highly active antiretroviral therapy for the treatment of HIV/AIDS; IMCI integrated management of infant and childhood illness. Note: A " " that appears in a shaded cell but not in the white cells beneath this area means that no details of the exact interventions were provided in the report. a. Addressed only interventions against risk factors. b. These and other disease prevention and control initiatives fell under a general item termed prevention and control of locally endemic diseases (HIV/AIDS was not an issue at the time). General Primary Care | 1207 ACKNOWLEDGMENTS Commission on Health Research for Development. 1990. Health Research: Essential Link to Equity in Development. New York: Oxford University This chapter draws considerably on the background paper by Press. Doherty and Govender (2004) titled "The Cost-Effectiveness of de Savigny, D., H. Kasale, C. Mbuya, G. Munna, L. Mgalula, A. Mzige, and G. Reid. 2002."Tanzania Essential Health Interventions Project: TEHIP Primary Care Services in Developing Countries: A Review of Interventions--An Overview." Ministry of Health, Dar es Salaam, the International Literature," commissioned by the Disease Tanzania. Control Priorities Project. It draws on productive discussions Doherty, J., and R. Govender. 2004. "The Cost-Effectiveness of Primary during the workshop "Integrating Interventions and Health Care Services in Developing Countries: A Review of the International Literature." Background paper for the Disease Control Priorities Systems" held in July 2004 in Johannesburg; the workshop was Project. supported by the project and convened by the School of Public Drummond, M. F., and A. Mills. 1987. Cost Effectiveness of Primary Health Health, University of the Witwatersrand, South Africa, and the Care: A Review of Evidence. London: Commonwealth Secretariat. Health Policy Unit, London School of Hygiene and Tropical Ewbank, D. C. 1993. "Impact of Health Programmes on Child Mortality in Medicine, United Kingdom. The authors acknowledge with Africa: Evidence from Zaire and Liberia." International Journal of Epidemiology 22 (Suppl. 5): S64­72. pleasure the critique, insights, and support provided by the Gilson, L. 1998. "The Lessons of User Fee Experience in Africa." In editors and chapter reviewers. Sustainable Health Care Financing in Southern Africa, ed. A. Beattie, J. Doherty, L. Gilson, E. Lambo, and P. Shaw,. Washington, DC: World NOTES Bank. ------. 2003. "Trust and the Development of Health Care as a Social 1. Bobadilla and others (1994) comment that separating interventions Institution." Social Science and Medicine 56: 1453­68. according to age group is artificial because benefits may accrue in later life, Gilson, L., J. Adusei, D. Arhin, C. Hongoro, S. K. Mujinja. 1997. "Should as in the case of hepatitis vaccine, and may improve well-being, such as African Governments Contract Out Clinical Health Services to Church cognitive abilities. Adult interventions, such as HIV prevention and pre- Providers?" In Private Health Providers in Developing Countries: Serving natal care, also benefit children. the Public Interest?, ed. S. Bennett, B. McPake, and A. Mills, 276­302. 2. TEHIP has functioned in a "high mortality" setting. Relevant evi- London: Zed Press. dence is related to mortality levels and trends, including cause-specific mor- tality, as well as to district-level financial allocations and changes over time. Jha, P., and A. Mills. 2002. "Improving Health Outcomes of the Poor: The Report of Working Group 5 of the Commission on Macroeconomics and Health." Geneva: World Health Organization. REFERENCES Kahn, K., S. M. Tollman, M. Garenne, and J. S. S. Gear. 1999. "Who Dies from What? Establishing Cause of Death in South Africa's Rural Almeida, C., P. Braveman, M. R. Gold, C. L. Szwarcwald, J. Mendes Ribeiro, Northeast." Tropical Medicine and International Health 4: 433­41. A. Miglionico, and others. 2001. "Methodological Concerns and Kurowski, C., K. Wyss, S. Abdulla, N. Yemadji, and A. Mills. 2003. "Human Recommendations on Policy Consequences of the World Health Report Resources for Health: Requirements and Availability in the Context of 2000." Lancet 357: 1692­97. Scaling Up Priority Interventions in Low-Income Countries." Health Bellagio Study Group on Child Survival. 2003. "Knowledge into Action for Economics and Financing Programme, London School of Hygiene and Child Survival." Lancet 362 (9380): 323­27. Tropical Medicine, London. Berman, P., and L. Rose. 1996. "The Role of Private Providers in Maternal Magnani, R. J., J. C. Rice, N. B. Mock, A. A. Abdoh, D. M. Mercer, and and Child Health and Family Planning Services in Developing K. Tankari. 1996. "The Impact of Primary Health Care Services on Countries." Health Policy and Planning 11 (2): 142­55. Under-Five Mortality in Rural Niger." International Journal of Epidemiology 25 (3): 568­77. Berwick, D. M. 2004. "Lessons from Developing Nations on Improving Health Care." British Medical Journal 328: 1124­29. McCord, C., and Q. Chowdhury. 2003. "A Cost Effective Small Hospital in Bangladesh: What It Can Mean for Emergency Obstetric Care." Blumenthal, D., E. Mort, and J. Edwards. 1995. "The Efficacy of Primary International Journal of Gynecology and Obstetrics 81: 83­92. Care for Vulnerable Population Groups." Health Services Research 30 (1): 253­73. Mills, A., R. Brugha, K. Hanson, and B. McPake. 2002. "What Can Be Done about the Private Health Sector in Low-Income Countries?" Bulletin of Bobadilla, J. L., P. Cowley, P. Musgrove, and H. Saxenian. 1994. "Design, the World Health Organization 80: 325­30. Content, and Financing of an Essential National Package of Health Services." Bulletin of the World Health Organization 72 (4): Mills, A., J. P. Vaughan, D. L. Smith, and I. Tabibzadeh. 1990. Health System 653­62. Decentralization: Concepts, Issues and Country Experience. Geneva: World Health Organization. Briggs, C. J., P. Capdegelle, and P. Garner. 2003. "Strategies for Integrating Primary Health Services in Middle- and Low-Income Countries: Narasimhan, V., H. Brown, A. Pablos-Mendez, O. Adams, G. Dussault, Effects on Performance, Costs and Patient Outcomes" (Cochrane G. Elzinga, and others. 2004. "Responding to the Global Human Review). In The Cochrane Library, Issue 4. Chichester, U.K.: John Wiley Resources Crisis." Lancet 363 (9419): 1469­72. and Sons. Newell, K. 1975. "Health by the People." World Health Organization, Bryce, J., S. el Arifeen, G. Pariyo, C. F. Lanata, D. Gwatkin, J.-P. Habicht, and Geneva. the Multi-Country Evaluation of IMCI Study Group. 2003. "Reducing Paalman, M., H. Bekedam, L. Hawken, and D. Nyheim. 1998. "A Critical Child Mortality: Can Public Health Deliver?" Lancet 362 (9378): Review of Priority Setting in the Health Sector: The Methodology of 159­64. the 1993 World Development Report." Health Policy and Planning 13 Coker, R. J., R. A. Atun, and M. McKee. 2004. "Health-Care System Frailties (1): 13­31. and Public Health Control of Communicable Disease on the European Palmer, N., A. Mills, H. Wadee, L. Gilson, and H. Schneider. 2003. "A New Union's New Eastern Border." Lancet 363 (9418): 1389­92. Face for Private Providers in Developing Countries: What Implications 1208 | Disease Control Priorities in Developing Countries | Stephen Tollman, Jane Doherty, and Jo-Ann Mulligan for Public Health?" Bulletin of the World Health Organization 81 (4): Travis, P., S. Bennett, A. Haines, T. Pang, Z. Bhutta, A. A. Hyder, and oth- 292­97. ers. 2004. "Overcoming Health-Systems Constraints to Achieve the Rannan-Eliya, R. P. 2001. "Strategies for Improving the Health of the Poor: Millennium Development Goals." Lancet 364: 900­906. The Sri Lankan Experience." Paper prepared for Health Systems Unger, J.-P., P. de Paepe, and A. Green. 2003. "A Code of Best Practice for Resource Centre, Department for International Development. Health Disease Control Programmes to Avoid Damaging Health Care Services Policy Programme, Institute of Policy Studies of Sri Lanka, Colombo. in Developing Countries." International Journal of Health Planning and Russell, F. 2003. "The Economic Burden of Illness for Households: A Management 18: S27­39. Review of Cost of Illness and Coping Strategy Studies Focusing on Vuori, H. 1985. "The Role of Schools of Public Health in the Development Malaria, Tuberculosis and HIV/AIDS." Paper presented at the Disease of Primary Health Care." Health Policy 4: 221­30. Control Priorities Project Workshop, Johannesburg, South Africa, Walsh, J. A., and K. Warren. 1979. "Selective Primary Health Care: An July 1­4. Interim Strategy for Disease Control in Developing Countries." New Shi, L. 1994. "Primary Care, Specialty Care, and Life Chances." England Journal of Medicine 301: 967­74. International Journal of Health Services 24 (3): 431­58. WHO (World Health Organization). 2000. The World Health Report 2000. Shi, L., and B. Starfield. 2000. "Primary Care, Income Inequality, and Self- Health Systems: Improving Performance. Geneva: WHO. Rated Health in the United States: A Mixed-Level Analysis." ------. 2002. The World Health Report 2002: Reducing Risks, Promoting International Journal of Health Services 30 (3): 541­55. Healthy Life. Geneva: WHO. ------. 2001. "The Effect of Primary Care Physician Supply and Income WHO (World Health Organization) and UNICEF (United Nations Inequality on Mortality among Blacks and Whites in U.S. Metropolitan Children's Fund). 1978. "Primary Health Care: A Joint Report by the Areas." American Journal of Public Health 91 (8): 1246­50. Director-General of the World Health Organization and the Executive Snow, R. W., V. O. Mung'ala, D. Forster, and K. Marsh. 1994. "The Role of Director of the United Nations Children's Fund." WHO, Geneva. the District Hospital in Child Survival at the Kenyan Coast." African World Bank. 1993. Investing in Health: World Development Report 1993. Journal of Health Sciences 1 (2): 71­75. New York: Oxford University Press. Starfield, B. 1994. "Primary Care: Is It Essential?" Lancet 344: 1129­33. ------. 1994. Better Health in Africa: Experience and Lessons Learned. ------. 1998. Primary Care: Balancing Health Needs, Services, and Washington, DC: World Bank. Technology. New York: Oxford University Press. Tollman, S. M. 1991. "Community Oriented Primary Care: Origins, Evolution, Applications." Social Science and Medicine 32: 633­42. General Primary Care | 1209 Chapter 65 The District Hospital Mike English, Claudio F. Lanata, Isaac Ngugi, and Peter C. Smith Health care comprises a continuum from home-based, self- picture of public district hospitals as underused, inefficient, administered treatment to highly specialized intervention and providing poor quality care (Barnum and Kutzin 1993) dependent on professionals with many years of training and a may reflect deficiencies in the entire health system as well as heavy capital investment. In principle, the role of the health at the hospital level. system planner is to balance the many separate components of · Second, optimizing the health system configuration is an the system to optimize the magnitude and distribution of active, continuing process that must often proceed incre- health benefits, subject to a variety of constraints such as budg- mentally, ideally tackling problems in order of priority. An etary levels, geography, and human resources capacity. While optimal balance is not likely to be achieved naturally recognizing that other paradigms are possible and valid, we through neglect or reliance on market mechanisms. generally adopt this optimization perspective in our discus- sions because it combines broad social (including user) and Hospitals are major consumers of health budgets. However, political dimensions with systematic economic principles when there is a paucity of good evidence--even in industrial decisions are made in a competitive, resource-constrained countries--on their effect (McKee and Healy 2002), whereas environment. Following such logic, it should be possible to the body of theory and opinion on their role is wide. This define the place, purpose, and size of the district hospital sector chapter can serve as only an introduction to topics that within a balanced system of care for any particular setting. include, among others, the political and social value of hospi- Although this view is theoretically appealing, the world of tals and their essential role in integrated health systems (Sachs real health systems that have evolved under different historical 2001; Van Leberghe, de Bethune, and de Brouwere 1997; WHO and political pressures is somewhat different. This perspective 1999; World Bank 1993). The chapter first introduces basic does, nevertheless, suggest some common principles involved concepts relevant to district hospitals that may affect their role in defining the optimum balance of care even within groupings and performance and a description of possible core services as diverse as "developing countries." Two further points are (see figure 65.1). For discussions of the evidence justifying worth considering: inclusion of an intervention or process as a core service at this level of care, the reader is referred to disease- and service- · First, although the focus of this chapter is the district hospi- specific chapters. Although recently attempts have been made tal, crucial links exist with many other aspects of the health to refine definitions of performance (WHO 2000b), the term is system. Choices made in relation to hospitals are likely to used in a general sense, referring to processes and outcomes affect the whole health system and vice versa. For example, that contribute to improved levels and distribution of health. programs to improve peripheral clinic referrals of women The chapter then summarizes currently available economic with high-risk pregnancies may result in a paradoxical data on hospital care, focusing where possible on the district decline in the quality of care if critical human and other level and acknowledging the difficulty in generalizing findings resources are inadequate at the hospital level. Thus, the from one setting to another. An illustration follows of some of 1211 Demand Supply (inputs) Outputs Peripheral health (Referral Home Hospital unit care) Policy and Direct: regulatory Institutional Individual Lives saved framework management patient care Quality of life Perceptions Facility size Disease averted, Facility size State of physical Knowledge individual and Knowledge Staffing Facility budget environment Competence population levels Resources Charges Charges Basic services Interpersonal (power and Physical Consumable Indirect: Accessibility skills water) accessibility supplies Political symbolism Human resources Safety Safety State of physical Sense of health Consumable Patient education environment Human resources security supplies management Basic services Environmental Degree of (power and Capital resources safety autonomy water) management Community status Knowledge Consumables Employment Competence Internal (local) Interpersonal skills market consumer Supervision Supervision Information Information Information Information Information Source: Authors. Note: Some of the factors that may influence a hospital's performance and its products or outputs, the value of which depends on one's perspective, are illustrated. The intrinsic roles of supervision and information flow are emphasized. Figure 65.1 Conceptual Framework for Delivery of Health Services at the District Hospital the factors that threaten district hospitals' performance, indi- What Is a District Hospital? cating the broad range of influences to which they are subject. Health systems are often organized in a "hub-and-spoke" Finally, possible strategies for improving performance are pro- arrangement, with a large district hospital (the hub) having posed, focusing on cross-cutting interventions, and highlight more and better-trained personnel and better equipment than areas where current knowledge is inadequate and research is more peripheral clinics (the spokes). Although variations fre- urgently needed. quently occur in practice (for example, a large district may have several relatively similar hospitals), this simple model of service provision is assumed throughout this chapter, with the district DEFINITIONS, BASIC CONCEPTS, hospital supplying first referral-level care for both outpatients AND FRAMEWORK and inpatients. District hospitals also, in theory, may serve a gatekeeping role for those patients with less common problems, The evolution of a hierarchical system of health care is readily for whom skills and resources are most effectively concentrated explained if one assumes the perspective of the provider, at even higher levels of care provided at a regional or national although less obvious if one's perspective is that of the com- level. Thus, from the perspective of provider efficiency, munity using the hospital or a government seeking to create economies of scale and economies of scope are important basic political capital. Concentrating skills and resources in one concepts in considering district and referral hospital functions. place for conditions that are often relatively uncommon or Such hierarchical health systems frequently overlap with that cannot easily be treated closer to the home environment wider political and administrative hierarchies that are based on is intuitively attractive. Such concentration also offers the geographically defined units. The district is, therefore, used in prospect of continued accumulation of experience and, thus, this chapter as a generic term for an administrative unit often skill and potentially benefits from system resources that may comprising a population of 100,000 to 1 million people for serve a variety of needs. whom one tier of local government is typically responsible. The 1212 | Disease Control Priorities in Developing Countries | Mike English, Claudio F. Lanata, Isaac Ngugi, and others shared administrative boundaries and frequent proximity of Average cost district hospitals to district political administrations often result in the district hospital's involvement in the much wider tasks of district health management and public health. The per- formance of these functions may be critical to the success of the health system as a whole, but this role is easily forgotten. Efficiency. Allocative efficiency deals with the desire to allo- cate resources to secure the maximum health benefit from the inputs available (Hensher 2001). Within this paradigm plan- Economies Diseconomies of scale of scale ners search for the balance between community care, primary care, and facility-based care that results in the greatest health benefit at the least cost. At the level of an individual hospital, the issue of allocative efficiency arises when decisions must be Scale of activity made to allocate resources to different services. In theory, cost- Source: Adapted from Posnett (2002). effectiveness studies with a global health status outcome meas- ure such as the disability-adjusted life year (DALY) should Figure 65.2 Theoretical Long-Term Average Cost Curve inform debate on allocative efficiency, because such studies provide a direct means of comparing alternative strategies. Technical efficiency deals with the extent to which specific of staff salaries and technology), and the pattern of diseases institutions are getting the most out of the resources available. vary widely, has not been examined. For example, is a district hospital deploying its given resources In discussing economies of scale, we must consider two fur- in the most effective manner to achieve the desired output? ther issues. First, considerable evidence suggests that the ability Technical efficiency is often measured using partial indica- to specialize and the experience gained with high volumes of tors such as cost per procedure. Interpreting such data often patients can lead to better outcomes for physicians practicing requires great care, but most fundamentally it requires some in larger hospitals. Second, although reducing the number (and comparator, because a way of knowing the resources needed to increasing the size) of hospitals may reduce health system costs produce the desired output rarely exists. Thus, technical effi- and improve outcomes, it may shift some costs to patients in ciency is usually a relative term, and performance indicators-- the form of increased travel time or even a reduction in the carefully interpreted--can be used to identify best current ability to reach the hospital and secure care. Thus, excessive practice. New technology or a change in the availability or price concentration of hospital services may compromise health and of resources may result in continual improvements in what is equity objectives, particularly in rural areas. The planner may achievable, so a process that was technically efficient can need to balance direct health system costs against the broader become relatively inefficient over time. Data on technical effi- population costs of securing access. In many circumstances, ciency often provide the basis for benchmarking hospital serv- this effort may give rise to an intermediate solution, such as ice providers and may identify poorly performing services for medium-sized hospitals, smaller local hospitals equipped to targeted improvement strategies. deal with common procedures, or dispersed clinics staffed by peripatetic specialist teams. Economies of Scale and Scope and Hospital Size. A central The hospital also offers the potential for improving effi- policy question is whether it is more efficient to concentrate ciency if different services use some of the same inputs. resources in a small number of large centers, where the planned Although the hospital might not be able to justify paying the number of procedures can be high, or to have a greater num- salary of a laboratory technician to perform hemoglobin meas- ber of smaller centers. The issue of economies of scale deter- ures and blood cross-matching only for the maternity unit, the mines the most efficient size hospital. Where the average costs fact that such a person also contributes to the work of the sur- of care can be shown to depend on hospital (or unit) size, gical, medical, and pediatric services makes that technician's economies of scale exist (see figure 65.2). Recent evidence sug- presence more cost-effective. This laboratory service, therefore, gests that, at least for industrial countries, large centers may offers an economy of scope. The concentration of inputs, both eventually suffer from diseconomies of scale, when the ineffi- human and technological, evident at the district hospital offers ciencies introduced in administering a very large facility begin major opportunities for unit-cost reductions and, therefore, to outweigh any advantages (Posnett 2002). The potential for economies of scope. Considering the mix of services provided diseconomies of scale in developing countries, where the mix- as hospitals are planned or augmented is important to antici- ture of cases, the costs of inputs (particularly the relative costs pate or account for economies of scope. The District Hospital | 1213 Equity. Equity is a fundamental principle guiding most public Such hospitals will usually provide 24-hour care and be health systems. It can embrace concepts such as equality of pro- integrated into the district health system at a wider level to pro- vision or equality of access (for equal need), equality of benefit vide or support a range of services: from health services, or equality of outcome.Although often not defined explicitly, many pro-poor policies, such as the Poverty · districtwide health information Reduction Strategy Papers that encompass health, are based on · implementation of peripheral primary health care policies some principle of equity. Loosely speaking, such policies aim to · administrative and logistics support to primary health care reduce disparities in access or overall health status observed efforts between different sections of a population, most obviously the · communication with the community differences between rich and poor sections of a community. · curative and chronic care for patients referred from periph- For health planners, however, equity principles pose some eral units hard challenges. For example, if an urban district has a public · district laboratory services hospital with adequate staff and resources providing a range of · training and continuing medical education of health work- acute services reasonably efficiently, should not every district ers and students hospital provide the same range of services? In practice, ensur- · links between health and other development agendas ing that a hospital in a poor, inaccessible rural district with a · development of local solutions to local health problems. highly dispersed, smaller population provides a similar level and breadth of service may be difficult and considerably more This menu of recommended services at the district hospital expensive. The result can be a hospital with apparently high level does not represent a rigorous attempt to optimize the unit costs of treatment that, because of late presentation or health system configuration to maximize its cost-effectiveness. resource constraints, secures poorer outcomes. The central pol- Indeed, the logic of the earlier discussion is that the precise mix icy question is: To what extent is society prepared to see of services provided should be informed by overall health sys- resources deployed to address such equity concerns at the tem design. Rather, the list represents what is perceived to be a expense of pure efficiency? fair minimum level of health provision for all, based on accu- Issues of efficiency, economies of scale and scope, and equity mulated knowledge and experience of the common demands have contributed in part to the development of strategies defin- for hospital care (the visible burden), the availability and ing an essential package of services that should be provided for simplicity of interventions, the perceived effectiveness of inter- an entire population (Bobadilla and others 1994). These pack- ventions, and their acceptability in an environment con- ages are often targeted at the most important causes of mortal- strained by limited information and limited availability of ity and morbidity, so the inefficiencies in providing an equitable human and financial resources (Van Leberghe, de Bethune, and service may be reduced. Nevertheless, the unit costs of reaching de Brouwere 1997). disadvantaged populations are often likely to be higher than An obvious logic supports the inclusion of many of these average unit costs, and planners need to recognize this fact core functions, sometimes supported by evidence of their when designing packages and set budgets accordingly. value. WHO's Commission on Macroeconomics and Health has attempted to define the services that small hospitals should offer as part of the close-to-client package on the basis of What Essential Services Should a District Hospital Provide? burden and likely cost-effectiveness (Sachs 2001). However, The World Health Organization (WHO 1992) envisages that a although useful for suggesting service priorities, the report district hospital should be able to offer diagnostic, treatment, considers primarily infectious diseases and maternal health. In care, counseling, and rehabilitation services provided by addition, it is not clear whether recommended services were predominantly generalist practitioners spanning the following included on the basis of data on condition-specific burden and disciplines: intervention cost-effectiveness or of the potential effect of the combined package of services considering potential economies · family medicine and primary health care of scale and scope. Future studies should perhaps address more · medicine clearly the issues of the incremental cost-effectiveness of new · obstetrics or additional interventions at the district hospital level when · mental health exploring the appropriateness of services. · eye care · rehabilitation · surgery (including trauma and orthopedics) Clinical Services · pediatrics The initial drive to implement primary health care (PHC) left · geriatrics. district hospitals sidelined. They were often grouped with 1214 | Disease Control Priorities in Developing Countries | Mike English, Claudio F. Lanata, Isaac Ngugi, and others expensive tertiary units; were labeled high cost, inequitable, Cross-Cutting Services at the District Hospital and relatively ineffective; and were rarely protected by powerful Some medical services provide support to a range of depart- professional groups based in the tertiary centers. Their position ments or users and are referred to as cross-cutting services. as an integral part of PHC was reestablished during the 1980s Such services include those aimed at recuperation and rehabil- (Canadian International Development Agency and the Aga itation (physiotherapy, occupational therapy, and so forth; see Khan Foundation 1981; WHO 1987). Currently the district chapter XX; laboratory services, and diagnostic imaging. hospital is envisaged as the apex of the pyramid of primary Whether and to what degree these services are provided may be health care, most obviously in such programs as Safe major determinants of the overall range of services that can be Motherhood and Integrated Management of Childhood offered, the fixed costs of providing care at district hospitals, Illness. In programs such as Integrated Management of and their cost-effectiveness. Their provision should, therefore, Childhood Illness, the expected role of district hospital­level be planned as part of the portfolio of care to be offered, taking care is explicit (WHO 2000a), with priority conditions reflect- into account expected use and estimates of the value added. ing burden-of-disease estimates (Black, Morris, and Bryce This strategy suggests a degree of flexibility that may conflict 2003). Although the effectiveness of this approach has yet to be with historical perspectives about what is important and "one established, evidence at the hospital level suggests that deliver- size fits all" national policies. Health information systems are ing a basic package of care may, in principle, cover the majori- also a critical cross-cutting service; they are discussed in the ty of admitted cases and improve service delivery (Ngoc Anh "Health Information Systems" section of this chapter. and Tram 1995). However, without tackling current difficulties It is worth noting here that the concentration in hospitals of at the hospital level, effectiveness cannot be assumed (see cross-cutting resources used by different activities often gives "Information and Integration" later in this chapter). rise to many accounting complications, such as allocating over- Other basic approaches to delivery of services at the district head costs, which bedevil attempts to secure meaningful cost hospital level, such as triage of new outpatient attendees and a comparisons across hospitals. basic package of neonatal care, also show promise (Duke, Willie, and Mgone 2000; Robertson and Molyneux 2001). Interventions such as the provision of basic trauma care can Wider Role in the District Health System effectively be offered only at this level of the health system (see District hospitals often house the technical expertise and chapter 68), while in other areas (for example, chapters 26, 31, professional authority essential for local implementation of and 67) hospital inpatient care should be considered together national policy, making them potentially key players in manag- with alternative means of delivering services if cost- ing, monitoring, and supervising district health plans. They effectiveness is to be maximized. These examples all serve to should also act as advocates for plans that address local health emphasize that close-to-client health services must be tightly needs. This section examines this wider role of the district hospi- integrated with district hospital­level care and demonstrate tal, the value of which is often hard to quantify, but which may be strong dependency on the referral system. Thus, cases too com- critical to the effectiveness of the local health system as a whole. plex or serious to be managed in the periphery are sent for care where skills and resources are more highly concentrated, in the Integration with Other Local Health-Related Services. A dis- expectation that health outcomes will be better. This attrac- trict hospital should, in most cases, be an integral part of a tively simple idea presupposes that the district hospital is able wider district health system. Although not specifically dis- to provide the care desired; although some evidence supports cussed here, part of the broader remit is often to link up with the likely effectiveness of this approach (Van Leberghe and other governmental and nongovernmental actors in health and Pangu 1988), clearly numerous potential obstacles exist along health-related programs, which may include water and sanita- this pathway (discussed in the later section "Factors Influencing tion, education, and social services. (A more specific discussion District Hospitals' Performance"). can be found in WHO 1990.) Those important coordination Additionally, although the focus has often been on district functions are hard to value in traditional examinations of cost hospitals as recipients of referrals, a much more dynamic rela- and cost-effectiveness but may be critical in sustaining a coor- tionship has been proposed (WHO 1987): for many PHC dinated health care approach, especially if greater autonomy is activities such as immunization programs the district hospital devolved to district administrations. is both a provider of services and a coordinating center for information and supplies. To permit early discharge, enhance Training. District hospitals often have a direct role in the pri- treatment compliance, and make home-based care possible-- mary training of health workers, particularly clinical assistants, all of which may improve cost-effectiveness--hospitals need to nurses, and health aides, as well as an ongoing role in providing play an active role in providing outreach services, supervision, continuing medical education. Their role in building human and support. resources capacity among those actively participating in health The District Hospital | 1215 care delivery and in ensuring that training and experience ECONOMICS OF DISTRICT HOSPITALS: reflect the real health needs of the community is potentially of A SUMMARY OF REPORTED EXPERIENCE great value. Additionally, as the focal point of outreach for many programs that aim to disseminate knowledge through The previous sections outline the suggested functions and the cascade mechanism, district hospitals are often relied on to extended role of a district hospital. Although some countries transmit knowledge to more peripheral levels of care. have adopted the principle of essential packages of services and defined detailed norms and standards for care at this level as Supervision. Together with their training function, district part of long-term health sector strategies, many countries lack hospital staff members are often supposed to provide supervi- any specific hospital strategy (WHO 1994). Even where a well- sion and support to health workers at more peripheral levels of articulated strategy exists, decades of different political, social, care and to act as part of the regulatory mechanism, sometimes economic, and historical influences on health system develop- in both the public and the private sectors.Although this function ment result in great variability of district hospitals, both is likely to be an important means of developing and refining the between and within developing countries. Thus, some district referral system through two-way exchange of information and hospitals of 500 beds have a full complement of specialist con- of seeing that policy decisions are implemented, the ability of the sultants and access to a wide range of diagnostic and therapeu- health staff to fulfill this function is often extremely limited. tic services, while other hospitals of as few as 30 beds, but more Because resources are scarce, activities with the least tangible often 80 to 150 beds, are run almost entirely by medical assis- benefit--such as supervision and monitoring--are frequently tants and nurses, sometimes lack reliable power or water sup- abandoned, breaking important chains of communication. plies, and often offer few or no high-quality modern diagnostic services. This variability makes it daunting to extrapolate find- Health Information Systems. Many national health informa- ings from one setting to another and may seriously undermine tion systems rely on district hospitals to coordinate data col- the value of attempts to provide useful general descriptions of lection in the district. In theory, for a number of diseases the hospitals. In particular, when interpreting calculated costs of district hospital may be the only source of information, for care at a national or individual level, we must remember several example, for severe diseases such as neonatal tetanus, acute critical points: flaccid paralysis, or operative deliveries. The district hospital is, thus, a core data source supposedly providing burden-of- · Relevant data may often be missing or inadequately defined disease data at greater resolution than is commonly available at a country level. and at a meaningful administrative level if action is required. · Because a number of accepted ways of calculating costs However, in many developing countries health information exist, particularly at the level of individual interventions, systems are inadequate and inaccurate; staff members are not different methods are likely to lead to different estimates. equipped with the skills necessary to interpret data The particular design used to estimate costs should be con- (Loevinsohn 1993) and are often unaware of their local value, sidered when interpreting any results. thus depriving the local staff of essential planning and · In particular, a central feature of the hospital is that many of monitoring tools. Introducing an information culture and the its resources are used for more than one activity, so unit necessary skills and infrastructure to support such a transition, cost estimates depend crucially on how the costs of these although of potentially enormous value, presents significant resources are allocated among activities. challenges even for middle-income countries. · The relative prices of inputs can vary substantially between regions and countries. Formulating a Package of Services to Maximize · In the majority of cases, only the cost of care is reported Cost-Effectiveness without reference to outcomes so that the cost per unit of Interventions identified as being cost-effective in particular health benefit (however defined) is unknown. service areas or necessary to preserve the integrity of an effec- · Calculated costs usually reflect the care offered; it may not tive and equitable health system should be a part of a basic be the same as the care that is necessary, of an acceptable package of services and responsibilities at the district hospital quality, or most effective. level. However, the way in which these individual components · Cost estimates cannot indicate the extent of unmet need or are combined and integrated is also critical. Factors, including other sources of inequity. economies of scale and scope, whether gains or losses in effi- · The costs of care will depend to some extent on the severity ciency result from integration, and the influence of use and of illness of the patients and, for average costs per bed day, resource availability, will all have a profound influence on on the variety and relative proportions of different illnesses whether the district hospital itself is as cost-effective as the sum (the case mix). These areas are rarely commented on or of its parts suggests it should be. adjusted for. 1216 | Disease Control Priorities in Developing Countries | Mike English, Claudio F. Lanata, Isaac Ngugi, and others Levels of Provision of Hospital Care If just government health expenditure is considered, the Data on the levels of service provision for many developing available data suggest that hospitals at every level taken together countries are crude. In the absence of any more meaningful consume 50 to 60 percent of recurrent national health budgets, data, the number of beds is most often used as a (poor) substi- with the proportion appearing to increase as countries become tute. Bearing this weakness in mind, sources estimate the richer (Barnum and Kutzin 1993). If private expenditure on average number of total hospital beds to be 1.3 per 1,000 pop- health care (insurance and out of pocket) is included, the pro- ulation in developing countries (World Bank 2002), a figure portion of total health expenditure consumed by all hospitals probably declining in many developing countries (Hensher and falls to 30 to 50 percent of the total in developing countries others 1999), with varying estimates of the average number of (excluding South America) (Mills 1990a). Whereas these figures doctors from 0.5 per 1,000 population in low-income countries reflect total hospital sector expenditure, the limited data avail- generally (World Bank 2002) to 0.09 doctors per 1,000 popula- able suggest that district hospitals may receive less than 50 per- tion in Sub-Saharan Africa (Peters and others 2000). These cent of this total in many countries, consuming fewer resources estimates are considerably lower than the averages for beds than secondary and tertiary referral facilities (Mills 1990a). and doctors of 7.2 per 1,000 and 2.9 per 1,000, respectively, in high-income countries (World Bank 2002). Although these The Nongovernmental and Private Sectors estimates provide some indication of the major disparities in In many countries (especially in Africa) nongovernmental service provision between rich and poor countries, their value institutions, often religious organizations, are major health is limited. Lack of information on the relative distribution of service providers, and private physicians are often as numerous beds and staff by geographic zone, or between district and as those in the public sector. In Kenya, for instance, the number higher referral levels of care in a single country, and the fact of private and nongovernmental hospitals is equal to the num- that bed and staff numbers are probably a poor reflection of ber of public hospitals (Government of Kenya 2001), while in activity make these figures a poor substitute for data on patient Indonesia, 32 percent of hospital beds are private (Gani 1996). throughput and outcomes, statistics rarely available for district This potentially important contribution to the hospital sector hospitals. Furthermore, with the concentration on provision of may also be underrecognized, particularly in urban settings, service, the demand for services may often be ignored. It is still where multiple, small facilities may operate without registra- true in many countries that most deaths, presumably many tion, resulting in inaccurate local, regional, and national data preventable, occur at home and that many chronic diseases are on levels of overall service provision. Although few data exist inadequately treated. The need for hospital care is largely unde- on the effectiveness and quality of these hospitals, the belief is termined, but some have argued that the lack of provision of widespread that they may be more efficient than public sector district hospital care, in Sub-Saharan Africa at least, is a signif- hospitals. This belief is not necessarily borne out by the limited icant impediment to improving overall health status (Van data available (Bitran 1996), and concerns exist about the qual- Leberghe, de Bethune, and de Brouwere 1997). ity of care provided by private as well as public providers (Brugha and Zwi 1998). What Do District Hospitals Cost at a National Level? Although it has been argued for some time that hospitals con- District Hospital Efficiency sume too much of health sector budgets, thereby depriving pri- Data on hospital efficiency in developing countries are scant. mary care of adequate resources, it is surprisingly difficult to Considerable variability has been observed in the technical effi- identify how much hospitals cost in low- and middle-income ciency with which surgical services were provided in a small countries. Even where data exist on health expenditure, such number of Indian hospitals, with differences in total salary data are often at a highly aggregated national level and the costs being the main explanatory variable (Purohit and Rai functions that are included (clean water and sanitation, for 1992). Also in India, some evidence has been provided that example) are not always clear (World Bank 2002). nongovernmental hospitals may be more efficient, on average, Furthermore, whether private or nongovernmental expendi- than public hospitals, although considerable variability existed ture, capital expenditure, or the value of noncash inputs--such within both groups (Bhat, Verma, and Reuben 2001). In Kenya, as donations of equipment or volunteers' time--are included is public hospitals were found to have an average inefficiency rarely apparent. Add to this ambiguity the nearly impossible level of 30 percent (that is, the same resources could have problem of separating what is spent at different levels of the achieved a 30 percent increase in output) with significant health or hospital system--for example, to distinguish between contributing factors including shortage of appropriate district and referral hospitals--and it should be clear that we professional staff members, poor combinations of inputs currently have only a crude understanding of the costs of dis- (resources), nonfunctioning theaters and laboratories, lack of trict hospitals as a unit of service provision (Mills 1990a). transportation, irregular distribution of drugs and supplies, The District Hospital | 1217 and frequent breakdowns in medical equipment (Owino and Tanzanian nongovernmental hospital data indicate that the Korir 1997). All these data highlight the critical role of human average cost per inpatient day derived from 1995 reports resources, often a hospital's principal recurrent input cost (see (including expenditure on maintenance and expatriate the next section). Underinvestment in or absence of staff or salaries) would equate now to approximately US$3.60 (range inadequate flexibility in reallocating roles between different US$2.60 to US$6.00) in district hospitals (Flessa 1998). health worker groups may prevent hospitals from functioning However, if care had actually been provided according to the efficiently (Hensher 2001). standards defined by the provider (including recommended staffing levels, building maintenance, and equipment), the esti- mated cost per day would have risen to the equivalent of What Are the Costs of Providing Care in District Hospitals? US$11.60 (range US$9.20 to US$15.90) (Flessa 1998). This cost In a detailed review of actual hospital expenditure, Mills compares with costs reported in Kenya in 1993­94 (Kirigia, (1990b) identified two input categories that together accounted Fox-Rushby, and Mills 1998), adjusted to 2004 prices of actual for two-thirds or more of recurrent expenditure in almost all inpatient costs per day from two district hospitals of US$8.30 settings. Salaries varied between 20 and 80 percent and medical to US$10.10, and adjusted 1995 data from a district hospital in supplies between 15 and 58 percent of reported hospital expen- Bangladesh of US$15.90 (McCord and Chowdury 2003). In a diture. These and other data also suggest that, in many coun- middle-income country, South Africa, the cost per inpatient tries, costs of referral hospital care are often more than double day calculated between 1996 and 1998 and adjusted to 2004 the cost of equivalent care at district hospitals, although with- prices in five district hospitals ranged from US$37.80 to out knowledge on case mix or illness severity such data are hard US$96.30 (Daviaud and others 2000). These data do not neces- to interpret (Barnum and Kutzin 1993; Mills 1990b). More sarily reflect the cost of optimal care, and the Tanzanian study recent data collected from seven church-supported hospitals in demonstrates that even in externally supported hospitals actual Tanzania also demonstrate considerable variability in the pro- expenditure may be insufficient to provide good-quality care portion of costs attributable to salaries and supplies even within and cover essential maintenance, resulting in steady deteriora- a single organization in the same country (Flessa 1998). The tion of capital stock and worsening efficiency in the long term. strong dependence of hospital costs on salaries particularly cau- Data describing costs of treating some specific conditions in tions against generalizations across countries. district hospitals are summarized in table 65.1. Given the diffi- In the following analysis, all original U.S. dollar costs have culties in extrapolating data across contexts and the potentially been adjusted to represent the U.S. dollar cost in 2004. The significant effect of exchange rate fluctuations, great caution Table 65.1 Costs of Delivering Care at the District Hospital Level Cost (original data Country and year Item costed adjusted to 2004 US$) Comment Kenya, 1993­94, two district Treatment of inpatient severe malaria US$41.50 to US$132.00 per case Step-down approach to allocate hospitals, research study in children treated all costs, including capital costs Zimbabwe, 1994­95, three district Medical inpatient stay; HIV/AIDS Non-HIV: US$49.20 to US$110.00 Bottom-up and step-down hospitals,a research study care HIV: US$133.00 to US$217.00 per approaches used, including inpatient stay capital costs Zimbabwe, 1999, six provincial Severe malaria inpatient care; Severe malaria, mean costs per case Overhead costs purposefully hospitals, research study Pulmonary tuberculosis inpatient US$26.60 to US$49.90; tuberculosis, omitted; 1999 exchange rates careb median costs per case US$22.20 to US$61.00 Uganda, modeling based on Aspects of safe motherhood Eclampsia: actual US$63.40; Attempt to estimate current pro- 1997­99 data factoring in program delivered at hospitalc; actual and recommended US$127.00 gram costs and costs if program expansion recommended practices Cesarean: actual US$53.20; implemented as recommended; recommended US$57.80 excludes facility costs Prenatal care: actual US$2.90; recommended US$8.30 Sources: Kenya--Kirigia and others 1998; Zimbabwe 1994­95--Hansen and others 2000; Zimbabwe 1999--Hongoro and McPake 2003; Uganda--Weissman and others 1999. Note: Shaded rows provide data from studies that did not include overhead or facility costs. a. Only data from district hospitals are shown. b. All hospitals had a median length of stay for tuberculosis cases of 10 days or less. c. Only selected items are shown. 1218 | Disease Control Priorities in Developing Countries | Mike English, Claudio F. Lanata, Isaac Ngugi, and others Table 65.2 Estimate of the Effectiveness of a Kenyan District Hospital in Preventing Childhood Deaths in a Rural Community with Good Access to the Hospital Study site and population: Kenyan rural community with access to Population 51,183; 52 percent younger than age 15 basic primary health care services provided by five clinics, three private Surveillance period 1991­93 Service provider Kenyan Ministry of Health district hospital supplemented by research unit Mortality rates: Neonatal 31.5 per 1,000 live births Infant 58.3 per 1,000 live births Child 12.4 per 1,000 children ages one to four years Observed number of admissions 2,223 Admission rate 45 per 1,000 children ages 1 to 59 months per year Proportion of deaths occurring in the hospital: Neonatal 28 percent Ages 1 to 59 months 30 percent Observed number of deaths 134 Expected number of deaths without inpatient care 349 based on expert estimates for case fatality rates Lives saved 215 Estimated cost per life saveda US$104.40 Source: Snow and others 1994. a. 2004 US$ equivalent, using admission cost data from Kirigia and others 1998. The estimated cost of the admissions in 2004 US$ would be 2,223 10.1 US$22,452.30. This expenditure prevented 215 deaths; average cost of life saved therefore 22,452.30/215 US$104.40. should be used in interpreting these data, which, it should be using a slightly modified DALY (McCord and Chowdury 2003). noted, derive in all cases from specific research rather than Over a three-month period, the total costs (including all staff, routine sources. capital, and hotel costs) of running the hospital were calculated and divided by the estimated total number of DALYs gained attributable to inpatient care over the same three months. The Measuring the Effect and Cost-Effectiveness authors report an average cost per DALY of approximately of District Hospitals US$11.00 in 1995, or US$13.30 in 2004 dollars (McCord and In the previous section, some limited data on the costs associ- Chowdury 2003; see table 65.3). This figure compares favorably ated with provision of care at the district hospital were pre- with costs per DALY of many primary care interventions sented. What of a hospital's cost-effectiveness? Ideally we would regarded as highly cost-effective (World Bank 1993). To what like to know the aggregate health output of a hospital in terms extent these results depend on the quality of primary care, the of improved health compared with a situation in which there is referral system, the inpatient care, the hospital administration, no hospital. Such data do not exist, even from industrial coun- and the commitment of health personnel working for a small tries, where the hospital has been the subject of intense aca- independent nongovernmental organization will remain demic study. uncertain until more such data become available. However, some attempts have been made to estimate the effect of a hospital by comparing the observed outcome of ill- ness treated with hospital care to consensus expert opinion on FACTORS INFLUENCING DISTRICT HOSPITALS' the likely outcome of illness in the absence of hospital care. PERFORMANCE Using this approach in Kenya, Snow and others (1994) esti- mated that a well-functioning rural district hospital might The overall macroeconomic policy framework, as illustrated reduce all-cause child mortality by 44 percent in a population here with reference to financing mechanisms, may often be over- with reasonable access to the hospital (see table 65.2). looked as a considerable influence on hospital performance. For Extending this approach, researchers in a small rural hospital in the sake of simplicity, other factors (not exhaustively described Bangladesh calculated the benefit of hospital admission for and illustrated in figure 65.1) are discussed as primarily affect- patients of all ages suffering from life-threatening conditions ing the demand for hospital services or their supply and may The District Hospital | 1219 Table 65.3 Estimate of the Cost-Effectiveness of a Nongovernmental District Hospital in Rural Bangladesh Study site and population: Rural Bangladesh, with Population 160,000 community served by four peripheral clinics Surveillance period July through October 1995 Service provider Independent nongovernmental organization Major causes of death 74 percent under-five mortality attributable to perinatal deaths; maternal mortality ratio high Admissions analyzed 541 (33 percent obstetric/gynecological problems) DALYs gained by hospital services: Adult medical 177.0 life years; 6.5 disability years Surgical 459.4 life years; 236.3 disability years Pediatric 371.5 life years; 10.8 disability years Obstetric/gynecological 897.5 life years; 125.4 disability years Newborn (resulting from ob/gyn interventions) 1,024.3 life years Total DALYs gained 3,308.7 Cost per DALY US$10.93 in 1995 ($13.30 in 2004 US$) Source: McCord and Chowdury 2003. operate at both national and local levels. The way some of these However, they demand technical skill and accurate data at the diverse factors affect people's daily lives is illustrated in box 65.1. central level, especially if capitation payments are adjusted for What is clear is that failure to tackle these many challenges all too differences in population health status or other needs. often results in facilities that fail their communities. Financing based on activity levels (such as the diagnosis- related group methods in widespread use in high-income coun- tries) are similarly demanding of central-level capacity and also Central Financing Mechanisms require considerable competence and probity at more periph- Three broad methods of government financing of public eral levels of the administration. However, such financing might district hospitals are generally used: prospective with a fixed be an essential prerequisite of insurance-based mechanisms. In budget, prospective with revenue depending on activity, and contrast to fixed budgets, it has the potential for encouraging retrospective in proportion to actual costs. The fixed budget is supplier-induced demand--the greater the hospital's income, widely used, often based on historical spending levels, with a the more services it provides. It produces some incentive to (frequently inadequate) provision for price changes. Such a sys- reduce unit costs. Expenditure control may be difficult unless a tem clearly can secure good expenditure control and is admin- cap is put on the aggregate hospital sector budget. istratively undemanding. However, it can often perpetuate Retrospective reimbursement of actual costs is a discredited historical inequities and fail to respond to new demands and system of financing hospitals because it offers no incentive to priorities. Moreover, fixed budgets offer few incentives to max- control costs or manage demand. In its favor, it may stimulate imize the effectiveness, quality, or quantity of care offered by higher-quality care. In practice, many health systems use a mix- hospitals (Barnum, Kutzin, and Saxenian 1995). ture of all three payment mechanisms, with broadly fixed Indeed, many budget systems continue to finance hospitals budgets, sometimes adjusted for changes in demand, and some through line-item budgets directly from the ministry of health element of retrospective reimbursement for unplanned activity. or finance. Such mechanisms allow central bureaucracies to In general, no one strategy is perfect. However, the considerable exert the maximum level of control over peripheral spending demands on management for some schemes imply that a with little or no capacity at peripheral levels for flexible use of global budget, ideally based on population needs, in conjunc- funds in response to local needs. Thus, centralized budget sys- tion with some form of quality-monitoring system may be the tems can contribute to technical inefficiency by preventing most appropriate way forward for many developing countries local managers from optimizing the deployment of inputs. In (Barnum, Kutzin, and Saxenian 1995). contrast, global fixed budgets provide for central control of Mechanisms permitting local income generation (cost total spending but may permit increased independence when recovery, cost sharing, facility improvement funds, and local allocating funds at a local level. Fixed budgets based on capita- taxes) may be superimposed on any of these schemes. Such tion payments can be more sensitive to local needs than incre- devices can help countries shift toward a local, more needs- mental budgeting and can contribute toward equity objectives. based allocation of financing and help promote accountability 1220 | Disease Control Priorities in Developing Countries | Mike English, Claudio F. Lanata, Isaac Ngugi, and others Box 65.1 Hospital Performance: Perspectives from a Sub-Saharan African Country Caretaker (C) and health worker (HW) experiences of then you sell it before you go." (C) (Peterson and others hospital care: 2004). "When the doctor realized my child was breathless he "I spent a long time in MCH [Maternal and Child Health]; quickly called us into the office even though I was at the the doctor wanted money before he would see me, and I back in the queue." (C) did not have any." (C) "The [nursing] sister came and talked to me and asked if I "There is a lot of suffering when it comes to drugs because had a problem, and I felt good and cared for." (C) they are usually not enough and most of the time the "Things here have greatly improved; the ward is clean and mothers do not have money." (C) the treatment prompt. We are happy and hope that this "I want to know everything about the illness; I asked the will continue." (C) nurses, but they refused to explain, so I got disheartened "I admitted a patient in very poor condition with malaria from asking anyone." (C) and anemia and I managed to remove blood for cross- "I had a patient with anemia and mild marasmus, and the match and fix a line, start on oxygen, and get the doctor. mother waited for three hours in the lab for an Hb only to Blood was started quickly, and the child rapidly be turned away as she had no money. Then I went to get improved." (HW) the child some milk, and I was turned away as the store- "I resuscitated a baby with severe asphyxia, and it success- man said it was too late. The child had to wait until the fully came up. The success was because I had attended a next day." (HW) course in basic life support skills for neonates." (HW) "A child with severe LRTI [lower respiratory tract infec- tion] was very dyspneic on admission. Only one cylinder Caretaker and healthworker descriptions of referral to of oxygen was available, but we started giving it to the hospital: child, and the condition improved. The condition became "If you do not have the money, you have to look for it first. worse when the oxygen ran out, and there was none left; he Sometimes you may even have to spend a day or two look- started gasping and died." (HW) ing for the money for the treatment. If you have coffee, Source: English and others 2004a, unless otherwise noted. by focusing local attention on the efficiency and quality of local attention and hotel services (sometimes referred to as amenity services. This flexibility presupposes that those empowered beds) as a means of generating profit; reports indicate that with authority have the skills and freedom to make and execute the fees levied may not even cover the cost of the enhanced plans. The experience of such a decentralized policy on district service, let alone generate extra revenue with which to cross- hospital or district health system performance is mixed, with a subsidize services for the poor (Flessa 1998; Suwandono and lack of real transfer of authority reducing effectiveness in some others 2001). areas (Blas and Limbambala 2001), while more balanced and carefully implemented mechanisms of decentralization may be productive (Bossert and others 2003). Demand for Services The specific effects of requiring out-of-pocket payments to Patients' demand for services may be influenced by a wide access health care are a matter of fierce debate. Although some variety of factors, many of which have little to do with the hos- data suggest an improvement in allocative or technical efficien- pitals themselves. Patients' perceptions of the severity of their cy, other data do not (Arhin-Tenkorang 2000; Van der Geest illness, cultural beliefs, physical accessibility, and financial and and others 2000). It has been suggested that an improved qual- opportunity costs together with the performance of the periph- ity of service may overcome the cost barrier to access (Van der eral health unit screening process all potentially limit the effec- Geest and others 2000). However, the likelihood that the poor tiveness of the referral mechanism and thus the hospital (Font will be excluded from hospital care is a major concern. There is and others 2002; Siddiqui and others 2001). Recent data high- also an increasing tendency to encourage district hospitals to lighting the inability of many families to meet the financial provide some beds with an enhanced level of professional costs of hospital referral (Peterson and others 2004) and the The District Hospital | 1221 potentially catastrophic consequences of severe illness (Xu and continue to operate within a fixed budget, thereby satisfying others 2003) underscore the importance of financial barriers, finance ministries but having little or no effect on health. Long- especially for the poor. Not only are there obvious implications term underinvestment in facilities and skilled, motivated staff for health generally, but underusing service capacity also may then condemn a health system to many years of under- reduces efficiency and increases the costs per case of hospital performance, given the time necessary to address these issues. care. Improving the efficiency and effect of a hospital may, This is the fundamental reason for seeking to measure system therefore, be best achieved by tackling factors that influence outputs and quality as well as costs. demand--for example, providing emergency transport and On a regional or national scale, the actual distribution of limiting out-of-pocket expenses. However, often a concern hospitals and personnel may work for or against effective serv- exists that the provision of free high-quality services may itself ice delivery. For political reasons (to reward a community or to promote unnecessary demand--the so-called moral hazard. In honor a powerful politician, for instance), hospitals may be sit- addition, the relative importance of demand factors may vary uated in areas that would not be chosen if purely rational plans considerably in different settings, for example, in urban and had been followed. Nongovernmental providers or philanthro- rural areas, making universal rules unhelpful. pists may build or alter hospitals without regard to the overall In the context of PHC, it is suggested that high demand for function of a health system or achieving either equity or effi- services provided by hospitals rather than peripheral clinics, ciency. Public, private, and nongovernmental hospitals may driven by a perception that hospitals provide higher-quality compete for patients, potentially reducing efficiency in some service and resulting in bypassing of the PHC level of care, is or all sectors. The crisis of inadequate personnel in low- inefficient. It has been proposed that hospitals be specifically income countries, which limits the range, quality, and quantity prevented from delivering PHC services (WHO 1990). of services that can be offered, has been described However, the view that patients who bypass PHC increase the (Narasimhan and others 2004). However, imbalances in the costs to the provider may not always be true (Siddiqui and oth- within-country distribution of staff members are less well ers 2001). Patients may also choose to bypass the district hos- publicized and equally damaging. All the factors mentioned pital and proceed directly to referral hospitals, often increasing and others are commonly encountered in health systems of the costs of care if the condition could have been treated in the developing countries and are major barriers to implementing lower-level facility. The perceived quality of care at the district potentially valuable interventions at an operational level level may be a major determinant of this behavior, with some (Oliveira-Cruz, Hanson, and Mills 2001). New interventions data suggesting that improved district services increase use must therefore often be considered in the light of existing rates (Barnum and Kutzin 1993), potentially making district (rather than optimal) levels of service provision and perform- hospitals more cost-effective but more costly. ance. Little literature is available on these public choice features of decision making. The Supply of Services A fundamental role of policy makers is to determine the geo- EFFECTING CHANGE WITH CROSS-CUTTING graphical distribution of hospitals and the functions they INTERVENTIONS should undertake. These decisions are often severely circum- scribed by topography, historical accident, and political imper- So far this chapter has outlined concepts fundamental to atives, as well as by the level and quality of resources that are understanding the position, functions, and performance of the available. Often, changes can be made only incrementally, district hospital and has presented some of the existing building on an existing structure of administration and capital (though limited) data on costs and cost-effectiveness. that may not be in any sense optimal. Operating at the interface between primary care--aimed often Nevertheless, many of the factors determining the quality of at the poor--and the more Western biotechnological model of supply are theoretically under the influence of local manage- care at secondary and tertiary levels--often more accessible to ment personnel, who are in a potentially powerful position to the better off--district hospitals are easy to ignore because they significantly affect a hospital's function. Lack of resources, low lack any advocates for their role. However, optimizing their role morale, inability to attract staff members to hardship areas, to maximize health benefits and promote equity does demand poor training, and inadequate supervision among many other the following: factors may all conspire to prevent health workers from execut- ing their duties effectively or even at all. Those factors may, in · explicit policy decisions about the services that should be turn, result in less demand for services from consumers, who offered at this level and about the balance between primary opt to avoid the hospital or go elsewhere for treatment. The care, district hospital care, and higher-level care services paradox resulting from this decline is that the hospital may provided 1222 | Disease Control Priorities in Developing Countries | Mike English, Claudio F. Lanata, Isaac Ngugi, and others · national strategies on the distribution of services that evidence of benefit particularly when part of a broadly based encompass all providers approach (Grol and Grimshaw 2003). District hospitals in · commitment to provision and equitable distribution of developing countries have largely missed out on this revolu- essential human resources and supplies tion, which may be of particular value in settings where care by · systems for monitoring hospital performance in terms of nonspecialists with little or no access to recent information is efficiency and quality and for intervention when perform- the norm. ance is poor. Information and Integration When a framework defining the district hospital is available, interventions that might improve performance can be consid- Although much focus is given to technological development in ered. The focus here is on cross-cutting interventions rather the fields of diagnosis, treatment, and imaging, relatively little than condition-specific or service area­specific interventions attention is paid to the potential for technology to change the described elsewhere. Cross-cutting interventions seem to be collection and use of information, despite the possibly major rarely prioritized but have the ability to add value in many effect on improving administrative and clinical management. areas and are perhaps critical when thinking of developing an As at the primary care level, where many of the interventions improved health system. are currently available to achieve significant reductions in mor- tality (Claesen and others 2003), many of the tools that could be used to improve health are well known at the district hospi- Human Resources tal level. Making better use of these tools through more reliable Key issues that affect district hospitals are the quantity and provision, better training, improved information collection, quality of personnel and their range of skills. Staff members on-the-spot analysis of data, and real-time use of the results for should be appropriate to the tasks they are asked to perform. service planning might be both possible and of considerable This approach may mean continuing to use nursing or auxil- benefit (Cibulskis and Hiawalyer 2002). Clearly, how a hospital iary staff members with more limited training in district hos- is performing as part of an integrated primary care system is pitals because they may be more cost-effective, running against also vital. Local information on population health, on use and the tide of rising academic requirements often demanded by referral patterns, and on success and the reasons underlying professional associations (AED 2003). Similarly, devolving successes and failures is invaluable if the hospital is to respond some tasks to lower cadres of staff may be practical and much to the particular needs of its locality. more efficient--for example, training and licensing clinical assistants to perform emergency surgery including cesarean Quality Improvement and Accreditation section. Such initiatives, too, may face opposition from power- ful professional vested interests. Although some tasks may be Quality improvement is a generic technique adapted from transferred downward, a problem often faced by district hospi- industry that involves a rolling approach to identifying prob- tals is an absence of high-quality senior staff members or lead- lems, solving them, and assessing the results of change (see fig- ers. Traditionally, running a district hospital has commanded ure 65.3) and that has been institutionalized in hospital care in less respect and remuneration than work at a secondary or ter- many developed countries (DiPrete-Brown and others 1993). tiary facility and has been regarded as a stage to be moved An essential first step is defining standards for service provi- through as rapidly as possible. Arguably, the challenges to a sion, which can span all areas, including the technical content district hospital professional are at least as great as those of a of care, the physical environment in which care takes place, and tertiary consultant specialist, and the development of appro- interpersonal relations between patients and health workers. priate skills-training programs, and parity of postgraduate This approach is often linked to formal systems for external qualifications and pay, might help foster the development of a assessment of hospitals' performance and accreditation. professional group that improves performance and fills a much Accreditation may serve as a goal for participating hospitals, a needed advocacy role. means of promoting positive competition, and a means of identifying poorly performing institutions. Potential advan- tages of such initiatives are empowerment of local service Improving Clinical Management providers to solve problems they feel are important and the For more than a decade, industrial countries have increasingly overall aim of working toward a systemwide standard of care. promoted the use of the best evidence in clinical management. However, although an obvious need exists for quality improve- Clinical guidelines, means to implement them, feedback on ment in hospitals in developing countries (English and others their use and value, clinical audit, and performance review 2004b; Nolan and others 2000), few examples exist of hospital- are all now the subject of considerable research, with some level interventions in industrial or developing countries that The District Hospital | 1223 cially in Sub-Saharan Africa. Reuse of needles and blood trans- fusion are the main sources of infection and also carry the risk Identify the of hepatitis B and C and other viral infections important in problem and their own right. It has been estimated that effective measures to the targets for improvement improve blood safety in particular are a highly cost-effective intervention at approximately US$8 or less per DALY (Creese and others 2002). Investigate and Nosocomial infection, another major adverse consequence understand of admission to hospital, is common in some settings in indus- the causes of trial countries, contributing significantly to hospital costs. the problem Historically, relatively simple approaches to prevention have proven reasonably effective with additional effect from dedi- Develop a cated prevention services (Ayliffe and English 2003). The strategy for potential effect of intervention in district hospitals in develop- intervention ing countries is largely unknown, although in China nosoco- mial infection rates of between 8 and 13 percent have been reported (Barnum and Kutzin 1993). Because overcrowding Plan, collect and lack of basic resources, even water, are common in some baseline data, and Communicate intervene effect of districts, the potential for simple cost-effective interventions to intervention, modify, and, if needed, prevent such infections seems high. reevaluate Monitor that intervention is Other Managerial Initiatives taking place In high-income countries, numerous other initiatives are being Collect data Evaluate effect tested to promote improved efficiency and quality. They often about the of change by rely heavily on having in place appropriate institutional change in comparison process with baseline arrangements, managerial capacity, and information systems, so their feasibility for local implementation is highly dependent Check data are complete and accurate on local circumstances. One of the most widely tested arrange- and confirm intervention ments within public national health systems has been the executed experiment with internal markets, in which a range of public Source: Adapted from Massoud and others (2001). hospitals compete for contracts from separate public service purchasers, such as local governments. The split of purchaser Figure 65.3 Quality Improvement Process and provider of public services is recognized as a potentially powerful instrument for securing efficiency improvements but can be demanding in terms of managerial skills (Le Grand, provide evidence of effect on major outcomes. One exception Mays, and Mulligan 1998). is a broadly based quality improvement intervention targeting A less direct way of introducing some form of competition maternal and child health in Peru that focused on the entire into the hospital market is to require hospitals to publish per- system of care. This project was associated with a 25 percent formance reports that allow direct comparisons to be made decrease in maternal deaths in program areas (see box 65.2 between hospitals. for details). However, the relatively poor progress of an An alternative to relying on indirect methods of influencing operational-level quality improvement and accreditation pro- behavior is to give physicians incentives or instructions to gram in Zambia's hospitals highlights the significant problems deliver care in line with guidelines reflecting best practice. In the of intervening in countries with poorly functioning health sys- United States, numerous experiments have been carried out tems that are severely constrained by lack of resources under the general banner of managed care (Glied 2000), and (Bukonda and others 2002). other systems have attempted analogous approaches to hospital regulation. At one extreme is the centralized U.K. system of per- formance management, under which hospitals are given chal- Hospital-Acquired Disease lenging and immediate targets and are rated according to meas- Probably the most important infection in developing countries ured outcomes (Smee 2002). At the other extreme is the system that can be acquired as a result of hospital care is HIV, espe- of guided self-regulation practiced in the Netherlands, under 1224 | Disease Control Priorities in Developing Countries | Mike English, Claudio F. Lanata, Isaac Ngugi, and others Box 65.2 Prevention of Maternal and Child Deaths from Improvements in the Quality of Health Services: An Example from Peru Recognizing the failure of previous training attempts to Proportion of quality indicators achieved by Peruvian health facilities improve the quality of health services, the Ministry of Proportion of quality indicators achieved (percent) Health, with support from the U.S. Agency for 100 International Development and the participation of local 90 institutions, developed an innovative program in Peru. Aiming to reduce maternal and perinatal deaths, the pro- 80 gram expected to increase use of health services by 70 improving quality and by strengthening links between 60 the health services and their communities by working 50 with midwives and community health workers. 40 Multidisciplinary teams implemented a quality improve- ment program in approximately 2,500 health facilities, 30 focusing on 20 · standardizing care 10 · ensuring the availability at all times of essential sup- 0 plies and equipment 1S 2S 1A FA 1S 2S 1A FA 1S 2S 1A FA 1S 2S 1A FA 1S 2S 1A FA · making use of existing information systems and doing Change in quality indicators over time small operational studies to generate data at the local Use of data in decision making Participatory working practices level to facilitate decision making Patient satisfaction Implementation of standardized care · promoting the participation of all personnel in a con- Essential supplies certed and agreed-on plan of action Note: Proportion of quality indicators (with 95 percent Confidence Index) measured in · measuring patients' satisfaction over time and address- the five domains achieved by health facilities at first supervision visit (1S), second supervision visit (2S), first accreditation visit (1A), and final accreditation visit (FA). ing the causes of complaint. Training activity mainly involved use of a participatory problem-solving technique. In parallel, health networks in By the end of the three-year program (1996­99), each health region participated in a program to work with demand for health services had increased considerably, the 1,143 midwives and 2,549 community health workers, success itself creating managerial problems in many under the coordination of a health facility member who instances. Motivation and satisfaction of patients and was part of the multidisciplinary team. health workers had also increased, and revenue collected Supervision and evaluation at each facility occurred (through fee-for-service payment) at the facilities rose. three and six months after training and before accredita- Maternal mortality in the regions included in the program tion visits. A tiered accreditation system was developed to was 60 percent higher than in other regions at the start of promote participation and provide an incentive for the intervention period and fell 25 percent after the inter- improving quality. Results of each evaluation were pre- vention, while no change was observed in the other health sented to the Ministry of Health, which made accredita- regions. The inequitable distribution of maternal mortal- tion decisions through an independent institution to gen- ity was narrowed to a 20 percent excess in intervention erate political support. Quality in five areas (correspon- areas. A national demographic and health survey examin- ding to the program aims) was assessed. Significant ing Peru between 1995 and 2000 found a significant over- improvements were observed in the proportion of indica- all reduction of maternal mortality, increases in prenatal tors achieved in all five aspects of quality evaluated (box care coverage, and a higher proportion of deliveries in figure). An evaluation one year after the end of the pro- health facilities or attended by health professionals. gram found that performance had declined but remained at 60 to 80 percent of the levels achieved at accreditation. Sources: Lanata, Butron, and Espino 2002; Ministerio de Salud, Peru 2001. The District Hospital | 1225 which hospitals are required to engage in quality improvement respond to a rapid change in needs, and health care financ- but are given no prescription as to what format that effort might ing. Research that permits hospitals to tackle these new chal- take (Klazinga, Delnoij, and Kulu-Glasgow 2002). lenges and develop efficient and cost-effective strategies to provide care for HIV-related disease while preventing a THE FUTURE: RESEARCH AND decline in care standards for HIV-uninfected patients is a INFORMATION NEEDS high priority. Accounting for Case Mix and Case Severity When A few fundamental and urgent needs must be met as a prereq- Measuring Hospital Performance uisite to improving understanding of district hospitals in low- and middle-income countries, although tackling these issues Overall inpatient-fatality rates and case-fatality rates of differ- may be far from simple: ent common diseases are often included in district hospital performance measures. These are crude measures unless some · developing and accepting meaningful performance indica- adjustment is made for case mix when describing inpatient tors in conjunction with developing appropriate standards fatality and for severity of illness when describing case fatality. of care Alternatively, hospital outcomes should perhaps be replaced as · collecting higher-quality routine data from district hospitals key indicators of performance by carefully chosen process indi- · improving understanding of the costs and health conse- cators, which are likely to be more generalizable tools of per- quences of different, evidence-based, service provision formance monitoring that offer the advantage of specifically portfolios proposed for district hospitals and improving identifying areas that require improvement (Lilford and others understanding of the marginal benefits of incremental 2004). additions and their implications for planning infrastructure Implications of Emerging and Existing Technologies and estimating human resources and technology needs. Technology has had an enormous effect on the amount of A solution to the first issue would perhaps pave the way for information available to clinicians and managers in industrial and enhance the value of further focused research in a number countries, from new rapid diagnostic tests to automated stock- of areas. checking and ordering procedures. A particularly exciting potential in developing countries may be the ability to under- Implications of a Changing Disease Spectrum take and interpret many diagnostic tests remotely, thereby In many middle-income and some low-income countries, the enabling district hospitals to operate without a skilled diagnos- demographic transition to noncommunicable diseases-- tic staff on site. It also seems probable that appropriately tar- notably cardiovascular, smoking-related, and malignant dis- geted technology could have a major effect, not least in the eases--will have considerable implications for the hospital generation, communication, and analysis of hospital use, cost, sector. Thus, hospital costs likely will rise as older patients with and outcome data, without which the health system cannot chronic diseases become an increasing proportion of inpa- identify and respond to needs. tients (Barnum and Kutzin 1993). In some cases, the relative Interventions That Improve Performance cost-effectiveness of hospital care will improve compared with further expansion of primary or preventive services that incur Interventions aimed at improving hospital administration and increasing marginal costs (Barnum and Kutzin 1993). clinical management at the district hospital level warrant inves- More immediately, in low-income countries in Africa, the tigation. For clinical management, interventions such as clini- massive impact of the HIV pandemic is most easily seen in cal guidelines, supervision, and feedback; audit and continuing the continent's hospitals. Bed occupancy is rising, and hospi- professional development; quality improvement strategies and tal stays appear to be lengthening, as an increasing propor- accreditation; and improvements in referral and integration tion of hospital admissions, now over 50 percent in some with PHC may improve district hospital performance and be countries' medical wards, have HIV-related disease (Mpundu relatively cost-effective. Such interventions deserve attention, 2000). Those diseases associated with HIV infection, notably along with more traditional research aimed at optimizing treat- tuberculosis, and changing demands for care, such as the ment of specific diseases. need for palliation, may change not only the workload but also the nature of the demands placed on the service. The REFERENCES advent of antiretroviral therapy, which might ameliorate some of these problems, will itself place great demands on the AED (Academy for Educational Development). 2003. "The Health Sector Human Resource Crisis in Africa: An Issues Paper." Support for hospital service provision mechanisms. With or without new Research in Africa, U.S. Agency for International Development, drugs, HIV will continue to tax both planners, who have to Washington, DC. 1226 | Disease Control Priorities in Developing Countries | Mike English, Claudio F. Lanata, Isaac Ngugi, and others Arhin-Tenkorang, D. 2000. "Mobilizing Resources for Health: The Case for ------. 2004b."An Evaluation of Inpatient Paediatric Care in First Referral User Fees Revisited." Commission on Macroeconomics and Health, Level Hospitals in 13 Districts in Kenya." Lancet 363 (9425): 1948­53. World Health Organization, Geneva. Flessa, S. 1998. "The Costs of Hospital Services: A Case Study of Ayliffe, G., and M. English. 2003. Hospital Infection--from Miasmas to Evangelical Lutheran Church Hospitals in Tanzania." Health Policy and MRSA. Cambridge, U.K.: Cambridge University Press. Planning 13 (4): 397­407. Barnum, H., and J. Kutzin. 1993. Public Hospitals in Developing Countries. Font, F., L. Quinto, H. Masanja, R. Nathan, C. Ascaso, C. Menendez, and Baltimore, MD: Johns Hopkins University Press. others. 2002. "Paediatric Referrals in Rural Tanzania: the Kilombero Barnum, H., J. Kutzin, and H. Saxenian. 1995. "Incentives and Provider District Study: A Case Series." BMC International Health and Human Payment Methods." International Journal of Health Planning and Rights 2: 4. Management 10 (1): 23­45. Gani, A. 1996. "Improving Quality in Public Sector Hospitals in Bhat, R., B. Verma, and E. Reuben. 2001. Hospital Efficiency: An Empirical Indonesia." International Journal of Health Planning and Management Analysis of District and Grant in Aid Hospitals in Gujarat. Ahmedabad: 11 (3): 275­96. Indian Institute of Management. Glied, S. 2000. "Managed Care." In Handbook of Health Economics, ed. J. P. Bitran, R. 1996. "Efficiency and Quality in the Public and Private Sectors Newhouse and A. J. Culyer. Amsterdam: Elsevier. in Senegal." Health Policy and Planning 10 (3): 271­83. Government of Kenya. 2001. "Health Management Information Systems Report for the 1996 to 1999 Period." Ministry of Health, Republic of Black, R., S. Morris, and J. Bryce. 2003. "Where and Why Are 10 Million Kenya, Nairobi. Children Dying Every Year?" Lancet 361 (9376): 2226­34. Grol, R., and J. Grimshaw. 2003. "From Best Evidence to Best Practice: Blas, E., and M. Limbambala. 2001. "The Challenge of Hospitals in Health Effective Implementation of Change in Patients' Care." Lancet 362 Sector Reform: The Case of Zambia." Health Policy and Planning 16 (9391): 1225­30. (Suppl. 2): 29­43. Hansen, K., G. Chapman, I. Chitsike, O. Kasilo, and G. Mwaluko. 2000. Bobadilla, J. L., P. Cowley, P. Musgrove, and H. Saxenian. 1994. "The Costs of HIV/AIDS Care at Government Hospitals in "Design, Content and Financing of an Essential National Package of Zimbabwe." Health Policy and Planning 15 (4): 432­40. Health Services." Bulletin of the World Health Organization 72 (4): 653­62. Hensher, M. 2001. "Financing Health Systems through Efficiency Gains." Commission on Macroeconomics and Health Working Paper WG3:2, Bossert, T., O. Larranaga, U. Giedion, J. Arbelaez, and D. Bower. 2003. World Health Organization, Geneva. "Decentralization and Equity of Resource Allocation: Evidence From Columbia and Chile." Bulletin of the World Health Organization 81 (2): Hensher, M., N. Fulop, J. Coast, and E. Jefferys. 1999. "The Hospital of the 95­100. Future: Better out Than in? Alternatives to Acute Hospital Care." British Medical Journal 319 (7217): 1127­30. Brugha, R., and A. Zwi. 1998. "Improving the Quality of Private Sector Delivery of Public Health Services: Challenges and Strategies." Health Hongoro, C., and B. McPake. 2003. "Hospital Costs of High Burden Policy and Planning 13 (2): 107­20. Diseases: Malaria and Pulmonary Tuberculosis in a HIV High Prevalence Context in Zimbabwe." Tropical Medicine and International Bukonda, N., P. Tavrow, H. Abdallah, K. Hoffner, and J. Tembo. 2002. Health 8 (3): 242­50. "Implementing a National Hospital Accreditation Program: The Zambian Experience." International Journal for Quality in Health Care Kirigia, J., J. Fox-Rushby, and A. Mills. 1998. "A Cost Analysis of Kilifi and 14 (Suppl. 1): 7­16. Malindi Public Hospitals in Kenya." African Journal of Health Sciences 5 (2): 79­84. Canadian International Development Agency and the Aga Khan Kirigia, J., R. Snow, J. Fox-Rushby, and A. Mills. 1998."The Cost of Treating Foundation. 1981. The Role of Hospitals in Primary Health Care. Paediatric Malaria Admissions and the Potential Impact of Insecticide Karachi: Canadian International Development Agency and the Aga Treated Mosquito Nets on Hospital Expenditure." Tropical Medicine Khan Foundation Canada. and International Health 3 (2): 145­50. Cibulskis, R., and G. Hiawalyer. 2002. "Information Systems for Health Klazinga, N., D. Delnoij, and I. Kulu-Glasgow. 2002. "Can a Tulip Become Sector Monitoring in Papua New Guinea." Bulletin of the World Health a Rose? The Dutch Route of Guided Self-Regulation toward a Organization 80 (90): 752­58. Community Based Integrated Health Care System." In Measuring Up: Claeson, M., D. Gillespie, H. Mshinda, H. Troedsson, C. G. Victoria, and Improving Health Systems Performance in OECD Countries, ed. P. Bellagio Study Group on Child Survival. 2003. "Knowledge into Action Smith. Paris: Organisation for Economic Co-operation and for Child Survival." Lancet 362 (9380): 323­27. Development. Creese, A., K. Floyd, A. Alban, and L. Guinness. 2002. "Cost-Effectiveness Lanata, C. F., B. Butron, and S. Espino. 2002. Mejorando la calidad de la of HIV/AIDS Interventions in Africa: A Systematic Review of the atención de salud en el Perú--El Programa de Capacitación Materno Evidence." Lancet 359 (9318): 1635­42. Infantil (PCMI) en su primera etapa: Cómo se hizo y a qué se debió su Daviaud, E., B. Engelbrecht, P. Molefakgotia, G. Crisp, D. Collins, éxito? Lima: Instituto de Investigación Nutricional. and P. Barron. 2000. "South African Health Report--District Le Grand, J., N. Mays, and J. Mulligan, eds. 1998. Learning from the NHS Health Expenditure Reviews." South African Ministry of Health, Internal Market. London: King's Fund Institute. Pretoria. Lilford, R., M. Mohammed, D. Spiegelhalter, and R. Thompson. 2004. "Use DiPrete-Brown, L., L. Franco, N. Rafeh, and T. Hatzell. 1993. Quality and Misuse of Process and Outcome Data in Managing Performance Assurance of Health Care in Developing Countries." Bethesda, MD: of Acute Medical Care: Avoiding Institutional Stigma." Lancet 363 Quality Assurance Project. (9415): 1147­54. Duke, T., L. Willie, and J. Mgone. 2000. "The Effect of Introduction of Loevinsohn, B. 1993. "Data Utilization and Analytical Skills among Mid- Minimal Standards of Neonatal Care on In-Hospital Mortality." Papua Level Health Programme Managers in a Developing Country." and New Guinea Medical Journal 43 (1­2): 127­36. International Journal of Epidemiology 23 (1): 194­99. English, M., F. Esamai, A. Wasunna, F. Were, B. Ogutu, A. Wamae, and Massoud, R., K. Askov, J. Reinke, L. Franco, T. Bornstein, E. Knebel, and C. others. 2004a. "Delivery of Paediatric Care at the First Referral Level in MacAuley. 2001. A Modern Paradigm for Improving Healthcare Quality. Kenya." Lancet 364 (9445): 1622­29. Bethesda, MD: Quality Assurance Project. The District Hospital | 1227 McCord, C., and Q. Chowdury. 2003. "A Cost Effective Small Hospital in Smee, C. 2002. "Improving Value for Money in the United Kingdom Bangladesh: What It Can Mean for Emergency Obstetric Care." National Health Service: Performance Measurement and International Journal of Gynecology and Obstetrics 81 (1): 83­92. Improvement in a Centralised System." In Measuring Up: Improving McKee, M., and J. Healy. 2002. "The Significance of Hospitals." In Hospitals Health Systems Performance in OECD Countries, ed. P. Smith. Paris: in a Changing Europe, ed. M. McKee and J. Healy. Buckingham, U.K., Organisation for Economic Co-operation and Development. and Philadelphia: Open University Press. Snow, R., V. Mungala, D. Forster, and K. Marsh. 1994. "The Role of the Mills, A. 1990a. "The Economics of Hospitals in Developing Countries-- District Hospital in Child Survival at the Kenyan Coast." African Part 1: Expenditure Patterns." Health Policy and Planning 5 (2): 107­17. Journal of Health Sciences 1 (2): 11­15. Mills, A. 1990b. "The Economics of Hospitals in Developing Countries-- Suwandono, A., A. Gani, S. Purwani, E. Blas, and R. Brugha. 2001. "Cost Part 2: Costs and Sources of Income." Health Policy and Planning 5 (3): Recovery Beds in Public Hospitals in Indonesia." Health Policy and 203­18. Planning 16 (Suppl. 2): 10­18. Ministerio de Salud, Peru. 2001. "La Experiencia del PCMI." Ministerio de Van der Geest, S., M. Macwan'gi, J. Kamwanga, D. Mulikelela, A. Mazimba, Salud, Lima. and M. Mwangelwa. 2000. "User Fees and Drugs: What Did the Health Sector Reforms in Zambia Achieve?" Health Policy and Planning 15 (1): Mpundu, M. 2000. "The Burden of HIV/AIDS on the Zambian Health 59­65. System." AIDS Analysis Africa 10 (5): 6. Van Lerberghe, W., X. de Bethune, and V. de Brouwere. 1997. "Hospitals Narasimhan, V., H. Brown, A. Pablos-Menedez, O. Adams, G. Dussault, G. in Sub-Saharan Africa: Why We Need More of What Does Not Work Elzinga, and others. 2004. "Responding to the Global Human as It Should." Tropical Medicine and International Health 2 (8): Resources Crisis." Lancet 363 (9419): 1469­72. 799­808. Ngoc Anh, N., and T. Tram. 1995. "Integration of Primary Health Care Van Lerberghe, W., and K. Pangu. 1988. "Comprehensive Can Be Effective: Concepts in a Children's Hospital with Limited Resources." Lancet 346 The Influence of Coverage with a Health Centre Network on the (8972): 421­24. Hospitalisation Patterns in the Rural Area of Kasongo, Zaire." Social Nolan, T., P. Angos, A. Cunha, L. Muhe, S. Qazi, E. A. Simoes, and others. Science and Medicine 26 (9): 949­55. 2000. "Quality of Hospital Care for Seriously Ill Children in Less- Weissman, E., O. Sentumbwe-Mugisa, A. Mbonye, E. Kayaga, S. Kihuguru, Developed Countries." Lancet 357 (9250): 106­10. and C. Lissner. 1999. Uganda Safe Motherhood Programme Costing Oliveira-Cruz, V., K. Hanson, and A. Mills. 2001. "Approaches to Study. Geneva: Ministry of Health Uganda and World Health Overcoming Health System Constraints at the Peripheral Level: Review Organization. of the Evidence." Commission on Macroeconomics and Health, World WHO (World Health Organization). 1987. "Hospitals and Health for All." Health Organization, Geneva. Report of the WHO Expert Committee on the Role of Hospitals at the Owino, W., and J. Korir. 1997. Public Health Sector Efficiency in Kenya: First Referral Level, WHO, Geneva. Estimation and Policy Implications. Nairobi: Institute of Policy Analysis ------. 1990. The Role of the Hospital in the District--Delivering or and Research. Supporting Primary Health Care. Geneva: WHO. Peters, D., A. Elmendorf, K. Kandola, and G. Chellaraj. 2000. ------. 1992. "The Hospital in Rural and Urban Districts." Report of a "Benchmarks for Health Expenditures, Services and Outcomes in WHO Study Group on the Functions of Hospitals at the First Referral Africa during the 1990s." Bulletin of the World Health Organization Level, WHO, Geneva. 78 (6): 761­69. ------. 1994. A Review of Determinants of Hospital Performance. Geneva: Peterson, S., J. Nsugwa-Sabiti, W. Were, X. Nsabagasani, G. Magumba, J. WHO. Nambooze, and G. Mukasa. 2004. "Coping with Paediatric Referral-- Ugandan Parents' Experience." Lancet 363 (9425): 1955­56. ------. 1999. The World Health Report--Making a Difference. Geneva: WHO. Posnett, J. 2002. "Hospitals in a Changing Europe." In European Observatory on Health Care Systems, ed. M. McKee and J. Healy. ------. 2000a. Management of the Child with a Serious Infection or Severe Buckingham, U.K., and Philadelphia: Open University Press. Malnutrition. Geneva: WHO. Purohit, B., and V. Rai. 1992. "Operating Efficiency in Inpatient Care: ------. 2000b. The World Health Report: Health Systems: Improving An Exploratory Analysis of Teaching Hospitals in Rajasthan, Performance. Geneva: WHO. India." International Journal of Health Planning and Management 7: World Bank. 1993. World Development Report 1993. Oxford, U.K.: Oxford 149­62. University Press. Robertson, M., and E. Molyneux. 2001. "Triage in the Developing World-- ------. 2002. World Development Indicators. Washington, DC: World Can It Be Done?" Archives of Disease in Childhood 85 (3): 208­13. Bank. Sachs, J. D. 2001. "Macroeconomics and Health: Investing in Health for Xu, K., D. Evans, K. Kawabata, R. Zeramdini, J. Klavus, and J. Murray. 2003. Economic Development." Commission on Macroeconomics and "Household Catastrophic Health Expenditure: A Multicountry Health, World Health Organization, Geneva. Analysis." Lancet 362 (9378): 111­17. Siddiqui, S., A. Kielmann, M. Khan, N. Ali, A. Ghaffar, U. Sheikh, and Z. Mumtaz. 2001. "The Effectiveness of Patient Referral in Pakistan." Health Policy and Planning 16 (2): 193­98. 1228 | Disease Control Priorities in Developing Countries | Mike English, Claudio F. Lanata, Isaac Ngugi, and others Chapter 66 Referral Hospitals Martin Hensher, Max Price, and Sarah Adomakoh The appropriate allocation of resources to referral hospitals of "pure" cost-effectiveness analysis is used to determine an within a national health system has long been a controversial appropriate or optimal resource allocation for referral hospital issue in health system planning in developing countries. services, several problems arise. To begin with, hospitals have Consensus appears to be widespread that referral hospitals complex economies of scope and scale. At the point when hos- consume an excessive share of health budgets and that their pitals offer a range of cost-effective interventions, the marginal contribution to improving health and welfare is low relative to cost-effectiveness of additional interventions may be much the expenditure on these facilities, but the literature does not greater than would be the case if these other interventions were indicate what percentage of budgets should ideally be allocated evaluated in isolation. Yet a standard disease-specific analysis of to referral hospitals. Presumably, except in the poorest coun- interventions would rarely be able to calculate the marginal tries, some referral facility is needed, but how much is required, costs of referral hospital­based interventions. Similarly, impor- and how should the proportion allocated to referral facilities tant and complex interdependencies exist between services and vary with increasing levels of health expenditure and health specialties within referral hospitals that may be almost impossi- system sophistication? ble to capture adequately using a cost-effectiveness analysis. One approach would be to review how much countries at A further limitation to a cost per DALY approach arises different levels of gross domestic product (GDP) currently because referral hospitals produce multiple outputs, many of spend on referral hospitals. However, as explained later, the def- which contribute so indirectly to DALYs that they cannot be inition of referral hospital varies widely; therefore, analyses of compared directly to individual health interventions, but national health accounts and studies of expenditure are rarely which are critical to the functioning of the health system. For comparable. Thus, although the chapter summarizes the litera- example, referral hospitals are arguably essential to the training ture on expenditure on referral hospitals, this evidence cannot of doctors, particularly specialists. If a country can justify train- provide guidelines for policy makers. ing its own doctors, then it must have a referral hospital. Yet the A second approach might be to undertake a detailed analysis value of this output in terms of DALYs probably cannot be cal- of the role of referral hospitals in treating disease to derive their culated. Indeed, many of the functions of a referral hospital contribution to total disability-adjusted life years (DALYs). A occur outside the hospital itself and involve enabling and facil- simple analysis of the cost-effectiveness of specific interventions itating the effective functioning of lower-level health services. offered by referral hospitals might allow the selection of those Although the referral hospital's contribution may constitute interventions that are justified given their marginal cost per only a small fraction of the total cost of an intervention pro- DALY gained. Those interventions, multiplied by expected vided at a lower level of care (which may perhaps be viewed as demand, would then be aggregated to give a total optimal allo- a fixed cost of the health system), the referral hospital's role cation for referral hospital services. This approach is precisely may nevertheless be essential, thereby justifying a considerable the one used for evaluating and prioritizing disease-specific premium on its valuation above and beyond the cost per DALY interventions throughout this volume. However, when this kind of the care directly provided within the hospital's own walls. 1229 Finally, strong arguments can be made that cost-effective- who are better equipped or specially trained to guide them in ness analysis fails to capture important dimensions of the indi- managing or to take over responsibility for a particular episode vidual utility--and thus the social welfare--that accrues from of a clinical condition in a patient (Al-Mazrou, Al-Shehri, and the provision of health services, especially those relating to Rao 1990). Furthermore, higher-level hospitals in developing high-cost and low-frequency conditions. countries do not treat only referred patients; tertiary hospitals We are, therefore, highly skeptical about the feasibility of are frequently the first point of contact with health services for proposing a formulaic and purely quantitative response to the many patients. question of how to achieve an appropriate allocation of Differentiating referral hospitals from district hospitals, resources to the referral hospital level. Although perhaps unsat- therefore, requires consideration of the different resources used isfying for some readers, this chapter attempts instead to provide by different levels of hospital. Such a differentiation will tend to an overview of the critical features of and challenges relating to revolve around three features--the availability of increasingly referral hospital care in developing countries and a guide to the specialized personnel, of more sophisticated diagnostic tech- many issues that decision makers face in setting policy for this nologies, and of more advanced therapeutic technologies-- level of care. We suggest that planners need to adopt a far more that permit the diagnosis and treatment of increasingly com- qualitative and intuitive approach to deciding on the appropri- plex conditions. ate allocation of resources for referral hospitals than for other This volume, including this chapter, uses a standard defini- health care interventions. Such an approach is informed by a tion of hospital levels (Mulligan and others 2003). Table 66.1 more extensive listing of the roles of referral hospitals and their presents some of the commonly used alternative terminology direct and indirect benefits and costs to society.We acknowledge for different levels of hospitals. Note that this chapter deals only that analysis of the value of referral hospitals is bedeviled by the with general--that is, multispecialty--secondary and tertiary fact that, when judged empirically, they do not work as they are hospitals. Specialized hospitals, such as psychiatric, substance supposed to. The chapter, therefore, considers the key problems abuse, tuberculosis, infectious diseases, and rehabilitation hos- faced in the real environment in which referral hospitals operate pitals, clearly have important roles to play in a well-functioning in poor countries before reviewing what needs to be done to referral system. However, they are attended by specific features improve their functioning, drawing in particular on the authors' and challenges, account for a relatively small share of overall knowledge of South Africa and the Caribbean. resources, and operate in a significantly different manner than general hospitals do. DEFINITION AND CHARACTERISTICS FUNCTIONS AND BENEFITS Any hospital, including a district hospital, will receive referrals from lower levels of care. Indeed, referral can be defined as any The functions of referral hospitals may broadly be categorized process in which health care providers at lower levels of the into (a) the direct clinical services provided to individual health system, who lack the skills, the facilities, or both to man- patients within the hospital and the community and (b) a set of age a given clinical condition, seek the assistance of providers broader functions only indirectly related to patient care. Table 66.1 Definitions and Terms for Different Levels of Hospital Disease Control Priorities Project: terminology and definitions Alternative terms commonly found in the literature Primary-level hospital: few specialties--mainly internal medicine, District hospital obstetrics and gynecology, pediatrics, and general surgery, or just general Rural hospital practice; limited laboratory services available for general but not Community hospital specialized pathological analysis General hospital Secondary-level hospital: highly differentiated by function with 5 to Regional hospital 10 clinical specialties; size ranges from 200 to 800 beds; often referred Provincial hospital (or equivalent administrative area such as county) to as a provincial hospital General hospital Tertiary-level hospital: highly specialized staff and technical equipment-- National hospital for example, cardiology, intensive care unit, and specialized imaging Central hospital units; clinical services highly differentiated by function; could have Academic or teaching or university hospital teaching activities; size ranges from 300 to 1,500 beds Source: Definitions from Mulligan and others 2003, 59. 1230 | Disease Control Priorities in Developing Countries | Martin Hensher, Max Price, and Sarah Adomakoh Range of Clinical Services Provided In South Africa, the National Department of Health is The primary function of the referral hospital is to provide attempting to improve the quality and accessibility of referral complex clinical care to patients referred from lower levels; hospital services through development plans that will try to however, no agreed international definition exists of which spe- ensure that hospitals at each level move toward providing a cific services should be provided in secondary or tertiary hos- comprehensive set of clinical services (National Department of pitals in developing countries. The exact range of services Health, South Africa 2003). The department has developed a offered tends to vary substantially, even between tertiary hospi- target template of services (table 66.2) for regional (secondary) tals within the same country, as much because of historical hospitals, tertiary hospitals, and so-called national referral accident as deliberate design. services (which will be offered at only a small number of the Table 66.2 Target Service Configurations by Level of Referral Hospital, South Africa Specialist services Components Specialist services Components available on site explicitly included available on site explicitly included Regional (secondary) hospitals Anesthetics -- Mental health (psychiatry Acute inpatient and outpatient treatment Diagnostic radiology X-ray, CT scan, ultrasound, fluoroscopy and psychology) Child and adolescent psychiatry General medicine Echocardiography, stress electrocardiogram Electroconvulsive therapy Specialist immunology nurse Liaison psychiatry Regional intensive care unit Satellite clinics Diabetes, endocrine clinic Obstetrics and gynecology Emergency obstetrics and gynecology Gastroenterology, including endoscopy, Ultrasound, prenatal diagnosis proctoscopy, sigmoidoscopy, colonoscopy Kangaroo mother care (with general surgery) Basic urogynecology Geriatric care Orthopedic surgery General orthopedic surgery Genetic nurse and counseling 24-hour trauma service, accident and Oncology palliation and basic care emergency Neurology basic care Pediatrics Neonatal low and high care Spirometry and oximetry General pediatric medicine service Basic rheumatology General pediatric surgery (general surgeon) General surgery Regional burns service Rehabilitation center Physiotherapy, occupational therapy, 24-hour trauma service, accident and orthotics and prosthetics, speech therapy, emergency dietetics, podiatry Acute rehabilitation team Tertiary hospitals Anesthetics -- General medicine As regional plus: Burns unit Specialized burns intensive care unit and Angiography operating theater Coronary care Clinical pharmacology -- Echocardiography, stress electrocardiogram specialist Endoscopy, proctoscopy, sigmoidoscopy, Critical care and intensive Full intensive care unit service colonoscopy (with general surgery) care unit Genetic nurse and counseling Dermatology Inpatient and ambulatory treatment Oncology palliation and basic care Diagnostic radiology X-ray, multislice CT scan, ultrasound, General surgery Complex and high-acuity care fluoroscopy, mammography, color Doppler ultrasound Infectious diseases -- Ear, nose, and throat surgery -- Mental health (psychiatry Child and adolescent psychiatry, old-age and psychology) psychiatry, forensic psychiatry, substance Gastroenterology -- abuse treatment, liaison psychiatry, treatment for eating disorders, inpatient psychotherapy, social psychiatry, acute psychotic (complicated) care, acute nonpsychotic (complicated) care (Continues on the following page.) Referral Hospitals | 1231 Table 66.2 Continued Specialist services Components Specialist services Components available on site explicitly included available on site explicitly included Tertiary hospitals (continued) Neonatology Neonatal intensive care unit Rehabilitation center Physiotherapy, occupational therapy, orthotics Nephrology Tertiary dialysis and nephrology service and prosthetics, speech therapy, dietetics, podiatry, audiology Obstetrics and gynecology As regional plus: service Acute rehabilitation team, including spinal beds Fetal and maternal medicine Stroke unit Ophthalmology -- Orthopedic surgery Subspecialty orthopedics Respiratory medicine -- Pediatric intensive care unit Full pediatric intensive care unit Trauma Tertiary major trauma center (protocol-based transfer only, no walk-in accident and Pediatric medicine Specialist general pediatricians emergency service) Pediatric surgery Specialist pediatric surgery service Urology -- Plastic and reconstructive -- Vascular surgery -- surgery National referral services Cardiology Echocardiography, ultrasound, electrocardiog- Obstetrics and gynecology Oncology raphy, stress testing, Holter pacemaker service Urogynecology follow-up, catheterization laboratory, Reproductive medicine electrophysiology, ablation Orthopedic surgery Orthopedic oncology Cardiothoracic surgery -- Pediatric cardiology -- Clinical immunology -- Pediatric endocrinology -- Cranio-maxillofacial surgery -- Pediatric gastroenterology -- Critical care and intensive Additional intensive care unit capacity care unit Pediatric hematology and -- oncology Diagnostic radiology MRI Pediatric infectious diseases -- Endocrinology -- Pediatric intensive care unit Additional pediatric intensive care Genetics -- unit capacity Geriatrics -- Pediatric nephrology Dialysis and renal transplant Hematology -- Pediatric neurology -- Medical and radiation -- Pediatric respiratory medicine -- oncology and allergology Neurology -- Renal transplant Renal transplant unit Neurosurgery -- Rheumatology -- Nuclear medicine -- Urology -- Source: National Department of Health, South Africa 2003. -- not available. largest tertiary hospitals). Although certainly not directly appli- heightened awareness about the important role of the hospital cable to all developing countries, the template does give a help- in reducing incidence and preventing disease outbreaks. For ful picture of how services "build up" from one level of care to example, hospitals scaled up services to prevent mother-to- another, and it can be used as a starting point for considering child transmission and initiated follow-up clinics for mothers the situation in different countries. and babies. In Barbados, the main hospital scaled up voluntary counseling and testing services to address the prevention of horizontal transmission from mothers to their partners, with Clinical Services within the Community positive outcomes. The program also served to increase access Referral hospitals may perform a number of functions that to obstetric services at the primary health care level because of provide population-level health benefits through direct the screening campaign initiated through the hospital's preven- involvement in public health interventions. Responding to the tion of mother-to-child transmission program (Adomakoh, HIV/AIDS epidemic in Latin America and the Caribbean has St. John, and Kumar 2002). 1232 | Disease Control Priorities in Developing Countries | Martin Hensher, Max Price, and Sarah Adomakoh Referral hospitals often prove to be a highly effective focal manage patients at the lowest level of care possible. Too often, point for disease-specific health promotion and education personnel in referral hospitals adopt an insular and inward- activities. Bermuda's diabetes education program serves all lev- looking perspective, focusing exclusively on the patients els of care and provides a strong link between the primary, directly under their care. However, referral hospitals should secondary, and tertiary health care levels. The program is cen- offer significant support to personnel in lower-level facilities, tered in the main referral hospital and serves not only and specialist staff members should ideally spend a significant diagnosed patients but also families at risk. Overall, hospitals in portion of their time providing advice and support beyond the Caribbean are recognizing that central coordination of the walls of their own hospital, either in person or through public health programs within hospitals can provide benefits various modes of telecommunication. Even in poor countries, by strengthening coordination with other services. a steady improvement in communications infrastructure means that such support functions should become easier to provide over time. Key dimensions of this support function Valuing the Benefit of Clinical Services include the following: Measuring the improvement in an individual's health status produced by the combined activities of a referral hospital, · availability by telephone or e-mail to advise referring prac- whether for patient care in the hospital or for population-based titioners on whether referral is required programs, would theoretically be possible, although practically · specialist advice to the patient's local practitioner on post- and methodologically demanding. To our knowledge, such an discharge care effort has not been attempted at the referral hospital level, · specialist advice on the long-term management of chronic though two studies have attempted to proxy the effect of conditions hospital interventions on health outcomes for small district · specialist attendance at lower-level facilities to provide regu- hospitals, focusing on survival only (McCord and Chowdhury lar outreach clinics 2003; Snow and others 1994). Both studies indicate that district · provision of expert diagnosis or consultation through hospitals appear to have a significant positive effect on health telemedicine outcomes. · coordination of discharge planning between levels of care Large numbers of patients receive care in referral hospitals, · coordination of the development of and training in the use and most survive with their suffering alleviated, having gained of shared care protocols and referral protocols substantial benefit from the care they receive. Therefore, the · provision of technology support by skilled technicians and aggregate direct personal health benefits from referral hospital scientists. care will almost certainly be high. The question of whether referral hospital care is cost-effective relative to other interven- Quality Assurance and Quality Improvement. Referral hos- tions delivered at lower levels of care is less easy to answer in pitals can and do play a pivotal role in quality assurance and aggregate. By its nature, appropriate care in a referral hospital improvement. The most important mechanism for quality will tend to require more complex input mixes and higher skill assurance and improvement is through the training that refer- levels and, hence, will be relatively expensive. Analysis of the ral hospitals provide. The other key mechanism is through the costs and cost-effectiveness of individual interventions offered setting of standards for treatment. For example, experts at at different levels is tackled directly by the disease-specific referral hospitals should review evidence of effectiveness and chapters in this volume. cost-effectiveness applicable to the local context, determine the formularies to be used at each level of the health system, and Wider Activities and Functions develop and amend treatment protocols. Referral hospitals can Aside from direct patient care, referral hospitals serve other improve the quality of peripheral services by giving advice, functions within the health system, some of which are offered offering on-site training, providing clinical services alongside within the facility, such as teaching and research, while others local practitioners, and monitoring the quality of the referrals reach out to the lower levels of the health services, such as tech- they receive. nical support and quality assurance. Education and Training. Many tertiary referral hospitals in Advice and Support to Lower Levels. The referral process developing countries are associated with universities and med- does not simply entail transferring a patient from a lower to a ical schools and may, therefore, also be regarded as teaching hos- higher level of care, nor does it end when a patient is dis- pitals.Any country wishing to train its own doctors will need one charged from a referral hospital. An effective referral system or more teaching hospitals. The number of doctors a country requires good communication and coordination between lev- needs will be influenced by its level of development, resources, els of care and support from higher to lower levels to help and personnel structure. Many will aim for a ratio of at least Referral Hospitals | 1233 Box 66.1 How Many Medical Students Should Be Trained Per 1 Million Population? In a steady state (that is, the number of doctors being the number needing to be trained is 1,000/40 25 per produced is equal to the number retiring from practice), year. If 30 percent of doctors leave the country or leave and if we assume that doctors practice, on average, for medical practice within 8 years of qualifying, then each 40 years after qualifying, the total number practicing will graduate, on average, contributes 30 years of service, and equal the number graduating in 1 year multiplied by 1,000 practicing doctors (1,000/30) 33 must qualify 40 years. If a population of 1 million needs 1,000 doctors, each year. Source: Authors. 2 per 1,000 population, though most developing countries have developed countries, 60 to 90 percent of doctors are specialists, 0.05 to 1.0 per 1,000 (Puzin 1996; WHOSIS 2004). If we whereas in developing countries the range is wider (for exam- assume a 40-year working life and loss through brain drain or ple, 76 percent of Indian doctors are specialists, 45 percent are other attrition of 25 percent, the number of doctors that must specialists in Tanzania, and 31 percent are specialists in be produced each year is between 16 and 67 per 1 million pop- Morocco). A World Health Organization expert workshop ulation, resulting in 0.5 to 2.0 doctors per 1,000 population agreed on a figure of 50 percent (Puzin 1996). Therefore, a (box 66.1). A population of 40 million would, thus, need med- country of 40 million would aim to train approximately 300 to ical schools able to graduate between 640 and 2,680 doctors per 1,300 specialists per year. On average, such training lasts four year. Medical schools possess economies of scale, and although years. Thus, at any time the academic referral hospital system some extremely small schools train 50 or so students a year, would need to supply 1,200 to 5,200 residents. A guideline agreement is widespread that a class size of about 150 to 200 is many countries use requires a ratio of postgraduate specialist optimal (see, for example, Harden and Davis 1998). A country supervision of not more than two residents per qualified spe- with fewer than 3 million population would really need to con- cialist. This ratio can be used to get some idea of the referral sider whether training doctors locally is justified on economic hospital capacity required to train specialists. and other grounds, but for larger countries, the arguments for Although basic doctors could spend most of their training training doctors locally are strong, and a teaching hospital time in primary care and district hospital facilities, with limited would, therefore, be required. exposure to tertiary care hospitals, the training of specialists-- Basic generalist doctors should be trained in a range of facil- as well as of other specialized allied staff members such as nurses ities across all levels of care, reflecting the facilities in which for intensive care or specialized psychiatry, physiotherapists they will work after graduation. Traditional approaches toward specializing in back injuries or burns, and pharmacists special- medical education have been widely criticized by educational- izing in oncology--can take place only in referral hospitals. ists and health planners for being dominated by training in ter- In recent years, continuing medical education has grown in tiary settings by specialists. Not only is this setting inappropri- importance as the need for professionals to continually update ate, but typical content and clinical experience do not reflect their knowledge and acquire new skills has been more clearly what the doctors will be doing or what they will need to know appreciated. The coordination and provision of appropriate after qualification. Nevertheless, the university teaching hospi- continuing medical education depends heavily on the specialists tal cannot be omitted from the basic training of doctors. If stu- and academics associated with referral and academic hospitals. dents and faculty were involved only in district-based services, they would miss many important advances in biomedical Management and Administration. Referral hospitals in many science and the care of complex problems (Husain 1996). developing countries play important roles in providing mana- Moreover, doctors need to know enough about what the vari- gerial and administrative support to other elements of the health ous tertiary specialties do to be able to refer patients appropri- system. These roles may include managing laboratory services ately and to make personal career choices. on behalf of the whole health system; serving as the location for The training of specialists, of course, depends far more on drug and medical supply depots and distribution systems and the existence and proper functioning of referral hospitals. managing procurement systems; hosting and managing health Again, a particular country will need to decide how many spe- information systems, often including epidemiological surveil- cialists it needs in which specialties and whether it should send lance systems; managing centralized transport fleets; and, on its doctors abroad to specialize or train them internally. In occasion, providing financial management, payroll, and human 1234 | Disease Control Priorities in Developing Countries | Martin Hensher, Max Price, and Sarah Adomakoh resource management services to other health units. Our intent Utility, or welfare, includes health as one of many important is not to consider whether such arrangements are "right" or outcomes, such as financial security, risk alleviation, and "wrong"--complex factors would have to be taken into account psychological reassurance. However, as Hammer and Berman in every individual circumstance--but to note that making (1995) note, health policy is typically conducted as if it has a changes to the functioning of referral hospitals may have unin- unidimensional objective--namely the maximization of health tended consequences. For example, moving referral hospitals (DALY) outcomes. Determining the appropriate resource allo- from funding based on a global budget to reimbursement sys- cation to referral hospitals purely on the basis of the cost of tems based on patient activity may unintentionally cause hospi- generating health (DALYs) may, therefore, seriously underesti- tals to cease to provide these wider support functions if explicit mate the optimum level of resources, because such measures alternative funding mechanisms are not established. will fail to capture the full welfare gains from the availability of higher-level health services. An example will highlight the dif- Research and Innovation. Referral hospitals tend to be where ference between valuing hospitals on the basis of their contri- most health research is undertaken. Whereas in developed bution to health status alone compared with including wider countries they may often be associated with the development of concepts of welfare in the valuation. new technologies, in developing countries they are more often Renal failure leading to the need for dialysis is relatively rare, the site of research for the initial piloting and introduction of and certainly rare in comparison to many other infectious and new technologies developed elsewhere and for the evaluation chronic diseases in lower- or middle-income countries. of their local suitability and field efficacy. Referral hospitals are Treatment is lifesaving, but must continue indefinitely (involv- also the vehicle for disseminating such technologies through ing visits two or three times every week) and is, therefore, the exposure of staff during training as well as through the role extremely expensive. In many cases, dialysis can be justified that referral hospitals frequently play in continuing profes- only if it is linked to a renal transplant program, which termi- sional education. nates the need for dialysis and frees the equipment for someone Research activities are vital in attracting and retaining else. The proportion of the total population who will benefit specialist staff members who are required not just for the from such a referral hospital program is small; therefore, the treatment of complex patients, but also for the training of new DALYs generated are low, and the program would not rank specialists. Research that is responsive to local conditions--that high among the priorities given a limited budget. However, is, local disease burdens and technology constraints--fills a every member of the population is at risk of renal failure and, critical gap because researchers in developed countries and if affected, would find that, in the absence of a publicly funded pharmaceutical companies do not generally pursue such program, he or she would either die or face extremely high research questions if they do not foresee sufficient returns to costs to secure treatment in the private sector or abroad. their investments. Even in poor countries, patients' price elasticity of demand is low when faced with life-threatening illnesses, particularly Valuing the Indirect Contribution to the Health System. when treatment can change the outcome. Studies on poverty From the enumeration of the many roles of referral hospitals have shown that a significant proportion of households that and their indirect effect on health through their contribution have become poor did so as a result of serious illness, which to the health system by way of supervision, administration, resulted in their liquidating assets to pay for health care (see, for training, research, and quality improvement, it is immediately example, Liu, Rao, and Hsiao 2003). Thus, people seek the peace evident that these benefits cannot readily be translated into of mind of knowing that they can obtain lifesaving treatment DALYs or any other metric to be used in a relative cost-benefit should they need it without the risk of incurring catastrophic analysis. costs of care. This additional welfare derives both from the financial security of not having to spend more than people can afford to save their lives and from the direct health benefits of Externalities and Intangible Benefits treatment itself. The utility from the former (financial security) The previous sections reviewed the various functions of refer- increases with the cost of the intervention required, whereas the ral hospitals within the health system, all of which contributed utility derived from the latter (direct health benefits) is unre- directly or indirectly to the health status of individuals. This lated to the cost of the intervention. Paradoxically, one could, section addresses other ways in which referral hospitals con- therefore, argue that the rarer a particular illness is--and the tribute to welfare and well-being, and comments on how they more costly the intervention required--the greater will be the complicate the issue of valuing the contribution of referral welfare gain from public spending on that intervention. hospitals in society. This argument, of course, is likely to stand in direct contrast Referral hospitals have a broader effect on overall societal to the conclusions drawn from prioritization based on cost- welfare than can be captured by measures of health outcomes. effectiveness. For most individuals, willingness to pay is far less Referral Hospitals | 1235 than the costs of the procedure to them; however, because the provide local health professionals with a good incentive to whole population benefits from the security of knowing that remain at home, whereas the absence of referral hospitals each individual would be entitled to referral hospital care would increase the propensity of local professionals to should he or she need it, in the aggregate the welfare value gen- emigrate. erated by public provision or funding may be many times greater than the value of the DALYs generated directly for those few patients who do receive treatment. This literature review DETERMINANTS OF AN APPROPRIATE BALANCE did not find evidence of studies on national willingness to pay OF REFERRAL-LEVEL CARE for referral hospital care in developing countries, but this area could be of interest for future research. When one considers the ideal level of resources to be provided In practice, too, the public--particularly an urban, middle- for referral hospital care and the appropriate balance between income public--expects the government to provide care of last resources for referral hospitals and for other levels of the health resort for complex trauma or diseases, especially for natural care system, no simple formula is available that can be applied and man-made disasters. Thus, even though referral hospitals to different countries and contexts. However, certain key fac- may provide care to a small number of people, often with lim- tors have an important influence on the need and demand for ited health benefits, politicians and the public alike may value referral-level care, the resources that may be available to the and prioritize them simply because they meet the public's health sector, and the ability of the health sector to provide ade- expectations for what the government must provide. In addi- quate and effective care in different settings. tion, politicians and the public often regard a country's ability to provide the kind of complex, high-tech care offered in a referral hospital as a measure of that country's level of devel- General Determinants opment and sophistication, and it is a source of national pride. Arguably the most important determinant of demand for and Whether economically rational or not, this nonhealth benefit ability to pay for referral hospital care is a society's level of eco- appears to drive public choices to some extent. nomic development and wealth, captured (albeit imprecisely) by measures of GDP per capita. Extensive international evi- dence indicates that national health expenditure displays an Negative Impacts unambiguously positive income elasticity both across countries The "negative" impact of referral hospitals is largely attributa- and over time; that is, as a country gets richer, it spends rela- ble to their potential to exert distortionary effects on the health tively more on health (see, for example, Getzen 2000; Schieber system by diverting resources from peripheral areas and from 1990). Studies in developed countries indicate that in the lower levels of care (Fiedler, Schmidt, and Wight 1998; Filmer, United States, every 1 percent long-run increase in GDP leads Hammer, and Pritchett 1997) for the following reasons: to a 1.6 percent increase in health expenditure, and in other countries the increase is between 1.2 and 1.4 percent (Getzen · Tertiary hospitals and specialists have a high political and 2000). Therefore, expecting developing countries to spend a public profile. higher proportion of their GDP on health care as they become · Urban and political elites are more likely to use referral hos- wealthier seems to be reasonable. If the poorest countries were pitals than rural primary care facilities or district hospitals. to focus their limited resources on highly cost-effective inter- · Harmful competition with lower levels of care may result ventions in primary health care, somewhat better-off countries from the maintenance of higher-level referral hospitals in might be expected to spend progressively more on the referral many poor countries, lowering use of the former. hospital level as resources became available. · Referral hospitals can be entry points for the introduction An overlapping set of demographic and geographical factors into the health system of inappropriate and unaffordable also plays an important role in determining the balance of technologies. referral care--namely, population size, population density, ter- · Skilled personnel frequently find referral hospitals far more rain, distances between main urban centers, and access. attractive to work at than rural and district hospitals for Populations of some millions are required to justify a major such reasons as preferences for a metropolitan location, tertiary hospital with a full range of tertiary services. Small better hospital resources allowing for a more rewarding countries with populations of less than 1 million will certainly professional experience, and better opportunities for pri- not be able to provide a full range of tertiary hospital services vate practice (official or unofficial). However, given the because of the need to achieve minimum volumes to ensure huge problem of global migration of health workers from service viability and to attract a critical mass of specialized per- poor to rich countries (Bundred and Levitt 2000), one sonnel. Countries with fewer than 100,000 inhabitants (gener- could argue that well-functioning referral hospitals might ally island states) may find even secondary hospital services 1236 | Disease Control Priorities in Developing Countries | Martin Hensher, Max Price, and Sarah Adomakoh beyond their means and capabilities. Supranational referral, Referral hospital services require a specialized staff to fulfill reliance on larger neighbors, or regional collaboration may be their mission. If specialized personnel are not available in a unavoidable for smaller countries, especially for tertiary care country, then attempting to develop referral hospitals on a provision, with the Caribbean and southern Africa providing large scale will clearly be infeasible. However, many countries clear examples of many smaller states relying on referral facili- arguably have too many specialized staff persons and too few ties in larger or wealthier neighbors. Within larger countries, well-trained generalists. Where large numbers of specialists population density can complicate the planning of referral exist, their presence will likely tend to draw resources dispro- services. Compact countries or regions with dense populations portionately toward the referral level and away from district can typically provide high levels of access to referral care at a health systems. Wherever such imbalances exist, positive relatively small number of sites, whereas countries or regions changes will require a substantial training or retraining agenda. with more dispersed populations face more complex tradeoffs The feasibility of such efforts is closely linked to the profes- regarding number of sites, volume thresholds, and transporta- sional and social status of different professional groups and tion systems. subgroups--for example, whether medical specialists are The other main influence on the appropriate balance of viewed as having a higher status than general practitioners-- referral services for a given country is its particular pattern and to the premium a society places on having access to and burden of disease. Although referral-level services will "advanced" medical care. always be needed, as a society passes through epidemiologic and demographic transitions, it is likely to require more of those services typically found at referral hospitals. For example, CURRENT BALANCE OF CARE IN PRACTICE rapidly increasing rates of heart disease and cancers are typi- cally encountered in industrializing nations and aging popula- In this section, we summarize data on the current balance tions, and these are diseases whose effective management between referral and lower levels of care. We first look at the requires access to the interventions, skills, and equipment that share of total health expenditure going to these different levels, will typically be concentrated at the referral hospital level. but given that referral care normally has much higher unit costs, we recognize that the balance also needs to be viewed in terms of volume of cases and access and equity. Health System Determinants A number of factors specific to the particular context of a country's health system will also influence the appropriate bal- Share of Health Expenditure ance between referral hospitals and lower levels of care. These Different health systems categorize hospitals and services ren- factors are especially important in considering the appropriate- dered differently. Methodologies in national health accounts in ness of plans to change the balance of care between levels. developing countries during the 1990s and early 2000s have Broadly, they can be summarized as follows: tended to use a simple, catch-all category of "hospitals" or "acute hospitals" (for example, WHO 2002). Even high-income coun- · capabilities of lower levels tries following the Organisation for Economic Co-operation · availability of specialized personnel and Development's system of health accounts provider classifi- · training capacity, organization, and needs cation (OECD 2000, 136) distinguish only between "general" · cultural issues, political issues, and traditions. hospitals and "mental health and substance abuse" and other "specialty" hospitals in their national health accounts. The first three factors are closely interrelated. If primary Consequently, making valid cross-country comparisons of health care and district hospital services are weak, cutting spending by levels of hospital care remains extremely difficult. resources for referral hospitals without destabilizing the system Mills (1990a) reviews published data on hospital expendi- will be more difficult. In such circumstances, rapid rebalancing ture patterns in developing countries, and Barnum and Kutzin of resources is unlikely to be possible because careful efforts will (1993) provide a comprehensive survey of expenditure on hos- be required to develop lower-level services first, while still main- pital services in a number of developing countries, drawing taining the referral service.Where lower-level services are strong, their information largely from World Bank sector reviews. devoting relatively fewer resources to referral hospitals may well These analyses remain the most authoritative assessment of the be possible. However, even though an effective district health proportion of public health expenditure absorbed by second- system will be able to treat a large proportion of patients at lower ary and tertiary hospitals, even though their data represent only levels of care, it will also be better able to identify patients who a handful of countries at different points in time. require referral for more complex care and, thus, may generate a Overall, Mills (1990a) finds that hospitals in developing greater appropriate demand for referral hospital care. countries appear to absorb from 30 to 50 percent of total Referral Hospitals | 1237 health expenditure. Public hospitals of all types absorb some Table 66.4 Cost Per Bed Day for Selected Specialties, 50 to 60 percent of public health expenditure, and secondary Tertiary Hospitals, Mauritius, 1995 and tertiary hospitals absorb about 60 to 80 percent of public (2001 U.S. dollars) hospital expenditure, with the remainder going to district hos- Specialty Minimum Maximum pitals. Her results are broadly similar to those of Barnum and Medicine 16 20 Kutzin (1993, 26­33), who find that public hospitals at all lev- els absorb a mean of approximately 60 percent of recurrent Orthopedics 18 23 public health expenditures. Across five countries (Belize, Pediatrics 29 43 Indonesia, Kenya, Zambia, and Zimbabwe), they find that Cardiothoracic surgery 36 39 tertiary hospitals account for between 45 and 69 percent of Burns 37 37 total public expenditure on hospitals. In South Africa, Thomas Intensive care unit 106 120 and Muirhead (2000) find that tertiary hospitals account for Source: Murray and others 1996. 28 percent of hospital expenditure and 17 percent of total public health expenditure, but when taken together with regional referral hospitals, constitute 59 percent of hospital similar at all levels. This phenomenon is explained by average expenditure. bed occupancy in Mauritian district hospitals of around 45 percent in 1995, compared with average bed occupancies of 90 percent or more in tertiary hospitals. Thus, the high cost of Unit Costs of Care district hospital care in this case reflects not inputs, which are One of the explanations for the high share of expenditure that much less extensive than in a tertiary hospital, but the effect of flows through higher-level referral hospitals is, of course, that low levels of utilization. Such a difference in utilization between the unit costs of a referral hospital are necessarily higher levels of hospital tends to be the norm in many developing than the unit costs of a district hospital. This difference results countries (Barnum and Kutzin 1993, 91­96). Note that the from the more complex case mix they treat, their more expen- regression-based unit costs of district hospitals used in the cost sive inputs, and the additional costs of their teaching functions analysis for this volume reflect an "optimized" bed occupancy (Barnum and Kutzin 1993, 26). Mills (1990b) reports that her of 80 percent (Mulligan and others 2003, 29). This assumption analysis of available data indicated that secondary-level hospi- is entirely defensible from a long-run perspective, assuming tals were typically twice as expensive per bed day as district hos- cost-minimizing behavior is necessary and appropriate. It does, pitals and that tertiary hospitals were typically between twice however, reflect quite a substantial shift from the levels of uti- and five times as expensive per bed day as district hospitals. lization and productivity commonly seen in rural district hos- Barnum and Kutzin (1993) find similar relationships between pitals in most developing countries. unit costs by level of hospital in a variety of different countries. The use of a simple unit cost hides important cost differ- This upward gradient in unit costs has also been found in ences between services and specialties within the same hospi- econometric studies of hospital costs (Adam, Evans, and tal, as demonstrated by the breakdown for Mauritian tertiary Murray 2003) and has been explicitly incorporated into the hospitals in table 66.4. Differences in length of stay for differ- regression-based unit cost estimates used in other chapters in ent specialties and conditions also obviously contribute to this volume. radically different costs per admission or patient; however, Table 66.3 shows data on unit costs by level of care from these differences should be captured by the condition Mauritius and highlights a commonly encountered contradic- and procedure costs used in the disease chapters in this tion of the preceding paragraph--namely, that costs appear volume. Table 66.3 Cost Per Bed Day in a Medical Ward by Level Appropriateness of Utilization of Referral Hospitals of Hospital, Mauritius, 1995 Perhaps the most frequent theme in the research literature on (2001 U.S. dollars) referral hospitals in developing countries is the inappropriate Level of hospital Cost utilization of higher-level facilities and the apparent failure of most referral systems in developing countries to function as District 17 intended. Broadly speaking, hospitals of all levels, up to and Regional 21 including national tertiary centers--especially in their outpa- Tertiary 20 tients departments--are overwhelmed by patients who could Source: Murray and others 1996. have been treated successfully at lower-level facilities, many of 1238 | Disease Control Priorities in Developing Countries | Martin Hensher, Max Price, and Sarah Adomakoh whom have self-referred, bypassing primary health care or dis- (Stefanini 1994), which reinforces the cycle and ensures trict hospitals in the process (Holdsworth, Garner, and that primary health care facilities remain underused and Harpham 1993; London and Bachmann 1997; Omaha and inefficient. others 1998; Sanders and others 2001). Atkinson and others (1999) describe an extreme manifesta- tion of this phenomenon, whereby the University Teaching Access and Equity Hospital is actually the only public hospital in Lusaka. By their nature, referral hospitals must be limited in number Combined with the bypassing of primary health clinics in the and will inevitably be sited in major towns and cities. As a city, this situation results in the University Teaching Hospital's result, a significant portion of the population, especially people functioning primarily as a glorified health center and first- living in rural areas, will tend to live at some distance from the contact provider for most of Lusaka's population. The problem nearest referral hospital. Studies of the accessibility of referral of bypassing typically seems to be driven by a number of fac- hospital care in countries such as Ethiopia (Kloos 1990) and tors, including patients' perception of superior quality of care Nigeria (Lyun 1983; Okafor 1983) have repeatedly confirmed and resource availability at referral hospitals, which often may the existence of a steep distance-decay function, indicating be entirely well founded and rational (see, for example, London that--other things being equal--individuals with a given need and Bachmann 1997; Nolan and others 2001); the desire to for a clinical service will be less likely to access that service the avoid delays in care if referral to a higher-level facility proves to farther away from the referral center they live. be necessary; and the fact that for many urban populations Compounding the impact of distance, investigators find a referral hospital may simply be the closest health facility. that problems relating to the availability, regularity, and cost of Grodos and Tonglet (2002) argue that many countries' failure transportation to referral centers also affect service utilization to develop an adequate urban equivalent of the district (Kloos 1990; Martey and others 1998). The same authors health concept greatly exacerbates inappropriate utilization of indicate that prohibitive hospital fees are often a significant hospitals. The urban phenomenon of widespread bypassing barrier to utilization, especially among poorer patients. Other and self-referral is frequently accompanied by low rates of important barriers included perceived lack of drugs and essen- formal referral from rural and outlying facilities (see, for tial supplies, even at referral centers; negative staff attitudes; example, Nordberg, Holmberg, and Kiugu 1996; Omaha and and cultural and linguistic differences (for example, where the others 1998). staff at a referral center does not speak the language of a These problems have a number of negative impacts and con- patient from a remote area). As noted earlier, peripheral dis- sequences. Simple conditions are unnecessarily treated in a trict hospitals also tend to have low rates of referral. These bar- high-cost environment; outpatient departments are congested riers, which all disproportionately affect rural patients, must by patients requiring primary care, thus causing long waiting be contrasted with the phenomenon noted earlier of excessive times; scarce staff time is diverted from specialized areas and and inappropriate use of referral hospitals for primary care by into inappropriate care; and more complex cases requiring spe- urban residents. cialized care are crowded out by more urgent but less technically In addition to finding that public hospitals favor urban res- demanding cases that could be cared for at lower levels. The lat- idents over rural dwellers, a number of studies have indicated ter has been a particular concern in those countries with more that public hospitals in many poor countries disproportion- serious HIV/AIDS epidemics. As the number of patients falling ately benefit the better off, leading their authors to argue that sick with AIDS increases rapidly, they start to occupy a signifi- diverting public funds from hospitals and toward primary cant proportion of beds in hospitals at all levels (Gilks and others health care would be pro-poor (see, for instance, Castro-Leal 1998), inevitably crowding out patients requiring other forms of and others 2000; Filmer, Hammer, and Pritchett 1997; Mahal care.Although AIDS cases may well require hospitalization, only and others 2002). Other studies find this tendency in some a small proportion of cases require specialized or tertiary care. countries but not in others (Makinen and others 2000). By con- Gilks and others (1998) find that this crowding-out effect may trast, in Latin American countries, Barnum and Kutzin (1993) fall over time as the health system adjusts to the pressures of find strong evidence that public hospitals are pro-poor in their AIDS, but countries facing impending AIDS epidemics should distributional effect. Even if referral hospital services are not be prepared for its initial appearance. currently pro-poor, policy makers face two contradictory alter- Taken together, this complex of problems undermines the natives: either to retarget public funds toward primary health effective delivery of both specialized care and appropriate pri- care for the poor, hence greatly reducing or abandoning public mary health care. Specialized care is pushed to the back- funding for referral hospitals, or to attempt to remove the bar- ground by the human wave of demand for primary care, while riers that prevent the poor from using higher-level services, hospitals unwittingly further undermine the credibility of the which would probably require increased spending on all levels primary health care system through one-sided competition of care. Referral Hospitals | 1239 GETTING BETTER VALUE FOR MONEY FROM · ensuring effective ordering, stock control, and distribution THE HOSPITAL SYSTEM systems to minimize theft and wastage of key supplies · undertaking planned preventive maintenance and pro- Although prescribing how resources should be allocated across grammed replacement of equipment and buildings. levels of care is hard, at least they should be efficiently used, wherever they are spent within the hospital system. The preceding analysis has highlighted how deficiencies at the lower Can Public-Private Interactions Improve Efficiency? levels of the hospital system render referral hospitals less effi- In the context of this discussion, privately owned hospitals that cient and how factors that affect access lead to skewed benefits provide subsidized care to public patients, such as nongovern- and inequity. Here we look more specifically at three areas for mental organization and mission hospitals, are regarded as improving the efficiency of the hospital system: interventions public hospitals. Private refers to for-profit hospitals that within the referral hospital, the use of public-private partner- are generally funded by paying patients and are minimally sub- ships, and strengthening of the referral chain. sidized. Few studies have been undertaken of how private hospitals operate in developing countries (see, for example, Muraleedharan 1999). Although the exact balance of and rela- Improving the Efficiency of Referral Hospitals tionship between the public and private health sectors varies Although space does not permit a lengthy discussion of greatly from country to country at all levels of the health sys- approaches to improve efficiency in the context of referral tem, a common theme in almost all low- and middle-income hospitals, this aspect is nonetheless important in planning and countries is that private hospitals do not follow the pyramidal system strengthening (for a more detailed discussion see referral form that public hospital systems have adopted almost Barnum and Kutzin 1993; Hensher 2001; Walford and Grant universally. Most private health sectors do not clearly delineate 1998). In summary, the key areas on which planners and man- district, secondary, or tertiary hospitals. Different private hos- agers should focus are as follows: pitals may offer different services and facilities on a more or less idiosyncratic basis, with independent medical specialists prac- · reducing inappropriate outpatient and inpatient use of ticing and admitting patients at various different hospitals. referral In most systems, scope exists for both positive collaboration · improving systems to allow early discharge from the and competition between public and private hospitals, espe- hospital cially for secondary and tertiary services. Competition between · ensuring that bed occupancy rates can be maintained as public and private sectors obviously has the potential to be close as possible to optimal rates--namely, 85 percent for beneficial by driving quality up and costs down, but it may also referral hospitals have negative effects by encouraging a duplication of services · developing systems for booked outpatient appointments, and resulting in the underutilization of fixed capital by creating admissions, and procedures to permit better planning of perverse incentives for physicians and patients and by compet- activity and staffing ing with the public sector for scarce human resources. In some · undertaking as much activity as possible on an ambulatory settings, the private sector may be able to offer services that the rather than an inpatient basis, supported by the use of public purse cannot afford to provide, thus allowing patients "step-down" beds and patient hotels who could not afford private care some chance of accessing · evaluating the staff skill mix and the potential for skill sub- sophisticated treatments through the government's paying pri- stitution, as well as efficient remuneration strategies, on a vate providers or by some pro bono provision of treatment for continuous basis poor patients. · evaluating and improving processes and systems, including In many countries, government hospitals are establishing cost-effective clinical guidelines for patient treatment, on a private wards as a vehicle for income generation. The fees for continuous basis such units are lower than those at private hospitals, offering · ensuring that new or replacement referral hospitals con- access to private facilities to patients who may not be able to form as much as possible to available evidence on afford private hospitals. The link with academic medicine often economies of scale--that is, that hospitals with fewer than adds to the appeal of such facilities. However, as is the case in 200 beds are likely to be scale inefficient and that disecon- South Africa, effectively only tertiary hospitals and a handful of omies of scale are likely to become increasingly evident in secondary hospitals are felt to be attractive enough to private hospitals with more than 600 beds patients to offer genuine opportunities as preferred providers. · adopting intelligent procurement processes and engaging in The mass of district and regional hospitals are unlikely to be effective negotiations with suppliers in relation to prices and attractive to private patients; therefore, the positive spinoffs of service levels these initiatives may be limited in their scale and reach. 1240 | Disease Control Priorities in Developing Countries | Martin Hensher, Max Price, and Sarah Adomakoh Contracting out services to private providers, particularly investment and funding to ensure the availability of appropri- high-cost, low-volume services, may be an efficient way to offer ate staff members and supervision, to ensure continuous drug such services to public patients. For example, the government supplies, and to provide basic laboratory tests (Walford and of Barbados contracts out surplus demand for dialysis to a pri- Grant 1998, 38). Given the pervasiveness of inappropriate use vate facility on the island. In some provinces of South Africa, of referral hospitals for primary health care problems by urban expensive imaging such as MRI has been contracted out to pri- residents, both urban and rural primary health care and district vate radiology practices. South Africa is also experimenting health systems must be adequately strengthened. Financing with contracting out the management of some academic refer- strategies that redistribute funds from urban to rural regions ral hospitals to a private hospital group that is assumed to have may unwittingly hamper such strengthening of the referral sys- greater management expertise and is free from certain public tem by failing to allow for the development of appropriate sector constraints, such as salary scales for senior managers. It lower-level facilities for urban residents. This risk is especially is too early to judge the success of this arrangement, but in all high when a country is pursuing a redistributive agenda against cases it is imperative that contracts be carefully regulated, mon- a background of limited or zero overall growth in expenditure. itored, and enforced. For a comprehensive review of contract- From a physical planning perspective, planners should con- ing, see Bennet, McPake, and Mills (1997). sider providing primary health care and district hospital walk- Particular problems may arise where the same doctors pro- in ambulatory services (emergency and general outpatients) in vide care in both public and private hospitals. Under fee-for- a physically distinct facility sited immediately next to the refer- service arrangements, physicians may focus on their more ral hospital. This arrangement not only enables triage and fil- lucrative private patients to the disadvantage of public hospi- tering of less severe cases (while proximity ensures that severe tal patients, refer patients with adequate insurance to their emergency cases can be transferred rapidly) but also enables private practices and private hospitals, and transfer patients rigorous enforcement of a referral-only policy within the refer- with expensive diseases or inadequate insurance to public ral hospital. hospitals. The development of effective patient transportation arrangements is also critical, not only to ensure that patients from remote areas have a fair chance of being successfully Improving the Functionality of Referral Systems referred to a center of excellence (bearing in mind that most An ideal referral system would ensure that patients can receive referral systems will almost certainly need to increase referral appropriate, high-quality care for their condition in the lowest- rates from rural areas), but also to ensure that patients can be cost and closest facility possible, given the resources available to discharged in a timely and well-planned fashion. the health system, with seamless transfer of information and Perhaps more challenging is the concurrent need to align responsibility as that patient is required to move up or down the incentives of referral hospitals, district hospitals, and pri- the referral chain. Although few referral systems anywhere in mary health care services. This goal may or may not be achiev- the world live up to this ideal fully, it does provide a target in able by means of an integrated management structure, but it relation to improving the current situation. Improving the certainly requires a good deal of communication, collaborative effective functioning of referral systems broadly requires planning, and collaborative development of shared care proto- progress in three areas: referral system design, facilitation of the cols, and senior personnel need to be given responsibility for smooth transfer of patients and information between levels, coordination and liaison across key interfaces of the referral and what Walford and Grant (1998, 38) refer to as effective network. A single, global budget controlled by an authority that "referral discipline." is concerned with optimizing the cost-effectiveness of health Improving referral system design must start with a detailed care delivery would seem to be a necessary condition to achieve attempt to assess which services should be provided at which alignment across service levels; however, a consideration of level of care, encompassing community- and home-based care, financing mechanisms is beyond the scope of this chapter. primary health care, district hospitals, secondary hospitals, ter- At the patient level, a number of mechanisms to improve tiary hospitals, and specialized hospitals. Such an assessment referral discipline can be considered. In situations in which must take local circumstances into account, requires a signifi- eliminating nonreferred patients entirely from the referral hos- cant analytical and consultative effort by planners and clini- pital is impossible, queuing systems should be redesigned to cians if it is to be credible, and must explicitly be open to revi- separate referred patients from nonreferred patients so that sion in light of practical experience. After such an exercise has referrals can be fast-tracked. Explaining to nonreferred patients identified which services can appropriately be provided at each why other patients are being fast-tracked past them is impor- level of care, adequate resources must be dedicated to strength- tant to encourage them to seek referral in future. Ideally, they ening lower levels of care to make them attractive and credible should be diverted to an on-site primary health care facility in the eyes of patients. This effort will require significant where they can be treated more quickly than in the referral Referral Hospitals | 1241 hospital. Another possibility may be to institute bypass fees for likely to undermine and destabilize the entire health system nonreferred patients, charging them a penalty fee for failing to than to liberate resources for primary health care. Clearly, use the referral system. Such a decision requires careful consid- countries must critically evaluate their health priorities and eration and planning. Credible lower-level care must be readily their balance of care and resources between levels, but they available, and substantial efforts to communicate the new pol- should do so carefully and thoroughly, with a clear under- icy to the public will be required if this approach is to be seen standing of the analytical effort required to draw meaningful as fair. More broadly, intensive public communication and conclusions, of the planning and managerial capacity that they education will be essential to inform the public how, where, will require to bring about successful change, and of the long and when they should seek health care at different levels and to time frames required to develop and implement robust plans build their confidence that lower-level facilities really will be for major system changes. able to offer acceptable quality care when they need it. ACKNOWLEDGMENTS CONCLUDING COMMENTS The authors gratefully acknowledge the crucial assistance of Etienne Yemek in undertaking literature reviews. This review of the available evidence indicates that referral hos- pitals frequently do command a large share of health sector resources and expenditure, yet no simple way exists of assessing REFERENCES what an appropriate share would be. Strong referral hospitals can distort priorities and undermine basic services, but they Adam T., D. B. Evans, and C. J. Murray. 2003. "Econometric Estimation of also provide important health benefits to large numbers of Country-Specific Hospital Costs." Cost Effectiveness and Resource Allocation 1:3. patients whom they treat successfully. Referral hospitals pro- Adomakoh, S., A. St. John, and A. Kumar. 2002. "Reducing Mother to vide essential support to lower levels of the system, which can- Child Transmission of HIV-1 in Barbados. Cost-Effectiveness of the not function effectively without access to upward referral, and PACTG 076 Protocol in a Middle Income, Low Prevalence Setting." they are frequently the most functional component of the Paper presented at the 14th International AIDS Conference, July 7­12, Barcelona, Spain. health system, paying greatest attention to quality of care. Al-Mazrou, Y., S. Al-Shehri, and M. Rao. 1990. Principles and Practice of Overall, we have argued that both national and international Primary Health Care. Riyadh: Al-Helal Press. policy makers should be cautious before demanding the reallo- Atkinson, S., A. Ngwengwe, M. Macwan'gi, T. J. Ngulube, T. Harpham, and cation of resources away from referral hospitals and should be A. O'Connell. 1999. "The Referral Process and Urban Health Care in still more cautious in allowing themselves to believe that such Sub-Saharan Africa: The Case of Lusaka, Zambia." Social Science and a reallocation is likely to be achievable in practice. In particu- Medicine 49: 27­38. lar, this chapter has made the case that a unidimensional focus Barnum, H., and J. Kutzin. 1993. Public Hospitals in Developing Countries: Resource Use, Cost, Financing. Baltimore: Johns Hopkins University on cost-effectiveness analysis and cost per DALY gained will fail Press. to capture the importance of referral hospital services ade- Bennet, S., B. McPake, and A. Mills, eds. 1997. Private Health Providers in quately. In reality, in most developing countries, the scope for Developing Countries. London: Zed Books. reallocation of resources from referral hospitals to lower levels Bundred, P., and C. Levitt. 2000. "Medical Migration: Who Are the Real of care is limited, and the managerial demands of achieving a Losers?" Lancet 356: 245­46. successful reallocation are great. Lower levels of care certainly Castro-Leal, F., J. Dayton, L. Demery, and K. Mehra. 2000. "Public Spending on Health Care in Africa: Do the Poor Benefit?" Bulletin of require strengthening, but this need is more likely to reflect the World Health Organization 78: 66­74. inadequate financing of the entire public health system than a Fiedler, J., R. Schmidt, and J. Wight. 1998. "Public Hospital Resource grossly excessive allocation to referral hospitals. Instead, refer- Allocations in El Salvador: Accounting for the Case Mix of Patients." ral hospitals should perhaps be seen as the capstone of the Health Policy and Planning 13 (3): 296­310. referral pyramid: they should not be too heavy, but if they are Filmer, D., J. Hammer, and L. Pritchett. 1997. "Health Policy in Poor Countries: Weak Links in the Chain." Policy Research Working Paper too light, the levels below them will lose cohesion. A restruc- 1874, World Bank, Washington, DC. turing of referral hospital services is certainly called for to Getzen, T. E. 2000. "Health Care Is an Individual Necessity and a National improve appropriate referral and utilization, especially by Luxury: Applying Multilevel Decision Models to the Analysis of Health remote and rural populations; to transform the inappropriate Care Expenditures." Journal of Health Economics 19: 259­70. use of referral hospitals as primary health care providers; to Gilks, C., K. Floyd, L. Otieno, A. Adam, S. Bhatt, and D. Warrell. 1998. improve efficiency; and to provide much better outreach and "Some Effects of the Rising Case Load of Adult HIV-Related Disease on a Hospital in Nairobi." Journal of Acquired Immune Deficiency support to lower levels of care. Syndrome and Human Retrovirology 18 (3): 234­40. This restructuring should not be confused with wholesale Grodos, D., and R. Tonglet. 2002. "Maîtriser un espace sanitaire cohérent demolition. Undermining referral services will be far more et performant dans les villes d'Afrique subsaharienne: Le district de 1242 | Disease Control Priorities in Developing Countries | Martin Hensher, Max Price, and Sarah Adomakoh santé à l'épreuve" (in French). Tropical Medicine and International Muraleedharan, V. R. 1999. "Characteristics and Structure of the Private Health 7 (11): 977­92. Hospital Sector in Urban India: A Study of Madras City." Small Applied Hammer, J., and P. Berman. 1995. "Ends and Means in Public Health Research Paper 5, Abt Associates, Partnerships for Health Reform Policy in Developing Countries." Health Policy 32 (1­3): 29­45. Project, Bethesda, MD. Harden, R. M., and M. H. Davis. 1998. "Educating More Doctors in the Murray, C., P. Mahapatra, R. Ashley, C. Michaud, A. George, P. Hrobon, U.K.: Painting the Tiger." Medical Teacher 20 (4): 306. and others. 1996. The Health Sector in Mauritius: Resource Use, Intervention Cost, and Options for Efficiency Enhancement. Cambridge, Hensher, M. 2001. "Financing Health Systems through Efficiency Gains." MA: Harvard Center for Population and Development Studies, Burden Working Paper WG3:5, Commission on Macroeconomics and Health of Disease Unit. Working Group 3. http://www.cmhealth.org/docs/wg3_paper2.pdf. National Department of Health, South Africa. 2003. "Strategic Framework Holdsworth, G., P. Garner, and T. Harpham. 1993. "Crowded Outpatient for the Modernisation of Tertiary Hospital Services". National Departments in City Hospitals of Developing Countries: A Case Department of Health, South Africa, Pretoria. http://www.doh. Study from Lesotho." International Journal of Health Planning and gov.za/mts/docs/framework.html. Management 8 (4): 315­24. Nolan, T., P. Angos, A. Cunha, L. Muhe, S. Qazi, E. Simoes, and others. Husain, M. 1996. "The University Hospitals in Pakistan: The Case of the 2001. "Quality of Hospital Care for Seriously Ill Children in Less- Aga Khan University Hospital." In The Proper Functioning of Teaching Developed Countries." Lancet 357 (9250): 106­10. Hospitals within Health Systems, ed. D. Puzin, 73­80. Geneva: World Health Organization. Nordberg, E., S. Holmberg, and S. Kiugu. 1996. "Exploring the Interface between First and Second Level Care: Referrals in Rural Africa." Kloos, H. 1990. "Utilization of Selected Hospital, Health Centers, and Tropical Medicine and International Health 1 (1): 101­11. Health Stations in Central, Southern, and Western Ethiopia." Social Science and Medicine 31 (2): 101­14. OECD (Organisation for Economic Co-operation and Development). 2000. A System of Health Accounts: Version 1.0. Paris: OECD. Liu, Y., K. Rao, and W. C. Hsiao. 2003. "Medical Expenditure and Rural Impoverishment in China." Journal of Health and Population Nutrition Okafor, S. 1983. "Factors Affecting the Frequency of Hospital Trips among 21 (3): 216­22. a Predominantly Rural Population." Social Science and Medicine 17: 591­95. London, L., and O. Bachmann. 1997. "Paediatric Utilisation of a Teaching Hospital and a Community Health Centre." South African Medical Omaha, K., V. Melendez, N. Uehara, and G. Ohi. 1998. "Study of a Patient Journal 87 (1): 31­36. Referral System in the Republic of Honduras." Health Policy and Planning 13 (4): 433­45. Lyun, F. 1983. "Hospital Services Areas in Ibadan City." Social Science and Medicine 17: 601­16. Puzin, D., ed. 1996. The Proper Functioning of Teaching Hospitals within Health Systems. Geneva: World Health Organization. Mahal, A., J. Singh, F. Afridi, V. Lamba, A. Gumber, and V. Selvaraju. 2002. Who "Benefits" from Public Sector Health Spending in India? New Delhi: Sanders, D., J. Kravitz, S. Lewin, and M. McKee. 2001 "Zimbabwe's National Council for Applied Economic Research. Hospital Referral System: Does It Work?" Health Policy and Planning 13 (4): 359­70. Makinen, M., H. Waters, M. Rauch, N. Almagambetova, R. Bitran, L. Gilson, and others. 2000. "Inequalities in Health Care Use and Schieber, G. J. 1990. "Health Expenditures in Major Industrialized Expenditure: Empirical Data from Eight Developing Countries and Countries, 1960­87." Health Care Financing Review 11: 159­68. Countries in Transition." Bulletin of the World Health Organization 78: Snow, R., V. Mung'ala, D. Forester, and K. Marsh. 1994. "The Role of the 55­65. District Hospital in Child Survival at the Kenyan Coast." African Martey, J., J. Djan, S. Twum, E. Browne, and S. Opoku. 1998. "Referrals for Journal of Health Sciences 1 (2): 71­75. Obstetric Complications from Ejisu District, Ghana." West African Stefanini, A. 1994. "District Hospitals and Strengthening Referral Systems Journal of Medicine 17: 58­63. in Developing Countries." World Hospitals 30 (2): 14­19. McCord, C., and Q. Chowdhury. 2003. "A Cost Effective Small Hospital in Thomas, S., and D. Muirhead. 2000. National Health Accounts Project: Bangladesh: What It Can Mean for Emergency Obstetric Care." The Public Sector Report. Cape Town, South Africa: University of International Journal of Gynaecology and Obstetrics 81: 83­92. Cape Town. Mills, A. 1990a. "The Economics of Hospitals in Developing Countries-- Walford, V., and K. Grant. 1998. "Health Sector Reform: Improving Part I: Expenditure Patterns." Health Policy and Planning 5 (2): 107­17. Hospital Efficiency." London: Department for International ------. 1990b. "The Economics of Hospitals in Developing Countries-- Development, Health Sector Resource Centre. Part II: Costs and Sources of Income." Health Policy and Planning 5 (3): WHO (World Health Organization). 2002. WHO National Health 203­18. Accounts 2002: Enhancing Country Templates-in-the-Making: Mulligan, J., J. Fox-Rushby, T. Adam, B. Johns, and A. Mills. 2003. "Unit Guidelines. Geneva: WHO, National Health Accounts Unit. Costs of Health Care Inputs in Low and Middle Income Regions." WHOSIS (World Health Organization Statistical Information System). Working Paper 9, Disease Control Priorities Project, Fogarty 2004."Physicians per 100,000 Population."World Health Organization, International Center, National Institutes of Health, Bethesda, MD. Geneva. http://www.who.int/GlobalAtlas/DataQuery/geoSelection. http://www.fic.nih.gov/dcpp/wps/wp9.pdf. asp. Referral Hospitals | 1243 Chapter 67 Surgery Haile T. Debas, Richard Gosselin, Colin McCord, and Amardeep Thind Countries with developing economies have not considered effective, with a cost per disability-adjusted life year (DALY) surgical care to be a public health priority, yet surgically treat- that is much lower than might have been expected, and can be able conditions--such as cataracts (Javitt 1993); obstructed on a par with other well-accepted preventive procedures, such labor (Neilson and others 2003); symptomatic hernias as immunization for measles and tetanus and home care for (Olumide, Adedeji, and Adesola 1976; Rahman and Mungadi lower respiratory infections (Armandola 2003; Dayan and 2000); osteomyelitis (Bickler and Rode 2002; Hilton 2003); others 2004; Moalosi and others 2003; Ruff 1999). otitis media (Smith and Hatcher 1992; Whitney and Pickering We have identified four types of surgically significant inter- 2002); and a variety of inflammatory conditions--add a ventions with a potential public health dimension: (a) the pro- chronic burden of ill health to populations. These acute and vision of competent, initial surgical care to injury victims, not chronic conditions take a serious human and economic toll only to reduce preventable deaths but also to decrease the num- and at times lead to acute, life-threatening complications. ber of survivable injuries that result in personal dysfunction Inadequacies in the initial care of injured patients (Hyder and impose a significant burden on families and communities; and Peden 2003; Jat and others 2004; Mock 2003; Mock and (b) the handling of obstetrical complications (obstructed labor, others 1995); of women with obstructed labor; and of children hemorrhage); (c) the timely and competent surgical manage- with treatable congenital anomalies, such as clubfoot (Ponseti ment of a variety of abdominal and extra-abdominal emergent 1999; Turco 1994) lead to preventable deaths or to chronic dis- and life-threatening conditions; and (d) the elective care of abilities that make productive employment impossible and simple surgical conditions such as hernias, clubfoot, cataract, impose dependency on family members and society. hydroceles, and otitis media. The role of surgery as a preventive strategy in public health needs to be studied and measured far more extensively than is currently the case. Another key reason for this study is that vir- NATURE, CAUSES, AND BURDEN tually all countries are developing their economies, and as a OF SURGICAL CONDITIONS result, developing nations are increasingly facing a double burden--that is, the infectious diseases that have historically Surgery is at the end of the spectrum of the classic curative been so relevant and the conditions that emerge with economic medical model and, as such, has not been routinely considered development (for example, trauma from motorcycle, truck, as part of the traditional public health model. However, no and car accidents). The inclusion of a surgery chapter in this matter how successful prevention strategies are, surgical condi- book recognizes that surgical services may have a cost-effective tions will always account for a significant portion of a popula- role in population-based health care. Recent studies (for tion's disease burden, particularly in developing countries instance, McCord and Chowdhury 2003) show that basic hos- where conservative treatment is not readily available, where the pital service, which requires no sophisticated care, can be cost- incidence of trauma and obstetrical complications is high, and 1245 where there is a huge backlog of untreated surgical diseases The World Health Report 2002 attributes 8,269, of a total (Murray and Lopez 1996). Some surgical procedures can 1,467,257,000 DALYs, to cataracts (0.56 percent), and all those certainly be perceived as forms of secondary or tertiary preven- DALYs are considered potentially surgical. Maternal conditions tion. Since the publication of the first edition of this book, (group I-C), perinatal conditions (group I-D), diabetes (group which did not have a chapter on surgery, the health care com- II-C), intentional injuries (group III-B), and unintentional munity has recognized that the surgical management of some injuries (group III-A), to name a few, are much broader cate- common conditions can indeed be a cost-effective intervention gories of conditions for which the demarcation between the (Javitt 1993; McCord and Chowdhury 2003). The purpose of surgical and nonsurgical burden is not as clear as for cataracts. this chapter is to explore this hypothesis in more depth. Faced with a near total lack of pertinent data, we decided that the next best approach was to try to obtain consensus on a "best educated guess" for the surgical burden of each condi- Methods for Determining Burden of Surgical Disease tion. We developed a survey instrument that listed all the We have arbitrarily decided to define a surgical condition as possible surgical conditions (all potential surgical DALYs any condition that requires suture, incision, excision, manipu- representing the maximum imaginable DALYs that could con- lation, or other invasive procedure that usually, but not always, ceivably be surgical). We sent the questionnaire to 32 surgeons requires local, regional, or general anesthesia. We prefer this in various parts of the world, asking them what was, in their definition for two main reasons, to one that would define opinion, the proportion of each condition that would require surgery as procedures performed by trained surgeons. First, surgery, which we have referred to as estimated surgical DALYs surgery does not have to be performed by qualified surgeons. or the conservative minimum. For each of the 18 completed Indeed, in developing countries with few doctors, nondoctors questionnaires, we discarded the two lowest and two highest can be trained to perform several types of operations satisfac- values for each condition, leaving a sample of 14 surveys. The torily. Second, we believe that the concept of surgery should lowest value of this sample was consistently chosen so as to err include minor surgical procedures that nurses or general prac- systematically on the conservative side. Note that more than titioners could perform along with nonoperative management 90 percent of all retained values were within 10 percent of the of surgical diseases (for example, certain types of abdominal, chosen value. We then applied this value to the DALY numbers thoracic, or head trauma and burns and infections). Any defi- provided by the World Health Report 2002 for each category of nition of surgery will have limitations, as has ours, and those potentially surgical conditions. limitations must be kept in mind when making interpretations, extrapolations, or estimates. Our broad definition is compati- ble with the concept of regionalized, coordinated, and interde- Findings pendent services provided at the community clinic level and at Table 67.1 presents our estimates of the actual surgical burden the district and tertiary hospital levels. The most difficult task for each category of potential surgical conditions for the world we then face is trying to determine the burden of surgical con- as a whole and by region. The table indicates that conditions ditions as measured in DALYs. To our knowledge, this meas- requiring surgery account for a significant proportion of urement has never been attempted. What we provide here is a DALYs. Developing more refined, region-specific information starting point, with the understanding that the calculations will to help policy makers will require more detailed data on the change as data are developed. burden of surgical diseases (diseases requiring surgical treat- Our methodology was based on data from the World ment) and on the cost-effectiveness of surgical therapy. To this Health Report 2002: Reducing Risks, Promoting Healthy Life end, an extremely helpful step would be for international (WHO 2002) and the global burden of disease study (Murray surgical associations to regularly monitor the disease burden and Lopez 1996). We began by listing all the conditions for attributable to surgical conditions throughout the world. which surgery might be indicated into three groups, with group A few salient points about the burden of surgical diseases I being communicable diseases, group II being noncommuni- can be made from data provided in table 67.1. We estimate very cable diseases, and group III being injuries. We then undertook conservatively that 11 percent of the world's DALYs are a comprehensive literature review for each condition to deter- from conditions that are very likely to require surgery. Our esti- mine the proportion of the total burden of disease attributable mated figures are as high as 15 percent for Europe and as low to it and the proportion of the burden that could be prevented as 7 percent for Africa. Estimated surgical DALYs for the world or treated by surgery. Essentially, we found no data of value are 27 per 1,000 population. The estimated figure is about twice except maybe for cataracts (group II-F), for which a single as much for Africa (38 per 1,000) as for the Americas (21 intervention (intraocular lens removal with or without per 1,000). implant) is or should ultimately be indicated for nearly 100 per- Table 67.2 summarizes the burden of common surgical con- cent of patients (Dandona and others 1999; Javitt 1993). ditions based on World Health Report 2002 data. A more 1246 | Disease Control Priorities in Developing Countries | Haile T. Debas, Richard Gosselin, Colin McCord, and others Table 67.1 Estimated Surgical DALYs by Region Estimated Estimated Estimated surgical surgical DALYs surgical DALYs Total DALYs DALYs as a percentage per 1,000 Region (millions) (millions) of total DALYs population World 1,468 164 11 27 Africa 358 25 7 38 Americas 145 18 12 21 Eastern Mediterranean 137 15 11 30 Europe 151 22 15 25 Southeast Asia 419 48 12 31 Western Pacific 258 37 15 22 Source: WHO 2002 and authors' estimates. Table 67.2 Burden of Common Surgical Conditions Estimated Estimated Surgical DALYs surgical DALYs surgical DALYs Estimated as a percentage per 1,000 Condition (millions)a Percentage of total DALYs population Injuries 63 38 4.3 10 Malignancies 31 19 2.1 5 Congenital anomalies 14 9 1.0 2 Obstetrical complications 10 6 0.7 2 Cataracts and glaucoma 8 5 0.5 1 Perinatal conditions 7 4 0.5 1 Other 31 19 2.1 5 Source: WHO 2002 and authors' estimates. a. Estimated surgical DALYs refers to our conservative estimate of DALYs averted by surgical treatment in the most likely diseases for the most likely indications. detailed look at these data allows us to make the following · Congenital anomalies refer to an ill-defined grouping of dis- observations: parate pathologies that includes congenital malformations such as cleft lip and palate, hernias, anorectal malforma- · Injuries account for 63 million DALYs, or about 4 percent of tions, and clubfoot. We estimate that some 50 percent of all DALYs and 38 percent of the world's estimated surgical congenital anomalies are surgical, representing about 14 mil- DALYs. lion DALYs, or 1 percent of all DALYs. · Surgical infections, including infected wounds, superficial · Malignancies account for 31 million surgical DALYs, or and deep abscesses, septic arthritis, and osteomyelitis, slightly more than 2 percent of all DALYs. undoubtedly account for a significant portion of surgical DALYs, but the available data do not permit quantification. Table 67.3 breaks down the burden of common surgical · Surgical DALYs pertaining to acute abdominal conditions, conditions by region, also showing rates per 1,000 population. including appendicitis, intestinal obstruction, gastrointesti- The absolute burden of injuries is highest in Southeast Asia, nal bleeding, hernias, and blunt or penetrating injuries also followed by the Western Pacific and Africa. In terms of popula- cannot be calculated because of the lack of data. tion rates, whereas injuries account for 10 DALYs per 1,000 · Approximately one-third of maternal conditions, including population for the world, the estimated figure is almost twice as hemorrhage, obstructed labor, and obstetrical fistulas, are much for Africa (15 per 1,000) as for Europe (8 per 1,000). surgical, and these represent 10 million DALYs, or 0.7 per- Similarly, rates of obstetrical complications are far higher in cent of all DALYs. Africa than elsewhere, at 6 DALYs per 1,000 population. In Surgery | 1247 Table 67.3 Estimated Surgical DALYs by Condition and Region (DALYs in millions followed by DALYs per 1,000 population in parentheses) Eastern Southeast Western Condition Africa Americas Mediterranean Europe Asia Pacific Injuries 10 7 6 7 20 13 (15) (8) (12) (8) (13) (8) Obstetrical complications 4 1 1 3 1 (6) (1) (2) ( 0.5) (2) (1) Cataracts and glaucoma 1 1 1 3 2 (2) ( 0.5) (2) (1) (2) (1) Malignancies 2 4 1 8 6 10 (3) (5) (2) (9) (4) (6) Perinatal conditions 2 1 1 3 1 (3) (1) (2) ( 0.5) (2) (1) Congenital anomalies 2 2 2 1 4 3 (3) (2) (4) (1) (3) (2) Other 4 3 3 5 9 7 (6) (4) (6) (6) (6) (4) Source: WHO 2002 and authors' estimates. contrast, Europe has the highest rate of surgical DALYS related those topics. A population-based approach to injury should to malignancies--9 per 1,000 population. not, however, be limited to injury prevention. Patients may sur- All these estimates are debatable. Work is needed to obtain vive their primary injuries only to become chronically disabled more valid, accurate, and reliable data, but in the meantime, we and a burden to their families and to society (Krug, Sharma, believe that our results represent a conservative and acceptable and Lozano 2000; Mock and others 1999; Nantulya and Reich baseline estimate of the burden of surgical conditions against 2002; Peden and Hyder 2002). The incidence and severity of which prospectively gathered data for given interventions can be the complications of survivable injury may be significantly compared in order to assess the extent to which they address the lessened by the provision of adequate surgical care during pre- burden.In addition,the burden needs to be monitored over time. hospital care and initial hospitalization. No published data Evidence suggests that the burden of intentional and uninten- from developing countries are available, however, either to tional injuries is rising, particularly in Sub-Saharan Africa and prove this plausible contention or to quantify the benefits of the Middle East. Some of the important contributing risk factors adequate initial surgical treatment. A strategy to prevent include (a) aging populations; (b) increased access to and use of chronic disability arising from survivable injury requires well- mechanized vehicles and tools without commensurate improve- coordinated services for resuscitation, evacuation, and early ments in roads, traffic control systems, or capacity for trauma and expert operative management of the initial injury. care; and (c) persistent armed conflicts (Kaya and others 1999; Many other surgical conditions that can be treated elec- Krug, Sharma, and Lozano 2000; Meyer 1998; Mock and others tively, such as hernias, hydroceles, and otitis media, are treated 1995, 1999; Nantulya and Reich 2002; Peden and Hyder 2002). when they present with complications requiring emergency surgery. Thus, a pertinent question is whether treating such conditions electively would be more cost-effective, but no reli- INTERVENTIONS able data are available to answer this question positively or negatively. Both population-based strategies and personal sevices pro- Population-based strategies could also be applied to prevent vided in community clinic, district, and tertiary hospitals are or treat some musculoskeletal conditions. For example, the considered in this section. incidence of clubfoot is estimated at 1 or 2 per 1,000 live births, but in developing countries these children are typically brought in for orthopedic care when it is too late for effective nonsurgi- Population-Based Strategies cal conservative management (Ponseti 1999; Turco 1994). Population-based approaches to the prevention of uninten- Because we have no baseline data for the burden of clubfoot tional and intentional injuries are discussed in the chapters on and other musculoskeletal conditions, we are unable to 1248 | Disease Control Priorities in Developing Countries | Haile T. Debas, Richard Gosselin, Colin McCord, and others quantify the DALYs that could be averted by comprehensive Table 67.4 shows the infrastructure requirements for this surgical care. type of hospital. Patients requiring more complex imaging The following sections describe the organization of surgical studies and laboratory tests would be referred to the tertiary services that we think would begin to provide coordinated sur- hospital. gical care in developing countries. The provision of surgical To the extent possible, all equipment and supplies services in developing countries requires organizational struc- (table 67.4) should be standardized, and an efficient and reliable ture and capacity at the level of community-based clinics, system for maintenance and replacement should be ensured. district hospitals, and tertiary care hospitals. Our concept of Operating room instruments and supplies should be available minimally adequate modules of surgical care is informed by to enable the performance of laparotomy, thoracotomy, obstet- our personal experiences, the experiences of others, and a rical and gynecological procedures, treatment of extremity frac- recent World Health Organization report (WHO 2003). We tures, skin grafting, and emergency burr hole of the skull. These recognize that to accommodate local needs and reality on the instruments should be available at least in duplicate. Table 67.4 ground, any proposed plan to develop surgical services must be also shows workforce needs and the types of surgical proce- flexible. Table 67.4 presents our estimates of the needs for dures that may be performed in a district hospital. infrastructure, equipment and supplies, and workforce for the The district hospital is assumed not only to serve as the three levels of surgical care: community clinic, district hospital, referral hospital for community clinics, but also to coordinate and tertiary hospital. the community clinics in its own region as a single operating unit, assuming responsibility for wireless communication, training the workforce, providing continuing medical educa- Community Clinics tion, and monitoring the quality and outcome of care. It would Table 67.4 shows resource and workforce requirements and also provide primary care to its contiguous population. types of surgical services a community clinic could provide for a population of around 20,000. We assume that surgical services in community clinics would be provided at no cost to Tertiary Hospitals patients. A cost-recovery system would be unlikely to succeed The tertiary hospital would function as the referral center for everywhere but, if implemented, should be equitable, with pay- all complex surgical care in a region, country, or group of ments adjusted to patients' ability to pay. A mechanism for countries. Ideally, but depending on the country's resource accountability and monitoring should be established to avoid constraints, it would provide the full range of care shown in the misuse of drugs and supplies. Simple patient records table 67.4. The tertiary hospital would also provide primary should be maintained, including outcomes of treatment and surgical care for its local population and could take on the role use of supplies. Even though the community clinic described of a teaching hospital for doctors, nurses, and other health care here is what we think it should be as opposed to what we know workers. it to be, our model may serve the needs of rural areas in devel- In the proposed organizational structure, the tertiary hospi- oping countries and could provide a starting point for estimat- tal is viewed as the top of a pyramid of surgical services, with ing costs. several district hospitals referring patients requiring complex surgical care to the tertiary hospital. As such, it should also take the primary responsibility for coordinating and collaborating District Hospitals with all the district hospitals and community clinics in its The next level of organization of surgical services is the district area of responsibility to ensure that surgical care is available hospital, which in addition to providing primary care for the throughout the region and that well-functioning wireless com- local population would also provide secondary-level surgical munication and ambulance systems are available. If a regional- services and serve as a referral center for a number of commu- ized system of separate ambulance services is not available, the nity clinics in a defined region. In turn, the district hospital tertiary hospital can provide the ambulance services required. would ideally refer patients requiring complex surgical care to Specialists in the tertiary hospital should provide telephone a tertiary-level hospital, but we recognize that such referral and electronic consultation for doctors and nurses in district cannot always be achieved in practice because of transporta- hospitals. The tertiary hospital should also coordinate and tion limitations, economic constraints, and prevalent social monitor the quality of care in the region that serves as its refer- and cultural contexts. District hospitals vary in size from as ral base, undertake clinical outcome studies, and provide con- small as 10 to 20 beds to as large as 200 to 300 beds and vary in tinuing medical education. In addition, it should be the main their degree of sophistication in relation to diagnostic and ther- teaching hospital for medical students, nurses, and technicians, apeutic capabilities. For this discussion, we have arbitrarily with the district hospitals and even the community clinics serv- chosen to focus on district hospitals with 100 beds or fewer. ing as clinical rotation sites for trainees. This organizational Surgery | 1249 Table 67.4 Resource Requirements for Surgical Services and Surgical Procedures by Level of Care Category of requirement Community clinic 100-bed district hospital Tertiary hospital Infrastructure Weatherproof building (100 square Inpatient facility of 100 beds including several wards A major facility providing: meters) and an isolation ward · Full emergency services with advanced diagnostic Storage space Outpatient facility including an emergency room services; Clean water supply Operating rooms (at least two: one clean, · Inpatient wards for complex general medical and one contaminated) surgical care Power supply Labor and delivery rooms · Various types of specialty services Recovery room or intensive care unit · Several delivery rooms and operating rooms Blood bank · One or more recovery rooms and intensive care units Pharmacy · Rehabilitation and occupational therapy facilities Clinical laboratory Radiology and ultrasonography suite Equipment and Furniture Anesthetic machines and inhalation gases All required equipment and supplies to undertake the supplies range of routine and complex services provided Refrigerator Monitors (electrocardiogram, blood pressure, pulse oximetry) Blood pressure machine Fully equipped operating room Minor surgical trays Fully equipped delivery room Sterile and burn dressings Fully equipped recovery room or intensive care unit Autoclave Respirators and oxygen supply Intravenous sets and solutions Blood products and intravenous fluids Bandages and splints Basic microbiology equipment Drugs: local anesthetics, nonsteroidal anti-inflammatory Pharmaceuticals (anesthetics, analgesics, antibiotics) drugs, antibiotics, tetanus toxoid, Surgical materials (drapes, gowns, dressings, gloves) silver nitrate ointment, oxytocin, and other consumables (disposable equipment and magnesium sulfate devices) Wireless communication equipment Materials for recordkeeping Human resourcesa Nurse or nurse equivalent Nurses (20) Nurses (50) Skilled birth attendant Midwives (2­3) Midwives (5) Orderly Anesthetists (2­3) Anesthetists (5) Anesthesiologist (1)b Anesthesiologists (3) Primary care physicians (4)c Primary care physicians (10) Obstetrician/gynecologist (1­2) Obstetricians/gynecologists (5) General surgeons (2) General surgeons (5) Pharmacy assistants (2) Orthopedic surgeon (1) Pharmacist (1)b Pharmacy assistants (2) Radiology technician (1) Pharmacist (1) Radiologist (1) Radiology technicians (5) Physiotherapist (1) Radiologists (2) Physiotherapists (5) Neurosurgeon (1)b Cardiac surgeonb Reconstructive surgeonb 1250 | Disease Control Priorities in Developing Countries | Haile T. Debas, Richard Gosselin, Colin McCord, and others Table 67.4 Continued Category of requirement Community clinic 100-bed district hospital Tertiary hospital Services provided Simple suturing and dressing of Emergency abdominal, thoracic, head injury Full emergency service wounds Management of all complex general surgery Incision and drainage of abscesses Full range of services in orthopedics, trauma, urology, Care of simple burns Uncomplicated general surgical operations for otolaryngology, ophthalmology, and obstetrics and Control of hemorrhage hernias, anorectal conditions, and biliary tract disease gynecology Splinting Surgical infection treatment and control Basic (and, if resources permit, advanced) neurosurgery and cardiovascular surgery Deliveries Obstetrical (including surgery for complications) Intensive care services Vacuum extraction and manual Simple orthopedic care: extremity fractures, vacuum aspiration dislocations, and amputations Major burn service Burn care Radiology services including CT and MRI imaging and angiography Physiotherapy and occupational therapy Full service clinical laboratory Education and training Physiotherapy Occupational therapy Training of doctors, nurses, and midwives Source: Authors. a. Because of the variability in size and the complexity of services provided by tertiary hospitals, it is difficult to describe a standard tertiary hospital; the human resource needs given in the table represent what we think are minimally adequate. b. Desirable, but not absolutely necessary. c. Can be a general internist, general practitioner, or general pediatrician. structure is ideally suited for the tertiary hospital to serve as the from around the world. Associations such as the International backbone of community-based surgical education. The extent Surgical Association could develop a Web portal tied to national to which this ideal function of a tertiary hospital can be imple- surgical associations to ensure greater success in this regard. mented will depend on the financial and other resources avail- A regionalized system for the purchase and delivery of able to the country. equipment and supplies is highly desirable. Such a system could ensure that all equipment and supplies were standardized and made available on demand in an efficient and predictable Coordinated Model System for Surgical Care manner. The proposed system for surgical services requires the coordi- Ground ambulance services are essential for patient transfer. nation and integration of the following: In some areas, collaboration with local taxi or bus services might offer the needed support. In some more economically · wireless communication advanced countries, tertiary hospitals might be able to provide · continuing education programs ambulance services using fixed-wing aircraft or helicopters. · regionalized supply system for equipment, essential drugs, If the proposed model for a surgical system is to be devel- and surgical materials oped, systems for ongoing data acquisition and for evaluation · ambulance service and monitoring should be built into the model. In this way, not · uniform data collection system only could information be captured, but also the quality and · coordinated and ongoing monitoring of quality and out- outcomes of care could be monitored on an ongoing basis. comes of care. A wireless system of communication could render costly COSTS AND COST-EFFECTIVENESS wired systems unnecessary and could connect community clin- OF INTERVENTIONS ics, district hospitals, and tertiary hospitals in a dependable way that facilitates consultation and referral. A Web-based system of In today's resource-constrained world, policy makers increas- communication could be particularly important for mentoring ingly need to be aware of the value of selective health care inter- and for providing continuing medical education. The Web can ventions. Cost-effectiveness analysis is one method that links also be used to enhance contributions by volunteer surgeons, inputs (costs) with the resulting health care gains measured anesthesiologists, surgical specialists, nurses, and technicians along a common metric, usually using DALYs. Surgery | 1251 Even though an extensive body of literature examines the defined as treatment of bruises, simple cuts requiring suturing, cost-effectiveness of a range of nonsurgical interventions in foreign body removal, drainage of abscesses, basic burn treat- developing countries (Jha, Bangoura, and Ranson 1998), the ment, normal deliveries, and simple trauma. literature examining surgical interventions in these countries is As far as possible, we used standardized regional cost esti- more sparse. Moreover, most of the available studies examine mates provided to the authors. When such information was surgical interventions for specific conditions (Marseille 1996; unavailable, we used our consensus judgment. Given the wide Singh, Garner, and Floyd 2000). A common criticism of such variation in costs between and within regions, we conducted studies is that they do not fully capture the choices policy mak- sensitivity analyses to capture the range of possible outcomes. ers face in real life. For example, policy makers must often When more than one source of cost estimates was available, the choose between allocating resources for constructing several mean of the estimates for that region were used as the best esti- community clinics or a single district hospital, both of which mate and a high-low range was noted. However, in many cases, provide a mix of surgical and nonsurgical services. Generally, only a single cost estimate could be obtained, in which case the the surgical ward in a district hospital will provide care for a data provide a point estimate,2 and we vary the cost estimate by wide range of conditions, such as trauma, childbirth, and ±20 percent to obtain a high-low range. abdominal conditions. We assume that for policy makers, Our calculation of the number of DALYs averted was based knowing the cost-effectiveness of the surgical service, ward, on the work of McCord and Chowdhury (2003), who calculate or clinic (as an intervention) is more useful than information the DALYs averted by a 50-bed hospital in Bangladesh, as about the cost-effectiveness of each condition-specific surgical described in box 67.1. We adjusted this figure to reflect the bed intervention. Unfortunately, no literature exists that examines size of our standard district hospital. In the absence of region- the cost-effectiveness of a surgical service or ward. This section specific data, we applied this figure to all six regions after attempts to fill that void with respect to district hospitals and making suitable adjustments. For the community clinic, we community clinics but not in relation to tertiary-level hospi- estimated that such a clinic averts approximately 200 DALYs tals, which vary in size, available resources, and role from per year as a result of surgical treatment, primarily from the region to region, making it difficult to describe the cost- incision and drainage of abscesses and the preliminary treat- effectiveness of a prototypical tertiary hospital. ment of burns. Because these DALY estimates are based on a single source, we vary the estimate by ±20 percent to obtain a high-low range and apply these estimates across the six regions. Method for Estimating Costs and DALYs On the basis of the resource requirements listed in table 67.4, we developed cost estimates for each of the six regions defined Results by the World Bank. Table 67.5 details the assumptions and Figure 67.1 presents the results of the cost per DALY averted cal- table 67.6 provides the regional costs. We defined the standard culations for a district hospital and community clinic. The low hospital in such a way as to facilitate comparisons across estimate represents the scenario in which the costs are the lowest regions, conceptualizing it as a 100-bed hospital with a male and the DALYs averted are the highest--that is, the best-case sce- ward and a female ward; two operating rooms; a recovery nario. In a similar vein, the high estimate is the worst-case sce- room, an intensive care unit, or both; an x-ray unit and an nario: the costs are highest and the DALY averted is the lowest. ultrasound machine; and a laboratory that can carry out basic The best estimates for cost per surgical DALY averted at blood chemistry tests, examine urine, and cross-match blood. a community health center (panel a of figure 67.1) hover in a This hospital also has an on-site laundry and kitchen and two narrow range between US$212 and US$241. The cost per sur- vehicles to serve as ambulances. The staff consists of 6 doctors gical DALY gained at a district hospital is cheapest for Sub- (4 primary care physicians, 1 obstetrician and gynecologist, Saharan Africa at US$33 (range of US$19 to US$102) and most and 1 general surgeon); 20 nurses; 6 midwives; 2 physiothera- expensive for Latin America and the Caribbean at US$94 pists; and 6 orderlies. The costs of an anesthetist and x-ray (range of US$47 to US$164). technician have been included in the operating costs of the Standard economic costs can differ from costs actually operating rooms and radiology area, respectively. The model incurred in service delivery, both because in practice not all time assumes that the hospital averages 80 percent occupancy and may need to be paid for (for example, hospitals may be able to that two-thirds of inpatients will be surgical cases.1 economize on staff because relatives help care for patients) and We defined a standard community clinic (see table 67.4) as because low-cost solutions may be found (for example, use of a facility of 100 square meters serving a population of approx- paramedical staff members in place of professionals). Box 67.2 imately 20,000, staffed by a nurse or nurse-substitute, a skilled describes some of these strategies and compares the standard birth attendant, and an orderly. Such a clinic treats approxi- economic cost presented above with the much lower financial mately 4,000 surgical cases per year, with a surgical case being cost of a nongovernmental organization (NGO) hospital. 1252 | Disease Control Priorities in Developing Countries | Haile T. Debas, Richard Gosselin, Colin McCord, and others Table 67.5 Costing Assumptions for District Hospital and Community Clinic Category Assumptions and comments 100-bed district hospital Inpatient hospital bed days The estimate includes "hotel" costs (capital, salaries, overhead, building, equipment, and food) for a hospital running at 80 percent occupancy. Assumption: two-thirds of all cases seen are surgical and costs are adjusted accordingly. Operating rooms The estimate includes all operating room­related costs (surgeon, nurses, equipment, and so on). Assumption: the operating room is running for 8 hours/day, 5 days/week, 52 weeks/year. Laboratory Assumption: 16 new admissions/day, all of whom will require a laboratory test once during their hospital stay. Yearly costs have been adjusted to reflect the percentage of surgical cases. X-ray Assumption: 16 new admissions/day, of whom half will undergo an x-ray examination once during their hospital stay. Yearly costs have been adjusted to reflect the percentage of surgical cases. Pharmacy Assumption: all admitted patients will use US$10 worth of drugs during their stay. Yearly costs have been adjusted to reflect the percentage of surgical cases. Blood transfusion Assumption: 400 units of blood transfused/year. Ambulance The estimate includes 2 ambulances; 15 percent markup for freight, insurance, and distribution; US$500/year running costs added; ambulance depreciated over 9 years. Staff Nurses Assumption: nurses devote two-thirds of their time to surgical patients and costs have been adjusted accordingly. Midwives Assumption: midwives devote 100 percent of their time to surgical patients. Doctors These are the four doctors whose costs have not been included in the operating room costs listed above. Because they provide ward coverage, two-thirds of their costs are attributed to surgical patients. Physiotherapists Assumption: physiotherapists devote two-thirds of their time to surgery patients. Orderlies Assumption: six orderlies, each devoting two-thirds of their time to surgical activities. Community clinic Building costs The estimate includes costs of lighting and power. Assumption: building size is 100 square meters, 20-year straight-line depreciation is used; 25 percent of costs are attributed to surgical patients. Water supply Assumption: US$20 for water supply, of which 25 percent attributed to surgery. Supplies The estimate includes surgical trays, sterile and burn dressings, intravenous sets and cannulas, bandages, splints, and plaster of Paris for 4,000 patients. Assumption: local costs of US$5/patient. Drugs The estimate includes local anesthetics, nonsteroidal anti-inflammatory drugs, tetanus toxoid, silver nitrate, and basic antibiotics. Assumption: 4,000 patients, US$5/patient. Furniture The estimate includes autoclave, surgical light, examination table, and beds. Assumption: total cost US$2,000, straight- line depreciated for 10 years. Refrigerator The estimate includes locally manufactured refrigerator, straight-line depreciated for 5 years; surgical cost attribution only. Wireless telephone Assumption: US$200 cost; 25 percent attributed to surgery. Staff Nurses and nurse-substitutes Assumption: nurses devote 25 percent of time to surgical duties and costs adjusted accordingly. Skilled birth attendants Assumption: attendants devote 25 percent of time to surgical duties and costs adjusted accordingly. Orderlies Assumption: orderlies devote 25 percent of time to surgical duties and costs adjusted accordingly. Sources: Authors. Note: The figures for the district hospital should be viewed with particular caution. First, we used a single source of data for our assumptions. Had we used other data sources, the results could conceiv- ably be different. Second, as a prototype we used a Sub-Saharan Africa district hospital that provides basic, low-tech surgical services. The provision of more sophisticated care can be expected to drive up the costs of care significantly. Discussion averted at a community clinic tend to be higher than those The data in figure 67.1, when compared with similar data for averted at a district hospital despite the lower costs of a com- other services presented in this book, indicate providing basic munity clinic. Although these observations may be taken as surgical services is relatively cost-effective. Figure 67.1 also evidence that surgical services are best provided at the district indicates that, from a surgical perspective, the costs per DALY hospital level, this goal may be impossible to put into practice. Surgery | 1253 Table 67.6 Annual Costs Attributable to Surgical Patients in a District Hospital and a Community Clinic, Best Estimates (2001 U.S. dollars) Latin America Middle East Sub- East Asia and Europe and and and North Saharan Category the Pacific Central Asia the Caribbean Africa South Asia Africa 100-bed district hospital Inpatient hospital bed days 204,042 363,277 640,071 498,904 156,826 148,635 Operating rooms 778,752 1,130,688 1,163,136 896,064 526,656 419,328 Laboratory 34,304 44,403 45,251 37,619 27,058 23,974 X-ray 18,578 26,788 27,501 21,276 12,700 10,195 Pharmacy 38,544 38,544 38,544 38,544 38,544 38,544 Blood transfusion 7,572 8,548 8,632 7,892 6,872 6,572 Ambulance 7,389 7,389 7,389 7,389 7,389 7,389 Staff Nurses 24,652 29,808 58,591 95,227 21,973 30,390 Midwives 11,100 13,422 26,382 42,870 9,894 13,680 Doctors 11,856 14,340 28,187 45,807 10,571 14,618 Physiotherapists 3,522 4,258 8,370 13,604 3,139 4,341 Orderlies 5,714 6,914 13,587 22,081 5,097 7,045 Total 1,146,026 1,688,380 2,065,641 1,727,277 826,718 874,551 Community clinic Building costs 1,280 1,321 601 1,324 569 574 Water supply 5 5 5 5 5 5 Supplies 20,000 20,000 20,000 20,000 20,000 20,000 Drugs 20,000 20,000 20,000 20,000 20,000 20,000 Furniture 200 200 200 200 200 200 Refrigerator 10 10 10 10 10 10 Wireless telephone 50 50 50 50 50 50 Staff Nurses and nurse-substitutes 667 807 1,585 2,577 595 822 Skilled birth attendants 667 807 1,585 2,577 595 822 Orderlies 361 437 858 1,394 322 445 Total 43,240 43,635 44,895 48,136 42,345 42,928 Sources: All district hospital costs and community clinic building and staff costs: DCPP guidelines; other community clinic costs: authors' estimates. The type of surgical care provided at the community clinic per surgical DALY averted ranging between US$33 and level, though not resulting in a very large DALY gain, is $US38; Europe and Central Asia, Middle East and North nevertheless important. It is inconceivable to think of a com- Africa, and Latin America and the Caribbean seem to be the munity clinic that does not have facilities for minor foreign most expensive, with the cost per surgical DALY averted rang- body removal, simple suturing of cuts and wounds, or splint- ing between US$77 and US$94; and East Asia and the Pacific ing of simple fractures. Furthermore, community clinics' refer- falls in the middle. This finding indicates that, from the per- ral and primary treatment functions, which are hard to evalu- spective of providing surgical care, a district hospital is an ate separately from the delivery of final treatment, are critical exceptional "buy" in Sub-Saharan Africa and South Asia, both for many conditions, notably trauma. areas with high disease burdens. Coupled with evidence that Costs per surgical DALY averted at the district hospital district hospitals are comparatively underfunded compared level seem to fall into three groups. Sub-Saharan Africa and with national (tertiary) hospitals (Fiedler, Wight, and Schmidt South Asia are the cheapest, with the best estimates of cost 1999), a prima facie case exists for increasing support for 1254 | Disease Control Priorities in Developing Countries | Haile T. Debas, Richard Gosselin, Colin McCord, and others Box 67.1 Estimation of the DALYs Averted by a Small Hospital in a Developing Country The DALY estimates in this chapter are based on a report Our estimates of the risk of death or serious disability from a 40-bed nongovernmental hospital in rural were based on tables McCord and Chowdhury created and Bangladesh in 1995 (McCord and Chowdhury 2003). were necessarily arbitrary, but we believe they are Obviously this experience was localized from one hospital, extremely conservative. For example in one country, at one time, but it is the only analysis avail- able from such a hospital that estimates effectiveness using · If the chance of death or disability was less than 5 per- DALYs or any other measure of the effect of a hospital on cent, it was considered to be 0 percent. Because normal the disease burden. The hope is that other similar studies out-of-hospital deliveries had less than a 5 percent done in a variety of hospital situations will permit gener- chance of death or disability, normal deliveries in the alizing with more confidence, but our personal experience hospital made no contribution to the estimated DALYs leads us to believe that the disease pattern presented to averted. small district hospitals in poor countries is remarkably · A cesarean section for obstructed labor was estimated constant, especially for surgical conditions. at 10 percent averted risk for the infant and 0 percent McCord and Chowdhury (2003) present the methods for the mother. for calculating DALYs in detail. They reviewed all dis- · A cesarean for transverse lie was estimated at 90 per- charges and deaths every week for three months, con- cent averted risk for the infant and 90 percent for the firmed the discharge diagnosis by means of a chart review, mother. and estimated the percentage chance that the hospital stay · An appendectomy for nonruptured appendicitis was had prevented death or disability. The review covered all estimated at 10 percent averted risk, because many patients discharged, classifying them as medical, surgical, cases of appendicitis respond to antibiotics outside the obstetrical and gynecological, or pediatric. Of the dis- hospital setting. charges, 62 percent were of surgical and obstetrical and · An appendectomy for a ruptured appendix with gynecological patients. Of the DALYs, 21 percent of those generalized peritonitis was estimated at 90 percent averted were for surgical patients and 61 percent from averted risk. obstetrical and gynecological patients. Eighty-nine percent · An elective herniorrhaphy for a small hernia was of the estimated DALYs averted were generated by averting estimated at 0 percent averted risk. premature death, and only 11 percent by preventing seri- · A herniorrhaphy for a large, disabling hernia was ous disability. Of the 192 surgical operations, 118 were estimated at 80 percent averted risk. emergencies. Of 137 obstetrical patients, 81 had compli- · A herniorrhaphy for a strangulated hernia that did not cated deliveries, complications of abortion, or ectopic reduce with conservative management was estimated at pregnancies. 80 percent averted risk. Source: Authors. district hospitals in developing countries. However, those pro- than 40 million people will be blind or almost blind because of viding such support have to be cognizant of realities on the cataracts (Brian and Taylor 2001). Box 67.3 describes a success- ground, especially political realities, because they have a sig- ful program in India. nificant effect on the direction of change (Blas and Limbambala 2001). Data on the cost-effectiveness of surgical interventions for specific conditions in developing countries are scarce. One RESEARCH AND DEVELOPMENT AGENDA notable exception is for the surgical treatment of cataracts (removal of the opaque lens with or without the insertion of an The literature on surgical care in developing countries is so intraocular implant). Blindness from cataracts is a significant meager that insufficient data are available to formulate an public health problem in many developing countries, and as agenda for research and development. Hence, of necessity, the their populations age, estimates indicate that by 2020 more research that needs to be done is extremely basic, much of it Surgery | 1255 a. Community Clinic specific DALYs that can be averted by means of surgical intervention. We have applied the DALYs averted from a sin- US dollars 450 gle study in a developing country (McCord and Chowdhury 2003) to other regions, a procedure that negates regional dif- 400 ferences in disease incidence, health care­seeking behavior, 350 case mix, and clinical practice variations. In addition, the calculation of DALYs averted should ideally be adjusted for 300 region-specific life expectancy and disability weights. 250 · Estimation of costs, both at a facility and regional level, is 200 needed, including reducing variability in estimation meth- ods (Adam and Koopmanschap 2003). In addition, multiple 150 estimates of costs are needed. For example, Mulligan and 100 others (2003) derive their operating room costs from a sin- 50 gle study of ambulatory surgery in Colombia (Shephard and others 1993). Even though they made adjustments to reflect 0 regional characteristics, further research is required to vali- and and and and Asia date their results, especially as they apply to different set- AsiaPacific Asia East Africa Africa Europe South America tings in different countries. East the Caribbean Sub-Saharan Central MiddleNorth Latin the · Better surgical data collection and analysis tools critical to needs assessment should be designed. Low estimate Best estimate High estimate · Development of appropriate surgical care models for all lev- b. District Hospital els of care based on local and regional characteristics and surgical needs would be useful. US dollars · Cost-effectiveness and cost-benefit analyses of health 200 systems implementation need to be determined, as do the 180 policy implications of creating the surgical care model 160 proposed in this chapter. The evaluation of surgery as a 140 prevention strategy in public health should include cost- 120 effectiveness analysis of adequate, prompt, initial surgical treatment of injury to prevent chronic disability from poorly 100 diagnosed and treated survivable injuries and of elective 80 treatment of hernia, hydrocele, otitis media, cataract, and 60 clubfoot to prevent complications and disability. 40 · The surgical workforce in developing countries requires more in-depth study to look at the mixes of workers needed, 20 the level of training required for the widely varying local sit- 0 uations of district hospitals, and the role for part-time sur- and and and and Asia gical talent. The thesis is that volunteer doctors, nurses, and AsiaPacific Asia East Africa Africa Europe South America anesthesiologists who now contribute considerably to surgi- East the Caribbean Sub-Saharan Central MiddleNorth Latin the cal care in developing countries in a relatively unstructured fashion could do so more effectively and in a manner that Low estimate Best estimate High estimate could help create sustainable local surgical workforces if a Source: Authors' calculations using costs in table 67.6 and methods for estimating well-coordinated system with extensive information and DALYs averted described in box 67.1. communication support could be developed. This concept Figure 67.1 Cost per Surgical DALY Averted for a Community Clinic merits in-depth study. If a well-planned, Web-coordinated, and a District Hospital global, highly integrated system could be developed, health care volunteers around the world could be organized information gathering. The following are some of the areas that strategically so as to deliver not only surgical care, but also require investment in research and development: training of local surgical workforces. The emphasis on training is crucial and would mitigate the complaints often · Estimates are needed of the burden of disease that requires heard that surgical volunteers too often contribute to the surgical intervention along with a determination of region- care of individual patients but fail to leave behind a 1256 | Disease Control Priorities in Developing Countries | Haile T. Debas, Richard Gosselin, Colin McCord, and others Box 67.2 Surgical Cost in a Bare-Bones Hospital The estimated economic costs in this chapter assume · Specialized services are provided by general physicians staffing and service levels generally derived from World or technicians trained to do surgery or give anesthesia. Health Organization recommendations for developing · Laboratory tests and x-rays are used sparingly. The only countries (Mulligan and others 2003), but in many places, laboratory procedure for an obstetrical patient could surgical services are delivered in much simpler and less be a hemoglobin determination. expensive facilities. Independent project hospitals (NGO · Only basic medicines are provided. More expensive or hospitals) often operate on remarkably tight budgets. complicated supplies are purchased outside the hospi- Private hospitals, often set up in private houses by indi- tal by the family. vidual surgeons, use locally trained staff with minimal For one independent, nongovernmental hospital in "hotel" service. Extremely poor countries operate hospi- Bangladesh, we were able to obtain the actual financial cost tals with a cost per patient per day of US$10 or less simply of all aspects of hospital operations during a three-month because they cannot afford more. Such hospitals achieve period. These costs included salaries, supplies, hotel costs, financial savings in several ways: and depreciated cost of equipment and buildings, as well as · Unpaid family members provide personal nursing care an overhead estimate to allocate a share of the total project and food, eliminating the need for a kitchen and many cost for administration, electricity, transportation, and trained nurses. so on. Separating the surgical service costs for 3 months, · Locally trained staff members substitute for profes- extrapolated to 12 months and a 100-bed hospital, we come sionally trained personnel. up with a much lower total cost than the low estimate in · Many staff members have duplicate functions. In the this chapter for the economic costs of a model district hos- operating room the same person may work as surgical pital, as shown in the table below. Part of the difference is assistant, scrub nurse, and orderly who cleans instru- caused by differing cost definitions (financial versus eco- ments or transports patients. nomic), but a good part is owing to the use of low-cost · Day staff members cover night calls for emergencies. approaches to the delivery of surgical care. Category NGO hospital (2001 US$) District Hospital in South Asia (2001 US$) Inpatient bed days 110,936 156,826 Operating time 178,508 526,656 Laboratory 11,788 27,058 X-Ray 6,676 12,700 Pharmacy n.aa 38,544 Blood transfusion 3,858 6,872 Ambulance n.a.a 7,389 Staff n.a.b 50,673 Total 311,766 826,718 Source: NGO hospital: McCord and Chowdhury 2003; district hospital: table 67.6. n.a. not available. a. Included in overhead, which is added to each cost center. b. Included in total cost of ward and operating room. mechanism for sustaining surgical care when they have left. countries. Convincing demonstrations of how much can be Those volunteers who come from the high-tech world of done without recourse to CT scans, ultrasound, and video- modern surgery should realize that the latest technology is assisted surgery could be the most useful contribution a vis- often more of a burden and diversion than a help in poor itor could make. Surgery | 1257 Box 67.3 Success Story: Cataract Surgery in India Prevention strategies aimed at known risk factors, such as for delivering cataract surgery in Mysore, India, by assess- tobacco use and exposure to the sun, are unlikely to have ing outcomes in a systematic sample of patients operated a significant effect on the need for surgical treatment of on in 1996­97. Patient satisfaction was 51 percent among cataracts in the foreseeable future (Ellwein and Kupfer those operated on in government mobile camps, 82 per- 1995). The benefits of cataract surgery have been well doc- cent among those treated at the medical college hospital, umented in many developing countries. and 85 percent among those treated in nongovernmental In 1993­94, with a World Bank credit of US$118 mil- hospitals. Cost-effectiveness was US$97 per patient treated lion, India expanded cataract surgery coverage to the dis- for the mobile camps, US$176 for the state medical college advantaged with the goal of reducing the prevalence of hospital, and US$54 for nongovernmental hospitals. Javitt blindness. District blindness control societies were set up (1993) estimates the cost of cataract surgery in India at and given the flexibility of financing different providers of less than US$25 per DALY averted. cataract surgery services to low-income groups at the dis- As the World Health Organization (Brian and Taylor trict level (Rose 1997). Mobile camps were a strategy 2001) has stressed, successful and sustainable surgical adopted for providing cheap and efficient cataract surgery treatment of cataracts is linked to a spectrum of other in rural districts. Because of these efforts, cataract surgery equally important activities, including ongoing training nationally increased from 1.2 million surgeries in 1991­92 of surgeons, nurses, and administrators; reliable and to 2.7 million in 1996­97. Singh, Garner, and Floyd (2000) affordable supply chains; and equipment purchase and analyze the cost-effectiveness of publicly funded options maintenance. Source: Authors. CONCLUSIONS · in the surgical care of several elective conditions that have a significant effect on the quality of life, such as cataract, The inclusion of this chapter indicates the evolving apprecia- otitis media, clubfoot, hernias, and hydroceles. tion that surgery has a role to play in public health strategies. Previous concerns that surgery is a curative intervention Few published data are available to enable reliable estimates performed in expensive, high-tech hospitals precluded appreci- of either the burden of surgical diseases or the cost- ation of the potential role of surgery in public health. Public effectiveness of surgical treatments in a region-specific manner health specialists now recognize not only that surgery has a to help policy makers and voluntary groups. This area merits a preventive role, but also that surgical treatment provided in great deal of attention in relation to research and development. low-tech community hospitals is cost-effective. In addition, a Nevertheless, the clear conclusion is that surgery must be significant number of surgical procedures, including cesarean considered a public health priority. sections and other abdominal operations, can be successfully performed by surgical technicians (Jamisse and others 2004; Pereira and others 1996). NOTES Surgery has an important role as a public health strategy in at least four areas: 1. This structure is based on the authors' personal experiences of prac- ticing in developing countries. We have defined surgical cases to include deliveries and cesarean sections. · in the prevention of death and chronic disability in injured 2. For example, operating room costs are based on the results of a sin- patients by the provision of timely, expert, and complete ini- gle study by Shepard and others (1993). tial surgical treatment · in the timely surgical intervention in obstructed labor, in pre- and postpartum hemorrhage, and in other obstetrical REFERENCES complications Adam, T., and M. A. Koopmanschap. 2003. "Cost-Effectiveness Analysis: · in the provision of competent surgery to treat a wide range of Can We Reduce Variability in Costing Methods?" International Journal emergency abdominal and nonabdominal conditions of Technological Assessment of Health Care 19 (2): 407­20. 1258 | Disease Control Priorities in Developing Countries | Haile T. Debas, Richard Gosselin, Colin McCord, and others Armandola, E. 2003. "Conference Report: Global Vaccines--What Are the Mock, C. N. 2003. "Improving Prehospital Trauma Care in Rural Areas of Challenges? Highlights from the Viral Vaccine Meeting, October Low Income Countries." Trauma 54 (6): 1197­98. 25­28, 2003, Barcelona, Spain." Medscape General Medicine 5 (4): 29. Mock, C. N., F. Abantanga, P. Cummings, and T. D. Koepsell. 1999. Bickler, S. W., and H. Rode. 2002. "Surgical Services for Children in "Incidence and Outcome of Injury in Ghana: A Community-Based Developing Countries." Bulletin of the World Health Organization 80 Survey." Bulletin of the World Health Organization 77 (12): 955­62. (10): 829­35. Mock, C. N., E. Adzotor, D. Denno, E. Conklin, and F. Rivara. 1995. Blas, E., and M. Limbambala. 2001. "The Challenge of Hospitals in Health "Admissions for Injury at a Rural Hospital in Ghana: Implications for Sector Reform: The Case of Zambia." Health Policy and Planning 16 Prevention in the Developing World." American Journal of Public (Suppl. 2): 29­43. Health 85 (7): 927­31. Brian, G., and H. Taylor. 2001. "Cataract Blindness: Challenges for the 21st Mulligan, J., J. A. Fox-Rushby, T. Adam, B. Johns, and A. Mills. 2003. Unit Century." Bulletin of the World Health Organization 79 (3): 249­56. Costs of Health Care Inputs in Low and Middle Income Regions. Working Dandona, L., R. Dandona, T. J. Naduvilath, C. A. McCarty, P. Mandel, Paper 9, Disease Control Priorities Project. Bethesda, MD. M. Srinivas, and others. 1999. "Population-Based Assessment of the Murray, C. J. L., and A. D. Lopez, eds. 1996. The Global Burden of Disease: Outcome of Cataract Surgery in an Urban Population in Southern A Comprehensive Assessment of Mortality and Disability from Diseases, India." American Journal of Ophthalmology 127 (6): 650­58. Injuries, and Risk Factors in 1990 and Projected to 2020. Cambridge, Dayan, G. H., L. Cairns, A. Mtonga, V. Nguyen, and P. Strebel. 2004. "Cost- MA: Harvard University Press. Effectiveness of Three Different Vaccination Strategies against Measles Nantulya, V. M., and M. R. Reich. 2002. "The Neglected Epidemic: Road in Zambian Children." Vaccine 22 (3­4): 475­84. Traffic Injuries in Developing Countries." British Medical Journal 324: Ellwein, L. B., and C. Kupfer. 1995. "Strategic Issues in Preventing Cataract 1139­41. Blindness in Developing Countries." Bulletin of the World Health Neilson, J. P., T. Lavender, S. Quenby, S. Wray. 2003. "Obstructed Labor." Organization 73 (5): 681­89. British Medical Bulletin 67: 191­204. Fiedler, J. L., J. B. Wight, and R. M. Schmidt. 1999. "Risk Adjustment and Olumide, F., A. Adedeji, and A. O. Adesola. 1976. "Intestinal Obstruction Hospital Cost-Based Resource Allocation, with an Application to El in Nigerian Children." Journal of Pediatric Surgery 11 (2): 195­204. Salvador." Social Science Medicine 48 (2): 197­212. Peden, M., and A. A. Hyder. 2002. "Road Traffic Injuries Are a Global Hilton, P. 2003 "Vesico-Vaginal Fistulas in Developing Countries." Health Problem." British Medical Journal 324 (7346): 1153­54. International Journal of Gynecology and Obstetrics 82 (3): 285­95. Pereira C., A. Bugalho, S. Bergstrom, F. Vaz, and M. Cotiro. 1996. "A Hyder, A. A., and M. Peden. 2003. "Inequality and Road Traffic Injuries: Comparative Study of Caesarian Deliveries by Assistant Medical Call for Action." Lancet 362 (9401): 2034­35. Officers and Obstetricians in Mozambique." British Journal of Jamisse, L., and F. Songane, and others. 2004. "Reducing Maternal Obstetrics and Gynecology 103 (6): 508­12. Mortality in Mozambique: Challenges, Failures, Successes and Lessons Ponseti, I. V. 1999. Idiopathic Clubfoot. New York: Oxford University Press. Learned." Int J Gynaecol Obstet 85 (2): 203­12. Rahman, G. A., and I. A. Mungadi. 2000. "Gangrenous Bowel in Jat, A. A., M. R. Khan, H. Zafar, A. J. Raja, Q. Hoda, R. Rehmani, and Nigerians." Central African Journal of Medicine 46 (12): 321­24. others. 2004. "Peer Review Audit of Trauma Deaths in a Developing Rose, J. 1997. "National Program for Control of Blindness." Indian Journal Country." Asian Journal of Surgery 27 (1): 58­64. of Community Health 3: 5­9. Javitt, J. C. 1993. "The Cost-Effectiveness of Restoring Sight." Archives of Ruff, T. A. 1999."Immunization Strategies for Viral Diseases in Developing Ophthalmology 111(12): 1615. Countries." Review of Medical Virology 9 (2): 121­38. Jha, P., O. Bangoura, and K. Ranson. 1998. "The Cost-Effectiveness of Shepard, D. S., J. Walsh, W. Munar, L. Rose, R. Guerrero, L. F. Cruz, and Forty Health Interventions in Guinea." Health Policy Planning 13 (3): others. 1993. "Cost-Effectiveness of Ambulatory Surgery in Cali, 249­62. Colombia." Health Policy Plan 8 (2): 136­42. Kaya, E., H. Ozguc, R. Tokyay, and O. Yunuk. 1999. "Financial Burden of Singh, A. J., P. Garner, and K. Floyd. 2000. "Cost-Effectiveness of Public- Trauma Care on a University Hospital in a Developing Country." Funded Options for Cataract Surgery in Mysore, India." Lancet 355 Journal of Trauma 47 (3): 572­75. (9199): 180­84. Krug, E. G., G. K. Sharma, and Q. Lozano. 2000. "The Global Burden of Smith, A., and S. Hatcher. 1992."Preventing Deafness in Africa's Children." Injuries." American Journal of Public Health 90 (4): 523­26. African Health 15 (1): 33­35. Marseille, E. 1996. "Cost-Effectiveness of Cataract Surgery in a Public Turco, V. 1994. "Present Management of Idiopathic Clubfoot." Journal of Health Eye Care Programme in Nepal." Bulletin of the World Health Paediatric Orthopedics. Part B 3: 149­54. Organization 74 (3): 319­24. Whitney, C. G., and L. K. Pickering. 2002. "The Potential of Pneumococcal McCord, C., and Q. Chowdhury. 2003. "A Cost-Effective Small Hospital in Conjugate Vaccines for Children." Pediatric Infectious Diseases Journal Bangladesh: What It Can Mean for Emergency Obstetric Care." 21 (10): 961­70. International Journal of Gynaecology and Obstetrics 81 (1): 83­92. WHO (World Health Organization). 2002. World Health Report 2002: Meyer, A. 1998. "Death and Disability from Injury: A Global Challenge." Reducing Risks, Promoting Healthy Life. Geneva: WHO. Journal of Trauma 44 (1): 1­12. ------. 2003. Surgical Care in a District Hospital. Geneva: WHO. Moalosi, G., K. Floyd, J. Phatshwane, T. Moeti, N. Binkin, and T. Kenyon. 2003. "Cost-Effectiveness of Home-Based Versus Hospital Care for Chronically Ill Tuberculosis Patients, Francistown, Botswana." Inter- national Journal of Tuberculosis and Lung Diseases 7 (9 Suppl. 1): S80­85. Surgery | 1259 Chapter 68 Emergency Medical Services Olive C. Kobusingye, Adnan A. Hyder, David Bishai, Manjul Joshipura, Eduardo Romero Hicks, and Charles Mock INTRODUCTION posed by such diverse circumstances will require innovation and a reorientation of public health planning. Emergency medical conditions typically occur through a sud- A number of misconceptions about emergency care are often den insult to the body or mind, often through injury, infection, used as a rationale for giving it low priority in the health sector, obstetric complications, or chemical imbalance; they may especially in low-income countries. These ideas include equat- occur as the result of persistent neglect of chronic conditions. ing emergency care with ambulance transportation, neglecting Emergency medical services (EMS) to treat these conditions the role of the community and facility care provided,and assum- include rapid assessment, timely provision of appropriate ing that emergency departments and physicians are the only interventions, and prompt transportation to the nearest appro- acute care resources. Such a narrow view ignores the important priate health facility by the best possible means to enhance contributions of other disciplines, skills, and personnel. Perhaps survival, control morbidity, and prevent disability (see the most common misperception is that emergency medical table 68.1). The goal of effective EMS is to provide emergency care is inherently expensive and requires high-technology medical care to all who need it. Advances in medical care and interventions as opposed to simple and effective strategies. technology in recent decades have expanded the parameters of Emergency care, which may be delivered in crisis situations what had been the traditional domain of emergency services. with poor planning and ineffective use of resources, may be These services, no longer limited to actual in-hospital treat- inefficient. In many countries, few resources are set aside for ment from arrival to stabilization, now include prehospital care possible emergencies, and when situations that demand emer- and transportation. gency care arise, they precipitate hurried and costly resource Despite the best efforts of primary care providers and public deployment. Efforts to improve emergency care, however, do health planners, not every emergency is preventable. Emergency not necessarily increase costs. This chapter shows that medical care is needed in diverse circumstances: prospective improved organization and planning for emergency care can patients range from rural farmers or fishers whose most com- be done at a reasonable cost and lead to more appropriate use mon mode of transportation may be canoes or animal-drawn of resources, improved care, and better outcomes (White, carts, to factory workers living in densely populated urban Williams, and Greenberg 1996). This chapter does not address slums, to residents of high-income cities and suburbs. Actual nonacute conditions, even though emergency care is often the provision of emergency care may range from delivery using only recourse for people with nonemergency conditions trained emergency professionals to delivery by laypeople and because of the failure of these other components of the system taxi drivers. Developing strategies to meet the range of needs (see figure 68.1). 1261 Table 68.1 Overview of Emergency Medical Services Acute event On-site management Transportation Health facility care Role of EMS Recognition Triage, stabilization, or both Safe and efficient transportation Prompt, appropriate, and quality care Key issues Surveillance and Trained personnel, equipment Safe transportation, equipment, Personnel, equipment, identification referral system organization of services Source: Authors. in the occurrence of mass casualties, may be necessary whenever a large number of patients requiring emergency Emergency care present at the same time. It typically entails categoriz- episodes: treatment delays ing patients into three groups: those very unlikely to survive, jeopardize life or even with treatment; those whose conditions are minor and human function who will recover without emergency care; and those with potentially lethal conditions who are likely to survive if they Treatable episodes: treatment delays still tolerated well receive timely emergency care. Patients in the last category form the highest priority for emergency care. · Stabilization. A distinction is often made between initial Self-limiting episodes: treatment unnecessary emergency care and stabilization on one hand and definitive medical care on the other. Initial emergency care and stabi- Source: Authors. lization are usually considered the domain of mobile EMS; lower levels of the health care system (for example, clinics Figure 68.1 The Emergency Medical Care Pyramid and smaller hospitals); and the emergency departments of any fixed facility. Definitive care is usually considered the domain of the hospital and of larger facilities and implies Definitions the resolution of the condition needing treatment. However, The current literature is often inconsistent in the use and inter- the distinction is somewhat arbitrary; a more accurate pretation of terminology. Accordingly, the specific terms used approach is to view care as a continuum. Many of the ele- in this chapter are defined as follows: ments of early care delivered in the course of emergency treatment, whether in the field or in fixed facilities, can be · Emergency care. Emergency medical care is that care deliv- considered definitive (McCord and Chowdhury 2003; ered in the first few hours after the onset of an acute med- McCord and others 2001). ical or obstetric problem or the occurrence of an injury, including care delivered inside a fixed facility. · Paramedical personnel. Paramedical personnel refers to all Burden of Disease persons with medical training who are involved in the care Investing in emergency medical care should become a priority. and transportation of patients in need of emergency med- Emergency medical systems address a diverse set of diseases ical care. The length and quality of training vary, from highly that span the spectrum of communicable infections, noncom- specialized personnel with capabilities for advanced life sup- municable conditions, obstetrics, and injuries. All of these port to those with simple first-aid training and limited field conditions may present to the EMS in their acute stages (for experience. example, diabetic hypoglycemia, septicemia, premature labor, · Hospital. A hospital is a geographically fixed facility in which or asthma), or they are acute in their natural presentation personnel with some acceptable level of training deliver (for example, myocardial infarction, acute hemorrhage, or emergency medical care. The distinctions between a clinic, injuries). Accordingly, defining the burden of disease addressed health center, and hospital are unclear, and the presence or by EMS can be problematic. absence of a doctor is not a determining factor in this dis- Malaria causes 300 to 500 million acute episodes worldwide tinction. A range of facilities from small, basic units up to annually and results in an estimated 1 million deaths, mostly in tertiary care hospitals provides an increasing level of capa- Sub-Saharan Africa. Effective emergency care can avert these bility for emergency and other care. deaths, as well as those from acute respiratory and diarrheal · Triage. Triage is the screening of patients in the field or in diseases in children and from noncommunicable diseases such the receiving area of a fixed facility to determine their rela- as diabetes, hypertension, and other cardiovascular diseases. In tive priority for treatment. Triage, which is usually necessary addition to the acute presentation of chronic conditions, the 1262 | Disease Control Priorities in Developing Countries | Olive C. Kobusingye, Adnan A. Hyder, David Bishai, and others lack of access to medical care and lack of sustained effective disability) that occur in low-income countries. The numbers treatment means that subacute episodes and flare-ups may be represent a conservative estimate of the potential burden, life threatening. Early recognition can prevent the emergency because they do not include all the conditions that could ben- precipitated by infectious disease and many other medical con- efit from emergency care and they do not include data from ditions or can limit the effects. high-income countries. More than 500,000 maternal deaths occur each year; 95 per- cent of these deaths are in low-income countries where emer- gency care is often lacking. It is estimated that 15 percent of INTERVENTIONS FOR EMERGENCY CARE: all pregnant women experience a potentially life-threatening SYSTEMS, STRUCTURES, AND ORGANIZATION condition and will need emergency care. Prenatal screening methods alone may not be effective in reducing this risk ratio. Emergency care must be appreciated as an entire system with Although identifying risk factors for acute complications is interdependent components. These components include pre- easy, identifying which of the at-risk women will actually hospital care, transportation, and hospital care. Each compo- develop a life-threatening condition is not possible (Graham nent is important, but all of them must work together to make 1997). The only way to prevent the deaths is by ensuring access a lasting effect on the health of a population. The organization to emergency obstetric care for all pregnant women. and operation of the prehospital care system will vary by coun- Injuries were responsible for 21.7 percent of global deaths try, but it should be linked to the local hospitals or facilities and 31.1 percent of disability-adjusted life years (DALYs) lost where patients are taken. When prehospital transportation is in 2001 (WHO 2002). Because both unintentional injuries poor or absent, deaths occur that could have been prevented (chapter 39) and injuries caused by interpersonal violence even by inexpensive procedures (Mock and others 1998). For (chapter 40) are by definition acute events, nearly all require example, the majority of maternal deaths may fall into this cat- emergency care (see table 68.2). In 2001, more than 80 percent egory. Poor quality of care at the hospitals will lead to in- of all deaths attributable to injury were in low-income coun- hospital deaths and may eventually discourage communities tries. Most injuries attributable to violence involve a predomi- that might have the capacity to promptly transfer patients to nantly young and productive population (WHO 2002) that is such facilities (Leigh and others 1997; Prevention of Maternal resilient and can respond well to appropriate emergency care. Mortality Network 1995). Skilled and motivated personnel, The conditions listed in table 68.2 represent 45 percent of appropriate supplies, pharmaceuticals, equipment, coordina- all deaths and 36 percent of the disease burden (including tion, and management oriented to the needs of the critically ill Table 68.2 Burden of Diseases Potentially Addressed by EMS in Low- and Middle-Income Countries (2001, all ages, both sexes) Deaths DALYs DALY YLL YLD Group Disease (thousands) (thousands) rates/1,000 population (thousands) (thousands) I: Communicable Diarrheal diseases 1,777 58,697 11.25 50,883 7,835 Childhood clustera 1,362 43,131 8.26 40,506 2,624 Maternal conditions 507 26,383 5.05 13,363 13,020 Meningitis 169 5,475 1.05 4,343 1,131 Malaria 1,207 39,961 7.66 35,461 4,438 II: Chronic Diabetes 757 15,804 3.03 10,140 5,661 Hypertensive heart 760 9,969 1.91 9,077 888 Ischemic heart 5,699 71,882 13.77 67,925 3,921 Cerebrovascular 4,608 62,669 12.01 51,539 11,100 Asthma 205 11,514 2.21 3,799 7,714 III: Injuries Unintentional 3,214 113,235 21.70 73,140 40,104 Intentional 1,501 42,615 8.16 34,374 8,242 Total EMS related 21,766 501,335 96.06 394,550 106,678 Percentage of LMIC total 45 36 43 22 Source: Global Burden of Disease Project 2001. YLL years of life lost to premature mortality; YLD years lived with disability; LMICs low- and middle-income countries. a. Pertussis, poliomyelitis, diphtheria, measles, and tetanus. Emergency Medical Services | 1263 all contribute to make emergency care effective in reducing outside of hospitals, and no transportation is dedicated to the death and disability. task of getting patients in need of emergency care into hospi- tals. There is a paucity of literature on the effect of first respon- ders, except for one study that evaluated a program to train a Prehospital Care core group of paramedics, who then trained thousands of lay Prehospital care encompasses the care provided from the com- first responders in northern Iraq and Cambodia. No data are munity (scene of injury, home, school, or other location) until available on the effectiveness of lay responders compared with the patient arrives at a formal health care facility capable of giv- the more trained paramedics. ing definitive care. This care should comprise basic and proven The following discussion introduces a scenario in which the strategies and the most appropriate personnel, equipment, and observed mortality rate reduction could be achieved in a devel- supplies needed to assess, prioritize, and institute interventions oping country's health system by a small group of paramedics to minimize the probability of death or disability. Most effec- working together with a large group of trained lay responders. tive strategies are basic and inexpensive, and the lack of high- The scenario uses only emergencies caused by trauma, although tech interventions should not deter efforts to provide good it is expected that both the paramedics and the lay first respon- care. Even where resources allow them, the more-invasive pro- ders would also save lives from medical or obstetric emergen- cedures performed by physicians in some prehospital settings, cies. Existing studies have not been large enough to document such as intravenous access and fluid infusion or intubations, do these effects, and they are not included in the estimates of cost- not appear to improve outcomes, and evidence suggests that effectiveness. they may, in fact, be detrimental to outcomes (Liberman and others 2003; Sampalis and others 1994, 1995, 1997). Trained Lay Responders A case is made for training lay per- Prehospital care should be simple, sustainable, and efficient. sons able to recognize threatening conditions--whether Because resource availability varies greatly among and within obstetric, traumatic, or medical. Cultural reasons may require countries, different tiers of care are recognized. Where no for- that those responding to obstetric emergencies be traditional mal prehospital system exists, the first tier of prehospital care birth attendants or similar persons in the community. Husum may be composed of laypeople in the community who have and colleagues demonstrated that lay people who are given been taught basic techniques of first aid. Recruiting and train- first-aid skills could effectively respond to emergencies in a ing particularly motivated citizens who often confront emer- community with a high trauma burden (Husum, Gilbert, and gencies, such as drivers of public transportation, to function as Wisborg 2003; Husum, Gilbert, Wisborg, and others 2003). In prehospital care providers can add to this resource. Ghana, it was demonstrated that commercial taxi and minibus The second tier comprises paramedical personnel who use drivers trained in first aid could provide effective prehospital dedicated vehicles and equipment and are usually able to get to care (Mock and others 2002). Lay responders are likely to be patients and take them to hospitals within the shortest possible successful where the burden of injuries and other emergencies time. This second tier may involve the performance of is high. Attrition of both the responders and the skills is a con- advanced procedures, the administration of intravenous and cern unless they are put to frequent use. other medications by physician or nonphysician providers, or both. This care is not always available in low-income countries; Paramedical Personnel In most middle-income countries and few trained personnel and inadequate funding make round- some cities of low-income countries, trained paramedical the-clock coverage impossible. Although providing advanced personnel render prehospital care (Mock and others 1998; life-saving measures in the prehospital environment may be Tannebaum and others 2001). As for lay responders, coverage beneficial in some cases, these benefits may be negated if such varies by country. In most of Sub-Saharan Africa and Asia, para- measures divert scarce resources from more basic interventions medical personnel and ambulances transfer patients between that can benefit far larger numbers of patients (Hauswald and health facilities and not from the scenes of injury or from homes Yeoh 1997). In most low- and middle-income countries of (Joshipura and others 2003). In middle-income countries, Africa, Asia, and Latin America, high maternal and child mor- though, they are a major component of existing emergency tality are linked to inadequate emergency care, especially poor medical systems (Arreola-Risa and others 2000; Mock 2002). access to quality hospital care. In these settings, it is imperative Their presence (coupled with vehicle ambulances) reduces the that resources be integrated, instead of one system for injuries interval between the recognition of an emergency and the and another for obstetric emergencies. arrival at the hospital, and some evidence suggests that training them in basic life-saving skills improves patient outcomes (Ali Personnel. Most of the world's population does not have and others 1997, 1998; Arreola-Risa and others 2000). access to formal prehospital care. No personnel are employed Effectiveness has also been demonstrated for well-placed for the sole purpose of dealing with medical emergencies dispatch sites in urban populations, where the vehicles and 1264 | Disease Control Priorities in Developing Countries | Olive C. Kobusingye, Adnan A. Hyder, David Bishai, and others personnel can be optimized. There is no evidence, however, for The recommended ratio of one unit for 50,000 people sug- the effectiveness of training prehospital care paramedical per- gested by McSwain (1991) results in response times as low as sonnel in advanced life-saving skills (Sethi and others 2003). four to six minutes. The ratio does not distinguish between Shorter prehospital times, in general, are considered an impor- basic life-support and advanced life-support capabilities. tant parameter of the quality of prehospital care. These times Traffic congestion, poor maps, and poor road signs may all have the following components: increase this response time in cities with poor infrastructure. In Monterrey, Mexico, one unit per 100,000 people manages an · Notification time is the time elapsed from occurrence of average response time of 10 minutes. Hanoi, Vietnam, with one injury or recognition of severe illness until the EMS system unit for 3 million people, has an average response time of 30 is notified. minutes (Mock and others 1998). · Response time is the time elapsed from notification until Where paramedical services exist alongside lay responder arrival of an ambulance to the site of the ill or injured services, these two could be managed by the same organiza- person. tional unit. The paramedical staff will be more successful in · Scene time is the time taken by prehospital providers from urban areas, where distances between dispatch sites, commu- arrival at until departure from the scene. nities served, and hospitals are short. Other enabling factors are · Transport time is the time elapsed from leaving the scene good telecommunications; rapid and dedicated transportation; until arrival at the hospital or other treatment facility. and coordinating capacity among the community, hospitals, and other emergency services. Notification time is influenced by the availability of telecom- munications. Response time is influenced by the capabilities of Intervention Cost and Effectiveness Training lay responders is a dispatch center to handle emergency calls and especially by the an intervention potentially available in low-income countries. geographic distribution of sites of ambulance dispatch. The Projections for costs and effectiveness have been made based greater in number and the more widely distributed the number on the following assumptions: of ambulance stations are, the shorter the response times are. Geographic distribution and associated response times can be · Serving a population of 1 million requires 7,500 trainees. improved in some circumstances by using a tiered or layered Sensitivity analysis ranges from 5,000 to 10,000 trained lay response system. This system requires having a relatively larger responders. number of basically trained and equipped first responders with · Trained lay responders can be trained in half a day (St. John wider geographic distribution and a smaller number of centrally Ambulance 1996). located, more highly trained and equipped second responders. · Training would have to be repeated every three years to This system allows the first responders to respond more rapidly, maintain efficacy. with second responder involvement only if needed (Arreola-Risa · Annually, 2,500 laypeople would be trained on a rolling basis. and others 1995). Accordingly, paramedical personnel should be introduced Using these assumptions (see annex 68.A for technical in large urban areas where they do not function at present and details on costing guidelines) would require the following: should be stationed at dispatch sites with dedicated vehicles, fast communications with the hospitals in the area, and links · 1,250 days of trainees' time (0.5 days each) valued at salary with other emergency services such as the fire and police level 1 departments. The communities served by the system should · 62.5 trainer days of time, with a ratio of 20 trainees per have a well-known and rapid method of calling the paramed- trainer, valued at wages for salary level 3 ical teams when an emergency arises. Because skills depreciate · one training facility with a classroom (100 square meters) with time, these personnel require refresher courses. Where valued at rent for basic space for 62.5 days paramedical personnel already exist as part of the EMS, their · 2,500 copies of photocopied curricula annually valued at numbers and organization (location, training, deployment, US$1 each. and monitoring) should be enhanced to improve response times and, hence, patient outcomes, especially for cardiac and The costs of providing trained paramedics can be estimated obstetric emergencies. Availability of EMS for a given popula- using the following assumptions: tion can be looked at either as number of units on duty or number of sites of ambulance dispatch. They are usually closely · Trained paramedic responders can be trained in 25 days related, with one or two units per site, but not in systems (St. John Ambulance 1996). where a large number of units are on duty from one central · Serving a population of 1 million requires 7,500 trainees. dispatch station. Sensitivity analysis ranges from 100 to 200. Emergency Medical Services | 1265 · Training would have to be repeated every three years to paramedics can lower mortality in trauma by 9 percent; thus, maintain efficacy. in 4,100 traumas, 370 lives can be saved. Dividing the sum of · Annually, 50 paramedics would be trained. the costs in table 68.3 by the 370 deaths averted provides a rough estimate of costs per death averted. These costs per death As a result, the following would be required: averted are divided by the regional life expectancy at age 20 (LE 20), with the assumption that the average age of trauma is · 25 50 1,250 days of paramedic trainee time (salary 20, to give the cost per life year saved. LE 20 is roughly 50 years level 1) in every region except Sub-Saharan Africa, where it is only 37 · 125 trainer days, with a ratio of 10 trainees per trainer years. Shortages of equipment and supplies may reduce the valued at salary level 3 effectiveness of the prehospital personnel. · one training facility offering a classroom (100 square It is possible to offer more refined "regional" estimates of the meters) valued at basic building values numbers of deaths averted by using local estimates of the bur- · 50 photocopied curricula annually, valued at US$1 each den of injuries instead of the global estimate of 4,100 injuries · one paramedic kit with stethoscope, gloves, bandages, and per million people. Yet given the uncertainty about regional splint material for each trainee (kits would be renewed by variation in the effectiveness of the intervention based on patient contributions). administrative, infrastructure, and human resource capacity, it would perhaps give a false impression that a firm and universal The trainees would offer volunteer services after training. basis exists to speculate quantitatively on the relative effective- We assume that volunteers are highly motivated individuals ness of the intervention in various regions. As a result, the who consider emergency medical service to their community above estimates, as used in table 68.4, serve to inform global as the most rewarding use of their leisure time (Fiedler 2000). dialogue rather than offer specific empirical numbers. The opportunity cost of their recurrent emergency service is assumed to be zero. Communities or cultures that have a short- Equipment and Supplies. Equipment and supplies should age of individuals with an ethic of volunteer service may have match the knowledge and skills of the personnel available to to devote funds toward maintaining incentives for "volunteer" use them. Even teams with the least resources should have the paramedics to serve. In such cases, this strategy may not be following: cost-effective. Table 68.3 shows the estimated costs of the lay first respon- · protective wear, especially gloves and aprons ders and paramedics intervention, drawing on the Disease · a stretcher or the equivalent Control Priorities Project's standardized input prices by region · pressure dressings (bandages--elastic, if possible--and for low, best, and high estimates. cotton or gauze dressings) · splints--various sizes, made out of local materials Outcomes According to the World Health Organization · radio, telephone, or other mode of rapid communication. burden-of-disease estimates, the global incidence of trauma is 410 per 100,000 or 4,100 per million. Husum, Gilbert, and Annex 68.A provides a comprehensive list for better- Wisborg (2003) indicate that first-level responders and trained resourced communities. Intervention cost and effectiveness Table 68.3 Cost of Using Trained Lay First Responders Together with Trained Paramedics (local currency converted to 2001 U.S. dollars using exchange rates) Region Low Best High East Asia and the Pacific 27,539 48,050 75,232 Europe and Central Asia 30,209 52,339 79,605 Latin America and the Caribbean 32,777 74,589 110,453 Middle East and North Africa 33,050 104,585 261,935 South Asia 27,183 45,637 116,456 Sub-Saharan Africa 30,765 52,339 115,171 Unweighted average 30,254 62,923 126,475 Source: Authors. Note: Cost of treating community of 1 million. 1266 | Disease Control Priorities in Developing Countries | Olive C. Kobusingye, Adnan A. Hyder, David Bishai, and others Table 68.4 Cost-Effectiveness of Combining Paramedics with Lay Responders (local currency converted to 2001 U.S. dollars using exchange rates) Region Low Best High Cost per death averted with trained lay first responders together with volunteer paramedics East Asia and the Pacific 74 130 203 Europe and Central Asia 82 141 215 Latin America and the Caribbean 89 202 299 Middle East and North Africa 89 283 708 South Asia 73 123 315 Sub-Saharan Africa 83 141 311 Unweighted average 82 170 342 Cost per life year saved with trained lay first responders together with volunteer paramedics East Asia and the Pacific 3 5 8 Europe and Central Asia 3 5 8 Latin America and the Caribbean 3 8 11 Middle East and North Africa 3 11 27 South Asia 3 5 12 Sub-Saharan Africa 4 6 14 Unweighted average 3 7 13 Source: Authors. Note: Cost of treating community of 1 million. cannot be estimated because no studies are available on this the community adopts traditional communication or more issue from low-income countries. modern communication systems. If radio communication were introduced, the purchase and maintenance of radio Traditional and Innovative Communications Systems. receivers, supplies, and government licensing costs would need Nowhere is the demand for efficient communication and rapid to be estimated. If cellular telephones options were being transportation more critical than in emergency medical care. explored, then the purchase of telephones, plans, licenses, and The best teams equipped with state-of-the-art technology and maintenance costs would need to be included. Where satellite supplies will be wasted if they cannot be reached quickly or if towers need to be installed, however, this cost will be much they have no contact with the hospitals where patients are to be higher than for all other components. Finally, if traditional taken. Most of the world's population lives in areas where there landline installation is being considered, the lines, equipment, is no telecommunications infrastructure, and this situation may and telephone bills will need to be taken into account. The not change significantly in the near future. Innovation is needed health sector may be able to share the costs of such interven- so that these populations can be enabled to access effective tions, especially traditional telephone lines, with other develop- emergency care interventions that already exist without waiting ment and infrastructure units of a national government. for traditional telephone lines to get to their rural homes. Radio communication is one solution in such settings. Equipping tra- Basic and Advanced Transportation Systems. Transporting a ditional birth attendants and remote health units with radio patient from the location of the acute event to a hospital facil- receiver sets linked to local hospitals has been used to shorten ity is a critical element of the prehospital component. Lack of the response time and reduce maternal deaths (Samai and transportation is often a major barrier to accessing emergency Sengeh 1997). Cellular telephones may offer communities that care (Lungu and others 2001; Samai and Sengeh 1997). In are remote from standard communications infrastructure an devising a prehospital system of transportation, one should opportunity to leap into a more modern and efficient mode. consider locally available resources and the range of viable alternatives for transportation. In some countries such trans- Intervention Costs and Effectiveness There are no studies portation may be part of a formal EMS system, whereas in from low-income countries on which to base intervention cost other cases it is entirely informal. For example, commercial and effectiveness estimates. The costs will depend on whether vehicles, the police, and relatives using private motorized or Emergency Medical Services | 1267 nonmotorized forms of transportation may bring seriously ill survivor cost an average of US$15,883, and the authors and injured patients to medical facilities (Andrews, acknowledged that the helicopter had to be used fully to spread Kobusingye, and Lett 1999; Joshipura and others 2003; the high fixed costs across many patients and trips. A review of Kobusingye and others 2002). A bicycle ambulance in Malawi civilian helicopter ambulance programs in the United States set up to improve emergency obstetric care was actually used concluded that the primary factor in the reduction of trauma more often for injuries and medical emergencies (Lungu and mortality was not the speed at which the patient was transported others 2001). but the administration of life-saving care by the helicopter med- Transportation should be accessible at short notice. A vehi- ical crew at the scene or at the outlying hospital (Moylan 1988). cle with a stretcher is ideal, but almost any transportation that In low-income countries, for the very few who benefit from such will get a patient to a facility where definitive care can be a high-end intervention, there are likely to be thousands who obtained is acceptable. Although a fully fitted and equipped cannot access care even using the most basic means. ambulance vehicle complete with trained paramedics delivers Intervention Costs and Effectiveness Costing transportation better outcomes, ethical and equity considerations dictate that systems requires the following assumptions: before this vehicle is made available to an elite population in the urban areas, basic transportation must be assured for all · In an urban population, one ambulance unit can serve a who need emergency transportation and care. population of 30,000 people. Thus, 1 million people would In a city setting, a vehicle ambulance can make as many as require 33 ambulance units (1 million/30,000). 20 trips per day. On average this schedule will require salaries · Each ambulance unit requires staffing of a rotation of six for an administrator and two crews, each comprising a driver paramedic-drivers and a seventh paramedic-driver to cover and two paramedics or nurses, as well as expenses for commu- vacation times and sick leaves. nication, supplies, pharmaceuticals, and the costs of operating · A supervisor oversees three ambulance units per year. the vehicle. A study of a decision to develop an EMS in Kuala · A garage for the ambulance and communications equip- Lumpur (1.1 million people, 243 square kilometers) estimated ment would be 100 square meters but would entail rental of that the purchase and staffing of 48 ambulances at US$53,000 office-style accommodations. each per year would be required, totaling US$2.5 million per · A vehicle to be outfitted as an ambulance can be purchased year (Hauswald and Yeoh 1997). The authors noted that, for as much as an off-road vehicle with a useful life of nine despite the paucity of ambulances, severely injured or ill years. patients did get to a hospital with only minor delays by using · The cost to modify the vehicle into a basic ambulance is taxis, family transportation, or the police. Ambulances need US$5,000 for a useful life of nine years. accurate maps, house numbers, and street or road signs, all of · The ambulance will require fuel and maintenance based on which might not be in place in low-income cities. It was esti- usage of 20,000 kilometers per year. mated that ambulances were unable to locate patients in 20 percent of calls in Kuala Lumpur because of mapping and sign- Given the preceding assumptions, age problems (Hauswald and Yeoh 1997). A study conducted in Turkey found that ambulance vehicle · The 231-member ambulance staff (33 ambulance units of costs were the leading component of capital costs (Altintas, 7 persons each) would be paid at salary level 2. Bilir, and Tuleylioglu 1999). The cost per trip was US$163, and · The 10 administrators would be paid at salary level 2. the cost per patient transported was US$180.50, which the We ignore the additional burden on the health system from authors thought were beyond the means of the private sector. additional visits. Quite possibly, hospital costs will rise as more In state-run ambulance services in New Delhi, India, the cost patients now get to the hospital with the help of ambulances. It per trip was approximately US$40, yet one in three of the is also possible that patients currently arrive in less desperate ambulances served only to transport patients, with no para- condition, so that the cost of care is lessened. No studies are medic staff on board (India, Government of Delhi 2001). available on which to base cost estimates to address those issues. The debates in high-income countries about helicopter Table 68.5 shows the estimated costs of the ambulance inter- ambulances provide lessons for low-income countries. In some vention, drawing on the Disease Control Priorities Project's cases, helicopter services have been discontinued because they standardized input prices by region for low, best, and high were not considered cost-effective (Hutton 1995). A study con- estimates. ducted in the United States, which concluded that helicopters were cost-effective, found that the cost per patient transported Outcome Based on the World Health Organization's 2001 bur- was US$2,214 and that for every 100 flights there were six den of disease estimates on epidemiology of trauma, ischemic survivors more than was predicted on the basis of injury severity heart disease, and obstetric emergencies, we estimate that for indices (Gearhart, Wuerz, and Localio 1997). Each additional each 10,000 population there will be 9 deaths from trauma,1 1268 | Disease Control Priorities in Developing Countries | Olive C. Kobusingye, Adnan A. Hyder, David Bishai, and others Table 68.5 Cost and Effectiveness of Ambulances (U.S. dollars) Region Low Best High Cost of treating community of 1 million with urban ambulances East Asia and the Pacific 691,603 871,208 1,090,032 Europe and Central Asia 839,468 1,024,235 1,220,888 Latin America and the Caribbean 849,556 1,550,521 1,747,630 Middle East and North Africa 894,379 2,435,000 4,960,705 South Asia 676,111 803,361 1,973,093 Sub-Saharan Africa 781,568 951,906 1,905,417 Unweighted average 788,781 1,272,705 2,149,628 Cost per death averted of treating community of 1 million with urban ambulances East Asia and the Pacific 988 1,245 1,557 Europe and Central Asia 1,199 1,463 1,744 Latin America and the Caribbean 1,214 2,215 2,497 Middle East and North Africa 1,278 3,479 7,087 South Asia 966 1,148 2,819 Sub-Saharan Africa 1,117 1,360 2,722 Unweighted average 1,127 1,818 3,071 Cost per life year saved of treating community of 1 million with urban ambulances East Asia and the Pacific 50 63 79 Europe and Central Asia 62 75 90 Latin America and the Caribbean 61 111 126 Middle East and North Africa 65 176 359 South Asia 50 60 147 Sub-Saharan Africa 67 81 163 Unweighted average 59 95 161 Cost per death averted of treating community of 1 million with rural ambulances East Asia and the Pacific 2,978 3,748 4,686 Europe and Central Asia 3,613 4,405 5,248 Latin America and the Caribbean 3,652 6,656 7,500 Middle East and North Africa 3,847 10,449 21,274 South Asia 2,911 3,457 8,470 Sub-Saharan Africa 3,361 4,092 8,178 Unweighted average 3,394 5,468 9,226 Source: Authors. 11 deaths from ischemic heart disease,2 and 2 deaths from trauma cases and 900 trauma deaths. With rapid resuscita- lethal obstetric emergencies.3 For modeling purposes, we con- tion and oxygen available through use of ambulances, we fine our attention to trauma, ischemic heart disease, and assume we can save 300 lives. obstetric cases. Although many possible lethal emergencies · Savings from myocardial infarction management. In one year may present, such as sepsis, malaria, snakebites, and asthma, by for 1 million people in low-income countries, 1,100 deaths confining attention to the major emergency conditions we will typically result from myocardial infarction. Low-dose locate 2,200 (900 1,100 + 200) potentially savable lives in a aspirin provided to myocardial infarction victims lowers population of 1 million. The savings are outlined as follows: mortality by 18 percent (Weisman and Graham 2002). In a population without ambulance services, rapid aspirin · Savings from trauma reductions. Saving lives from trauma administration cannot be ensured; with EMS, aspirin use depends on the quality of trauma care at the destination can potentially be increased from about 0 percent to 100 per- facility. In one year for 1 million people, there will be 4,100 cent for heart myocardial infarction. Therefore, instead Emergency Medical Services | 1269 of 1,100 deaths, there will be 1,100 (1 0.18) deaths, sav- table, we assume that the 500 deaths averted from obstetric ing 200 lives, but with only an average of five life years per emergencies and trauma occur at age 20, but the 200 deaths life saved. averted from ischemic heart disease save only five additional · Savings from emergency obstetrics management. Obstetric life years per case. Regional life expectancy at age 20 years (esti- deaths for medically attended patients are approximately mated) is used as before. 100 times lower than for patients who do not receive med- ical care. Accordingly, an ambulance system essentially saves Costs for a Rural Ambulance Service The key difference deter- all of the obstetric emergencies from death; this saving mining higher costs for rural ambulances is that more ambu- would amount to 200 deaths averted in the case described lance units are necessary to cover the population. We assume previously. As a result, in the hypothetical population of that in rural areas one ambulance unit can cover a population of 1 million people in low-income countries, 700 lives can be 10,000, although variation will occur, depending on population saved by an ambulance system focusing on three causes density and geographical topography. On the basis of this only: ischemic heart disease (200), obstetric (200), and assumption, all of the cost estimates for rural ambulances are trauma (300). essentially three times higher for a population of 1 million, as are the costs per death averted and per life year gained. They are The middle section of table 68.6 displays the cost per death assumed to have the same effectiveness as the urban services averted. To compute life years saved in the last section of the because the increase in units aims at delivering the same quality Table 68.6 Summary of Cost and Effectiveness (U.S. dollars) Intervention Trained lay first responders Staffed community Staffed community Region and paramedic responders ambulance, urban ambulance, rural Costs for a population of 1 million East Asia and the Pacific 48,050 871,208 2,623,392 Europe and Central Asia 52,339 1,024,235 3,083,637 Latin America and the Caribbean 74,589 1,550,521 4,659,017 Middle East and North Africa 104,585 2,435,000 7,314,544 South Asia 45,637 803,361 2,419,607 Sub-Saharan Africa 52,339 951,906 2,864,062 Unweighted average 62,923 1,272,705 3,827,376 Cost per death averted for a population of 1 milliona East Asia and the Pacific 130 1,245 3,748 Europe and Central Asia 141 1,463 4,405 Latin America and the Caribbean 202 2,215 6,656 Middle East and North Africa 283 3,479 10,449 South Asia 123 1,148 3,457 Sub-Saharan Africa 141 1,360 4,092 Unweighted average 170 1,818 5,468 Cost per life year gained for a population of 1 milliona East Asia and the Pacific 5 63 190 Europe and Central Asia 5 75 227 Latin America and the Caribbean 8 111 335 Middle East and North Africa 11 176 530 South Asia 5 60 180 Sub-Saharan Africa 6 81 245 Unweighted average 7 94 284 Source: Authors. a. Personnel consist of lay first responders together with paramedics. 1270 | Disease Control Priorities in Developing Countries | Olive C. Kobusingye, Adnan A. Hyder, David Bishai, and others of care as in the urban centers. This assumption may not hold Because the goal of an effective EMS is the provision of true if the quality of care at the receiving facilities is lower than emergency care to all who need it--universal emergency care-- that in the urban areas. the following section presents guidelines on the necessary inputs at different levels. Two of the components in hospital Uncertainty of Estimates Substantial uncertainty remains emergency care are discussed in more detail: (a) training and over the actual effectiveness of the interventions in emergency (b) equipment and supplies. The first level of formal health medicine. The tables in this chapter should be approached with care is often staffed by nonphysician clinicians; the second level due caution because ultimately the projections of the effective- is staffed by at least one physician and other trained health care ness of interventions have been patched together from a hand- professionals; and the third is staffed by specialists in addition ful of intervention trials whose success may or may not be to other health care professionals. Some middle-income coun- similar in other contexts. The projections include the following: tries have additional levels (major emergency care centers), and some hospitals are specialized (chapters 65 and 66). · Ambulance services can and do save lives by performing Training. Most in-service training for emergency care profes- field stabilization and by hastening the arrival of critical sionals working in hospitals is designed to address a particular patients when time makes a difference in the outcome. problem, such as severe injuries, emergency pediatrics, or · Only several dozen ambulance runs per year for a unit serv- obstetric emergencies. Yet because of the resource constraints ing a population of 10,000 will actually have the potential to of low-income countries, the same personnel will be confronted save lives. with all these problems. Few courses in emergency care have · Ambulance services are more cost-effective in denser popu- been rigorously evaluated (Black and Brocklehurst 2003; Sethi lations and when roads are more passable, making trips and others 2003). The Advanced Trauma Life Support (ATLS) shorter. course for doctors has resulted in improved patient outcomes · Training lay responders and paramedics can be relatively in some settings, although it may be too expensive for most cost-effective. low-income countries and inappropriate in settings where doc- tors do not see the majority of patients. In a tertiary hospital in The financial support of an ambulance unit may rely on the Trinidad and Tobago, injury mortality was reduced by 50 per- value perceived by the hundreds of patients who are comforted cent following ATLS training (Ali and others 1993). by having rapid access to care or by knowing ambulances are Life-saving obstetric skills training was found to contribute there if needed, even though their lives and health are not actu- to a reduction in maternal deaths. In Kebbi state in Nigeria, this ally improved by ambulances. training led to a drop in case-fatality rates among women with Table 68.6 summarizes the best estimates of cost, cost per obstetric complications from 22 percent to 5 percent. Similar death averted, and cost per life year. trends were observed in other sites where the intervention was implemented (Oyesola and others 1997; Prevention of Maternal Mortality Network 1995). Emergency Triage Assessment and Health Facility­Based Subsystem Treatment has been used in many countries to improve pediatric This subsystem refers to a level within the health care system emergency care (WHO 2000). Other examples are Primary where appropriate definitive care is delivered. Formal health Trauma Care, which is a trauma management course to train facilities vary immensely between and within countries. In some doctors and other health workers in district hospitals and countries, this subsystem may be a regional hospital with spe- remote locations (Wilkinson and Skinner 2000), and Advanced cialists; in others, a district hospital with general practitioners or Life Support in Obstetrics (see http://www.aafp.org/also). These nonspecialist doctors; and in still others, a health center with courses have been beneficial in standardizing protocol-based competent nonphysician clinicians. In some low-income coun- emergency care, but their outcome evaluations are still awaited. tries, some types of emergency medical care, for conditions such Low-income countries need to identify training models for their as acute diarrhea or severe malaria, may be effectively delivered versatile emergency care personnel, especially those working at a health center staffed by nondoctor clinicians. However, such at middle-level facilities, who respond to different types of a facility will be grossly inadequate for the management of a emergencies. severe multiple injury or obstructed labor. The triage process in The costs of this intervention are not available in the litera- the prehospital subsystem should determine which patients ture and will require an estimation of trainer costs, space, receive transportation to which facility, instead of merely trans- materials, and refresher courses. portation to the nearest facility. Precious time and lives may be lost when patients are taken to facilities where the desired defin- Equipment and Supplies. A specimen list of resources for itive care is not available. emergency care required at different levels of facilities is Emergency Medical Services | 1271 provided in annex 68.A. This template is flexible, and countries Documentation and Quality Assurance can customize it to suit local conditions, such as existing facility Quality of care is critical to the interaction of the poor with the levels and prevailing burden of emergency disease conditions. EMS. Lack of funds, lower-paid jobs, social class distinctions, Equipment and supplies at each level should match the knowl- ethnicity, and other affiliations make the already vulnerable edge and skills of the personnel available to use them. poor susceptible to receiving poor-quality care. For an EMS to maintain and improve the care of patients, systematic docu- mentation and periodic audits, or other processes to ensure Systems Organization quality of care, need to be incorporated. Quality management Emergency care needs to be planned and implemented care- systems that are simple, are continuous, and allow for rapid fully. The various components that make up the EMS should be changes in the system need to be implemented. linked to ensure that the entire system operates as a unit. A Because of scarcity of resources, expensive machines and coordinator should be responsible for monitoring and coordi- elite specialists should not be advocated for the urban privi- nating all emergency medical care in the community or district leged at the expense of the majority of the rural poor. The most and should work with a committee to which the other compo- difficult decisions concern balancing funds invested in the nents send representation. A community representative should emergency care capacity of secondary and primary care centers be a key member of this committee. with support for referral and transportation networks to feed Coordination costs are very important and should not be tertiary care centers. These decisions are too variable and too overlooked in the development of a new EMS. Such costs system specific to yield to a uniform policy prescription. Two include the salary of the coordinator, an efficient telephone or principles are advocated to inform these difficult decisions: communication system, a vehicle, fuel, and a budget to organ- ize meetings of stakeholders at least twice a year (Bazzoli, · Collect data on costs, capacities, and outcomes. Harmata, and Chan 1998; Nurok 2001). · Tighten the integration of the emergency care system to improve function and lead to wiser decisions on where to invest. Health Financing for Emergency Care Emergency medical systems in low-income countries must be Legislative Instruments to Ensure Emergency Care pro-poor in their orientation, which requires explicit consider- The issues discussed in the preceding sections supply the ation of how the poor interact with an EMS and how barriers rationale for countries to have specific legislation addressing to acute care for the poor can be overcome. Issues of access to the provision of emergency care. This area requires major an EMS become critical because the lack of money often keeps cooperation between public health and law, which together people from using emergency services. Direct payment of costs provide the legal framework for ensuring that all individuals for transportation, medical treatment, and drugs makes lack of who deserve emergency care can receive it, irrespective of their money a major barrier to EMS for the poor in every country. personal characteristics or their ability to pay. Having laws that As a result, emergencies frequently cripple individuals and protect trained individuals and laypeople as they provide such families financially in poor communities, often for many years care is also important. in the future. Financial protection for emergency health care is a necessity in low-income countries and has not received adequate atten- IMPLEMENTATION OF CONTROL STRATEGIES: tion. The goal of such protection is to ensure that individuals LESSONS OF EXPERIENCE and families do not spiral down the pathway to abject poverty as a result of interaction with the national health system. Such A large proportion of trauma patients in developing countries financial protection may be achieved by a number of different do not have access to formal emergency medical services. means, including community financing (Ande and others Boxes 68.1 and 68.2 contain examples of interventions to pro- 1997; Desmet, Chowdhury, and Islam 1999; Macintyre and vide appropriate emergency care in such countries. Hotchkiss 1999). Community loan funds to cover transporta- tion and other requirements for emergencies, especially for obstetrics, have been explored with mixed results (Essien and THE RESEARCH AND DEVELOPMENT AGENDA others 1997; Shehu, Ikeh, and Kuna 1997). Some experience seems to indicate that these approaches can indeed overcome The research and development priorities for emergency care several of the barriers to accessing emergency medical services are challenging to define because emergency care is a neglected and should be explored further. area of research in low-income countries and the needs are 1272 | Disease Control Priorities in Developing Countries | Olive C. Kobusingye, Adnan A. Hyder, David Bishai, and others Box 68.1 Improving Trauma Care in the Absence of a Formal Ambulance System Background: The efficacy of a program that builds on the first aid since taking the course. There was considerable existing, although informal, system of prehospital trans- improvement in the provision of the components of portation in Ghana was assessed. In Ghana, the majority first aid in comparison to what was reported before the of injured persons are transported to the hospital by some course: type of commercial vehicle, such as a taxi or bus. Methods: A total of 335 commercial drivers were Component of first aid Before (percent) After (percent) trained using a six-hour basic first-aid course. The efficacy of this course was assessed by comparing the process of Crash scene management 7 35 prehospital trauma care provided before and after the Airway management 2 35 course, as determined by self-reporting from the drivers. Bleeding control 4 42 The course was conducted with moderate amounts of Splint application 1 16 volunteer labor and gifts in kind, such as transportation to Triage 7 21 the course. The actual cost of the course amounted to US$3 per participant. Conclusions: Even in the absence of a formal EMS, Results: Follow-up interviews were conducted on 71 of improvements in the process of prehospital trauma care the drivers a mean of 10.6 months after the course. In the are possible by building on existing, although informal, interviews, 61 percent indicated that they had provided prehospital transportation. Source: Mock and others 2002. Box 68.2 Training for Emergency Care in India The training of personnel working in emergency medical Course is an indigenous two-day course developed in services is crucial to the success of efficient delivery of India by the Academy of Traumatology (India) with the care. Evidence exists to support usefulness of life-support help of international peers. The curriculum takes into training for emergency caregivers in low- and middle- account local conditions and capabilities. The cost is income countries. Courses such as the ATLS are available US$50 per trainee, and the course is taught by local train- and well established in some high- and middle-income ers to a group of 100 trainees at a time. Animal specimens countries. In most low-income countries, such training is are used to teach life-saving procedures instead of expen- not available, mainly because of prohibitive costs. The sive commercially produced manikins. More than 2,000 three-day ATLS course costs on average US$700 per health professionals have been trained in less than three trainee and is taught by six trainers to a group of 20 years. The course has now become a national training trainees at a time. The National Trauma Management standard for immediate trauma care in India. Source: Authors. great. As a neglected topic, emergency care is part of the figure 68.2. The rectangle is a schematic representation of the "10-90" gap of health research: less than 10 percent of global totality of the global burden of disease that can be potentially research investments are for problems affecting 90 percent of addressed by EMS (see "Burden of Disease"). A portion of this the world's population (Global Forum 2000). potential burden is being addressed (or reduced) by those existing interventions that have been implemented, defined by Approach to Research and Development for EMS box A. If the efficiency of current interventions were to be The spectrum of research required is diverse and may be enhanced and their coverage increased, then another portion of easily understood with the help of the schematic shown in the burden defined by box B could be addressed; this increase Emergency Medical Services | 1273 in global knowledge. Following on the illustration of fig- D: No emergency care interventions currently available to ure 68.2, there is a need to better understand the epidemiology address this burden of those conditions that can be addressed by an EMS in a low- A: Currently implemented B: Existing emergency C: Potential emergency income country and which interventions currently in place are emergency care care interventions care interventions addressing them. We have little knowledge of how to enhance interventions that that are able to that could address are addressing this address existing this burden if they the efficiency of these existing interventions and reduce their burden burden if efficiency were made cost- costs. Most important, the lack of intervention trials done in enhanced effective low-income countries creates a major research priority for the field of EMS. Well-designed, locally appropriate interventions Figure 68.2 Schematic Illustration of the Burden of Disease that establish their effectiveness are urgently needed and Potentially Addressed by EMS should include both interventions that may be available in high-income countries and new ones. Economic analysis is in efficiency will require operations research, policy research, another area for major research input in the field of EMS, and social science research. If existing interventions, which where cost and cost-effectiveness information from low- have not been implemented because of their high costs, were income countries is scant. These gaps only reflect the need for made more cost-effective, then another portion of the burden a more systematic analysis of where research investments defined by box C could be reduced. This process of making should be directed in the next five years for EMS. interventions more cost-effective will require economic analy- sis and clinical research in many instances. Finally, a portion of the burden exists for which we do not have existing interven- CONCLUSIONS: PROMISES AND PITFALLS tions, defined by box D; it requires basic and clinical research to Emergency medical systems are a critical component of develop and pilot interventions that can address other deter- national health systems in low-income countries. Governments minants of the emergency care­related burden in the future. and ministries of health in low-income countries need to pay This schematic representation is thus useful to demonstrate specific attention to the development of EMS in their countries two critical needs: and to ensure that the evolution of any EMS is both evidence based and appropriate to their national needs. More important, · essential research on emergency care in low-income the context and implementation of EMS should help health countries equity and not widen existing health disparities. · a diverse set of research studies and approaches to reduce This chapter highlights not only the urgent need for more the burden that can be addressed by EMS. attention to EMS in low-income countries but also points out an opportunity for these countries in defining better EMS for Priority Setting for Research and Development their needs. In promoting the systematic development of an of Emergency Care evidence-based EMS, low-income countries could help define Setting priorities for research and development on EMS needs more effective and cost-effective emergency systems than cur- to be a region-specific, if not country-specific, process. No cur- rently exist in high-income countries. This opportunity should rent list exists of global research and development priorities for not be lost as a result of political inattention or lack of funds; EMS, again reflecting the need for more attention and invest- the international and national stakeholders must move forward ments in this area. This chapter does not intend to prescribe a to stem the preventable loss of life from the lack of an EMS. list of issues or topics for global research and development Emergency care needs to be planned as an integral compo- efforts, but rather to highlight the gap in global research and nent of public health systems in low-income countries. Too lit- development and to suggest possible issues and topics that may tle is known about the true extent of the need for emergency be broadly relevant to low-income countries for these efforts in care, the designs of EMS that would work well for different the short to middle term. communities and populations, and the costs and benefits of A number of methods exist for setting research priorities delivering emergency care in low-income countries. These gaps in the health sector, such as the Combined Approach Matrix are a call for more investments in the research, development, promoted by the Global Forum for Health Research (Global and implementation of EMS, especially in these countries. Forum 2002) and the Essential National Health Research Universal emergency care is consistent with the right to health process promoted by the Council for Health Research for care because, by definition, emergency care is a matter of life Development. Countries and regions can use these approaches and death. Society should endeavor to ensure that prompt to help develop their research agenda for EMS. appropriate care is available in critical moments when a delay The review of evidence available in the field of emergency in care--or the delivery of inappropriate care--could mean care as applicable to low-income countries reveals many gaps loss of lives. 1274 | Disease Control Priorities in Developing Countries | Olive C. Kobusingye, Adnan A. Hyder, David Bishai, and others Annex 68.A Essential Resources for the Delivery of Emergency Care in Hospitals Major Regional District Primary emergency emergency emergency emergency care care care care Resources center center center center Organization and administration Multidisciplinary emergency care team Maintenance of statistical data Resources Immediate access to radiology or CT and ultrasound scan facility on site Blood bank on site Access to blood bank Radiological technician on site 24 hours a day Radiological services available promptly Clinical laboratory services Laboratory services on site available 24 hours a day (including, but not limited to, the following tests) Hemoglobin, glucose, gram stain, blood slide test Bacterial cultures Quality improvement Evidence of quality improvement program in accident and emergency department Monthly morbidity and mortality review Medical nursing audit and utilization review Personnel Designated doctor in charge, member of the emergency care team, with special competence in care of critically ill and injured patients, present in the emergency care unit 24 hours a day Designated doctor in charge, member of the emergency care team, with special competence in care of critically ill and injured patients, available on call Nursing personnel with special competence in the care of the critically ill and injured patients, designated member of the emergency care team, present in the emergency unit 24 hours a day All personnel trained in airway, breathing, and circulatory support techniques Equipment required for resuscitation per station shall include but not be limited to Bag valve resuscitator with reservoir Sphygmomanometer (blood pressure cuff) Cervical collars Chest decompression set (Continues on the following page.) Emergency Medical Services | 1275 Annex 68.A Continued Major Regional District Primary emergency emergency emergency emergency care care care care Resources center center center center Cut down set Delivery pack Diagnostic peritoneal lavage set open (1) Dressing trolley Drip stand Laryngoscope and blades (adult) Laryngoscope and blades (pediatric) McGills forceps (adult and pediatric) Ophthalmoscope Overhead x-ray gantry (full access to all beds) Portable ventilator capable of pediatric vent Resuscitation patient trolley Scissors to cut clothing Scoop stretcher (1) Spine board (1) Spot lamp (1) Sterile basic packs (2 per station) Stethoscope Suction apparatus Wheelchair (1) X-ray gowns (staff) X-ray viewing box Consumables (adult and pediatric) Catheters (all sizes) Central lines Chest drains Diathermy Endotracheal tubes Eye protection for staff Gloves Humidification filters Intraosseous needles Intravenous cannulas, fluids, lines Introducers and endotracheal tubes (all sizes) Lumbar puncture set 1276 | Disease Control Priorities in Developing Countries | Olive C. Kobusingye, Adnan A. Hyder, David Bishai, and others Annex 68.A Continued Major Regional District Primary emergency emergency emergency emergency care care care care Resources center center center center Malaria test kits Masks Medical waste disposal systems Nasal cannula Nasogastric tubes (all sizes) Nebulization masks Oropharyngeal airways Oxygen mask Suction catheters Syringes (assorted) Tracheotomy tubes Urine dipstick Wound care products Drugs shall include but not be limited to the following Activated charcoal Adrenaline Flumazenil (or similar benzodiazepine) Antihistamine (such as diphenhidramine) Atropine Ciprofloxacin or equivalent Beta-2 antagonist (such as propranolol) Calcium chloride Calcium gluconate Dextrose, 50 percent Diazepam Dopamine Emetic (ipecac) Metronidazole IV Furosemide or equivalent Heparin, 1,000 g/ml Hydrocortisone Lidocaine IV Magnesium sulfate IV Midazolam Morphine (Continues on the following page.) Emergency Medical Services | 1277 Annex 68.A Continued Major Regional District Primary emergency emergency emergency emergency care care care care Resources center center center center Naloxone Nitroglycerin Crystalloids (such as normal saline) Phenytoin Polyvalent snake venom Potassium chloride Scoline (suxamethonium chloride) Sodium bicarbonate Streptokinase Tetanus toxoid Vitamin K Note: Additional information on recommended drugs is available in the Model List of Essential Medicines (WHO 2002) and the WHO's Complementary Model List. NOTES Andrews, C. N., O. C. Kobusingye, and R. R. Lett. 1999. "Road Traffic Accident Injuries in Kampala." East African Medical Journal 76 (3): 1. There are 4,715,000 trauma deaths in low- and middle-income 189­94. countries per population of 5,219,401,000; thus, there are 9.033 trauma Arreola-Risa, C., C. N. Mock, L. Lojero-Wheatly, O. de la Cruz, C. Garcia, deaths per 10,000 people. F. Canavati-Ayub, and G. J. Jurkovich. 2000. "Low-Cost Improvements 2. There are 5,699,000 ischemic heart disease deaths per 5,219,401,000 in Prehospital Trauma Care in a Latin American City." Journal of population; thus, there are 10.9 ischemic heart disease deaths per 10,000 Trauma 48 (1): 119­24. people. 3. There are 2,000 to 4,000 births among 10,000 people based on crude Bazzoli G. J., R. Harmata, and C. Chan. 1998. "Community Based Trauma birth rates of 26 (South Asia), 39 (Sub-Saharan Africa), 22 (Latin Systems in the United States: An Examination of Structural America), and 17 (East Asia). Maternal mortality runs at 1 per 1,000 Development." Social Science and Medicine 46 (9): 1137­49. people. Black, R. S., and P. Brocklehurst. 2003. "A Systematic Review of Training in Acute Obstetric Emergencies." International Journal of Gynecology and Obstetrics 110 (9): 837­41. REFERENCES Desmet, M., A. Q. Chowdhury, and M. K. Islam. 1999. "The Potential for Social Mobilization in Bangladesh: The Organization and Functioning of Two Health Insurance Schemes." Social Science and Medicine 48: Ali, J., R. Adam, A. K. Butler, H. Chang, M. Howard, D. Gonsalves, and 925­38. others. 1993. "Trauma Outcome Improves Following Advanced Trauma Life Support (ATLS) Program in a Developing Country." Essien E., D. Ifenne, K. Sabitu, A. Musa, M. Alti-Mu'azu, V. Adidu, and Journal of Trauma 34: 890­99. others. 1997. "Community Loan Funds and Transport Services for Ali, J., R. Adam, T. J. Gana, and J. I. Williams. 1997. "Trauma Patient Obstetric Emergencies in Northern Nigeria." International Journal of Outcome after the Pre-hospital Trauma Life Support Program." Gynecology and Obstetrics 59 (Suppl. 2): S237­44. Journal of Trauma 42 (6): 1018­21; discussion: 1021­22. Fiedler, J. L. 2000. "The Nepal National Vitamin A Program: Prototype to Ali, J., R. Adam, D. Josa, I. Pierre, H. Bedsaysie, U. West, and others. 1998. Emulate or Donor Enclave?" Health Policy and Planning 15 (2): 145­56. "Effect of Basic Pre-hospital Trauma Life Support Program on Gearhart, P. A., R. Wuerz, and A. R. Localio. 1997. "Cost-Effectiveness Cognitive and Trauma Management Skills." World Journal of Surgery Analysis of Helicopter EMS for Trauma Patients." Annals of Emergency 22 (12): 1192­96. Medicine 30: 500­6. Altintas, K. H., N. Bilir, and M. Tuleylioglu. 1999. "Costing of an Global Forum for Health Research. 2002. The 10/90 Report on Health Ambulance System in a Developing Country, Turkey: Costs of Ankara 1700­2100. Geneva: Global Forum for Health Resarch. Emergency Aid and Rescue Services' (EARS) Ambulance System." European Journal of Emergency Medicine 6: 355­62. Graham, W. 1997. "Every Pregnancy Faces Risk." Presentation at a Safe Motherhood Technical Consultation, Sri Lanka, October 18­23. Ande, B., J. Chiwuzie, W. Akpala, A. Oronsaye, O. Okojie, C. Okolocha, and others. 1997. "Improving Obstetric Care at the District Hospital, Hauswald, M., and E. Yeoh. 1997. "Designing a Pre-hospital System for a Ekpoma, Nigeria." International Journal of Gynecology and Obstetrics Developing Country: Estimated Cost and Benefits." American Journal 59 (Suppl.): S47­53. of Emergency Medicine 15: 600­3. 1278 | Disease Control Priorities in Developing Countries | Olive C. Kobusingye, Adnan A. Hyder, David Bishai, and others Husum, H., M. Gilbert, and T. Wisborg. 2003. "Training Pre-hospital Oyesola, R., D. Shehu, A. T. Ikeh, and I. Maru. 1997. "Improving Trauma Care in Low-Income Countries: The `Village University' Emergency Obstetric Care at a State Referral Hospital, Kebbi Experience." Medical Teacher 25 (2): 142­48. State, Nigeria." International Journal of Gynecology and Obstetrics 59 Husum, H., M. Gilbert, T. Wisborg, Y. Van Heng, and M. Murad. 2003. (Suppl. 2): S75­81. "Rural Pre-hospital Trauma Systems Improve Trauma Outcome in Prevention of Maternal Mortality Network. 1995. "Situation Analysis of Low-Income Countries: A Prospective Study from North Iraq and Emergency Obstetric Care: Examples from Eleven Operations Cambodia." Journal of Trauma 54 (6): 1188­96. Research Projects in West Africa." Social Science and Medicine 40 (5): Hutton, K. C. 1995. "The End of an Era: The Demise of Life Flight San 657­67. Diego." Air Medical Journal 10: 11­13. Samai, O., and P. Sengeh. 1997. "Facilitating Emergency Obstetric Care India, Government of Delhi. 2001. "Report of Evaluation Study on CATS." through Transportation and Communication, Bo, Sierra Leone." Int J Evaluation Unit, Planning Department, New Delhi. Gynecol Obstet 59 (Suppli. 2): S157­64. Joshipura, M. K., H. S. Shah, P. R. Patel, P. A. Divatia, and P. M. Desai. 2003. Sampalis, J. S., S. Boukas, A. Lavoie, A. Nikolis, P. Frechette, R. Brown, and "Trauma Care Systems in India." Injury 34 (9): 686­92. others. 1995. "Preventable Death Evaluation of the Appropriateness of the On-site Trauma Care Provided by Urgences-Sante Physicians." Kobusingye, O., D. Guwatudde, G. Owor, and R. Lett. 2002. "Citywide Journal of Trauma 39 (6): 1029­35; comment: 1027­28. Trauma Experience in Kampala, Uganda: A Call for Intervention." Injury Prevention 8: 133­36. Sampalis, J. S., A. Lavoie, M. Salas, A. Nikolis, and J. I. Williams. 1994. "Determinants of On-scene Time in Injured Patients Treated by Leigh, B., H. B. S. Kandeh, M. S. Kanu, M. Kuteh, I. S. Palmer, K. S. Daoh, Physicians at the Site." Prehospital and Disaster Medicine 9 (3): 178­89. and F. Moseray. 1997. "Improving Emergency Obstetric Care at a District Hospital, Makeni, Sierra Leone." International Journal of Sampalis, J. S., H. Tamim, R. Denis, S. Boukas, S. A. Ruest, A. Nikolis, and Gynecology and Obstetrics 59 (Suppl. 2): S55­65. others. 1997. "Ineffectiveness of On-site Intravenous Lines: Is Prehospital Time the Culprit?" Journal of Trauma 43 (4): 608­15; Liberman, M., D. Mulder, A. Lavoie, R. Denis, and J. S. Sampalis. 2003. discussion: 615­17. "Multicenter Canadian Study of Prehospital Trauma Care." Annals of Surgery 237 (2): 153­60; comment: 161­62. Sethi, D., I. Kwan, A. M. Kelly, I. Roberts, and F. Bunn. 2003. "Advanced Trauma Life Support Training for Ambulance Crews." Cochrane Lungu, K., V. Kamfose, J. Hussein, and H. Ashwood-Smith. 2001. "Are Library 2, Oxford, U.K.: Update Software. Bicycle Ambulances and Community Transport Plans Effective in Strengthening Obstetric Referral Systems in Southern Malawi?" Shehu, D., A. T. Ikeh, and M. J. Kuna. 1997. "Mobilising Transport for Malawi Medical Journal 12 (2): 16­18. Obstetric Emergencies in Northwestern Nigeria." International Journal of Gynecology and Obstetrics 59 (Suppl. 2): S17­80. Macintyre, K., and D. Hotchkiss. 1999. "Referral Revised: Community Financing Schemes and Emergency Transport in Rural Africa." Social St. John Ambulance. 1996. First on the Scene: Emergency and Standard Science and Medicine 49: 1473­87. Levels. Instructor's Guide, 2nd ed. Ottawa: St. John Ambulance Canada. McCord, C., and Q. Chowdhury. 2003. "A Cost Effective Small Hospital in Bangladesh: What It Can Mean for Emergency Care." International Tannebaum R. D., J. L. Arnold, A. De Negri Filho, and V. S. Spadoni. 2001. Journal of Gynecology and Obstetrics 81 (1): 83­92. "Emergency Medicine in Southern Brazil." Annals of Emergency Medicine 37 (2): 223­28. McCord, C., S. Premkumar, S. Arole, and R. Arole. 2001. "Efficient and Effective Emergency Obstetric Care in a Rural Indian Community Weisman, S. M. and D. Y. Graham. 2002. "Evaluation of the Benefits and Where Most Deliveries Are at Home." International Journal of Risks of Low-Dose Aspirin in the Secondary Prevention of Gynecology and Obstetrics 75 (3): 297­307. Cardiovascular and Cerebrovascular Events." Arch Intern Med. 162: 2197­2202. McSwain, N. E. 1991. "Prehospital Emergency Medical Systems and Cardiopulmonary Resuscitation." In Trauma, 2nd ed., ed. E. E. Moore, White, K. L., T. F. Williams, and B. G. Greenberg. 1996. "The Ecology of K. L. Mattox, and D. V. Feliciano, 99­107. Norwalk: Appleton & Lange. Medical Care. 1961." Bulletin of the New York Academy of Medicine 73 (1): 187­205; discussion: 6­12. Mock, C. N., G. J. Jurkovich, D. nii-Amon-Kotei, C. Arreola-Risa, and R. V. Maier. 1998. "Trauma Mortality Patterns in Three Nations at Different Wilkinson, D. A., and M. W. Skinner. 2000. Primary Trauma Care Manual: Economic Levels: Implications for Global Trauma System A Manual for Trauma Management in District and Remote Locations. Development." Journal of Trauma 44 (5): 804­12. Oxford, U.K.: Primary Trauma Care Foundation. Mock, C. N., M. Tiska, M. Adu-Ampofo, and G. Boakye. 2002. WHO (World Health Organization). 2000. "Management of the Child "Improvements in Prehospital Trauma Care in an African Country with Serious Infection or Severe Malnutrition: Guidelines for Care with No Formal Emergency Medical Services." Journal of Trauma 53 at the First Referral Level in Developing Countries." Integrated (1): 90­97. Management of Childhood Illness, Department of Child and Adolescent Health and Development, WHO, Geneva. Moylan, J. A. 1988. "Impact of Helicopters on Trauma Care and Clinical Results." Annals of Surgery 208 (6): 673­78. ------. 2002. World Health Report 2002: Reducing Risks, Promoting Healthy Life. Geneva: WHO. Nurok, M. 2001. "The Death of a Princess and the Formulation of Medical Competence." Social Science and Medicine 53 (11): 1427­38. Emergency Medical Services | 1279 Chapter 69 Complementary and Alternative Medicine Haile T. Debas, Ramanan Laxminarayan, and Stephen E. Straus The objective of medicine is to address people's unavoidable such as cardiovascular diseases, diabetes, hypertension, needs for emotional and physical healing. The discipline has depression, and use of tobacco and other addictive sub- evolved over millennia by drawing on the religious beliefs and stances. Because lifestyle, diet, obesity, lack of exercise, and social structures of numerous indigenous peoples, by exploit- stress are important contributing factors in the causation of ing natural products in their environments, and more recently these noncommunicable diseases, CAM and TM approaches by developing and validating therapeutic and preventive to these factors in particular will be increasingly important approaches using the scientific method. Public health and for the development of future health care strategies for the medical practices have now advanced to a point at which developing world. people can anticipate--and even feel entitled to--lives that are longer and of better quality than ever before in human history. Yet despite the pervasiveness, power, and promise of con- DEFINITIONS AND DOMAINS OF temporary medical science, large segments of humanity either COMPLEMENTARY AND ALTERNATIVE cannot access its benefits or choose not to do so. More than MEDICINE AND TRADITIONAL MEDICINE 80 percent of people in developing nations can barely afford the most basic medical procedures, drugs, and vaccines. In the We refer to medical practices that evolved with indigenous industrial nations, a surprisingly large proportion of people peoples and that they have introduced to other countries opt for practices and products for which proof as to their safety through emigration as traditional medicine. We refer to and efficacy is modest at best, practices that in the aggregate are approaches that emerged primarily in Western, industrial known as complementary and alternative medicine (CAM) or as countries during the past two centuries as scientific or Western traditional medicine (TM). medicine, although we acknowledge that not all Western Much of this book considers the formidable challenges to medicine is based on scientifically proven knowledge. The advancing human health through the further dispersion of terms complementary and alternative describe practices and effective and economical medical practices. This chapter con- products that people choose as adjuncts to or as alternatives to siders both proven and unproven but popular CAM and TM Western medical approaches. Increasingly, the terms CAM and approaches and attempts to portray their current and potential TM are being used interchangeably (Kaptchuk and Eisenberg place in the overall practice of medicine. 2001; Straus 2004). With globalization, the pattern of disease in developing Endless varieties of practices are scientifically unproven and countries is changing. Unlike in the past, when communica- poorly accepted by medical authorities. For the sake of organ- ble diseases dominated, now 50 percent of the health burden izing an agenda for research into these approaches, the in developing nations is due to noncommunicable diseases, U.S. National Institutes of Health has grouped them into five 1281 somewhat overlapping domains (http://nccam.nih.gov/health/ Table 69.1 Estimated Use of CAM and TM by Patients and whatiscam) as follows: Practitioners Worldwide · Biologically based practices. These include use of a vast array Region or country Extent of use of vitamins and mineral supplements, natural products such Africa Used by 80 percent of the population for primary as chondroitin sulfate, which is derived from bovine or shark health care cartilage; herbals, such as ginkgo biloba and echinacea; Australia Used by 49 percent of adults and unconventional diets, such as the low-carbohydrate China Accounts for 30 to 50 percent of total health care approach to weight loss espoused by the late Robert Atkins. Fully integrated into the health system · Manipulative and body-based approaches. These kinds of 95 percent of Chinese hospitals have TM units approaches, which include massage, have been used India Widely used throughout history. In the 19th century, additional formal 2,860 hospitals provide TM manipulative disciplines emerged in the United States: chi- Indonesia Used by 40 percent of the entire population ropractic medicine and osteopathic medicine. Both origi- Used by 70 percent of the rural population nated in an attempt to relieve structural forces on vertebrae Japan 72 percent of physicians practice TM and spinal nerve roots that practitioners perceived as evok- ing a panoply of illnesses beyond mere musculoskeletal pain. Thailand TM integrated into 1,120 health centers · Mind-body medicine. Many ancient cultures assumed that Vietnam Fully integrated into the health care system the mind exerts powerful influences on bodily functions and 30 percent of the population is treated with TM vice versa. Attempts to reassert proper harmony between Western countries CAM and TM not strongly integrated into the health these bodily systems led to the development of mind-body care system medicine, an array of approaches that incorporate spiritual, France: at least 75 percent of the population has used meditative, and relaxation techniques. CAM at least once · Alternative medical systems. Whereas the ancient Greeks Germany: 77 percent of pain clinics provide acupuncture postulated that health requires a balance of vital humors, United States: 29 to 42 percent of population uses Asian cultures considered that health depends on the bal- CAM ance and flow of vital energies through the body. This latter Source: WHO 2002. theory underlies the practice of acupuncture, for example, which asserts that vital energy flow can be restored by summarized numerous surveys of use (table 69.1). In developing placing needles at critical body points. nations, TM is the sole source of health care for all but the privi- · Energy medicine. This approach uses therapies that involve leged few. By contrast, in affluent countries individuals select the use of energy--either biofield- or bioelectromagnetic- CAM approaches according to their specific beliefs. For example, based interventions. An example of the former is Reiki ther- as many as 60 percent of those living in France, Germany, and the apy, which aims to realign and strengthen healthful energies United Kingdom consume homeopathic or herbal products. through the intervention of energies radiating from the hands Only 1 to 2 percent of Americans use homeopathy,but 10 percent of a master healer. of adults use herbal medicines, 8 percent visit chiropractors, and 1 to 2 percent undergo acupuncture every year (Ni, Simile, and Alternative systems of medicine use elements from each of Hardy 2002). Use of CAM and TM among patients with chron- these CAM and TM domains. For example, traditional Chinese ic, painful, debilitating, or fatal conditions, such as HIV/AIDS medicine incorporates acupuncture, herbal medicines, special and cancer, is far higher, ranging from 50 to 90 percent diets, and meditative exercises such as tai chi. Ayurveda in India (Richardson and Straus 2002). similarly uses the meditative exercises of yoga, purifying diets, There is remarkably little correlation between the use of and natural products. In the West, homeopathic medicine and CAM and TM approaches and scientific evidence that they are naturopathic medicine each arose in the late 19th century as safe or effective. For many CAM and TM practices, the only reactions to the largely ineffectual and toxic conventional evidence of their safety and efficacy is embodied in folklore. approaches of the day: purging, bleeding, and treatments with Beginning more than 1,500 years ago, data on the use of heavy metals such as mercury and arsenicals. thousands of natural products were assembled into impressive monographs in China, India, and Korea, but these compendi- DEMOGRAPHY, USE, TOXICITY, AND EFFICACY ums--and similar texts from Arabic, Egyptian, Greek, and Persian sources and their major European derivatives--are The use of CAM and TM varies widely between and within coun- merely catalogs of products and their use rather than formal tries.The World Health Organization (WHO) has published and analyses of safety and efficacy. 1282 | Disease Control Priorities in Developing Countries | Haile T. Debas, Ramanan Laxminarayan, and Stephen E. Straus Table 69.2 Some Natural Products That May Alter Drug This section describes the socioeconomic determinants of Actions seeking treatment from traditional healers and providers of CAM; reviews the evidence on the cost-effectiveness of CAM Herbal product Class of drug and TM; and discusses cost-effective approaches to regulating, Ephedra (ma huang) Alpha and beta adrenergics improving, and expanding the use of CAM and TM. Much of Garlic Anticoagulants; some HIV protease inhibitors this evidence is from industrial countries; few studies have been Ginkgo biloba extract Anticoagulants conducted in or are applicable to low- and middle-income Glucosamine Antidiabetics countries. This caveat is important for two reasons. First, the Saw palmetto Androgens CAM and TM modalities discussed in this section may not be St. John's wort HIV protease inhibitors; some chemotherapy used in many developing countries. Second, the limited data on drugs; cyclosporine A; birth control cost-effectiveness may not be applicable in the case of those Valerian Sedatives countries. Nevertheless, the data give a rough picture of the rel- ative cost-effectiveness of a number of CAM and TM practices. Source: Niggemann and Gruber 2003. Many people who today choose herbal products in lieu of Economic Factors That Influence the Use of Complementary prescription medications assume that because these products and Alternative Medicine and Traditional Medicine are natural, they must be safe, even when the evidence for this Users of CAM and TM approaches choose health practices that assertion is essentially anecdotal. Recent studies have shown resonate with their beliefs about health (Astin 1998). Although that herbals are highly variable in quality and composition, economic factors play a role in this choice, the underlying with many marketed products containing little of the intended incentives are not always predictable. For instance, a common ingredients and containing unintended contaminants, such as misconception is that patients opt for CAM and TM services heavy metals and prescription drugs. A few herbals are banned because they are cheaper alternatives to conventional medical outright in several countries. Comfrey and kava have been care. Even though there are certainly instances when the cost of associated with liver failure, aristolochia with genitourinary treatment using CAM or TM is much cheaper than the cost of cancer (De Smet 2002), and ephedra with heart attacks and accessing a conventional medical service, several studies have strokes (Shekelle and others 2003). More important, herbals found that CAM and TM cost the same or more than conven- contain ingredients that can accelerate or inhibit the metabo- tional treatments for the same conditions (see, for example, lism of prescription drugs (table 69.2). The most notorious of Muela, Mushi, and Ribera 2000). these is St. John's wort, which affects the metabolism of nearly At least one study has shown that financial considerations 50 percent of all prescription drugs (Markowitz and others are rarely the primary factor in choosing a traditional healer, 2003). The cumulative data on the pharmacological and poten- ranking behind such reasons as confidence in the treatment, tial adverse effects of herbal supplements now dictate that ease of access, and convenience (Winston and Patel 1995). In patients discuss their use of supplements with knowledgeable the United States, the average cost of a single visit to a Navajo practitioners before initiating treatment. healer was US$388, and the average annual cost of using a As to evidence of the efficacy of CAM and TM approaches, traditional healer represented roughly a fifth of the reported thousands of small studies and case series have been reported annual income of respondents in a survey (Kim and Kwok over the past 50 years. Few were rigorous enough to be at all 1998). The high cost of using a healer was cited as the most compelling, but they are sufficient to generate hypotheses that common barrier to seeking care from this source. In Kenya, the are now being tested in robust clinical trials. The existing body average charge per patient per visit to a TM practitioner was of data already shows that some approaches are useless, that for K Sh 46 (US$4 in 1981), which was significantly greater than many the evidence is positive but weak, and that a few are the average charge per visit even in private health care facilities highly encouraging (table 69.3). (Mwabu, Ainsworth and Nyamete 1993). Finally, a survey in Zimbabwe reported that the median cost of consulting an herbalist was Z$23 per visit, compared with Z$1 for a govern- ECONOMICS OF COMPLEMENTARY AND ment clinic and Z$29 for a private doctor (Winston and Patel ALTERNATIVE MEDICINE AND TRADITIONAL 1995). The same survey found that outcomes tended to be bet- MEDICINE ter when patients went to government clinics (67.3 percent of visits resulted in a good outcome) than when patients consulted Although social, medical, and cultural reasons may account for herbalists (50 percent of visits resulted in a good outcome). why people in a given country prefer CAM and TM to conven- TM is not always more expensive than conventional medi- tional (Western) medicine, economic forces are also at play. cine, however. Survey respondents in Ghana reported that the Complementary and Alternative Medicine | 1283 Table 69.3 Levels of Evidence for the Efficacy of Selected CAM and TM Approaches CAM or TM Study Level of approach Potential use outcome evidence Source Artemisia annua Treating drug-resistant malaria Positive A van Agtmael, Eggelte, and van Boxtel 1999 Black cohosh Controlling menopausal symptoms Mixed B Kronenberg and Fugh-Berman 2003 Cranberry Preventing urinary tract infection Positive B Jepson, Mihaljevic, and Craig 2000 Echinacea Preventing or treating viral colds Mixed B Barrett 2003; Taylor and others 2003 Garlic Lowering blood cholesterol Positive C Le Bars and others 1997 Ginkgo biloba Preventing or treating dementia Mixed B Kanowski and Hoerr 2003 extract Ginseng Improving energy and immunity Mixed C Richy and others 2003 against infection Glucosamine Relieving osteoarthritis Positive A Reginster, Deroisy, and Rovalty 2001 Hawthorn Improving cardiac function Mixed B Pittler, Schmidt, and Ernst 2003 Milk thistle Improving liver function Positive C Jacobs and others 2002 St. John's wort Treating moderate to severe depression Negative A Hypericum Depression Trial Study Group 2002 Treating mild depression Positive B Di Carlo and others 2001 Saw palmetto Relieving symptoms of benign prostatic hypertrophy Positive B Gerber and others 2001 Acupuncture Relieving arthritis pain Positive B Berman and others 1999 Relieving the pain of tooth extraction Positive B Lao and others 1995 Treating hypertension Mixed C Chiu, Chi, and Reid 1995 Relieving nausea from chemotherapy Positive A Shen and others 2000 Relieving addiction withdrawal Mixed B Margolin and others 2002 Treating asthma Negative B Linde, Jobst, and Panton 2000 Meditation Decreasing anxiety Positive B Speca and others 2000 Decreasing blood pressure Mixed B Schneider and others 1995 Biofeedback Preventing migraine Positive B Holroyd and Penzien 1990 Homeopathy Treating asthma Mixed B White and others 2003 Treating gastroenteritis Positive C Jacobs and others 2003 Magnet therapy Treating plantar fasciitis Negative B Winemiller and others 2003 Chiropractic Treating lower back pain Positive B Cherkin and others 2003 Source: Authors. A multiple high-quality, randomized, controlled trials; B single high-quality trials or smaller, less rigorous trials; C weaker clinical trials; Mixed conflicting results among studies of similar quality. cost of malaria treatment at a health clinic ranged from ¢1,900 exchange for labor), and payment may be contingent on out- to ¢3,000 (US$1.30 to US$2.00 in 1997), treatment at home come. The availability of an outcome-contingent contract using drugs bought from pharmacies or health care workers favors TM over Western medicine when the disease condition ranged between ¢200 and ¢1,000 (US$0.10 to US$0.70), and requires providers to both exert effort in curing patients and treatment by an herbalist was virtually free (Ahorlu and others induce patients to comply with their recommendations. 1997). Nonetheless, this strategy may be difficult to apply to the larger Another common misconception is that the poor are more health care system. likely to use TM. At least one study shows that this may not be Furthermore, patients tend to seek care from traditional true. In Zimbabwe, the mean monthly income of households healers for conditions such as mental illness, impotence, and visiting an herbalist, Z$877, was greater than the mean monthly chronic disorders, which they perceive as requiring greater income of households using government clinics, Z$718 involvement by the extended family and kinship group. (Winston and Patel 1995). Accordingly, the availability of financial support for seeking Although some traditional healers charge more than con- treatments for these disorders is greater than it is for illnesses ventional practitioners, their fees may be negotiable, the such as malaria or diarrhea, for which patients more often seek method of payment may be flexible (often on credit or in conventional treatment. 1284 | Disease Control Priorities in Developing Countries | Haile T. Debas, Ramanan Laxminarayan, and Stephen E. Straus Few published data are available on the financial costs of Economic Evidence TM in low- and middle-income countries. The data presented Although most studies tend to focus on a specific CAM or TM here on the use of traditional healers are extracted from the practice, Sommer, Burgi, and Theiss (1999) looked more World Bank's living standards surveys in Vietnam to provide broadly at whether the provision of CAM and TM services one nationally representative snapshot of the situation. Of through prepaid health plans or government insurance 28,254 individuals in the sample, 10,033 had consulted a health reduces the overall costs of health care and found that it does care provider in the four weeks preceding the survey. These not. A possible reason is that few individuals who are offered consultations included both home visits and visits to a access to CAM use them, and those who do might access those provider. Of the 10,033, 1,829 had been to a public provider, services in addition to, not in place of, more conventional 1,431 to a private provider, 7,650 to a pharmacy, and 259 to a health services. traditional healer.1 The most common reasons for visiting a Studies that compare the cost-effectiveness of different traditional provider were headache, followed by cough and CAM and TM approaches using the same analytical framework fever. The per visit drug cost for consulting a traditional healer are rare. One such study in Peru looked at the costs and cost- was D 46, and the total cost per visit was D 51, compared with effectiveness of treatment using conventional medicine and drug costs of D 38 and total costs of D 41 for going to a private TM (EsSalud andOPS 2000). Complementary medical prac- clinic. tices evaluated included acupuncture, homeopathy, tai chi, One commonly cited motivation for using CAM and TM is meditation, reflexology, hydrotherapy, naturopathy, and mas- that their use might lower the incidence and costs of side effects sage. Patients were enrolled in either the Western medicine associated with conventional treatments, but the published evi- group or the CAM group. Patients were not randomized dence on this point remains mixed. There is some evidence that between the two treatment groups, but they were matched by CAM is used in addition to conventional treatments (Thomas disease pathology and severity, age, and sex. Furthermore, and others 1991), but CAM may also have the effect of displac- selected patients had completed at least one year in the health ing conventional treatments. An outpatient survey found that, system, as the investigators reasoned that this would enable of 246 patients who had been receiving conventional treatment them to evaluate their follow-up. Overall, the investigators from the Royal London Homeopathic Hospital since the onset found that complementary medicine was between 53 and of care, a third had halted their conventional treatment and 63 percent less expensive than conventional medicine for another third had reduced their intake of conventional med- achieving equivalent levels of effectiveness. Complementary ication (van Haselen 2000).2 The extent to which homeopathic medicine was especially cost-effective for osteoarthritis, hyper- treatment displaced conventional treatment varied by indica- tension, facial paralysis, and peptic ulcers. tion. The use of homeopathic treatment often replaced con- The rest of this section looks at the economic evidence on ventional treatments in patients with skin and respiratory specific forms of CAM or TM. infections; in patients with cancer, its use was purely comple- mentary and therefore added to overall health care costs. Acupuncture. Lindall's (1999) study finds that an acupunc- Thomas and others (1991) observe that patients who use ture referral for musculoskeletal conditions costs a mean of CAM and TM also commonly access conventional medical US$422, roughly 60 percent less than the cost of referral to a care. In industrial countries, most CAM usage complements Western practitioner. However, this study was not randomized, conventional care, but this is also common in developing and patients had to have failed first-line drug treatment before nations. For instance, Mwabu (1986) provides evidence from being offered the choice of second line-treatment, either with Kenya that patients are likely to use more than one type of acupuncture or with Western medicine. provider from the range of those available, such as government facilities, mission clinics, private clinics, pharmacies, and tradi- tional healers. Furthermore, the choice of provider depends on Homeopathy. Evidence indicates that the cost of homeopathic patients' illness, condition, socioeconomic status, and educa- medication is lower than the average cost of allopathic prod- tion. If an initial visit to one kind of provider did not resolve ucts, which would be an economic factor in favor of its use if the disease satisfactorily, a follow-up visit was made to a differ- homeopathy were proven to be effective. A study by the ent kind of provider. Finally, the quality of care--including National Health Service in the United Kingdom found that the efficiency of service and waiting time at government and pri- drug costs associated with homeopathy were lower than those vate clinics--is an important determinant of whether patients of allopathic practitioners (Swayne 1992). A four-year study of choose to go to traditional healers. Most traditional healers sur- 100 patients that compared homeopathic drug costs with those veyed in a second study referred patients to Western practices of conventional drugs found an average cost saving of US$96 for treatment when necessary (Mwabu, Ainsworth, and during the study period for those using homeopathic drugs Nyamete 1993). (Jain 2003).3 Complementary and Alternative Medicine | 1285 Ayurveda. A study that compared medical expenditures over a Beyond Cost-Effectiveness: Ancillary Benefits and Costs four-year period for participants in a comprehensive program of CAM and TM of ayurvedic-based natural medicine (which included antioxi- Although cost-effectiveness is one guiding rationale for deter- dant strategies, mind-body medicine, and other techniques) mining resource allocations for expanding (or restricting) with participants whose expenditures were covered through a access to CAM and TM, additional societal benefits and costs, BlueCross BlueShield health insurance plan found that the such as effects on biodiversity, must also be considered. CAM expenditures for the ayurvedic group were 50 percent lower per and TM could provide a rationale for conserving species, but person (Orme-Johnson and Herron 1997). However, the study overharvesting of endangered species for medicinal purposes is was not randomized and failed to control for the inclination of also a concern. According to WHO, 85 percent of the world's only a subset of people to accept and remain compliant with population (principally those in developing countries) ayurvedic approaches. depends on plants for medicine, and 25 percent of prescription drugs have an active ingredient derived from a flowering plant Chiropractic. Some studies found that spinal manipulation is (Cox 2001). The possible extinction of medicinal plants is of less expensive than conventional treatments for episodes of concern not only to developing countries but also to industrial back pain. One nonrandomized study found that the cost of countries, as in the cases of poaching of American ginseng and chiropractic treatment over a five-year period, including both overharvesting of native saw palmetto. Similarly, the reliance of provider costs and equipment costs (US$28,902), was 24 per- Chinese TM on tiger genitals, bear gallbladders, and black rhi- cent less than the cost of Western pain therapy (US$38,029) noceros horns has played an important role in poaching and (Kumar, Malik, and Demeria 2002). Moreover, 15 percent of threatens to wipe out these mega fauna. patients in the chiropractic group were able to return to work, Local knowledge and culture regarding the uses of medici- compared with none in the control group. nal plants may be important determinants of whether a certain However, other larger and better-controlled studies failed species will survive (Etkin 1998). In addition to the biodiversity to find a difference between chiropractic and physical therapy value of these saved species, scientists may be able to analyze in terms of either outcomes or costs (Cherkin and others these plants for potential clinical application on a broader scale 1998; Skargren and others 1997; Skargren, Carlsson, and than TM permits. Although preserving traditional knowledge Oberg 1998). A study of adults with low back pain who were of healing practices helps preserve the culture and identity of randomly assigned to physical therapy or chiropractic manip- indigenous populations, CAM and TM may impose significant ulation or were just given an educational booklet found no costs. In such instances, promoting conventional treatments significant differences in either the mean costs of care or the that do not depend on endangered species may bring impor- outcomes between the physical therapy and chiropractic tant benefits to society. groups (Cherkin and others 1998). Three-quarters of the par- ticipants in these groups--who incurred costs of roughly US$430 over the two-year period of the study--reported that EXPANDING THE BENEFICIAL USE OF their outcome was either good or excellent, compared with a COMPLEMENTARY AND ALTERNATIVE MEDICINE third of those who were assigned booklets; however, the mean AND TRADITIONAL MEDICINE cost of care for the booklet group was only US$153 for the two-year period. Despite the uncertainty about the clinical efficacy and cost- effectiveness of certain CAM and TM practices, expansion of Mind-Body Treatments. Little evidence is available on the their use in instances in which moderate evidence of their effi- cost-effectiveness of practices such as meditation and yoga, but cacy and good evidence of their safety exists could yield health, the cost of acquiring the skills required for these practices, as well social, and economic benefits. A number of surveys show that as the time costs of practicing them, are so low relative to con- local pharmacies are the primary source of treatment for many ventional medicine that evidence of their clinical effectiveness ailments, especially in rural areas where government or private might suffice to justify their use on economic grounds. Available clinics are less accessible. In these situations, improving the evidence from clinical studies suggests that mind-body treat- quality of TM might serve as an effective substitute for allow- ments can be cost-effective (Caudill and others 1991; Friedman ing the unregulated use of conventional medical treatments. and others 1995; Hellman and others 1990; Sobel 1995). Training traditional healers is substantially less expensive than Blumenthal and others (2002) find significant declines in coro- training doctors or nurses. A study of 52 traditional healers nary events and in predicted costs of care for patients who were interviewed as part of a survey in Kenya estimated that the assigned to a one-and-a-half-hour long weekly class on stress average out-of-pocket (cash) costs of training to be a tradi- management, relative to usual care for each of the first two years tional healer were K Sh 418 (US$40 in 1981) (Mwabu, of follow-up and after five years. Ainsworth, and Nyamete 1993). 1286 | Disease Control Priorities in Developing Countries | Haile T. Debas, Ramanan Laxminarayan, and Stephen E. Straus Traditional healers can also be recruited into a more broadly Recognizing the redistributive nature of investment in TM based system for delivering public health; for example, with is important. Indigenous people will seek the help of tradi- additional training, traditional healers can serve as primary tional healers because of proximity, familiarity, and trust. health care workers (Hoff 1997) and provide advice on such Investments in TM could therefore be used strategically to matters as sexually transmitted diseases and oral rehydration increase access to conventional preventive and therapeutic therapy (Nations and de Souza 1997; Nations and others 1988; care. Including the traditional healer as part of the health care Ndubani and Hojer 1999). In addition, permitting access to team may thus be an important strategy both to attract CAM and TM within the context of the conventional health patients and to upgrade the skills and training of traditional care system would facilitate access to multiple health services at healers. one location. How equity is affected by the proportions in which different Comprehensive policy on CAM and TM is lacking in most condition-specific interventions are combined and how other countries, including the United States. According to the 1994 interventions (regulations, tax policy, managerial changes) are Dietary Supplement, Health, and Education Act, the U.S. Food likely to affect equity need to be studied. Given that the major- and Drug Administration cannot require proof that dietary ity of indigenous populations in developing countries use TM supplements and herbal products are safe and effective before for their primary health care, the availability, safety, and afford- they are sold, although it is charged with requiring good man- ability of TM, including herbal medicines, should be ensured as ufacturing practices. The quality of herbal products is not reg- a matter of equity. One way to do this is by supporting local ulated, and herbal products typically differ from source to production of safe and effective herbals such as artemisia at source and from batch to batch in terms of their component affordable prices. In addition, rigorous research on TM should ingredients and respective amounts and in terms of whether be supported. WHO is currently conducting collaborative they contain contaminants. In the United States, no single studies on herbal treatments for HIV/AIDS, malaria, sickle cell entity is responsible for all aspects of CAM and TM control, anemia, and diabetes. Ineffective or unsafe herbal products education, information, and research, and no national, volun- identified by such studies should be removed from use, while tary system of self-regulation exists. National nongovernmen- those with proven efficacy and safety should be made available tal organizations, such as the Accreditation Commission for for therapeutic use. Acupuncture and Oriental Medicine, the American Board of Medical Acupuncture, the Council of Chiropractic Education, the Council of Homeopathic Education, and the Commission LESSONS LEARNED AND IMPLEMENTATION on Massage Therapy Accreditation, accredit education in some CAM and TM fields, but such accreditation bodies do The pervasiveness of different modalities of TM and CAM not exist in many developing countries. Nearly all countries varies greatly from country to country. For example, in China, lack rigorous research training programs in CAM and TM. where traditional Chinese medicine is well integrated into the A common misperception is that in the developing world health system, many different modalities may be used to treat a CAM and TM is used primarily by poorer, uneducated popu- given condition. In the United States, by contrast, CAM pro- lations, while in industrial countries it is used more by affluent grams are slowly being integrated with conventional medicine. and better-educated segments of the population (Eisenberg Several medical schools have nascent CAM programs and have and others 1998). In both settings, relatively little evidence integrated them into medical school curricula to differing supports this view. Many investigators have failed to critically degrees. One of the more acclaimed programs of this kind assess the use of CAM and TM by minority and immigrant in the United States is that developed by Andrew Weil at populations in Western nations. In Africa, nearly 85 percent of the University of Arizona Health Sciences Center. His the population uses TM, often as the only way to obtain pri- Integrative Medicine Fellowship Program trains physicians in mary health care, and wealthier people in developing coun- CAM and TM and strives to produce a new delivery model tries often use TM (WHO 2002). Investments in improving whereby physicians, patients, and nurses form a healing team the quality and consistency of TM could reduce the cost of for the care of the patient. However, this program needs to be health care delivery, especially for chronic conditions such critically evaluated before its adoption by more institutions can arthritic pain and AIDS, where TM interventions may be urged. improve patients' sense of well-being, appetite, and energy. At Despite the complexity, diversity, and controversy the same time, in the absence of resources to extend the public surrounding CAM/TM approaches, some notable success health infrastructure, a network of certified CAM and TM stories reveal the influence of globalization, whereby providers could provide the infrastructure for delivering modalities discovered in the developing world have been other care, such as immunizations and maternal-child health adopted in the West, with or without modifications, and vice programs. versa. Complementary and Alternative Medicine | 1287 Artemisinin The second success story is research showing that chiro- Artemisinin is a recently developed, active metabolite of practic manipulation for low back pain is superior to bed rest, artemisia, an herbal extract that has been used in China for physical therapy, or provision of an educational booklet centuries to treat fever. Chinese scientists determined the active (Cherkin and others 1998). Chiropractic manipulation has also ingredient of the herbal in the 1970s, and Western pharmaceu- shown results comparable to those achieved with nonsteroidal, tical companies have developed several derivatives as drugs for anti-inflammatory drugs in alleviating back pain (Straus 2004). use against resistant Plasmodium malaria (Li and others 2000). Randomized clinical trials have shown that one such drug, Homeopathy dihydroartemisinin-piperaquine, is effective against drug- resistant Plasmodium falciparum malaria (Hien and Dolecek Homeopathy is a success in terms of its broad appeal and use, 2004). Another artemisinin derivative, artesunate, was shown not because of the strength of evidence supporting it. Indeed, to increase parasite clearance and reduce the gametocyte count few conventional scientists and physicians find homeopathy to when added to existing drugs to combat malaria (Adjuik and be plausible. According to the "principle of similars" underly- others 2004). ing homeopathy, practitioners choose remedies that, when given in high concentrations, produce symptoms similar to those that the patient presents with. The substance is then put Acupuncture in solution and serially diluted by as much as 1060, well beyond Another CAM and TM modality that has considerable accept- the point defined by Avogadro's number (at which a single ance is acupuncture. Many pain management clinics, hospitals, molecule of the original substance could remain in the solu- and academic centers in the West now provide acupuncture tion). Homeopathy claims that the acts of serial diluting and services, and some insurance companies reimburse for vigorous shaking imprint information into water so that acupuncture services. Rigorous clinical trials have demonstra- medicinal properties are retained even when no or few mole- tive positive efficacy in two areas: (a) management of postop- cules of the starting medicine are present. erative nausea and emesis (Shen and others 2000) and (b) ame- As implausible as this claim may seem, homeopathy is used lioration of the pain of chronic osteoarthritis (Ezzo and others worldwide with reported success (Jonas, Kaptchuk, and Linda 2001; Soeken 2004; Tukmachi and others 2004). Studies pro- 2003). Randomized controlled trials have suggested that it viding rational explanations of the mechanisms whereby might be effective for treating influenza (Vickers and Smith acupuncture might be achieving its effects complement the evi- 2000), allergies (Taylor and others 2000), and postoperative dence about its efficacy; for example, one mechanism of action ileus (Barnes, Resch, and Ernst 1997). However, critics have appears to involve opioid-dependent brain pathways. This kind questioned the quality and analyses of these trials. Some have of two-step process--that is, initial demonstration of clinical questioned the validity of pooling data from trials of different efficacy followed by scientific research into the mechanism of populations, interventions, and outcome measures, as several action--is one way that CAM and TM will gain scientific reviews of homeopathy have done. Jonas, Kaptchuk, and Linda acceptance and integration into conventional medicine. (2003, 393) assert that "there is a lack of conclusive evidence on the effectiveness of homeopathy for most conditions. Homeopathy deserves an open-minded opportunity to Chiropractic Medicine and Osteopathy demonstrate its value by using evidence-based principles, but it Chiropractic medicine was invented in the American heartland should not be substituted for proven therapies." during the waning years of the 19th century. It uses spinal manipulation to treat an array of conditions thought to arise because of abnormal alignment of or stresses on vertebrae, Mind-Body Intervention most often in patients with musculoskeletal complaints. Two The work of David Spiegel at Stanford University on group aspects of chiropractic medicine are success stories. First, even support for breast cancer patients excited wide interest in the though practitioners of conventional medicine ostracized prac- potential value of mind-body interventions (Spiegel and others titioners of chiropractic medicine in the late 19th century and 1989). The study was a randomized controlled trial with a the first half of the 20th century, it has gradually evolved into a 10-year follow-up involving 86 women with metastasized viable healing discipline that is increasingly accepted by the breast cancer. A one-year psychosocial intervention consisting conventional medicine community. The evolution of chiro- of weekly supportive group therapy with self-hypnosis for pain practic can be compared with that of osteopathy. Osteopathy showed that the mean survival time in the treated group was was developed in the United States in parallel with chiroprac- 37 months, compared with 19 months for the control group. tic, but the field elected to accommodate rather than reject allo- Moreover, Spiegel (1994) notes that appropriate psychotherapy pathic techniques. (both group and individual) not only reduced depression and 1288 | Disease Control Priorities in Developing Countries | Haile T. Debas, Ramanan Laxminarayan, and Stephen E. Straus anxiety and improved coping skills, but also saved money by should be no different from those used in conventional bio- reducing the number of office visits, diagnostic tests, medical medical research. procedures, and hospital admittances. Although Spiegel's find- Both CAM and TM and biomedical practitioners need to ings have not been replicated, they do illustrate the potential understand the strengths, limitations, and contributions of benefits of mind-body intervention and have led to studies of their particular approaches so that they can work together in possible mechanisms through which such interventions may ways that ensure the best possible care for their patients and the operate. achievement of their shared goals of improved individual and public health. Once these issues have been addressed, countries could devote additional resources to studying those CAM and THE RESEARCH AND DEVELOPMENT AGENDA TM approaches that appear to be the most promising in rela- tion to their most pressing public health problems. Some pri- The lack of product quality and consistency and the absence of ority areas for CAM and TM research are widely applicable, compelling data on the safety and efficacy of most CAM and including studies of approaches to palliate chronic pain and TM approaches present major challenges to any effort to opti- suffering, relieve depression, help release the grip of addictive mize the distribution of precious health resources. These diffi- substances, and slow the progression of degenerative disorders culties also pose opportunities for research. Other formidable such as arthritis and dementia. challenges include the variability in training, credentialing, and licensing CAM and TM practitioners. Increasingly, efforts are being made in several countries to regulate both products and NOTES practitioners. Ultimately, stringent controls on training, prac- tices, and products must be complemented by rigorous 1. Because some individuals had gone to more than one provider dur- ing the four-week period, the total comes to more than 10,033. research to ascertain which approaches are safe and effective-- 2. The median duration of current treatment at the Royal London and for which indications. Homeopathic Hospital was three years. The global use and potential effect of CAM and TM 3. This study did not take into account the costs of physician time, the costs of laboratory tests, or patients' costs. practices, the lack of adequate data validating their safety and efficacy, and the existence of highly effective conventional alter- natives for many of them dictate that resources should be devoted to fuller characterization and standardization of CAM REFERENCES and TM approaches. Investing precious resources in integrat- Adjuik, M., A. Babiker, P. Garner, P. Olliaro, W. Taylor, and N. White. 2004. ing such approaches further into health care infrastructures "Artesunate Combinations for Treatment of Malaria: Meta-Analysis." can be justified only on the basis of compelling data. This point Lancet 363 (9402): 9­17. leads to the question of what constitutes a rational agenda for Ahorlu, C. K., S. K. Dunyo, E. A. Afari, K. A. Koran, and F. K. Nkrumah. this work. 1997. "Malaria-Related Beliefs and Behavior in Southern Ghana: Implications for Treatment, Prevention, and Control." Tropical For resource-rich industrial nations, one model for CAM Medicine and International Health 2 (5): 488­99. and TM research is that being implemented by the National Astin, J. A. 1998. "Why Patients Use Alternative Medicine: Results of a Center for Complementary and Alternative Medicine National Study." Journal of the American Medical Association 279 (19): (NCCAM) of the U.S. National Institutes of Health (http:// 1549­53. nccam.nih.gov). In 2004, NCCAM planned to invest US$117 Barnes, J., K. L. Resch, and E. Ernst. 1997. "Homeopathy for Postoperative Ileus? A Meta-Analysis." Journal of Clinical Gastroenterology 25: 628­33. million in research and research training. It is supporting some Barrett, B. 2003. "Medicinal Properties of Echinacea: A Critical Review." 800 individual projects at present, including studies of the Phytomedicine 10: 66­86. composition of natural products and their pharmacological Berman, B. M., B. B. Singh, L. Lao, P. Langenberg, H. Li, V. Hadhazy, and effects, studies of the neurobiological mechanisms of acupunc- others. 1999. "A Randomized Trial of Acupuncture as an Adjunctive ture and the placebo effect, and clinical trials with 30 to 30,000 Therapy in Osteoarthritis of the Knee." Rheumatology 38: 346­54. participants. NCCAM now has a strategic plan for its interna- Blumenthal, J. A., M. Babyak, J. Wei, C. O'Connor, R. Waugh, E. Eisenstein, and others. 2002. "Usefulness of Psychosocial Treatment of Mental tional programs that emphasizes research, training, and efforts Stress-Induced Myocardial Ischemia in Men." American Journal of to learn about the rich, indigenous TM heritage. Australia, Cardiology 89 (2): 164­68. through a government agency similar to NCCAM, is conduct- Caudill, M., R. Schnable, P. Zuttermeister, H. Benson, and R. Friedman. ing research and training programs in collaboration with its 1991. "Decreased Clinic Use by Chronic Pain Patients: Response to indigenous people. Although the scope of NCCAM's research Behavioral Medicine Interventions." Clinical Journal of Pain 7: 305­10. agenda is larger than what most other nations could accom- Cherkin, D. C., R. A. Deyo, M. Battie, J. Street, and W. Barlow. 1998. "A Comparison of Physical Therapy, Chiropractic Manipulation, and modate, its underlying philosophy should be universal. That is, Provision of an Educational Booklet for the Treatment of Patients with the standards for research into CAM and TM approaches Low Back Pain." New England Journal of Medicine 339 (15): 1021­29. Complementary and Alternative Medicine | 1289 Cherkin, D. C., K. J. Sherman, R. A. Deyo, and P. G. Shekelle. 2003. "A Kanowski, S., and R. Hoerr. 2003. "Ginkgo Biloba Extract EGb 761® Review of the Evidence for the Effectiveness, Safety, and Cost of in Dementia: Intent-to-Treat Analyses of a 24-Week, Multicenter, Acupuncture, Massage Therapy, and Spinal Manipulation for Back Double-Blind, Placebo-Controlled, Randomized Trial." Pharm- Pain." Annals of Internal Medicine 138: 898­906. acopsychiatry 36: 297­303. Chiu, Y. J., A. Chi, and I. A. Reid. 1995. "Cardiovascular and Endocrine Kaptchuk, T. J., and D. M. Eisenberg. 2001. "Varieties of Healing. 2: A Effects of Acupuncture in Hypertensive Patients." Clinical and Taxonomy of Unconventional Practices." Annals of Internal Medicine Experimental Hypertension 19: 1047­63. 135: 196­204. Cox, P. A. 2001. "Biodiversity and Pharmacology." In Encyclopedia of Kim, C., and Y. S. Kwok. 1998. "Navajo Use of Native Healers." Archives of Biodiversity, vol. 4, ed. S. A. Levin. San Diego, CA: Academic Press. Internal Medicine 158 (20): 2245­49. De Smet, P. A. 2002. "Herbal Remedies." New England Journal of Medicine Kronenberg, F., and A. Fugh-Berman. 2003. "Complementary and 347: 2046­56. Alternative Approach for Menopause: A Review of Randomized, Di Carlo, G., F. Borrelli, E. Ernst, and A. A. Izzo. 2001. "St. John's Wort: Controlled Trials." Reproductive Toxicology 17: 137­52. Prozac from the Plant Kingdom." Trends in Pharmacologic Science 22: Kumar, K., S. Malik, and D. Demeria. 2002. "Treatment of Chronic Pain 292­97. with Spinal Cord Stimulation versus Alternative Therapies: Cost- Eisenberg, D. M., R. B. Davis, S. G. Ettner, and S. Appel. 1998. "Trends in Effectiveness Analysis." Neurosurgery 51 (1): 106­15. Alternative Medicine Use in the United States, 1990­1997: Results of a Lao, L., S. Bergman, P. Langenberg, R. H. Wong, and B. Berman. 1995. Follow-Up National Survey." Journal of the American Medical "Efficacy of Chinese Acupuncture on Postoperative Oral Surgery." Oral Association 28 (18): 1569­75. Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and EsSalud and OPS (Organización Panamericana de Salud [Pan American Endodontics 79: 423­28. Health Organization]). 2000. "Estudio costo-efectividad: Programa Le Bars, P. L., M. M. Katz, N. Berman, T. M. Itil, A. M. Freedman, and A. F. Nacional de Medicina Complementaria." Seguro Social de EsSalud. Shatzberg. 1997. "A Placebo-Controlled, Double-Blind, Randomized Lima: EsSalud and OPS. Trial of an Extract of Ginkgo Biloba for Dementia, North American Etkin, N. L. 1998. "Indigenous Patterns of Conserving Biodiversity: EGb Study Group." Journal of the American Medical Association 278: Pharmacologic Implications." Journal of Ethnopharmacology 63 (3): 1327­32. 233­45. Li, Y., Y. M. Zhu, H. J. Jiang, J. P. Pan, G. S. Wu, J. M. Wu, and others. 2000. Ezzo, J., V. Hadhazi, H. Birch, L. Lao, G. Kaplan, and M. Hochberg. 2001. "Synthesis of Antimalarial Activity of Artemisinin Derivatives "Acupuncture for Osteoarthritis of the Knee: A Systematic Review." Containing an Amino Group." Journal of Medical Chemistry 43 (8): Arthritis and Rheumatism 44 (4): 819­825. 1635­40. Friedman, R., D. Sobel, P. Myers, M. Caudill, and H. Benson. 1995. Lindall, S. 1999. "Is Acupuncture for Pain Relief in General Practice Cost- "Behavioral Medicine, Clinical Health Psychology, and Cost Offset." Effective?" Acupuncture Medicine 17: 97­100. Health Psychology 14: 509­18. Linde, K., K. Jobst, and J. Panton. 2000."Acupuncture for Chronic Asthma." Gerber, G. S., D. Kuznetsov, B. C. Johnson, and J. D. Burstein. 2001. Cochrane Database Systematic Review CD000008 [PMID:11416076]. "Randomized, Double-Blind, Placebo-Controlled Trial of Saw Margolin, A., H. D. Kleber, S. K. Avants, J. Konefal, F. Gawin, E. Stark, and Palmetto in Men with Lower Urinary Tract Symptoms." Urology 58: others. 2002. "Acupuncture for the Treatment of Cocaine Addiction: A 860­64. Randomized Controlled Trial." Journal of the American Medical Hien, T. T., and C. Dolecek. 2004. "Piperaquine against Multidrug- Association 287: 55­63. Resistant Plasmodium falciparum Malaria." Lancet 363: 18­22. Markowitz, J. S., J. L. Donovan, C. L. DeVane, R. M. Taylor, Y. Ruan, J. S. Hellman, C. J., M. Budd, J. Borysenko, D. C. McClelland, and H. Benson. Wang, and K. D. Chavin. 2003. "Effect of St. John's Wort on Drug 1990. "The Study of the Effectiveness of Two Group Behavioral Metabolism by Induction of Cytochrome P450 3A4 Enzyme." Journal Medicine Interventions for Patients with Psychosomatic Complaints." of the American Medical Association 290: 1500­4. Behavioral Medicine 16: 165­73. Muela, S. H., A. K. Mushi, and J. M. Ribera. 2000. "The Paradox of the Cost Hoff, W. 1997. "Traditional Health Practitioners as Primary Health Care and Affordability of Traditional and Government Health Services in Workers." Tropical Doctor 27 (Suppl. 1): 52­55. Tanzania." Health Policy Planning 15 (3): 296­302. Holroyd, K. A., and D. B. Penzien. 1990. "Pharmacological versus Non- Mwabu, G. 1986. "Health Care Decisions at the Household Level: Results Pharmacological Prophylaxis of Recurrent Migraine Headache: A of a Rural Health Survey in Kenya." Social Science and Medicine 22 (3): Meta-Analytic Review of Clinical Trials." Pain 42: 1­13. 315­19. Hypericum Depression Trial Study Group. 2002. "Effect of Hypericum per- Mwabu, G., M. Ainsworth, and A. Nyamete. 1993. "Quality of Medical foratum (St. John's Wort) in Major Depressive Disorder: A Care and Choice of Medical Treatment in Kenya." Journal of Human Randomized Controlled Trial." Journal of the American Medical Resources 28 (4): 838­62. Association 287: 1807­14. Nations, M. K., and M. A. de Souza. 1997. "Umbanda Healers as Effective Jacobs, B. P., C. Dennehy, G. Ramirez, J. Sapp, and V. A. Lawrence. 2002. AIDS Educators: Case Control Study in Brazilian Urban Slums "Milk Thistle for the Treatment of Liver Disease: A Systematic (Favelas)." Tropical Doctor 27 (Suppl. 1): 60­66. Review and Meta-Analysis." American Journal of Medicine 113: Nations, M. K., M. A. de Sousa, L. L. Correia, and D. M. da Silva. 1988. 506­13. "Brazilian Popular Healers as Effective Promoters of Oral Rehydration Jain, A. 2003. "Does Homeopathy Reduce the Cost of Conventional Drug Therapy (ORT) and Related Child Survival Strategies." Bulletin of the Prescribing? A Study of Comparative Prescribing Costs in General Pan American Health Organization 22 (4): 335­54. Practice." Homeopathy 92 (2): 71­76. Ndubani, P., and B. Hojer. 1999. "Traditional Healers as a Source of Jepson, R. G., L. Mihaljevic, and J. Craig. 2000. "Cranberries for Preventing Information and Advice for People with Sexually Transmitted Diseases Urinary Tract Infections." Cochrane Database Systematic Review in Rural Zambia." Tropical Doctor 29 (1): 36­38. CD001321 [PMID:15106157]. Ni, H., C. Simile, and A. M. Hardy. 2002. "Utilization of Complementary Jonas, W. B., T. J. Kaptchuk, and K. Linda. 2003. "Critical Overview of and Alternative Medicine by United States Adults: Results from the Homeopathy." Annals of Internal Medicine 138: 393­99. 1999 National Health Interview Survey." Medical Care 40: 353­58. 1290 | Disease Control Priorities in Developing Countries | Haile T. Debas, Ramanan Laxminarayan, and Stephen E. Straus Niggemann, B., and C. Gruber. 2003. "Side Effects of Complementary and Speca, M., L. E. Carlson, E. Goodey, and M. Angen. 2000. "A Randomized, Alternative Medicine." Allergy 58: 707­16. Wait-List Controlled Clinical Trial: The Effect of a Mindfulness-Based Orme-Johnson, D. W., and R. E. Herron. 1997. "An Innovative Approach Stress Reduction Program on Mood and Symptoms of Stress in Cancer to Reducing Medical Care Utilization and Expenditures." American Patients." Psychosomatic Medicine 62: 2613­22. Journal of Managed Care 3 (1): 135­44. Spiegel, D. 1994. "Health Caring, Psychosocial Support for Patients with Pittler, M. H., K. Schmidt, and E. Ernst. 2003. "Hawthorn Extract for Cancer." Cancer 74 (4): 1453­56. Treating Chronic Heart Failure: A Meta-Analysis of Randomized Spiegel, D., J. R. Bloom, H. C. Kraemer, and E. Gottheil. 1989. "Effect of Trials." American Journal of Medicine 114: 665­74. Psychosocial Treatment on Survival of Patients with Metastatic Breast Reginster, J. Y., R. Deroisy, and L. Rovalty. 2001. "Long-Term Effects of Cancer." Lancet 2 (8668): 888­91. Glucosamine Sulphate on Osteoarthritis Progression: A Randomised, Straus, S. E. 2004. "Complementary and Alternative Medicine." In Cecil Placebo-Controlled Clinical Trial." Lancet 357: 251­56. Textbook of Medicine, 22nd ed., ed. L. Goldman and D. Ausiello, Philadelphia: Saunders. Richardson, M. A., and S. E. Straus. 2002. "Complementary and Alternative Medicine: Opportunities and Challenges for Cancer Swayne, J. 1992. "The Cost and Effectiveness of Homeopathy." British Management and Research." Seminars in Oncology 29: 531­45. Homeopathic Journal 81: 148­50. Richy, F., O. Bruyere, O. Ethgen, M. Cucherat, Y. Henrotin, and J. Y. Taylor, J. A., W. Weber, L. Standish, H. Quinn, J. Goesling, M. McGann, and Reginster. 2003. "Structural and Symptomatic Efficacy of Glucosamine C. Calabrese. 2003. "Efficacy and Safety of Echinacea in Treating Upper and Chondroitin in Knee Osteoarthritis: A Comprehensive Meta- Respiratory Tract Infections in Children: A Randomized Controlled Analysis." Archives of Internal Medicine 163: 1514­22. Trial." Journal of the American Medical Association 290: 2824­30. Schneider, R. H., F. Staggers, C. N. Alexander, W. Sheppard, M. Rainforth, Taylor, M. A., D. Reilly, R. H. Llewellyn-Jones, C. McSharry, and T. C. K. Kondwani, and others. 1995. "A Randomized Controlled Trial of Aitchison. 2000. "Randomised Controlled Trial of Homeopathy versus Stress Reduction for Hypertension in Older African Americans." Placebo in Perennial Allergic Rhinitis with Overview of Four Trial Hypertension 26: 820­27. Series." British Medical Journal 321: 471­76. Shekelle, P. G., M. L. Hardy, S. Morton, M. Maglione, W. A. Mojica, M. J. Thomas, K. J., J. Carr, L. Westlake, and B. T. William. 1991. "Use of Suttorp, and others. 2003. "Efficacy and Safety of Ephedra and Nonorthodox and Conventional Health Care in Great Britain." British Ephedrine for Weight Loss and Athletic Performance: A Meta- Medical Journal 302 (6770): 207­10. Analysis." Journal of the American Medical Association 289: 1537­45. Tukmachi, E., R. Jubb, E. Dempsy, and P. Jones. 2004. "The Effect of Shen, J., N. Wenger, J. Glaspy, R. D. Hays, P. S. Albert, C. Choi, and P. G. Acupuncture on Symptoms of Knee Osteoarthritis: An Open Shekelle. 2000. "Electroacupuncture for Control of Myeloablative Randomized Controlled Study." Acupuncture Medicine 22 (1): 14­22. Chemotherapy-Induced Emesis: A Randomized Controlled Trial." van Agtmael, M. A., T. A. Eggelte, and C. J. van Boxtel. 1999. "Artemisinin Journal of the American Medical Association 284: 2755­61. Drugs in the Treatment of Malaria: From Medicinal Herb to Registered Skargren, E. I., P. G. Carlsson, and B. E. Oberg. 1998. "One-Year Follow-Up Medication." Trends in Pharmacologic Science 20: 199­205. Comparison of the Cost and Effectiveness of Chiropractic and van Haselen, R. 2000. "The Economic Evaluation of Complementary Physiotherapy as Primary Management for Back Pain: Subgroup Medicine: A Staged Approach at the Royal London Homoeopathic Analysis, Recurrence, and Additional Health Care Utilization." Spine Hospital." British Homeopathic Journal 89 (Suppl. 1): S23­26. 23 (17): 1875­84. Vickers, A. J., and C. Smith. 2000. "Homeopathic Oscillococcinum for Skargren, E. I., B. E. Oberg, P. G. Carlsson, and M. Gade. 1997. "Cost and Preventing and Treating Influenza and Influenza-Like Symptoms." Effectiveness Analysis of Chiropractic and Physiotherapy Treatment Cochrane Database Systematic Review CD001957 [PMID:10796675]. for Low Back and Neck Pain: Six-Month Follow-Up." Spine 22 (18): White, A., P. Slade, C. Hunt, A. Hart, and E. Ernst. 2003. "Individualized 2167­77. Homeopathy as an Adjunct in the Treatment of Childhood Asthma: A Sobel, D. S. 1995. "Rethinking Medicine: Improving Health Outcomes Randomized Placebo Controlled Trial." Thorax 58: 317­21. with Cost-Effective Psychosocial Interventions." Psychosomatic WHO (World Health Organization). 2002. "Fact Sheet No. 271" (June). Medicine 57: 234­44. WHO, Geneva. Soeken, K. L. 2004. "Selected CAM Therapies for Arthritis-Related Pain: Winemiller, M. H., R. G. Billow, E. R. Laskowski, and W. S. Harmsen. 2003. The Evidence from Systematic Reviews." Clinical Journal of Pain 20 (1): "Effect of Magnetic versus Sham-Magnetic Insoles on Plantar Heel 13­18. Pain: A Randomized Controlled Trial." Journal of the American Medical Sommer, J. H., M. Burgi, and R. Theiss. 1999. "A Randomized Experiment Association 290: 1474­78. of the Effects of Including Alternative Medicine in the Mandatory Winston, C. M., and V. Patel. 1995. "Use of Traditional and Orthodox Benefit Package of Health Insurance Funds in Switzerland." Health Services in Urban Zimbabwe." International Journal of Complementary Therapeutic Medicine 7 (2): 54­61. Epidemiology 24 (5): 1006­12. Complementary and Alternative Medicine | 1291 Chapter 70 Improving the Quality of Care in Developing Countries John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and Julio Frenk Although the quantity rather than quality of health services has Increasing evidence, much of it developed since the mid been the focus historically in developing countries, ample evi- 1990s, shows that quality can be improved rapidly. However, to dence suggests that quality of care (or the lack of it) must be at improve clinical practice--and thus quality of care--quality the center of every discussion about better health. The follow- must be defined and measured, and appropriate steps must be ing examples are illustrative: In one study evaluating pediatric taken (Silimper and others 2002). This chapter highlights care in Papua New Guinea, 69 percent of health center workers approaches to improving clinical practice and quality of care reported that they checked for only two of the four examina- that take place over months instead of years. Indeed, better tion criteria for pneumonia cases. Only 24 percent of these quality can improve health much more rapidly than can other workers were able to indicate correct treatment for malaria. drivers of health, such as economic growth, educational When clinical encounters were observed at aid posts, providers advancement, or new technology. met minimal examination criteria in only 1 percent of cases (Beracochea and others 1995). In a study in Pakistan, only 56 percent of providers met an acceptable diagnostic standard Definition and Framework for viral diarrhea, and only 35 percent met the acceptable stan- Health systems provide health actions--activities to improve dard for treatment (Thaver and others 1998). or maintain health. These actions take place in the context of and are influenced by political, cultural, social, and institu- tional factors (shown along the edges of figure 70.1). QUALITY DEFINITION AND POPULATION Demographic and socioeconomic makeup, including genetics FRAMEWORK and personal resources, affect the health status of individuals seeking care. Access to the health care system is required to These deficiencies in quality of care represent neither the obtain the care that maintains or improves health, but simple failure of professional compassion nor necessarily a lack of access is not enough; the system's capacities must be applied resources (Institute of Medicine 2001). Rather, they result from skillfully. Thus, quality means optimizing material inputs gaps in knowledge, inappropriate applications of available and practitioner skill to produce health. As the Institute of technology (Murray and Frenk 2000), or the inability of organ- Medicine defines it, quality is "the degree to which health izations to change (Berwick 1989). Local health care systems services for individuals and populations increase the likeli- may have failed to align practitioner incentives and objectives, hood of desired health outcomes and are consistent with to measure clinical practice, or to link quality improvement to current professional knowledge" (Institute of Medicine 2001, better health outcomes. 244). 1293 Political Institutional Factors Factors Health Policy Reforms Demographic Health Health and Socio- Care Structure Process Outcomes economic Access Factors The Quality of Care Cultural Social Factors Factors Source: Peabody and others 1999. Figure 70.1 Quality-of-Care Framework Elements of Quality. Quality comprises three elements: better health outcomes (Javitt, Venkataswamy, and Sommer 1983). At best, however, structure is a blunt approximation of · Structure refers to stable, material characteristics (infra- process or outcomes; structural improvements by themselves structure, tools, technology) and the resources of the organ- rarely improve the health of a population. izations that provide care and the financing of care (levels of Process, by contrast, can be measured with every visit to a funding, staffing, payment schemes, incentives). provider. Measuring process is difficult, however, particularly · Process is the interaction between caregivers and patients in developing countries. The private nature of the doctor- during which structural inputs from the health care system patient consultation, a lack of measurement criteria, and the are transformed into health outcomes. absence of reliable measurement tools have limited the ability · Outcomes can be measured in terms of health status, to assess process (Peabody, Tozija, and others 2004). However, deaths, or disability-adjusted life years--a measure that new methods are being developed that can provide valid meas- encompasses the morbidity and mortality of patients or urements of clinical practice (Thaver and others 1998). In addi- groups of patients. Outcomes also include patient satisfac- tion, evidence-based clinical studies have steadily revealed tion or patient responsiveness to the health care system which process measures lead to better health outcomes. This (WHO 2000). combination of ubiquity, measurability, and linkage to health outcomes makes the measurement of process the preferred way Structural measures are the easiest to obtain and most com- to assess quality. monly used in studies of quality in developing countries. Many Although good outcomes are the objective of all health evaluations have revealed shortages in medical staff, medica- actions, outcomes alone are not an efficient way to measure tions and other important supplies, and facilities, but material quality for two reasons. The first is the quality conundrum. A measures of structure, perhaps surprisingly, are not causally patient may receive poor-quality care but may recover fully, or related to better health outcomes (Donabedian 1980). Although a patient may receive high-quality care for an illness such as higher technology or a more pleasant environment may be con- cerebral malaria and still not recover. Second, adverse health ducive to better-quality care, the evidence indicates only a weak outcomes are relatively rare and obviously do not occur with link between such structural elements and better health out- every encounter. comes (Donabedian 1988). The notable exceptions are cases in The classic framework of structure-process-outcome is well which physical improvements either increase access to primary established. However, in recent years the concept of quality has care in very poor settings or increase the volume of a clinical been expanded to include specific aims for improvement. For procedure, such as cataract surgery, that is specifically linked to example, the Institute of Medicine's (2001) landmark report, 1294 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others Box 70.1 The Institute of Medicine's Six Elements of Quality 1. Patient safety. Are the risks of injury minimal for 4. Timeliness. Are delays and waiting times minimized? patients in the health system? 5. Efficiency. Is waste of equipment, supplies, ideas, and 2. Effectiveness. Is the care provided scientifically sound energy minimized? and neither underused nor overused? 6. Equity. Is care consistent across gender, ethnic, geo- 3. Patient centeredness. Is patient care being provided in a graphic, and socioeconomic lines? way that is respectful and responsive to a patient's pref- erences, needs, and values? Are patient values guiding clinical decisions? Source: Institute of Medicine 2001. Crossing the Quality Chasm, broadens the concept to include had not recently received care was higher (Peabody, Tozija, and other, more contextual elements to illuminate how process others 2004). changes can improve care. It focuses on six aims: patient safety, Our quality-of-care framework supports these findings. effectiveness, patient centeredness, timeliness, efficiency, and When process is improved among groups of providers,the aggre- equity (see box 70.1). gate improvement in quality leads to better health outcomes for the entire patient population. In addition, resources can be allo- Quality Assessment Perspectives. We can look at the Institute cated among clinical interventions based on actual effectiveness of Medicine's aims from two perspectives: patient perception, and the overall impact of care on the population. For example, and technical or professional assessment. Patients' perceptions cancer chemotherapy may be available and may prolong the lives of quality depend on their individual characteristics and affect of cancer patients.However,it may result in fewer lives saved than their compliance, follow-up decisions, and long-term lifestyle the expansion of coverage of directly observed treatment short- changes (Zaslavsky and others 2000). Interpersonal relation- course coverage for tuberculosis patients. ships, cultural appropriateness, and gender sensitivity--long thought to be luxuries of wealthier countries--are also major QUALITY OF CARE IN DEVELOPING COUNTRIES determinants of patient access and utilization in developing countries. These findings have led to the inclusion of patient The process of providing care in developing countries is often satisfaction and patient responsiveness as outcome measures. poor and varies widely. A large body of evidence from indus- Technical assessment concerns whether providers meet nor- trial countries consistently shows variations in process, and mative standards for appropriateness of care or adherence to these findings have transformed how quality of care is per- explicit evidence-based criteria. Although patient perception or ceived (McGlynn and others 2003). A 2002 study found that satisfaction is important, researchers increasingly rely on objec- physicians complied with evidence-based guidelines for at least tive, evidence-based quality criteria that can be more readily 80 percent of patients in only 8 of 306 U.S. hospital regions linked to better health outcomes at both the individual and the (Wennberg, Fisher, and Skinner 2002). It is important to note population levels. that these variations appear to be independent of access to care or cost of care: Neither greater supply nor higher spending Population-Level Considerations. Quality is typically resulted in better care or better survival. Studies from develop- assessed through the interaction between individual doctors ing countries show similar results. For example, care in tertiary and patients. However, emerging evidence shows that the and teaching hospitals and care provided by specialists may be average quality of care given by groups of doctors and other better than care for the same cases in primary care facilities and providers is an important determinant of overall community by generalists (Walker, Ashley, and Hayes 1988). health status. For example, in a cross-sectional analysis in the One explanation for variation and low-quality care in the former Yugoslav Republic of Macedonia, researchers found not developing world is lack of resources. Limited data indicate, only that patients' heath status was significantly higher in areas however, that high-quality care can be provided even in environ- where quality was higher but also that the overall self-reported ments with severely constrained resources. A study in Jamaica, health status of those members of the general population who which used a cross-sectional analysis of government-run Improving the Quality of Care in Developing Countries | 1295 primary care clinics, showed that better process alone was linked by changing the socioeconomic, legal and administrative, to significantly greater birthweight (Peabody, Gertler, and cultural, and information context of the health care system. Liebowitz 1998). A study in Indonesia attributed 60 percent of all perinatal deaths to poor process and only 37 percent to eco- Legal Mandates, Accreditation, and Administrative Regula- nomic constraints (Supratikto and others 2002). tions. Legal mandates, accreditation, and administrative regu- Cross-system or cross-national comparisons provide the lations affect quality by controlling entry into the practice of best examples of the great variation in clinical practice in devel- health care. These policies include the licensing of profession- oping countries. In one seven-country study, researchers als and facilities, their accreditation or certification to perform directly observing clinical practice found that 75 percent of certain procedures, and the formal delineation of functions cases were not adequately diagnosed, treated, or monitored and that various types of health workers can legally perform. that inappropriate treatment with antibiotics, fluids, feeding, Although these policies assume that providers' prior qualifica- or oxygen occurred in 61 percent of cases (Nolan and others tions are good predictors of actual performance in health care 2001). Another study compared providers' knowledge and delivery, there is little evidence that such policies have a positive practice in California and FYR Macedonia, using vignettes to effect on process or outcomes. They are more successful at bar- adjust for case-mix severity. Although the quality of the overall ring unqualified persons from practicing than at ensuring or aggregate process was lower in FYR Macedonia, a poor quality among those who are allowed to practice. A review of country, the top 5 percent of Macedonian doctors performed as health sector regulations in Tanzania and Zimbabwe, for exam- well as or better than the average Californian doctor (Peabody, ple, revealed that the regulations primarily control entry into Tozija, and others 2004). the market and ensure a minimum standard of quality In a study commissioned for this chapter, an international (Kumaranayake and others 2000). team measured quality in five developing countries (China, Hospital accreditation, with its periodic reviews of health El Salvador, India, Mexico, and the Philippines), using the facility performance standards, can potentially provide ongo- same clinical vignettes at each site. The team evaluated the ing regulatory pressure for improvement. To date, research has process for common diseases according to international, not demonstrated that hospital accreditation programs are evidence-based criteria. Quality varied only slightly among linked to improvements in health outcomes. In a randomized countries. The within-country range of quality of doctors was controlled trial of a hospital accreditation program in the 10 times as great as the between-country range. Such wide KwaZulu-Natal province of South Africa, researchers showed a variation strongly suggests that efforts to improve health conclusive link between the implementation of the program status must involve policies that change the quality of clinical and improvements in the accreditation standard indicators. care. However, they were unable to link those indicators to improve- ments in health outcomes (Salmon and others 2003). POLICY INTERVENTIONS TO IMPROVE QUALITY Malpractice Litigation to Enforce Legal Mandates To be The success of quality improvement policies can be measured effective in promoting quality, malpractice litigation must rely by their ability to raise the average level of health and reduce on adequate legal and judicial systems, which are deficient in variation in quality. Two types of policies are intended to most developing countries. In India, one of the few developing improve quality and thus health outcomes: countries with the appropriate legal structure in place, inclu- sion of the medical sector under the Consumer Protection Act · those that influence provider behavior by altering the struc- of 1986 allows victims to receive redress for negligent medical tural conditions of organization and finance or that involve practice. Although improvements have resulted, some argue the design and redesign of health care systems that the system needs greater involvement of professional · those that directly target provider behavior at the individual organizations to be effective (Bhat 1996). or the group level. Professional Oversight Peer review is as old as professional Within each category, the evidence is examined to see the societies. The power and the influence of such societies vary effect of the policy on the health outcomes of populations. widely among countries (Heaton 2000). Large provider organ- izations, such as hospitals or public health institutions, often Interventions Affecting Provider Practice by Changing routinely collect information on provider practices and patient Structural Conditions outcomes and use those data to guide, educate, supervise, dis- Although structural components such as materials and staff cipline, or recognize providers. In the Philippines, public health are not strongly linked to outcomes, other components of managers used a checklist of 20 observable behaviors against structure--organization and finance--can influence process which health workers in remote provinces were rated. The 1296 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others performance of providers in facilities where workers were which promotes, facilitates, and evaluates the Essential reviewed was significantly better than in comparable facilities National Health Research strategy in such countries as Benin, that did not adopt the reviews (Loevinsohn, Guerrero, and the Arab Republic of Egypt, and Indonesia. COHRED aims to Gregorio 1995). Others, however, assert that the "quality by develop a system of effective health research to improve health inspection" environment engendered by oversight leads to an services, including quality of care. The Quality Assurance proj- antagonistic relationship between workers and managers and ect funded by the U.S. Agency for International Development precludes cooperative problem solving and continuous has studied and shared information about quality in the devel- improvement (Berwick 1989). A qualitative study evaluating oping world since 1990. Under the Quality Assurance project supervisor-provider interactions in health care facilities in umbrella, researchers have studied and implemented quality Zimbabwe found that supervisors were adept at giving techni- measurement and improvement interventions and have used cal feedback but were not as proficient at making suggestions these case studies to develop a library of tools and articles to for improvement or at working with providers and patients to promote global quality improvement. solve problems (Tavrow, Kim, and Malianga 2002). Public-Private Provision of Care. In most health care systems, National and Local Clinical Guidelines In many industrial a professional regulatory framework governs the network of countries, evidence-based clinical guidelines are used to ensure civil servants delivering health care. These civil servants operate high-quality care, better health outcomes, and cost-effective alongside autonomous, self-governed, private providers-- treatments. (Examples of institutions supporting this approach independent for-profit physicians and health clinics and non- are the U.K. National Institute for Clinical Excellence, the U.S. profit nongovernmental organizations (NGOs). Two conclu- Agency for Healthcare Research and Quality, and the Dutch sions arise from the often heated debate about the right balance College of General Practitioners.) Guidelines are typically between public and private services. First, private practitioners developed for a clinical disease or symptom. They should be provide a significant amount of care in developing countries. derived from evidence-based criteria resulting from well- Second, though there is no one prescription for striking the designed clinical investigations or expert opinion. Because they right public-private mix, in some cases the public regulatory are derived from empirical studies, guidelines in developing framework has led to private provision of higher-quality care. countries can, in principle, be identical to those in industrial The government of Senegal successfully contracted with countries. When resource constraints limit transferability, diag- community-based groups for preventive nutrition services. nostic and treatment guidelines may have to be modified. Eighteen months after nutrition services were implemented, Technologies such as x-ray studies have gained widest accept- severe malnutrition disappeared among children age 6 to ance in preventive and primary care services, such as integrated 11 months (Marek and others 1999). The success of the pro- management of childhood illness, where they serve both as gram has led to its expansion nationwide. clinical standards and as educational guides. Including physi- cians in the development and review of guidelines has proved Targeted Education and Professional Retraining. Continuing particularly effective in the challenging process of implement- medical education is a common approach to improving clini- ing guidelines. cal practice, but it neither changes clinical practice nor advances health outcomes (Davis and others 1995). Newer Sharing Information on Quality Improvement Technology. techniques--targeted education, case-based learning, and Worldwide interest in quality has given rise to new professional interactive and multimodel teaching techniques--have had bodies, scientific publications, and institutions dedicated to some success. In Guatemala, distance education targeting diar- sharing ideas and innovations in quality improvement. rhea and cholera case management increased accurate assess- Organizations such as the Robert Wood Johnson Foundation, ment and classification of diarrhea cases by 25 percent. the Nuffield Trust, and the Institute for Healthcare Rehydration did not improve, however, and improvements in Improvement cultivate ideas for improvement, bring people counseling were insignificant (Flores, Robles, and Burkhalter and organizations together to learn from each other, and take 2002). In Tanzania, training staff in the control of acute respi- action to achieve results. Although the sharing of information ratory infections of young children yielded reductions in on quality health care practices has long been an established under-five mortality within two years (Mtango and Neuvians part of provider education and training networks, the sharing 1986). of information on successful systemwide policies for process improvements could potentially accelerate the scale-up of Organizational Change. In recent years, organizational quality practice. change in the health care system has been shown to influence One organization active in developing countries is the quality of care and to further the six aims of the Institute of Council on Health Research for Development (COHRED), Medicine by focusing on the continual design and redesign of Improving the Quality of Care in Developing Countries | 1297 systems. The emphasis is on developing organizational and ment. Team members using the PDSA model design a individual capabilities where they most profoundly affect the quality-improvement intervention (plan), implement it on process of care. Design and redesign interventions assume that a small scale (do), evaluate the results (study), and imple- simply adding a new resource or a new process in isolation will ment or alter the intervention accordingly (act). Often mul- not improve care because better care is the product of many tiple PDSA cycles are necessary before the appropriate processes working together. Although change interventions improvement method can be identified. All improvement have not been widely used in the developing world because techniques that involve the design and redesign of systems they require large investments to plan and implement, four use some form of the PDSA cycle. Successful scale-up of a related models of organizational change have been successful in PDSA prototype is possible with careful leadership over- changing provider practice in developing nations: sight. A team of investigators in Russia's Tula province developed a series of successful interventions for adults who · Total Quality Management in health care Advances in have poorly controlled hypertension. The interventions, business management practices to continually design and which were started in 20 clinics, were expanded to 500 clin- redesign systems for quality improvement have been effec- ics within 18 months. The scale-up resulted in a sevenfold tively adapted for health systems. In Total Quality Manage- increase in patients receiving hypertension management at ment, also known as Continuous Quality Improvement, the primary care level and an 85 percent reduction in admis- teams use mutually reinforcing techniques in a cycle of plan- sions for hypertension. In Tver province, the same group ning, implementing, evaluating, and revising to improve the addressed problems related to prenatal care. They began quality of clinical and administrative processes. These tech- with 5 hospitals and scaled up to cover all 42 hospitals and niques include process mapping, statistical quality control, all maternity clinics in the province. The result was a 99 per- and structured team activities. In rural Bihar, India, private cent reduction in newborns with hypothermia and a reduc- practitioners who treat sick children were provided with tion in pregnancy-induced hypertension from 44 percent standard case-management information, were given feed- to 6 percent (Berwick 2004). Although the experience of back on their performance, and were tracked and monitored researchers implementing interventions that are based on over time. This strategy produced significant improvements system redesign in the developing world has been largely in practitioners' case-management skills (Chakraborty, positive, it is not clear whether the resources and leadership D'Souza, and Northrup 2000). In Malaysia, anesthesia safety exist to bring these interventions to scale through country has been improved through the implementation of or regional policies. Further evidence is needed concerning consensus-based protocols that emphasize (a) communica- the real-world feasibility and cost-effectiveness of system tion among the operating, recovery, and ward team mem- redesign. bers; (b) individual feedback; and (c) frequent monitoring · Internal enabling environment Creating the right environ- to identify areas for improvement (Tan 1999). ment for change involves leadership and leadership training, · Collaborative Improvement Model The early success of Total clinicians empowered to make quality improvement deci- Quality Management techniques has given rise to a related sions, and resources for quality improvement planning model, the Collaborative Improvement Model. It addresses activities (Silimper and others 2002). The internal enabling broad and complex systemic processes within health care environment in Costa Rica promoted strong leadership that systems and has facilitated the scale-up of quality improve- led to the adoption of structural adjustment loans in the ments. This model, designed to continuously improve organ- early stages of health sector reforms. The loans were used to izational and individual performance, comprises four ele- maintain such public health programs as mother and child ments: definition of an aim, measurement, innovation, and nutrition, even though public spending dropped and prices testing to see whether the innovation meets the original aim. increased dramatically (Peabody 1996). An enabling envi- This approach strikes a pragmatic balance between the need ronment can also be created by teams of individuals, each for action and the need to be scientifically grounded. It has representing different stakeholder groups (physicians, been used with success in Peru and the Russian Federation. In nurses, staff members, patients, and so forth) or simply by a Peru,the collaborative improvement model was used by mul- strong leader with an interest in teamwork and the resources tidisciplinary teams in 41 clinics to design changes aimed at to support a discrete quality improvement function for team achieving world-class tuberculosis care. The preliminary members. results have led to impressive changes in the process of care, but it is too early to determine whether they have been effec- tive in improving quality (Berwick 2004). Interventions Directly Affecting Provider Practice · Plan-Do-Study-Act cycle The Plan-Do-Study-Act (PDSA) Practitioners are often forced to provide care in uncertain set- cycle calls for action-oriented learning in quality improve- tings. Technical limitations may reduce the ability to diagnose 1298 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others or predict outcomes, or they may have only probabilistic contractor's obligation, the methods of monitoring perform- knowledge about the efficacy of their proposed treatment for a ance, and the sanctions for nonperformance were only mini- particular patient. The nature of clinical practice is often soli- mally specified (Broomberg, Masobe, and Mills 1997). tary, and physicians have few available ways to gauge their clin- ical acumen and skills. Performance-based feedback, however, High Volume of Care. Evidence exists that a high volume of can reward high-quality care and increase knowledge about care by individuals or institutions leads to better health out- appropriate actions. If the feedback mechanism is effective, it comes (Habib and others 2004). Physician experience (learning) can also serve as the basis for establishing systemwide incen- and practice (repetition) lead to fewer complications, less tives for improving quality of care. resource use, and better quality for a variety of procedures, such as cataract surgery and laparoscopy (Brian and Taylor 2001). Training with Peer Review Feedback. In Mexico City, physi- More complex procedures, including endarterectomy, cancer cian retraining on treatment of diarrhea, combined with the surgery,and coronary bypass surgery,have shown similar effects. concurrent creation of a peer-review structure, decreased the Volume effects leading to better health outcomes are not use of antibiotics and increased the use of oral rehydration confined to surgical procedures (Zgibor and Orchard 2004). therapy (ORT). These improvements continued to be seen in Facilities specializing in the care of chronic diseases such as dia- a follow-up evaluation 18 months later (Gutierrez and others betes, myocardial infarction, and heart failure are also associ- 1994). The approach has been effectively expanded to pre- ated with better outcomes. Debate exists over how much of the scribing practices for rhinopharyngitis among primary care volume effect is due to specialist care. The benefits of high- physicians, using an interactive training workshop and a man- volume care persist, however, even after controlling for referral agerial peer-review committee (Perez-Cuevas and others and case-mix biases. When carefully trained nonphysicians are 1996). substituted for physicians, volume effects persist but can be accomplished at significantly lower costs. In one study, nurse Performance-Based Remuneration. A potentially powerful practitioners and physician's assistants were able to provide instrument for accelerating quality improvements involves high-quality care for common outpatient conditions such as making payments directly to providers who meet quality stan- hypertension, diabetes, asthma, otitis media, pharyngitis, and dards that are based on process indicators associated with back pain at substantially lower costs than that of physicians favorable patient outcomes. Systems that tie performance to (Douglas and others 2004). remuneration use relatively small incentives--equivalent to 3 to 10 percent of the provider's total compensation. Performance-Based Professional Recognition. Providers Performance-based remuneration has been successfully used work in a community of peers in which professional status, in the United States to compensate both private and public prestige, and recognition are often as valuable as material providers (McBride, Neiman, and Johnson 2000). rewards. Nonmonetary incentives, such as public recognition Examples of performance-based incentives come from or disclosure, administrative privileges, and awards from pro- developing countries too. The Nicaraguan Ministry of Health fessional organizations, can promote improvements in quality. has implemented a pilot program in six hospitals that offers an Uganda, for example, implemented the Yellow Star Program as incentive bonus (a maximum average of 17 percent of hospital part of a broader health services improvement project. This revenue) for facilities that achieve performance targets that program evaluated health facilities on a quarterly basis, using include quality measures (Jack 2003). In Haiti, a performance- 35 indicators of technical and interpersonal quality, and based payment scheme was set up for NGOs that provided awarded a large yellow star to facilities that scored 100 percent services to the population. The scheme resulted in all three par- in two consecutive quarters. The star was then prominently dis- ticipating NGOs reaching target immunization coverage rates played outside the facility. (Eichler, Auxila, and Pollock 2001). Thus, payment for specified The Mexican Ministry of Health has implemented a strategy and observable performance (in terms of provider effort, client that combines the accreditation and the training strategies coverage, or health impact on the population) can be usefully discussed earlier with nonmonetary incentives. The National applied to NGOs and private providers. Crusade for Quality in Health Care introduces quality-oriented The specific features of performance-based remuneration incentives to health facilities and medical schools. It also are crucial. A study evaluating the South African government's includes public recognition in an effort to encourage learning experience in contracting with private organizations to operate and to change practice. The National Crusade has already gen- district hospitals found no cost savings--in fact, the govern- erated measurable improvements in the responsiveness of ment was spending more than if it provided the services itself. state-level health systems (Secretaría de Salud de Mexico 2003). The contracting may have failed because remuneration was not Both types of policies examined in this section are associ- based on specific process or outcome measures. Instead, the ated with better quality and better health outcomes--lower Improving the Quality of Care in Developing Countries | 1299 premature mortality and avoidable morbidity, increased or obtaining payments, for example) and thus lack crucial patient satisfaction, and more health-seeking behaviors. When clinical details. One prospective study showed that charts iden- effective, these policies result in increased coverage rates, better tified only 70 percent of items performed during the clinical prescribing patterns, and increased adherence to clinical guide- encounter (Luck and others 2000). In a related analysis, 6.4 per- lines. They can spell the difference between an individual's cent of the items recorded in the chart were false and had never survival or death, between an individual benefiting from the really occurred. encounter with the health sector or being harmed by it, and Where resources and infrastructure are sufficient, the elec- between an individual and society rising from poverty or sink- tronic medical record (EMR) is becoming a priority for health ing deeper into it. systems worldwide. EMR technology promotes uniformity, leg- ibility, and communication, which can lead to guideline use MEASURING QUALITY and reduce prescription errors. It also holds the promise of managing populations rather than individuals by aggregating Improving quality requires that we measure it accurately. The patients into groups. However, the EMR has not always lived up successful outcomes discussed in the previous section rely on the to its potential. In many countries, some impressive successes links between policy and changes in clinical practice. Such links, have occurred--as have spectacular failures, costing billions of however, can be created and demonstrated only when valid and dollars (McConnell 2004). The great heterogeneity in record- reliable measures of process are easily understood, inexpensive keeping practices, problems with medical records (both paper to obtain, resistant to manipulation, and related to better health and electronic), and costs of trained medical abstractors have outcomes. led to a search for other reliable ways to measure quality. Measuring Structure Direct Observation and Recording of Visits. Direct observa- Material measures of structure abound. Numerous facility- tion and recording of visits is a commonly used approach in based surveys in developing countries have cataloged capital developing countries (Nolan and others 2001). Ethically, the equipment and staffing levels, and financial reports track budg- provider and the patient must be informed of the observation ets and expenditures (but rarely production costs). Facility or recording, which introduces participation bias because inventories of drugs and supplies are generally available; service provider behavior may change as a result of being evaluated. In utilization figures are routinely reported to national-level addition, trained observers are costly, and variation between authorities. Such measurements, however, are often beside the observers is difficult to remedy. point. Even when material structural deficiencies are corrected, they are not reliably linked to changes in health outcomes. Administrative Data. Administrative data, collected for pur- Measuring the organization and financing of health care is more poses of managing the delivery of care, are available in all but difficult. Although descriptions of the organization and financ- the poorest settings. A data collection system, once established, ing of health systems exist, objective functional assessments of is ubiquitous and can provide information on charges and systems (such as patient flows, the patient referral system, or many cost inputs. Administrative data, however, lack sufficient details of the relative pricing of services) are less often available. clinical detail to be useful in evaluating process. In a 2003 study, an incorrect diagnosis was recorded in the data 30 percent of Measuring Process the time (although the diagnosis was made correctly). Overall, Technical advances have mitigated longstanding difficulties in these data reflected the actual clinical diagnosis only 57 percent measuring process. Five approaches and their strengths and of the time (Peabody, Luck, Jain, and others 2004). As informa- weaknesses merit consideration: chart abstraction, direct tion systems advance, accuracy problems may be mitigated, observation and recording of visits, administrative data, stan- although the lack of adequate clinical detail will continue to dardized patients, and clinical vignettes. limit the use of administrative data. Chart Abstraction. Chart abstraction, or review of the med- Standardized Patients. Standardized patients can be a gold ical record, has long been used to measure technical quality. standard for process measurement (Luck and Peabody 2002). Such familiar quality evaluations as clinical audits, physician Trained to simulate illness, standardized patients present them- report cards, and profiles are based on chart abstraction. The selves unannounced into a clinical setting to providers who core strength of the medical record is that it is ubiquitous and have previously given their consent to participate in the study. can generally be obtained after each encounter. Chart reviews, At the conclusion of the visit, the standardized patient reports however, suffer from problems of legibility when notes are on the technical and interpersonal elements of process. handwritten. Often they are generated for reasons other than Standardized patients are reliable over a range of conditions recording the actual events of the clinical visit (legal protection and provide valid measurements that accurately capture 1300 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others variation in clinical practice among providers over time. physiological assessments (such as blood pressure), and mone- However, they are expensive and useful only for adult condi- tarily by measuring income. Other things being equal, a healthy tions and only those conditions that can be simulated. Thus, individual generates more income than one who is often sick. they are not practical for routinely evaluating quality. This benefit goes beyond the period of illness. Research on early childhood development has shown that higher-quality prenatal Clinical Vignettes. Clinical vignettes were developed explicitly and postnatal care not only decreases mortality but also for measuring quality within a group of providers and improves subsequent school performance, which is critical to evaluating quality at the population level. Vignettes are respon- future labor productivity (Van der Gaag 2000). The monetary sive to variation in quality, and providers readily accept them if benefits of better individual health can be assessed by examining they are given anonymously (Peabody, Luck, Glassman, and the individuals' expected income in the context of a life cycle others 2004). More than 20 vignettes have been used in 13 model. Expected income depends on the risk of death at various countries around the world. They can be administered on points in time and the corresponding opportunities for educa- paper, by computer, or over the Internet. Providers are typically tional attainment. This scenario can be simulated by improving presented with several cases. When process is being measured quality and then estimating how much the higher quality lowers for many providers, each provider is presented with the same mortality and increases education attainment, both of which case or set of cases, thus eliminating the need for case-mix increase an individual's future income (see figure 70.2). adjustment. The provider completing the vignette is asked to take a history, do an examination, order the necessary tests, make a diagnosis, and specify a treatment plan. The questions Social Macroeconomic Benefits are open ended and include interactive responses that simulate Societies that have healthier populations also have higher levels the visit and evaluate the physician's knowledge. In two sepa- of human capital and a greater capacity to generate wealth. rate, prospective validation studies among randomly selected Higher quality of care for the individual increases society's providers, vignettes consistently demonstrated greater predic- human capital by reducing both the number of premature tive validity of process than did the abstracted medical record. deaths (thus increasing the labor force) and the amount of tem- Vignettes have been validated against the gold standard of porary or permanent disability (thus improving worker pro- standardized patient visits, and they reflect actual clinical prac- ductivity). Providers and insurers also benefit from lower costs tice, not just physicians' knowledge. Vignettes have several by avoiding unnecessary or inappropriate care. Thus, society other advantages. Because exactly the same case can be given to benefits from both better health and lower public expenditures many providers, vignettes are useful for comparison studies. for treatment, which can then be reallocated to other productive They are also useful for pre- and postevaluations of policy uses. Interventions that improve quality have an especially high interventions designed to improve quality. Finally, they are social value when they have large positive externalities (for inexpensive to administer and straightforward to score, making instance, when better process reduces the incidence of a com- them particularly useful in developing countries. municable disease). Sometimes, however, society benefits but some stakeholders do not. For example, physicians who provide better preventive care may experience less demand for their cur- ECONOMIC BENEFITS AND COSTS ative services and associated resources. OF QUALITY CARE Several attempts have been made to estimate the correlation between health outcomes and long-term economic growth. Policy interventions can lead to higher-quality process of care The high prevalence of such diseases as malaria has been linked and can rapidly improve a population's health outcomes, but is in some studies to a slowing of economic growth by one to two quality improvement cost-effective? This section shows that it percentage points per year. These studies were severely limited is. We compare the economic benefits of better quality of care by the number of countries and by the many unobserved fac- at the individual and population levels with the costs of imple- tors excluded from the models (Sachs 2001). These limitations menting quality improvement interventions. We then discuss suggest another way to estimate the benefits of higher quality why these interventions not only increase individual and social on health outcomes and long-term economic growth. Because welfare but also are cost-effective in the long run. diagnostic accuracy and treatment of malaria can be improved with better-quality care, improving quality should increase national income through reductions in mortality rates. Individual Economic Benefits Indeed, cross-country data suggest that a one-year increase Individuals benefit from better quality of care because they are in life expectancy is associated with an increase in the gross physically, emotionally, and mentally healthier. These benefits domestic product (GDP) growth rate of 1 to 4 percentage points can be quantified subjectively by self-report, objectively by (Bloom, Canning, and Sevilla 2001). Our own simulations show Improving the Quality of Care in Developing Countries | 1301 Increase in present value expected income (percent) 120 100 80 60 40 9 7 20 5 Effect of 0 nine policy 3 3 intervention 8 scenarios on 13 1 18 quality 23 Unemployment rate (percent) Source: Authors' calculations. Note: These results model the effect on income of a policy intervention that leads to higher quality. The effect is determined as an increase in the present value of income for varying rates of unemployment. Higher quality is based on nine different scenarios of quality improvement. For each successive scenario, infant mortality rates are reduced by 1 percent and educational attainment is increased by 5 percent. The baseline possibilities are for the Islamic Republic of Iran. Figure 70.2 Economic Benefits of a Quality Intervention That Reduces Child Mortality Rates and Leads to Higher Educational Attainment Accumulated gains in GDP during a 50-year period relative to current GDP 0.20 0.18 0.16 0.14 0.12 0.10 0.08 5.0 0.06 4.0 Yearly 0.04 3.0 decline 0.02 in child 2.0 0 mortality rates 0.01 0.015 0.02 1.0 (percent) 0.025 0.03 0.035 0.04 Change in GDP growth rate from one year increase in life expectancy Source: Authors' calculations. Figure 70.3 Gains in GDP Resulting from Reductions in Child Mortality Rates that quality improvements can result in as much as a 5 percent Economic Costs annual reduction in child mortality rates, which can generate, Policies that improve the quality of care have both direct and over 50 years, economic gains equivalent to 18 percent of cur- indirect costs. Direct costs relate to the human and physical rent GDP (see figure 70.3). Similar results would be obtained if resources needed to implement the intervention. Indirect costs the effect of better quality on morbidity and disability were come from more subtle changes, including alterations in the simulated. quantity of health services provided, in provider demand for 1302 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others various inputs (such as equipment and medication), in the impact levels were considered, which were based on two previ- market prices of heath care, in government health budgets, and ous studies (Chakraborty, D'Souza, and Northrup 2000; ultimately in the macroeconomy. For interventions at the local Mtango and Neuvians 1986). level, such as training doctors in a particular region, it is usually The analysis showed that, under average conditions, sufficient to measure direct costs. Although the level of improving quality of care for conditions of acute respiratory detail required can be overwhelming when the interventions illness can be very cost-effective. When the baseline quality is are complex, the calculations are usually straightforward. The low and the disease prevalence is high, an intervention that costs of local interventions depend on local prices of such raises quality has a cost-effectiveness ratio of US$132 to inputs as labor, transportation, training kits, food, space rental, US$800 per life saved; if the policy intervention is ineffective or and accommodations. The cost of training providers in the the prevalence of pneumonia is low, the average cost of saving appropriate treatment of childhood illnesses ranges from a a life could be more than US$2,000. When 60 percent of cases low of US$1 to a high of US$430 (Santoso, Suryawati, and are already appropriately diagnosed and treated, the cost- Prawaitasari 1996). effectiveness ratio rises to US$5,000 per life saved.1 The direct and indirect costs of interventions at the central or local government level are harder to quantify. Expanding Better Treatment of Diarrhea. Diarrhea remains one of the training programs to all public providers, enforcing standards leading causes of childhood morbidity and mortality in the for private and public providers, changing payment sys- developing world. The diarrhea incidence rate among children tems, and developing policies to protect consumers against in resource-constrained countries can reach six to seven malpractice are macro-level interventions that have direct episodes per year (Thapar and Sanderson 2004). ORT is the program-level costs. They affect the economy as a whole by accepted standard of care for acute diarrhea. Unfortunately, a changing the allocation of public resources and the relative large proportion of cases are still treated with nonrehydration prices of goods and services. Macroevaluations of health medication, including antibiotics and antidiarrheals. Improved policy interventions are seldom conducted, even though sys- diagnosis of dehydration and reduced use of unnecessary med- temwide interventions are likely to have the highest effect on ications, however, lead to better outcomes. quality and health-related benefits. Various interventions can make sizable changes in the diag- nostic and prescribing patterns of providers. Verbal case review, combined with a package of additional intervention referred Cost-Effectiveness of Improved Process to as INFECTOM (Information, Feedback, Contracting with Two interventions that vividly illustrate the cost-effectiveness Providers to Adhere to Practice Guidelines, and Ongoing of improvements in clinical practice and outcomes have been Monitoring), increased the proportion of cases treated correctly chosen: detection and treatment of acute respiratory illnesses from 16 percent to 48 percent (Bloom, Canning, and Sevilla and appropriate drug use and treatment for diarrhea. 2001). One study reports that small group, face-to-face inter- ventions reduced antimicrobial prescriptions by 16 percent Better Treatment of Pneumonia in Children. Part of the high and antidiarrheal prescriptions by 7 percent among a group mortality from childhood pneumonia in the developing world of providers treating acute diarrhea in Indonesia (Santoso, can be explained by poor-quality care, which is defined as the Suryawati, and Prawaitasari 1996). The same study showed that inability either to accurately diagnose or to treat the disease. formal seminars reduced antimicrobial use by 10 percent and Our prototype intervention has two cost components: the cost antidiarrheal use by 7 percent. On the basis of these studies, an of implementing an educational activity for providers and the average cost per intervention was used, ranging from US$25 to cost of treating nonsevere and severe childhood pneumonia. US$125. The former component is based on a study and uses conserva- The savings from switching to a less costly treatment (instead tive high-end cost estimations (Kelley and others 2001); the of antibiotics, for example) were subtracted from the direct latter is the midpoint from another study (Stansfield and costs that are related to implementing the training activity. Shepard 1990). The number of lives saved depends on the Because other savings, such as those related to a lower use of effect of the intervention--that is, the change in the percent- inpatient services, were ignored, the estimates are conservative. age of cases diagnosed and treated; the prevalence rate of both Savings could be greater: Two years after an ORT unit was estab- types of pneumonia; the population covered by each provider; lished at the Kamuza Central Hospital in Malawi, 50 percent the case-fatality ratio; and the effectiveness of the treatment. fewer children with diarrhea were admitted to the pediatric Both the case-fatality ratio and the effectiveness ratio were ward, and those admitted required 56 percent less intravenous fixed at middle values suggested by earlier work (Stansfield fluid for rehydration (Martines, Phillips, and Feachem 1990). and Shepard 1990). For the other parameters, a large range of Again, the number of lives saved depends on the disease variation was considered, producing 450 scenarios. Finally, six prevalence; the effect of the policy on treatment quality; the Improving the Quality of Care in Developing Countries | 1303 population covered by each provider; the average case-fatality Such studies should be complemented by cost-benefit and ratio, which was set at 6 per 1,000 on the basis of Snyder and cost-effectiveness analyses. Sometimes, in public health emer- Merson (1982); and the effectiveness of the treatment. For the gencies, for example, control groups may not be practical or latter parameter, reductions in mortality rates following ORT ethical, in which case real-time operations research is an treatment of 40 to 60 percent and reductions in effectiveness acceptable substitute. ratios of 5 to 100 percent have been reported (Shepard, Brenzel, In the area of research topics, top priority should be given and Nemeth 1986). Accordingly, the effectiveness ratio was set to quality monitoring and assurance strategies to gain an at 80 percent. As before, alternative values for the other param- understanding of exactly what the health system is contribut- eters were adopted, generating 450 scenarios. ing to society and at what cost. Quantifying the associated Educational interventions to improve the quality of care for costs of different variants of quality monitoring and assurance treatment of diarrheal diseases are also highly cost-effective. In strategies should also be a high-priority item on the quality general, the cost of saving a life through educational interven- research agenda. The second priority should be to increase the tions is less than US$500 and could be as low as US$14. evidence base regarding the effects on provider behavior of Scenarios with high cost per life saved (more than US$6,000) public policies concerning quality of care and whether they are when prevalence rates are low or when implementation lead to better health outcomes. We need to learn more about costs for quality-related interventions are high. the long-term effects of different contracting and remunera- Although the data available to estimate the costs and tion policies on providers' practices and the consequent benefits of health outcomes and process are limited, these results of such policies for health outcomes. Finally, we need simulations, combined with published reports of successful to understand how contracting and remuneration policies policy interventions, clearly show the cost-effectiveness of affect problems unique to the developing world, such as the interventions that improve health outcomes through better use of doctor substitutes and the migration of skilled quality of care. However, reliable measures of quality are neces- providers to wealthier countries. sary to design and evaluate these interventions. CONCLUSION RESEARCH AGENDA ON QUALITY Good quality means that providers are able to manage an indi- Most of the issues discussed throughout this chapter represent vidual's or a population's health care by timely, skillful applica- important topics for research. Establishing a research agenda tion of medical technology in a culturally sensitive manner requires prioritizing both the type of research and the topics to within the available resource constraints. Eliminating poor be studied. Quality-of-care research must also strike a balance quality involves not only giving better care but also eliminating between relevance to decision making and excellence in scien- underprovision of essential clinical services (systemwide tific rigor (Frenk 1992). microscopy for diagnosing tuberculosis, for example); stopping Observational studies are needed to document the extent overuse of some care (prenatal ultrasonography or unnecessary and correlates of quality at various levels: individual providers, injections, for example); and ending misuse of unneeded serv- institutional providers, health care systems, and whole popula- ices (such as unnecessary hysterectomies or antibiotics for viral tions. Apart from offering much-needed basic descriptions infections). A sadly unique feature of quality is that poor qual- (especially in developing countries), these studies can test spe- ity can obviate all the implied benefits of good access and effec- cific indicators of the dimensions of quality and can compare tive treatment. At its best, poor quality is wasteful--a tragedy in the measurement approaches discussed earlier. severely resource-constrained health care systems. At its worst, Intervention studies introduce planned changes into health it causes actual harm. care settings and assess their consequences. It is fundamentally Despite the urgency of improving health in developing important that intervention studies compare one provider countries, quality of care has been largely ignored. Both group or policy alternative with another. In addition, control providers and patients agree this must change, but how can this groups must be used so that any observed change can be goal be reached? From the information marshaled for this attributed to the intervention itself rather than to another chapter, we can draw five conclusions: source of variation. The external validity of studies is often undermined by the choice of highly specific sites, making it dif- · Better quality leads to better health outcomes in developing ficult to generalize the findings and to build a body of sound countries. evidence. If randomized trials cannot be conducted, the pre- · Process, the proximate determinant of health outcomes, can ferred option is quasi-experimental studies with clear control be measured in valid and reliable ways, such as clinical groups and longitudinal designs (Peabody and others 1999). vignettes and electronic medical records. 1304 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others · Measured in the above ways, the process of care in develop- based choice market that many have envisaged. Instead, it ing countries is poor. motivates managers and providers to undertake changes that · The process of care can be improved in the short term. improve the delivery of care (Schneider and Lieberman 2001). · Policies affecting structural conditions, including the actual Outside pressure--perceived or real--appears to extend the process of care or the continual design and redesign of the quality debate beyond traditional boundaries, allowing for health care system, have been shown to be effective in devel- innovative collaborations and "out of the box" thinking oping countries. (Devers, Pham, and Liu 2004). Nongovernmental and private organizations involved in health care delivery should also be We believe that two broad strategies would help to rapidly required to report basic quality measures, perhaps as a condi- improve health care quality in developing countries: tion for funding, thus ensuring that similar pressure to improve quality is exerted outside the public sector. · encouraging explicit comparative research on outcomes and Public dissemination can create shock waves when poor process quality is "discovered," leading to popular demand to increase · disseminating empirical findings on quality variation. quality. For example, findings of widespread medical errors in Encouraging Explicit Comparative Research on Outcomes the United States, estimated to have resulted in as many as and Process 98,000 deaths per year, launched the medical safety revolution (Institue of Medicine 2001). Dissemination among physicians Comparisons highlighting different outcomes can be com- and surgeons by means of report cards and ratings has been pelling. For example, when 30-day mortality rates for coronary effective at changing clinical practice. One advantage of dis- artery bypass surgery at various facilities were disclosed in the semination among providers is that the results can be more United States, care started to shift from many low-volume hos- refined and technical than ratings meant for wider audiences. pitals to high-volume hospitals (Chassin 2002). In developing Dissemination is the responsibility of public research and pub- countries, comparisons show that the insured are more likely lic initiative. Because dissemination is inherently controversial, to have cesarean sections than are the uninsured (Barros and it requires public financing--even more than other public others 1991). Although critics of comparative analysis are jus- goods (Jamison, Frenk, and Knaul 1998). tified in saying that systems and populations vary, such criti- Ultimately, improving quality is about value. In health care, cism misses an important point: Differences in outcome high- price is not a reliable proxy for quality and cannot be used as light possibilities that help in the search for the underlying a guide. Because patients and consumers cannot directly causes of poor quality. Although poor quality may have many observe quality, their ability to demand high-quality services is causes, one of them is almost always poor clinical practice, limited, and they are often left to settle for a market that has which can be remedied. We also favor a league or summary suboptimal equilibrium and poor quality of care. In addition, table approach to making comparisons. In this approach, the providers often lack knowledge of optimal treatments and providers being compared agree on criteria before prospective technologies and thus are not aware of how they can produce assessments are done. The data for the comparison should be higher-quality care. Because the provider-patient interaction of the highest quality; the league tables themselves should is so private and personal, quality of care is hard to observe be easy to interpret; and the findings should be rapidly avail- and to measure. New measurement tools, however--such as able (Devers, Pham, and Liu 2004). The league table itself clinical vignettes and the electronic medical record--are being should be set up at the regional, national, and international developed and improved. As research links care with out- levels so that a variety of benchmarks are available. comes and cost inputs, we can expect to have more accurate Implementing quality comparisons will greatly facilitate the and reliable data about clinical practice for use in making process of policy evaluation and cost-benefit analysis and help quality assessments. indicate directions for future research. Access to accurate, con- Investments in quality, however, must be judged critically as sistent quality-of-care data will compel external funders, such well. When we invest in quality, an investment can be beneficial as the World Bank, to build quality assurance into their lending but can come at a cost. So while quality goes up, value can go and development programs. As major health programs such as up or down--or costs can go up while quality actually goes the Global Fund to Fight AIDS, Tuberculosis, and Malaria are down or stays the same, thus pushing the value of care down scaled up around the world, mechanisms to measure and and undermining other efforts to improve quality. Finally, as improve care quality will grow more important. we showed for acute respiratory illness and diarrhea, quality can go up and costs go down, thus increasing overall value. Disseminating Empirical Findings on Quality Variation Examples of this optimal outcome must be actively sought out Public dissemination of information on quality, particularly in and reported, because the success of a given investment cannot low-literacy countries, does not seem to create the individual- be known in advance (Berwick 2004). Improving the Quality of Care in Developing Countries | 1305 Improving health status does not have to rely solely on Davis, D. A., M. A. Thomson, A. D. Oxman, and R. B. Haynes. 1995. macroeconomic growth or other long-term development indi- "Changing Physician Performance: A Systematic Review of the Effect of Continuing Medical Education Strategies." Journal of the American cators. Health outcomes can be rapidly improved in the short Medical Association 274 (9): 700­5. term by ensuring the appropriateness of the circumstances Devers, K. J., H. H. Pham, and G. Liu. 2004. "What Is Driving Hospitals' or setting under which the health care encounter occurs Patient-Safety Efforts?" Health Affairs 23 (2): 103­16. (structural improvement) or by increasing the likelihood that Donabedian, A. 1980. The Definition of Quality and Approaches to Its health care providers behave in ways most beneficial to Assessment. Ann Arbor, MI: Health Administration Press. patients under the prevailing circumstances (process improve- ------. 1988. "The Quality of Care: How Can It Be Assessed?" Journal of the American Medical Association 260 (12): 1743­48. ment). However, this improvement will not occur sponta- Douglas W. R., D. H. Howard, E. R. Becker, E. K. Adams, and M. H. neously or routinely, despite the best intentions of beneficiar- Roberts. 2004. "Use of Midlevel Practitioners to Achieve Labor Cost ies, providers, and governments. Quality improvement tools Savings in the Primary Care Practice of an MCO." Health Care and technologies and information on successful quality Economics 39 (3): 607­25. improvement policies must be consistently shared among Eichler, R., P. Auxila, and J. Pollock. 2001. "Performance-Based Payment to developing countries to build local capacity. Funding and Improve the Impact of Health Services: Evidence from Haiti." World Bank Institute Online Journal. incentives must also be consistent with high quality. Finally, Flores, R., J. Robles, and B. R. Burkhalter. 2002. "Distance Education with the political will to ensure that quality becomes a top priority Tutoring Improves Diarrhea Case Management in Guatemala." on the health reform agenda must be sustained. International Journal of Quality in Health Care 14 (Suppl. 1): 47­56. Frenk, J. 1992. "Balancing Relevance and Excellence: Organizational Responses to Link Research with Decision Making." Social Science and NOTE Medicine 35 (11): 1397­404. Gutierrez, G., H. Guiscafre, M. Bronfman, J. Walsh, H. Martinez, and O. 1. As a reference, if the average expected value of a life is close to Munoz. 1994. "Changing Physician Prescribing Patterns: Evaluation of US$30,000, even the highest cost-effectiveness ratio found in the analysis an Educational Strategy for Acute Diarrhea in Mexico City." Medical (US$12,000 per life saved), would imply a cost-benefit ratio below 50 per- Care 32 (5): 436­46. cent, assuming an initial average wage of US$1,000 growing at 2 percent Habib, M., K. Mandal, C. V. Bunce, and S. G. Fraser. 2004. "The Relation of per year, a 5 percent discount rate, and unchanged mortality rates. Volume with Outcome in Pachoemulsification Surgery." British Journal of Ophthalmology 88: 643­46. Heaton, C. 2000. "External Peer Review in Europe: An Overview from REFERENCES the ExPeRT Project. External Peer Review Techniques." International Journal of Quality in Health Care 12 (3): 177­82. Barros, F. C., J. P. Vaughan, C. G. Victora, and S. R. Huttly. 1991. "Epidemic Institute of Medicine. 2001. Crossing the Quality Chasm. Washington, DC: of Caesarean Sections in Brazil." Lancet 338 (8760): 167­69. National Academy Press. Beracochea, E., R. Dickson, P. Freemand, and J. Thomason. 1995. "Case Jack, W. 2003. "Contracting for Health Services: An Evaluation of Recent Management Quality Assessment in Rural Areas of Papua New Reforms in Nicaragua." Health Policy and Planning 18 (2): 195­204. Guinea." Tropical Doctor 25 (2): 69­74. Jamison, D. T., J. Frenk, and F. Knaul. 1998. "International Collective Berwick, D. 1989. "Continuous Improvement as an Ideal in Health Care." Action in Health: Objectives, Functions, and Rationale." Lancet 351 New England Journal of Medicine 320 (1): 53­56. (9101): 514­17. ------. 2004. "Lessons from Developing Nations on Improving Health Javitt, J., G. Venkataswamy, and A. Sommer. 1983. "The Economic and Care." British Medical Journal 328 (7448): 1124­29. Social Aspects of Restoring Sight." In 24th International Congress of Bhat, R. 1996. "Regulating the Private Health Care Sector: The Case of the Ophthalmology, ed. P. Henkind, New York: J. B. Lippincott. Indian Consumer Protection Act." Health Policy and Planning 11 (3): Kelley, E., C. Geslin, S. Djibrina, and M. Boucar. 2001. "Improving 265­79. Performance with Clinical Standards: The Impact of Feedback on Bloom, D. E., D. Canning, and J. Sevilla. 2001. "The Effect of Health on Compliance with the Integrated Management of Childhood Illness Economic Growth: Theory and Evidence." NBER Working Paper 8587, Algorithm in Niger, West Africa." International Journal of Health National Bureau of Economic Research, Cambridge, MA. Planning Management 16 (3): 195­205. Brian, G., and H. Taylor. 2001. "Restoring Sight to the Millions: The Kumaranayake, L., P. Mujinja, C. Hongoro, and R. Mpembeni. 2000. "How Aravind Way." Bulletin of the World Health Organization 79 (3): Do Countries Regulate the Health Sector? Evidence from Tanzania and 270­71. Zimbabwe." Health Policy and Planning 15 (4): 357­67. Broomberg, J., P. Masobe, and A. Mills. 1997. "To Purchase or to Provide? Loevinsohn, B. P., E. T. Guerrero, and S. P. Gregorio. 1995. "Improving The Relative Efficiency of Contracting-Out versus Direct Public Primary Health Care through Systematic Supervision: A Controlled Provision of Hospital Services in South Africa." In Private Health Field Trial." Health Policy and Planning 10 (2): 144­53. Providers in Developing Countries: Serving the Public Interest?, ed. S. Luck, J., and J. Peabody. 2002. "Using Standardised Patients to Measure Bennett, B. McPake, and A. Mills. London: Zed Press. Physicians' Practice: Validation Study Using Audio Recordings." British Chakraborty, S., S. A. D'Souza, and R. S. Northrup. 2000. "Improving Medical Journal 325 (7366): 679. Private Practitioner Care of Sick Children: Testing New Approaches in Luck, J., J. Peabody, T. R. Dresselhaus, M. Lee, and P. Glassman. 2000."How Rural Bihar." Health Policy and Planning 15 (4): 400­7. Well Does Chart Abstraction Measure Quality? A Prospective Chassin, M. R. 2002."Achieving and Sustaining Improved Quality: Lessons Comparison of Standardized Patients with the Medical Record." from New York State and Cardiac Surgery." Health Affairs 21 (4): 40. American Journal of Medicine 108 (8): 642­49. 1306 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others Marek, T., I. Diallo, B. Ndiaye, and J. Rakotosalama. 1999. "Successful Santoso, B., S. Suryawati, and J. E. Prawaitasari. 1996. "Small Group Contracting of Prevention Services: Fighting Malnutrition in Senegal Intervention vs. Formal Seminar for Improving Appropriate Drug and Madagascar." Health Policy and Planning 14 (4): 382­89. Use." Social Science and Medicine 42 (8): 1163­68. Martines, J., M. Phillips, and R. G. Feacham. 1993. "Diarrheal Diseases." Schneider, E. C., and T. Lieberman. 2001. "Publicly Disclosed Information In Disease Control Priorities in Developing Countries, ed. D. Jamison, about the Quality of Health Care: Response of the U.S. Public." Quality W. H. Mosley, A. R. Measham, and J. L. Bobadilla, New York: Oxford in Health Care 10 (2): 96­103. University Press. Secretaría de Salud de Mexico. 2003. Salud: México 2002. Mexico City: McBride, A. B., S. Neiman, and J. Johnson. 2000. "Responsibility-Centered Secretaría de Salud. Management: A 10-Year Nursing Assessment." Journal of Professional Shepard, D. S., L. E. Brenzel, and K. T. Nemeth. 1986. "Cost-Effectiveness Nursing 16 (4): 201­9. of Oral Rehydration Therapy for Diarrhoeal Diseases." PHN Technical McConnell, H. 2004. "International Efforts in Implementing National Note 86-26, Population, Health, and Nutrition Development, World Health Information Infrastructure and Electronic Health Records." Bank, Washington, DC. World Hospitals and Health Services 40 (1): 33­37, 39­40, 50­52. Silimper, D. R., L. M. Franco, T. Veldhuyzen van Zanten, and C. MacAulay. McGlynn, E. A., S. M. Asch, J. Adams, J. Keesey, J. Hicks, A. DeCristofaro, 2002. "A Framework for Institutionalizing Quality Assurance." and E. A. Kerr. 2003. "The Quality of Health Care Delivered to Adults International Journal for Quality in Health Care 14 (Suppl. 1): 67­73. in the United States." New England Journal of Medicine 348 (26): Snyder, J. D., and M. H. Merson. 1982. "The Magnitude of the Global 2635­45. Problem of Acute Diarrhoeal Disease: A Review of Active Surveillance Mtango, F. D., and D. Neuvians. 1986. "Acute Respiratory Infections in Data." Bulletin of the World Health Organization 60 (4): 605­13. Children under Five Years: Control Project in Bagamoyo District, Stansfield, S. K., and D. S. Shepard. 1990. "Acute Respiratory Infection." Tanzania." Transactions of the Royal Society of Tropical Medicine and In Disease Control Priorities in Developing Countries, ed. D. Jamison, Hygiene 80 (6): 851­58. W. H. Mosley, A. R. Measham, and J. L. Bobadilla. New York: Oxford Murray, C. J., and J. Frenk. 2000. "A Framework for Assessing the University Press. Performance of Health Systems." Bulletin of the World Health Supratikto, G., M. E. Wirth, E. Achadi, S. Cohen, and C. Ronsmans. 2002. Organization 78 (6): 717­31. "A District-Based Audit of the Causes and Circumstances of Maternal Nolan, T., P. Angos, A. J. Cunha, L. Muhe, S. Qazi, E. A. Simoes, and others. Deaths in South Kalimantan, Indonesia." Bulletin of the World Health 2001. "Quality of Hospital Care for Seriously Ill Children in Less- Organization 80 (3): 228­34. Developed Countries." Lancet 357 (9250): 106­10. Tan, S. K. P. 1999. "Safety with Anaesthesia--The Paradigm of Continuous Peabody, J. 1996. "Economic Reform and Health Sector Policy: Lessons Improvement." Medical Journal of Malaysia 54 (1): 1­3. from Structural Adjustment Programs." Social Science and Medicine Tavrow, P., Y.-M. Kim, and L. Malianga. 2002. "Measuring the Quality 43 (5): 823­35. of Supervisor-Provider Interactions in Health Care Facilities in Peabody, J., P. Gertler, and A. Liebowitz. 1998. "The Policy Implications of Zimbabwe." International Journal of Quality in Health Care 14 Better Structure and Process on Birth Outcomes in Jamaica." Health (Suppl. 1): 57­66. Policy 43 (1): 1­13. Thapar, N., and I. Sanderson. 2004. "Diarrhoea in Children: an Interface Peabody, J., J. Luck, P. Glassman, S. Jain, J. Hansen, M. Spell, and M. Lee. between Developing and Developed Countries." Lancet 363 (9409): 2004. "Measuring the Quality of Physician Practice by Using Clinical 641­53. Vignettes: A Prospective Validation Study." Annals of Internal Medicine Thaver, I. H., T. Harpham, B. McPake, and T. Garner. 1998. "Private 141 (10): 771­80. Practitioners in the Slums of Karachi: What Quality of Care Do They Peabody, J., J. Luck, S. Jain, D. Bertenthal, and P. Glassman. 2004."Assessing Offer?" Social Science and Medicine 46 (11): 1441­49. the Accuracy of Administrative Data in Health Information Systems." Van der Gaag, J. 2000. "From Child Development to Human Medical Care 42 (11): 1066­72. Development." In From Early Child Development to Human Peabody, J., M. Rahman, P. J. Gertler, J. Mann, D. O. Farley, and G. M. Development, ed. M. E. Young, Washington, DC: World Bank. Carter. 1999. Policy and Health: Implications for Development in Asia. Walker, G. J., D. E. Ashley, and R. J. Hayes. 1988. "The Quality of Care Is Cambridge, UK: Cambridge University Press. Related to Death Rates: Hospital Inpatient Management of Infants Peabody, J., F. Tozija, J. Muñoz, R. J. Mordyke, and J. Luck. 2004. "Using with Acute Gastroenteritis in Jamaica." American Journal of Public Vignettes to Compare the Quality of Care Variation in Economically Health 78 (2): 149­52. Divergent Countries." Health Services Research 39: 1937­56. Wennberg, J. E., E. S. Fisher, and J. S. Skinner. 2002. "Geography and the Perez-Cuevas, R., H. Guiscafre, O. Munoz, H. Reyes, P. Tome, V. Libreros, Debate over Medicare Reform." Health Affairs (Millwood) Suppl. Web and G. Gutierrez. 1996. "Improving Physician Prescribing Patterns to Exclusives: W96­114. Treat Rhinopharyngitis: Intervention Strategies in Two Health Systems WHO (World Health Organization). 2000. World Health Report 2000 of Mexico." Social Science and Medicine 42 (8): 1185­94. Health Systems: Improving Performance. Geneva: WHO. Sachs, J. D. 2001. Macroeconomics and Health: Investing in Health for Zaslavsky, A. M., J. N. Hochheimer, E. C. Schneider, P. D. Cleary, J. J. Economic Development. Report of the Commission on Macro- Seidman, E. A. McGlynn, and others. 2000. "Impact of Socio- economics and Health. Geneva: World Health Organization. demographic Case Mix on the HEDIS Measures of Health Plan Salmon, J., J. Heavens, C. Lombard, and P. Tavrow. 2003. "The Impact of Quality." Medical Care 38 (10): 981­92. Accreditation on the Quality of Hospital Care: KwaZulu-Natal Province, Zgibor, J. C., and T. J. Orchard. 2004. "Specialist and Generalist Care for Republic of South Africa." Operations Research Results 2 (17). Published Type 1 Diabetes Mellitus--Differential Impact on Processes and for the U.S. Agency for International Development (USAID) by the Outcomes." Disease Management and Health Outcomes 12 (4): 229­38. Quality Assurance Project, University Research Co., Bethesda, MD. Improving the Quality of Care in Developing Countries | 1307 Chapter 71 Health Workers: Building and Motivating the Workforce Charles Hongoro and Charles Normand Policy on human resources for health should support health Economics predicts that employers will employ workers as policy objectives and be a means for achieving policy goals. The long as the additional value of their services is at least as great implication of such a focus is that health systems development as the cost of employing them, and workers will work if the should start by identifying the tasks that must be carried out rewards are of greater value than those accruing to other uses and the skills needed to perform them. Meeting policy goals of their time. If key professionals are in short supply, higher depends on being able to recruit, train, and retain staff with the salaries will be needed to attract them. Workers will invest in necessary bundles of skills. Traditionally, skills are defined by training if they value higher future incomes and more interest- membership of a profession, especially medicine, nursing, mid- ing work above the costs of income lost during training and of wifery, and the allied health professions. Low- and middle- fees paid for training programs. This chapter focuses on how income countries (LMICs), often from necessity, have widened health systems might build and improve HR capacity. the range of health care workers to meet the service needs, with Appropriate HR capacity is critical for the effective imple- some people trained in extremely basic skills and others receiv- mentation of disease control interventions. Salaries account ing enhanced training, such as nurses trained in emergency for 50 to 80 percent of health sectors' recurrent costs (Bach obstetrics. What is meant by a doctor or nurse also varies. 2000). Table 71.1 shows the number of physicians and nurses Even though structures and institutions vary widely, some per 100,000 population in selected countries. The number of problems are common to most LMICs. First, persuading health workers is related to the level of development because doctors to work in remote rural areas is difficult, and they typ- of the tight resource constraints facing LMICs and because of ically do not remain long in such posts. Second, emigration of supply constraints, often exacerbated by migration of skilled doctors and nurses is extensive. Third, it is common for doctors workers (Awases, Gbary, and Chatora 2003) and prevalence of to work in both the public and the private sectors (referred to AIDS. In Africa, where the disease burden is high and increas- as dual practice), sometimes harming public services. Dual ing rapidly, the number of health workers is particularly low. practice may encourage doctors to skimp on their public health Most African countries export health professionals to high- efforts, to pilfer supplies, and to induce demand for their income countries. private services (Bir and Eggleston 2003). A study in six African countries showed that most health Many health sector human resource (HR) problems are workers intend to migrate for higher salaries. In Ghana, 70 per- predictable from a simple labor market perspective, given the cent of 1995 medical graduates had emigrated by 1999 (Awase, combinations of incentives confronting health care workers Gbary, and Chatora 2003). Pay differentials provide strong and the constraints policy makers face. Experience in LMICs incentives to migrate. For example, a junior doctor in the shows how problems have arisen and what policies have United Kingdom averages a monthly salary of US$3,029 and a succeeded. registered nurse averages US$1,500, compared with US$300 1309 Table 71.1 Numbers of Physicians and Nurses, Selected HEALTH CARE PROVISION AND ASSOCIATED Countries and Country Groups, 1998 HUMAN RESOURCE NEEDS Physicians Nurses Physician- Studies on developing services to meet the Millennium per 100,000 per 100,000 nurse Development Goals emphasize the importance of making Country population population ratio health workers with the appropriate skills available and moti- Angola 5 100 0.05 vating them (Jha and Mills 2002). The problems include lack Bangladesh 19 11 1.7 of technical skills, low motivation, and poor support networks Boliviaa 29 14 2.1 (Kurowski and others 2003). This chapter, therefore, focuses Botswana 20 100 0.2 on HR planning, training and professional development, Brazil 136 44 3.1 incentives for workers to accept and stay in posts and to Burkina Faso 3 20 0.15 deliver services, and alternatives to conventional professional Central African Republic 3 10 0.3 groups. Equatorial Guinea 20 50 0.4 India 106 94 1.1 Incentives and Motivation Nepal 4 5 0.8 The labor market model outlined earlier provides a framework Pakistan 57 34 1.7 for analyzing the role of incentives. A health worker will accept Papua New Guinea 7 67 0.1 a job if the benefits of doing so outweigh the opportunity cost. Perub 10 7 1.4 Improving recruitment and retention requires either offering South Africa 20 100 0.2 higher rewards that make alternative employment less attrac- Sri Lanka 37 103 0.4 tive or making qualifications less "portable"--that is, less likely to be recognized in other countries. The development of new Low-income countries 73 132 0.6 health professions in many countries is a way of reducing the Middle-income countries 142 278 0.5 portability of qualifications, thereby reducing the opportunity High-income countries 286 750 0.4 cost of jobs at home. Another advantage is that training can be Global average 146 334 0.4 more specific to local health system needs, but ensuring quality Global median 114 233 0.5 and safety are important issues. Sources: PAHO 2003; Support for Analysis and Research in Africa 2003. Health workers will choose to train and increase their skills a. 1999 data. if the rewards of doing so exceed the cost. In general, the sup- b. 1996 data. ply of skilled professionals rises as rewards increase, because more will seek training, more will return to the workforce, and fewer will move to other jobs or other countries. Because health per month for a Ugandan medical officer and US$180 for a workers value both financial and nonfinancial rewards, they registered nurse. will work for lower salaries if other job characteristics are Scaling up service provision using current provision models attractive. would require large increases in resources and could require a The causes of health HR problems in developing countries change in strategy by development partners toward supporting are complex, and attempts to address them must reflect this recurrent costs (Jha and Mills 2002). The labor market model complexity. Table 71.2 suggests a framework for exploring indicates that higher salaries would be needed to attract addi- links between factors at individual, organizational, and tional staff members, so funding would have to increase more health system levels. The framework is inspired by a systems than in proportion to the number of staff members employed. approach, which gives prominence to the roles of and rela- Many LMICs pay health workers on civil service scales, tionships between different component parts in influencing which they control to contain overall government spending. the whole. This practice further widens the gap between salaries for The individual health worker level serves as a starting professionals at home and abroad. point for exploring the determinants of health worker behav- Although improved economic performance and increased ior and performance (Kyaddondo and White 2003). development assistance may allow some increases in health Performance here means productivity and quality of services. spending, in most LMICs it is not plausible that such increases Individuals respond to individual concerns through coping would be sufficient to make the necessary skills available with- strategies, such as informal and dual practices, with associated out a range of strategies, including better regulation, stronger consequences. There are multiple links between individual incentives, and initiatives to make key skills available at lower health worker behavior and organizational and systemic cost. factors. Organizational and system arrangements define 1310 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand Table 71.2 Framework for Diagnosing HR Issues Financial Incentives in the Health Sector Most of the comparatively scarce evidence on the relative Level Issue importance of financial and other incentives for health workers at the individual level comes from developed countries. Two Individual health worker findings emerge from recruitment and turnover studies. First, Internal capacity Knowledge, skills, competencies, attitude at extremely low salaries, financial incentives are particularly Remuneration Salary and perquisites, payment methods important (Normand and Thompson 2000). Second, at least Work environment half of the variation in turnover can be attributed to financial Immediate environment See organizational-level issues incentives (Gray and Phillips 1996). These findings leave con- Distant environment See systemic-level issues siderable scope for improving retention using organizational Productivity Outputs or outcomes per given unit of changes, but such changes will be only partially successful if time or per individual or group of much better financial rewards are available elsewhere. individuals International migration has increased as restrictions on Responses to organizational Low motivation, morale, productivity, and moves to high-income countries have been eased (Bach 2000). and systemic constraints quality of services Many developed countries have shortages of health profession- Informal practices, for example, unofficial als and actively recruit from low-income countries, thereby fees, dual practice, misuse of public raising the opportunity cost of remaining at home. resources, and unethical practices Emergence of unskilled informal practitioners Health Care Systems' Responses to Health Worker Issues Organizational level Presence or absence of an appropriate Health sector reforms have been widespread in recent years, physical and operational context--that is, availability of materials and facilities often with international support. These reforms tended to to deliver services of acceptable quality, focus more on structures and financing and less on resource workload management, and organiza- issues (Martineau and Buchan 2000). Other government tional norms and practices (organizational reforms aimed mainly at improving efficiency and reducing culture, management and leadership the cost of government administration have often had large styles) effects on the health workforce (Adams and Hicks 2000; Organizational autonomy across key Corkery 2000). Some changes have attempted to introduce strategic and functional issues better incentives, such as performance-related pay and renew- Systemic level Level of bureaucracy and decentralization able contracts, and to remove underperformers and ghost in the health system, and funding and workers. Evidence on the effects of these reforms suggests that regulatory arrangements more emphasis should have been placed on designing incen- Policy context--for example, pro-market tives to improve performance and retention and on moving policies and diversification of the health care market in the spirit of a public- further away from workforce quotas and norms. Using the private mix three levels of analysis, the following sections consider policies, Socioeconomic and political context management, and incentives and how they can help match Medico-legal policy and enforcement skills to needs. Source: Authors. Note: This chapter uses a narrow definition of health system that considers socioeconomic and political issues as part of a wider context. HEALTH CARE STAFF the incentive context for health workers and influence both Workforce planning should be dynamic and should link policy organizational and individual performance. goals to staff members' skills and numbers and to performance- Therefore, the configuration of the health system must enhancing incentives. create incentives for appropriate supply and deployment of health workers. HR development experts tend to focus more on problems encountered in the lower tiers of this framework. Workforce Planning to Meet Policy Goals Political pressure for short-term solutions partly explains Several factors make workforce planning in health particularly why many countries do not address HR problems compre- difficult, including changing needs as service models change, hensively. The wider context can also be important. Good long training time for some professions, and lack of direct gov- governance at the national level is necessary to make policy ernment control over the number of professionals being interventions at the health system level or below effective. trained, for example, because of the growth of private medical Health Workers: Building and Motivating the Workforce | 1311 schools, such as in Bangladesh, or because of people going countries, although there have been attempts to change this overseas for training (although Singapore has addressed this balance. For example, more than 25 percent of Malawi's problem by restricting the colleges that the government recog- curriculum for medical students focuses on community health. nizes for registering doctors). The greatest difficulty comes Given that any training program can cover only a portion of rel- from the unpredictable loss of skilled staff members to private evant knowledge, focusing on locally relevant topics is increas- health sector jobs, jobs abroad, and jobs outside health. Thus, ingly important. Educational reforms in many medical schools a close link exists between HR planning and incentives and reg- in Africa and elsewhere are based on the community-based ulation. In Ghana, nurse training has often been the only avail- educational model (Jinadu, Olofeitime, and Oribador 2002). able form of tertiary education for women, and many of those who are trained do not practice. A key to more rational workforce planning is better coordi- Numbers and Types of Health Professionals nation between health planning and planning for training and Categories of workers result from combinations of previous education. Powerful interest groups can oppose the expansion and current needs, national traditions, interest group pressures, of training. Training establishments often oppose change be- and historical accidents. Doctors, nurses, and some paramed- cause it may disrupt existing arrangements and threaten current ical professions have wide international recognition but vary staff members. The development of new professional groups in definition. Professional traditions and professional bodies faces particular resistance from existing professional groups, bring some safeguards for quality and safety, and at best, which, quite correctly, perceive the new groups as posing a threat professionals champion the needs of patients. Membership of to their interests. For example, some dentists in South Africa a recognized profession can bring desirable independence from expressed concern over training of dental technicians who carry management. However, internationally recognized qualifica- out a wide range of preventive and restorative dentistry at lower tions make it easy for professionals to migrate to countries fees (Matomela 2004). offering higher incomes and better careers. Models for HR needs are easy to devise, but determining Most developing countries have new categories of staff that the appropriate model parameters is difficult. For example, do not match internationally recognized professions (Buchan health planners must estimate the length of a nursing career-- and Dal Poz 2003). Examples include nurses with extended potentially up to 45 years but often much less, especially if training and roles and people working at subnurse levels with nurses are willing to work outside nursing (Phillips and others training of a few weeks to three years. Bangladesh has family 1994). Good data are needed on dropout rates from training. welfare visitors, health assistants, and medical assistants who HR sections of health ministries are usually poorly resourced, might elsewhere be classified as nurses or auxiliary nurses; in have low status, and work with poor-quality data, and this Uganda, clinical officers have three years of training and work situation must be changed if planning is to improve. as subdoctors; and nursing aids in Uganda have three months of training. Training is for specific roles without the generic training in conventional professions. Typically, such employees Basic Skills Training and Continuing Skill Development are mobile nationally, but they do not transfer easily across Whereas the quality of basic training of health professionals countries. varies widely in LMICs, the provision of continuing education In the labor market model, employers want to employ staff and development is almost universally inadequate. Hence, skill members if their contribution to service provision is of greater levels of staff members fall over time. Evidence indicates that value than the cost of their employment. Because those with good-quality continuing professional development is a positive portable qualifications can work in other countries, salary lev- incentive and helps to retain staff members. Requirements to els needed to retain workers reflect that possibility. Theory sug- undertake continuing education can be made a condition of gests that staff members will develop new skills if such an continued professional registration and can thereby provide investment of their time and money produces significantly some guarantee of competence. increased salary or benefits. Many countries cannot fulfill their Good basic education includes development of both profes- requirements for health workers, but normally this difficulty sional skills and learning skills. Basic training and continuing reflects salaries that are too low to attract staff. However, rais- development should be planned together. In many cases, the ing salaries may make employment of the full complement of large investment in basic training is lost because of lack of staff members unaffordable. maintenance, so that shifting some resources to updating and Staffing norms serve little useful role if the salaries needed renewing skills is efficient. to fill the posts are unaffordable. Decisions about how many A further challenge is to align the content of training to the people should be employed and in what capacities should be skills professionals need. Many training programs in LMICs based on the contributions those employees will make and the provide skills oriented toward service needs in developed costs of employing them. Staffing norms can be useful for 1312 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand planning, but they require careful analysis of affordability of HEALTH WORKER INCENTIVES care, the skills needed, and the way to provide those skills most efficiently. Several countries have of necessity turned to new The World Health Report 2000 defines incentives for health models of provision using staff skilled in the delivery of key workers as "all the rewards and punishments that providers face elements of high-priority services, such as immunization and as a consequence of the organizations in which they work, the emergency obstetric care. institutions under which they operate, and the specific inter- ventions they provide" (WHO 2000, p. 61). Health workers face a hierarchy of incentives or disincentives generated by the work Safety and Effectiveness of New Health Professions they do, the way they are paid, and the organizational and sys- Research on the new professions is limited, and much of the tem context in which they work. Incentives are generally material is anecdotal (Buchan and Dal Poz 2003). A growing designed to accomplish the following: literature from developed countries indicates that nurses can be safe and effective in place of doctors in primary care (Venning · to encourage providers to furnish specific services and others 2000). The fear is that the absence of a formal pro- · to encourage cost containment fession and the lack of internationally recognized training · to support staff recruitment and retention could damage quality and safety. This issue is important, but · to enhance the productivity and quality of services even if new professionals are less safe than doctors, they may be · to allow for effective management. much safer than the absence of a service such as emergency obstetric care. In some countries, new professions play a major Responses of providers to incentives depend on context and part in the provision of services. A good example is Malawi, on the stage of their career. Incentives that induce productivity where clinical officers with extensive training (but much less vary with experience, stage in a career path, and changes in than that of doctors) are a major resource, carrying out surgi- providers' social responsibilities. Ideally, incentive structures cal procedures and administering anesthetics as well as provid- should recognize the evolutionary nature of work expectations. ing medical care. In some countries, regulations govern such Typically, incentives vary by type of employer: nongovern- extended roles (McAuliffe and Henry 1995). mental organization, public, or private. Public sector incentives Fenton, Whitty, and Reynolds's (2003) study of emergency tend to be the weakest because resource constraints and cesarean sections carried out by clinical officers in Malawi bureaucratic rules on civil servant employment constrain the found that the overall maternal death rate was 1.3 percent, use of both financial and nonfinancial incentives. which is high, but much lower than if services had not existed. Perinatal deaths were 13.6 percent. None of the anesthetists was medically qualified, but outcomes were better when these prac- Typology of Incentives titioners had received anesthetics training (maternal deaths Extrinsic incentives can be individual and organizational, mon- were 0.9 percent compared with 2.4 percent). The researchers etary and nonmonetary (table 71.3). Discussions of provider found no significant difference in outcome between medically behavior in LMICs have focused mainly on financial incentives, qualified surgeons and those trained as clinical officers. Care partly because of their low income levels compared with indus- should be taken in interpreting the results of one study, but it trial countries. The challenge is to establish an optimal mix of does suggest that well-trained clinical officers can safely substi- financial and nonfinancial incentives that generate the desired tute for doctors in providing some important procedures. behavior of health workers. Experience from vertical programs for priority diseases or services--for example, poliomyelitis, malaria, family planning, Human Resource Policy and New Staff Groups and sexually transmitted diseases--provide evidence about dif- New staff groups are increasingly providing essential services in ferent incentives. Programs often offered staff members better LMICs. In Zimbabwe, a new cadre called primary health care pay and incentive packages than those other public health nurses, whose qualifications are lower than general nurses, workers received (Beith and others 2001). The exact effects of was introduced in 2003 to curb external migration by nurses stronger incentives are unknown, but these programs generally (Chimbari 2003). At the system level, such a development succeeded, as evidenced by the eradication of leprosy, the near requires regulation and standard setting; at the service provision eradication of poliomyelitis in many countries, and the large level, appropriate supervision and management is needed; and drop in average fertility in developing countries in the 1990s. at the individual level, incentives and training need to be con- Successful vertical programs used combinations of incen- sidered. Employing fully qualified doctors and nurses might be tives, including better salaries, field and transportation the safest option, but failing to provide services because of allowances, streamlined management, specialized training, staffing constraints is unlikely to be the next best option. better facilities and material resources, and results-oriented Health Workers: Building and Motivating the Workforce | 1313 Table 71.3 Typology of Incentives of health system capacity, which depends in part on the align- ment of health workers' objectives with policy and with system Individual Organizational Environmental goals. incentives incentives incentives Aligning health worker and system objectives is difficult. Financial Internal Amenities The aim is to have satisfied health workers who are motivated Salary Autonomy Transportation to work harder (Hicks and Adams 2001). Evidence is limited, Pensions Accountability Job for spouse but financial and nonfinancial incentives are mutually rein- Illness, health, accident, Market exposure School for children forcing, and changing the culture of the health system to make and life insurance Financial responsibility goals more readily understood and shared can make financial Travel and transport incentives more powerful. Such change in the organization of External allowances health care can be politically sensitive because it can give health Governance Child care allowance sector workers advantages over other public employees. Public finance policy Rural location allowance Incentives may have conflicting effects. For example, decen- Regulatory mechanisms Heat allowance tralization might create the autonomy needed for effective Retention and professional management, but without transparent management and career allowances structures and job security, providers might view such a change Subsidized meals, as a threat (Kyaddondo and White 2003). Getting the balance clothing, and right requires understanding the socioeconomic and political accommodation circumstances and may be helped by using participatory Nonfinancial approaches to policy making and implementation. Vacation days Flexible working hours Context Access to training and education Context is defined here from an individual or an organizational Sabbatical and study leave provider's perspective. It constitutes what Adams and Hicks Planned career breaks (2000) refer to as external incentives--that is, methods used by health systems to control the activities of health organizations Occupational health or funders. Functional and professional autonomy The power of incentives depends on context. Health systems in developing countries have varying cultural and economic Technical support and feedback systems histories that shape providers' expectations and responses to Transparent reward incentives. Financial incentives are strong when health workers' systems incomes are low, as in most developing countries. Nevertheless, Valued by the organization examples of strong nonfinancial incentives exist in countries such as Thailand, where family ties and kinship affect health Source: Adapted from Zurn 2003. workers' decisions on where to work. Such nonfinancial incen- Note: These are mostly extrinsic incentives. tives affect the size of the financial incentives needed to change where people choose to work. History and experience determine a country's working cul- management to support improved health worker productivity ture and norms. In developing countries, most health systems and program performance. Goals were clearly specified, were are large bureaucracies whose management is driven centrally understood and shared by the staff, and were often linked to by guidelines, standards, and reporting systems. Incentives incentives. The choice of vertical structures also reflects the in such systems work against innovation, risk taking, and perceived difficulties of using existing health systems, with improved efficiency. A possible approach is to introduce their excessive bureaucracy, underfunding, and lack of capacity changes that are based on the ideas of so-called new public to implement integrated disease control. management. New public management replaces line manage- Vertical programs must eventually be reintegrated into the ment with contracts or agreements between funders and policy system. The HIV/AIDS pandemic is a good example of a dis- makers on the one hand and providers on the other. Providers ease that might require targeted interventions until the capacity are given more managerial autonomy and are controlled by of health systems in LMICs improves to a level that allows the means of contracts and regulation. This approach can more disease to be managed like other diseases. The success in inte- easily embody new financial incentives, and autonomous grating vertical programs depends on the parallel development providers can develop cultures that are more innovating. Such 1314 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand a radical change in managerial context can, in principle, make The effect of incentives can be assessed in terms of their other incentives easier to use. objectives (Adams and Hicks 2000). Table 71.4 summarizes Other dimensions of context are the regulatory framework incentive packages used in selected countries. The results shown and its enforcement. Most developing countries have regula- should be interpreted with caution, because of problems of tions governing the activities of the health sector. These regula- attribution and poor data. Adams and Hicks (2000) argue that tions tend to be outdated or poorly enforced (Bloom, Han, economic incentives in payment mechanisms for physicians and Li 2001). The main reason for regulatory ineffectiveness is conform to economic logic, but little is known about the low institutional capacity and widespread corruption. The response of other categories of health workers to such incentives. symptoms of regulatory failure are widespread informal activi- Experience in Thailand illustrates the labor market model ties, dual practice, malpractice and medical negligence, and the outlined earlier. In general, public doctors prefer to practice in presence of unqualified drug sellers (for example, in Bangladesh urban areas, where conditions are usually more attractive and and Tanzania) and practitioners (as in India) (Bhat 1996; opportunities for private practice are better. Thailand pays pub- Killingsworth and others 1999; McPake and others 1999). lic doctors who work in rural and remote areas significantly Where the regulatory system is dysfunctional, providers tend to more than those working in urban areas, and this incentive has pursue their individual interests, often in private practice, to the persuaded some to move (Wibulpolprasert and Pengpaiboon detriment of organizational and system performance. Effective 2003). The government also added nonfinancial incentives, such incentive systems that are based on performance require regula- as changing physicians' employment status from civil servants to tion and governance structures that minimize the common contracted public employees, providing housing, and intro- problems of patronage and corruption (Rasheed 1995). ducing a system of peer review and recognition. These initiatives Health system organization factors include governance and were coupled with significant environmental changes, including the degree of decentralization. Links exist between working sustained rural development. In most developing countries, culture and norms and the structural aspects of health system providers in rural areas are paid less than those in cities, and it is organization. The locus of control and decision making play an hard to recruit and retain health workers in rural areas. important part in health worker behavior. In theory, designing China provides another example of how changes in the incentive schemes that are responsive to health workers' environment--for example, the introduction of pro-market needs is much easier in a decentralized system. This theory is policies--can change provider behavior, in this case from rely- based on the belief that subnational units are better placed to ing on government salaries alone to the use of "red packages" make effective decisions on funding, regulating, and organizing (Bloom, Han, and Li 2001). These red packages were gifts that frontline activities than are centralized units. However, experi- were traditionally exchanged as an expression of mutual ence in developing countries shows that lack of capacity at appreciation, but they have now evolved into informal cash subnational levels has constrained decentralization, sometimes payments from patients to health workers. leading to unintended effects such as wrong priorities (Bloom, Health systems have a spectrum of workers with different Han, and Li 2001). Any move toward decentralization requires skills and expectations, and incentives for one group can have investment in new management skills and capacities. negative effects on others (Adams and Hicks 2000). Policy makers must strike a balance between competing interests of professional groups and system goals. The unionization of Incentives in Practice labor and the growth of professional associations or councils Many countries have attempted to reform their economies and can give health workers considerable bargaining power. health sectors to improve general economic and health system Solving one problem can create others. This situation often performance. For example, Cambodia, the Arab Republic of occurs when governments respond to the grievances of the Egypt, Uganda, and Zambia have attempted civil service reforms most vocal professional groups, usually doctors, and neglect (Corkery 2000). These reforms include attempts to reduce the other groups. This piecemeal approach has caused HR crises, size of the civil service to lower costs and to improve productiv- such as strikes and go-slows. Although health workers are nor- ity using incentives such as formal employment contracts and mally somewhat motivated to pursue health policy goals, their performance-based pay and promotion. Such reforms have been own interests can conflict with those goals. Providing higher largely unsuccessful in developing countries because of the polit- salaries to health workers, by increasing costs, can reduce access ical difficulties in reducing the size of the civil service. Structural to services by some social groups (Bloom, Han, and Li 2001). and organizational changes are typically unpopular with labor unions, especially if union members perceive them as threaten- ing their well-being. Experience also underscores the difficulties Compensation of aligning system and organizational objectives with individual Provider payment systems transfer resources from payers (gov- providers' objectives (Martineau and Buchan 2000). ernments, insurers, and patients) to providers (Maceira 1998) Health Workers: Building and Motivating the Workforce | 1315 Table 71.4 Incentive Packages for Health Workers, Selected Countries Complementary Objectives Incentives measures Constraints Results Recruiting and retaining staff Pay competitive salaries Fiscal policies that increase Budget limitations Helped retain physicians in in the country Include seniority awards in the after-tax marginal value Low public service salaries Bahrain pay scales of salaries Policies to reduce salaries as a share of operating costs Allow after-hours private Service standards and con- Work effort that may be con- Considered successful practice in public institutions trols to prevent reduced work centrated in private practice, in Bahrain effort in the public system leading to a deterioration of In some countries, resulted quality in public practice in deterioration of public systems where providers also engage in independent private practice (McPake and others 1999) Tolerate informal payments Not applicable Informal charges that limit Resulted in widespread use access and may impede of informal payments in reforms that involve formal Eastern and Central Europe, user fees and exemptions Sub-Saharan Africa, and some East Asian and Pacific countries (Balabanova and McKee 2003; Chakraborty and others 2002; Thompson and Witter 2000) Recruiting and retaining staff Provide higher salaries or Decentralized administration Overall staff shortages Premium payments for in rural areas location allowances Freedom to allocate institu- Budget limitations working in rural areas found (Wibulpolprasert and tional revenues or savings successful in Thailand Professional and lifestyle Pengpaiboon 2003) from operational efficiency (Wibulpolprasert and disadvantages Base remuneration on to fund incentives Pengpaiboon 2003) Smaller potential for earn- workload Improved infrastructure and ings from private practice staff competence than in urban areas Conflicting financial incen- tives (for example, loss of housing allowance in Bangladesh) Risks posed by internal conflicts and civil wars (for example, Colombia and Uganda) Require service in defined Consistent application of Loss of confidence if health Aided retention of profes- areas as condition of licens- policies on transfers and workers perceive the selec- sionals in Ghana and ing or specialty training tenure tion process as arbitrary Zimbabwe Provide opportunities for Providers' concerns that a (Chimbari 2003) government-sponsored temporary posting may further education become indefinite Provide housing and good- Adequate salary Budget limitations Found successful for nurses quality educational opportu- but not doctors in Nepal nities for health workers' families Recruit trainees from rural Emphasis on public health Traditionally, overrepresenta- Found successful in Thailand areas and family practice in tion of urban area students training curricula in student populations 1316 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand Table 71.4 Continued Complementary Objectives Incentives measures Constraints Results Enhancing the quality and Provide training and promo- Clear job descriptions and Opposition by professional Resulted in successful availability of primary care tion opportunities for nurses criteria for promotion associations to expanded retraining of health and medical auxiliaries roles for multifunction health assistants and other health Train multifunctional health workers in Nepal workers in rural areas in workers Limited training capacity in Nepal to make them eligible for promotion Mobilize women volunteers Uganda from communities, traditional Resulted in regrading of birth assistants, and local state-certified nurses to leaders state-registered nurses in Zimbabwe (Chimbari 2003; Pannarunothai, Boonpadung, and Kittidilokkul 2001) Encouraging teaching and Pay health workers more if None Allowances perhaps uncom- In Nepal, found successful in research and reducing the they do not practice privately petitive with private practice basic medical sciences but internal brain drain earnings resulted in massive resigna- tions in clinical departments Uncommon incentive, although a few countries (for example, Thailand) do pay professional allowances or nonpractice allowances Improving the quality of care Specify clinical guidelines in Leadership role by profes- Vested interests of profes- Uncommon in developing provider contracts sional organizations sional associations countries Inclusion in the curricula of Weak peer review systems Some success recorded in medical schools Low consumerism and weak Cambodia's contracting advocacy experiment License institutions and Tradition of professional Regulatory capture and a Reduced number of hospitals professionals based on self-regulation culture of self-protection and unqualified doctors in defined standards Acceptance of civil and legal Low capacity to enforce laws Estonia Pass laws requiring the authority and regulations Resulted in limited success registration of drugs and according to evidence from other potentially dangerous most developing countries substances (Bhat 1996) Source: Adapted from Adams and Hicks 2000. and can be structured to provide financial incentives. Most hours work. This method creates incentives for providers to do studies focus on payment mechanisms for doctors and their extra work and increase throughput, but providers may divert effect on productivity, costs, and quality of services (Bitran and patients to after-hours services, and the method's feasibility Yip 1998). Table 71.5 summarizes common payment mecha- depends in part on monitoring and governance standards. nisms and the desired incentives. The evidence shows that the The challenge is to find payment combinations that motivate operation of payment mechanisms is sensitive to the payment providers to provide desired volume and quality of services structure and how it is implemented (Berman and others 1997; while containing costs. Bitran and Yip 1998; Chomitz and others 1998). Payment systems are more successful when built on existing traditions and culture (that is, when they take into account gift Empirical Evidence on Payment Methods systems or, indeed, levels of corruption). It is normally best to Evidence of provider payment systems that have successfully use a combination of payment methods. For instance, if there aligned system and provider incentives is still limited (Bitran is a shortage of public providers, they might be paid a basic and Yip 1998). Interesting findings come from small-scale salary for normal working hours and fees for service for after- experiments such as Cambodia's New Deal (box 71.1). Health Health Workers: Building and Motivating the Workforce | 1317 Table 71.5 Major Payment Mechanisms Introducing financial incentives for health workers is costly. Policy makers in governments and development partners need Payment mechanism Key incentives for providers to ensure that adequate funding is available and sustainable. Fees for service Increase the number of cases seen Resources are also needed to improve working environments Increase service intensity and system capacities. Both financial incentives and other Provide more expensive services incentives are important, but services are likely to improve only Case payment (for example, Increase the number of cases seen if financial incentives are strengthened. diagnosis-related groups) Decrease service intensity Provide less expensive services Group Incentives Daily charge Increase the number of bed days Health workers typically work in teams. This system weakens through longer stays or more cases financial incentives because the efforts of individuals may have Flat rate (bonus payment) Provide specific bonus services and neglect other services little influence on overall performance. Indeed, individual incentives can worsen team cooperation. For example, if pro- Capitation Attract more patients to register while minimizing the number of motion is competitive and depends on measures of individual contacts with each and minimizing productivity, this approach can be a disadvantage for those who service intensity work for system goals in cooperative ways. Salary Reduce the number of patients and Designing effective group incentives is difficult. Paying the number of services provided group bonuses for achieving a given level of output can work Global budget Reduce the number of patients and only if individual team members feel adequately rewarded for the number of services provided their efforts and if there is no perceived free-rider problem. Source: Bennett, McPake, and Mills 1997. Most of the limited evidence on group incentives is for developed countries and shows that much depends on the production process and the organization of the teams (Ratto, workers' salaries were considered by many to be below the min- Propper, and Burgess 2002). Group financial incentives tend to imum required for a decent life, and workload is increasing be weak, and using other approaches such as team building, because of HIV/AIDS. better sharing of information, and improved working condi- The Cambodian experiment attempted to align individual tions is probably better. health workers' and system goals through performance-based bonus payments and a set of internal regulations. Regulations can alter the working and organizational culture in a way that Influence of System Capacities and Sustainability allows individual-based incentives to work. There were prob- Issues on Incentives lems in enforcing penalties for violating regulations. Failure to The theoretical merits and demerits of different incentives are enforce regulations may lead providers to lose confidence in the well understood, but system capacities and financial con- system. Countries with limited administrative and institutional straints may limit their applicability. Few developing countries capacity should use simple payment mechanisms that are have health systems that are capable of effectively implement- enforceable within their capacity constraints (Barnum, Kutzin, ing and operating some of the payment systems shown in and Saxexian 1995). A lesson from the experiment is that the table 71.5. The overall funding for the health sector may be context matters, and any strategy for offering incentives to too low to pay providers more. Also, the skills and expertise workers must be embedded in traditions and cultural practices. needed to design and implement contract- and case-based pay- In a competitive environment, contracts are a useful tool for ment methods may be inadequate, and the country may lack aligning health workers' behavior with organizational and sys- the information technology needed to capture relevant data to tem objectives. In the Cambodian example, contracts between support such contract- or case-based payment methods. Most the purchaser and district-level facilities--and between district- health workers in developing countries are civil servants, and level facilities and management committees--were an attempt the particular needs of health workers may be lost in a general to establish accountability structures that specify targeted activ- public service. Some countries are considering delinking ities. More interesting was the attempt to transfer some man- health workers from public service commissions and setting up agement risk and responsibilities to individual health workers independent health commissions to run the health sector. In using subcontracts that permitted management committees to Zambia, however, delinking failed because of a lack of capacity monitor their activities and pay them accordingly, though at both the national and the local levels to implement the nec- whether the contracts were well specified is not clear, and the essary HR changes (Martineau and Buchan 2000). Evidence administrative and transaction costs are unknown. The use of from Trinidad and Tobago suggests that insufficient govern- contracts requires management and monitoring capacity. ment commitment impeded the transfer of staff members 1318 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand Box 71.1 Cambodia's New Deal Experiment: The First Year The New Deal experiment in Sotnikum district, Siem The district (referral) hospital, health centers, and Reap province, was launched in 2000 by the Ministry of operational office were each managed by an elected man- Health, Médicins Sans Frontiéres, and the United Nations agement committee, and individual contracts were signed Children's Fund. It is an example of a concerted attempt to between staff members and management committees. The break the vicious circle of underpayment of health staff contracts stipulated that a bonus would be paid in members and underuse of public health services by tack- exchange for strict adherence to internal regulations. The ling the problem of low official income. The New Deal was benchmark bonus level was set at an average of US$60 to developed following wide consultations and consensus US$90 per person per month. The management commit- building, and locally credible management structures were tee was responsible for enforcing the new framework of established to monitor and enforce the new framework. accountability. Official fees were also introduced on the Staff motivation was a major problem among health assumption that the population would agree to pay for workers, as manifested by high levels of absenteeism from better public service. work, low time input at work (an average of one to two The district got its funding from government appropri- hours a day), and poor quality of services. Informal ations, user fees, and external subsidies from various charges, drug thefts, and dual practice by public health sources; however, given the overall lack of funding avail- workers were common, largely because of their low public able for the scheme, Médicins Sans Frontiéres and the salaries: government staff received US$10 to US$12 per United Nations Children's Fund had to provide an initial month, compared with a minimum of US$100 required injection of funds to support the bonus system. for a basic standard of living. At the same time, because of At the hospital level, individuals signed contracts with informal charges and extensive use of unregulated private the management committees, and compliance with inter- services, households spent more than US$30 per inhabi- nal regulations improved. The staff was generally present, tant per year on health services, equivalent to 11 percent of fees were transparent, emergencies were attended at night, total household expenditure. patients received drugs, and informal payments were not The New Deal was seen as a vehicle for improving demanded. Use of health services increased significantly services by enhancing personal income, and its overall after the arrangement had been introduced. The number objectives were (a) to improve access to quality health of documented violations was limited, though problems care, (b) to build up the health system, and (c) to act as a were encountered in sanctioning penalties. Staff members catalyst for changes in national health policy. The princi- started receiving bonuses that gradually grew beyond the ple underlying the improvements in the personal income negotiated maximum, creating a hospital debt crisis by of public health workers was that they would better com- midyear, compounded by understaffing and underem- ply with internal regulations governing (a) job descrip- ployment problems, which meant that most staff mem- tions and working hours; (b) payment of informal fees; bers worked overtime. Nevertheless, the quality of services (c) misappropriation of drugs, materials, and funds; and improved significantly, and per capita family expenditures (d) diversion of patients to private practice. on health fell 40 percent. Source: Soeters and Griffiths 2003. from the public service, leading to disillusionment among likely to encourage private practice among public workers. workers and effective opposition from unions (England 2000). Thus, the effects of methods and levels of payments are influ- In countries with thriving private sectors, devising strong enced by what is happening in the private sector. incentives for public sector workers is difficult. For instance, in Uganda, the private not-for-profit sector used to have better working conditions and pay than the public sector and conse- Optimal Combination of Health Worker Compensation quently had better staffing levels. The government had to and Incentives increase public sector salaries significantly in the 1990s to Although the optimal mix of provider compensation depends attract health workers back. The use of fees for service in the on context and policy objectives, some general policy guidelines private sector when public health workers are paid a salary is on the design of payment methods to achieve organizational Health Workers: Building and Motivating the Workforce | 1319 and system goals are available. Linking compensation to per- ADVICE FOR GOVERNMENTS formance makes intuitive sense, but care is needed in working out the details. Health workers respond to both financial and Governments in developing countries face huge challenges in nonfinancial incentives, but the extent of the effect varies, and strengthening their health systems, especially their HR capacity, the two can interact. if cost-effective disease control interventions are to achieve For new payment systems to work well, health workers their desired results. Strengthening their systems will entail must be governed by effective managerial authority. Because developing self-sustaining systems for the supply, use, and new payment systems aim to encourage particular behaviors retention of health workers. The following considerations are and hold providers accountable, clear responsibility must be important in relation to putting effective policies and incentive delineated within provider organizations. This delineation structures in place: may be easier to achieve if the management of providers has · Countries should explicitly link the planned number of each some autonomy. Evidence from developing countries that category of staff members to health policy goals and set pri- have attempted to introduce managerial autonomy and corpo- orities, taking overall resources into account when planning ratization of health service institutions, such as public hospi- HR needs. tals and medical stores, indicates that delinking health workers · Countries should recognize that the salaries necessary to from government control is politically sensitive. Nevertheless, recruit and retain staff members will depend on the oppor- such organizational or system changes are desirable if new tunities such workers have for other employment within the payment methods are to create the right incentives and country and abroad, and planned numbers in each category achieve the desired changes. should be based on this reality. Part of the context for incentive systems is what type of dis- · Countries should understand both that qualifications that ease control activities are best provided through markets or are recognized internationally are likely to attract higher hierarchies. Traditionally, the public sector has been domi- salaries and that such qualifications may only be partially nant. The economic arguments for government involvement suited to the needs of essential health services in LMICs. are well understood, but delivery of services within the frame- They should focus on developing the most important skills work of government policy objectives can be by private (both by training new types of health workers, taking into account for-profit and not-for-profit) providers. Thus, the private sec- evidence that use of such health workers can be safe when tor is increasingly involved in the social marketing of condoms properly trained. Many countries will be unable to prevent and bednets, franchising, and contracting (Bennett, McPake, the loss of professionals with portable qualifications, because and Mills 1997). salaries offered will be far below those available elsewhere. From an economic viewpoint the only issues are the cost, · Countries' training policies should take into account the quality, and sustainability of such arrangements. Emerging decline in skills over time and the need to allocate scarce evidence on private sector involvement in health services sug- resources between basic training and continuing staff gests that the private sector is willing to participate in non- development. clinical disease control activities if the incentive structure is · Countries should adapt and not imitate compensation and right. Private not-for-profit providers, such as hospitals and incentive structures, given the evidence that effective incen- clinics associated with churches, have traditionally comple- tive structures depend on local conditions and traditions as mented government health care activities, especially in poor well as on universal principles. and peripheral populations (Gilson and others 1997). In · Policy makers should remember that the availability and recent years, Bangladesh has experimented with contracting cost of suitably qualified human resources will affect feasi- nongovernmental organizations to provide primary care serv- bility and cost-effectiveness of disease control interventions. ices in urban areas. Lessons from this experience are still · When developing vertical disease control programs, program emerging and indicate that, despite many early mistakes, managers must avoid introducing powerful incentives that this form of provision can be innovative and can help make damage existing services by drawing away key personnel. a break from bureaucratic traditions. Such contracting · Policy makers should identify potentially harmful, unin- depends on having contracting skills in both parties to the tended consequences when designing regulation and incen- contract. A good understanding of context and incentives is tive systems. For example, if doctors are allowed to practice also crucial. in both public and private services, the effects of private In summary, incentive or payment packages should attempt practice on incentives in public practice tend to be negative to link payment with individual or group performance and unless carefully monitored. should be assisted by supportive organizational and system · Countries should recognize that the use of incentives to changes if the desired provider behavior is to be achieved. No improve performance normally requires good regulatory single best combination of payment methods exists. frameworks and skilled managerial resources. 1320 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand RESEARCH AND DEVELOPMENT AGENDA in resource-constrained environments and on what is needed to retain professionals in such settings. We know New staff categories are emerging in many LMICs, and these little about how health care workers make decisions about a workers are an important part of the workforce. Such staff cat- range of incentives and disincentives generated by organiza- egories are likely to increase, given migration and the high cost tions and the systems in which they work. For example, what of employing people with portable qualifications, but little does it take to convince doctors and nurses to work in rural research is available on the appropriateness and safety of the and remote parts of a country? To what extent are financial new sets of skills, and little is known about the range of new and nonfinancial incentives important in attracting people professions, the content of and approach to training, the extent into training as health workers, deploying them to needy of professional supervision, and the outcomes of treatment. areas, motivating them, and retaining them in the system? Sharing experience of such staff categories would be valuable. Priorities, therefore, include a study to map the different new To a significant extent, current problems in improving staff groups in health systems in LMICs and to classify their access to care, in widening the range of effective services that tasks, roles, and training, and studies to compare the outcomes are provided, and in improving the quality of care depend on of conventional and new staff groups. better matches of skills to needs, better motivation of staff, and In addition to gaining a better understanding of the pat- clearer understanding of how improved structures and incen- terns, roles, and performance of new staff groups, data are tives will work. Perhaps as important is that much of the debate needed on the length of time such workers remain in their focuses on developments within traditional patterns of staffing posts, the extent to which their new qualifications are portable, of services, but new patterns are increasingly emerging, and the and their migration patterns. Information is also lacking on extent of evaluative research is inadequate for drawing strong how best to provide professional supervision for these new staff conclusions on how such developments can alleviate the con- groups and how to encourage such employees to be profes- straints facing health systems. The development of incentive sional in their work. systems should be coupled with the development of organiza- Limited evidence is available on the relationship of different tional and institutional capacity that supports sustainable HR health care compensation methods to individual and organiza- development in general. tional behavior in developing countries. The following are pos- sible research areas (and some practical steps) that might help fill information gaps and further understanding of the role of REFERENCES health worker compensation and incentives in disease control in developing countries: Adams, O., and V. Hicks. 2000. "Pay and Non-pay Incentives, Performance, and Motivation." Paper prepared for the World Health Organization's December 2000 Global Health Workforce Strategy Group, World · Databases. A useful step would be to set up HR databases Health Organization, Department of Organization of Health Services for developing countries as the Pan American Health Delivery, Geneva. Organization has done for its region. Awases, M., A. Gbary, and R. Chatora. 2003. Migration of Health · Literature review. A review of unpublished materials on Professionals in Six Countries: A Synthesis Report. Brazzaville: World Health Organization, Regional Office for Africa. countries' experiences with using different payment and Bach, S. 2000. "Human Resources and New Approaches to Public Sector compensation mechanisms at national or subnational levels Management: Improving Human Resource Management (HRM) would also be useful. Failed experiments are seldom pub- Capacity." Paper prepared for the World Health Organization's lished, but they provide useful lessons. December 2000 Global Health Workforce Strategy Group, World Health Organization, Department of Organization of Health Services · HR supply. Traditional HR planning models are no longer Delivery, Geneva. effective in handling health system dynamics in developing Balabanova, D., and M. McKee. 2003. "Understanding Informal Payments countries. More research is required to develop HR models for Health Care: The Example of Bulgaria." Health Policy 62 (3): in health that include the effects of HIV/AIDS, migration, 243­73. scaling-up of existing interventions, new technology, and Barnum, H., J. Kutzin, and H. Saxexian. 1995. "Incentives and Provider Payment Methods." International Journal of Health Planning and reforms. The underlying question should be how HR supply Management 10 (1): 23­45. mechanisms can meet health systems' needs in terms of Beith, A., R. Eichler, J. Sanderson, and D. Weil. 2001. "Can Incentives and numbers, knowledge, skills mix, and competencies. Enablers Improve Performance of Tuberculosis Control Programs? · Demand and utilization. Getting the size of the health Analytical Framework, Catalogue of Experiences and Literature workforce right is important in its own right, but that Review." Unpublished paper, Management Sciences for Health, Rational Pharmaceutical Management Project Plus, and Stop alone is insufficient for improving health workers' motiva- Tuberculosis Partnership, November, 2004. tion and productivity. Research needs to focus on how to Bennett, S., P. McPake, and A. Mills, eds. 1997. Private Health Providers in improve the motivation and performance of health workers Developing Countries: Serving the Public Interest. London: Zed Books. Health Workers: Building and Motivating the Workforce | 1321 Berman, P. A., A. K. Nandakumar, J. J. Frere, H. Salah, M. El-Edawy, Kyaddondo, D., and S. R. White. 2003. "Working in a Decentralized S. El-Saharty, and N. Nassar. 1997. A Reform Strategy for Primary Care System: A Threat to Health Workers' Respect and Survival in Uganda." in Egypt. Technical Report 9. Bethesda, MD: Partnership for Health International Journal of Health Planning and Management 18 (4): Research, Abt Associates. 329­42. Bhat, R. 1996. "Regulation of the Private Health Sector in India." Maceira, D. 1998. "Provider Payment Mechanisms in Health Care: International Journal of Health Planning and Management 11 (3): Incentives, Outcomes, and Organizational Impact in Developing 253­74. Countries." Major Applied Research 2, Working Paper 2, Partnership Bir, A., and K. Eggleston. 2003. "Physician Dual Practice: Access for Health Reform, Abt Associates, Bethesda, MD. Enhancement or Demand Inducement." Working Paper, Tufts Martineau, T., and J. Buchan. 2000. Three Diverse Case Studies on the University, Department of Economics, Medford, MA. Importance of Human Resources to Successful Health System Reforms. Bitran, R., and W. C. Yip. 1998. "A Review of Provider Payment Reform in Washington, DC: American Public Health Association. Selected Countries in Asia and Latin America." Major Applied Matomela, N. 2004. "Department Outlines Amended Dental Act." Research 2, Working Paper 1, Partnership for Health Reform, Abt BuaNews Pretoria, October 7. Associates, Bethesda, MD. McAuliffe, M. S., and B. Henry. 1995. Nurse Anaesthesia Worldwide: An Bloom, G., L. Han, and X. Li. 2001. "How Health Workers Earn a Living in Analysis of Practice, Education, and Legislation. Geneva: World Health China." Human Resources for Development Journal 5 (1­3): 25­38. Organization. Buchan, J. M. D., and M. R. Dal Poz. 2003. "Role Definition, Skill Mix, McPake, B., D. Asiimwe, F. Mwesigye, A. Turinde, M. Ofumbi, Multi-Skilling, and `New Workers.'" In Towards a Global Workforce L. Ortenblad, and P. Streefland. 1999. "Survival Strategies of Public Strategy: Studies in Health Services Organization and Policy, vol. 21, ed. Health Workers in Uganda: Implications for Quality and Accessibility P. Ferriho and M. Dal Poz, 275­300. Antwerp, Belgium: ITG Press. of Care." Social Science and Medicine 49 (7): 849­65. Chakraborty, S., R. Gatti, J. Klugman, and G. Gray-Molina. 2002. "When Is Normand, C., and C. Thompson. 2000. "Review of the Primary Care `Free' Not So Free? Informal Payments for Basic Health Services in Rehabilitation Project in Azerbaijan." Report prepared for the United Bolivia." Unpublished paper, World Bank, Washington, DC. Nations Children's Fund. Chimbari, M. J. 2003. "A Report on Health Care Providers in Zimbabwe." PAHO (Pan American Health Organization). 2003. Bangladesh Health Disease Control Priorities Project Working Paper. Labor Market Study. Washington, DC: PAHO. Corkery, J. 2000. Public Service Reforms and Their Impact on Health Sector Pannarunothai, S., D. Boonpadung, and S. Kittidilokkul. 2001. "Paying Personnel in Uganda. Geneva: International Labour Organization and Health Personnel in the Government Sector by Fee-for-Service: A World Health Organization. Challenge to Productivity and Quality, and a Moral Hazard." Human England, R. 2000. "Health Sector Reform Experiences: Lessons for Belize Resources for Health Development (electronic journal) 1 (2) from Trinidad and Tobago?" Issues Note, Institute for Health Sector Phillips, V., A. M. Gray, D. Hermans, and C. Normand. 1994. "Health and Development, London. Social Service Manpower in the U.K.: A Review of the Research Fenton, P. M., C. J. Whitty, and F. Reynolds. 2003. "Caesarean Section in 1986­1992." Public Health and Policy Department Publication 7, Malawi: Prospective Study of Early Maternal and Perinatal Mortality." London School of Hygiene and Tropical Medicine, London. British Medical Journal 327 (7415): 587. Rasheed, S. 1995. "Ethics and Accountability in the African Civil Service." Gilson, L., J. Adusei, D. Arhin, C. Hongoro, P. Mujinja, and K. Sagoe. 1997. DPMN Bulletin 3 (1): 12­14. "Should Governments Contract out Clinical Health Services to Ratto, M., C. Propper, and S. Burgess. 2002. "Using Financial Incentives to Church Hospitals?" In Private Health Provider in Developing Countries: Promote Teamwork in Health Care." Journal of Health Services Serving the Public Interest, ed. S. Bennett, B. McPake, and A. Mills, Research Policy 7 (2): 69­70. 276­302. London: Zed Books. Soeters, S., and S. Griffiths. 2003. "Improving Government Services Gray, A., and V. L. Phillips. 1996. "Labour Turnover in the British National through Contract Management: A Case from Cambodia." Health Health Service: A Local Labour Market Analysis." Health Policy 36 (3): Policy and Planning 18 (1): 74­83. 273­89. Support for Analysis and Research in Africa. 2003. "The Health Sector Hicks, C., and O. Adams. 2001."Pay and Non-pay Incentives, Performance, Human Resources Crisis: An Issues Paper." Academy for Educational and Motivation." Paper prepared for the World Health Organization's Development, Washington, DC. December 2001 Global Health Workforce Strategy Group, World Thompson, R., and S. Witter. 2000. "Informal Payments in Transitional Health Organization, Geneva. Economies: Implications for Health Sector Reform." International Jha, P., and A. Mills. 2002. Improving Health Outcomes of the Poor: Report Journal of Health Planning and Management 15 (3): 169­87. on Working Group 5 of the Commission of Macroeconomics and Health. Venning, P., A. Durie, M. Roland, C. Roberts, and B. Leese. 2000. Geneva: World Health Organization. "Randomised Controlled Trial Comparing Cost-Effectiveness of Jinadu, M. K., E. O. Olofeitime, and P. Oribador. 2002. "Evaluation of an General Practitioners and Nurse Practitioners in Primary Care." British Innovative Approach to Community-Based Medical Undergraduate Medical Journal 320 (7241): 1048­53. Education in Nigeria." Education for Health 15 (2): 139­48. WHO (World Health Organization). 2000. The World Health Report Killingsworth, J. R., N. Hossain, Y. Hedrick-Wong, S. D. Thomas, A. 2000--Health Systems: Improving Performance. Geneva: WHO. Rahman, and T. Begum. 1999. "Unofficial Fees in Bangladesh: Price, Wibulpolprasert, S., and P. Pengpaiboon. 2003. "Integrated Strategies to Equity, and Organizational Issues." Health Policy and Planning 14 (2): Tackle the Inequitable Distribution of Doctors in Thailand: Four 152­63. Decades of Experience." Human Resources for Health 1: 12. Kurowski, C., K. Wyss, S. Abdulla, N. Yémadji, and A. Mills. 2003. "Human Zurn, P. 2003. "Incentives for Human Resource Management." Paper pre- Resources for Health: Requirements and Availability in the Context of sented at the Workshop on Human Resource for Health Development: Scaling-Up Priority Interventions in Low-Income Countries--Case The Joint Learning Initiative, Veyrier-du-Lac, France, May 8­10. Studies from Tanzania and Chad." London School of Hygiene and Tropical Medicine, London. 1322 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand Chapter 72 Ensuring Supplies of Appropriate Drugs and Vaccines Susan Foster, Richard Laing, Bjørn Melgaard, and Michel Zaffran In 1988, the World Health Organization (WHO) reported that America, Europe, and Japan. Lipitor (atorvastatin), a 30 percent of the world's population, some 1.725 billion people, cholesterol-reducing drug and the world's best-selling drug in lacked regular access to essential medicines. By 1999, the 15 per- 2002, had sales of US$8.6 billion, and growth was 20 percent cent of the population who lived in high-income countries pur- annually. Zocor (simvastatin), another cholesterol reducer and chased and consumed 90 percent of all medicines, by value the second-best seller, had sales of US$6.2 billion and was grow- (WHO 2004f). Again as of 1999, a recent WHO report estimates ing at 13 percent (IMS Global Learning Consortium 2003). that 30 percent of the world's population, including 47 percent These figures contrast with the dearth of research on neglected of Africans, 65 percent of people in India, 29 percent of people diseases prevalent in developing countries. For example, of the in the Eastern Mediterranean, and 26 percent of Southeast 1,325 new medicines launched between 1975 and 1997, only 11 Asians (excluding those from India), had no access to essential were specifically for tropical diseases (Trouiller and others medicines (WHO 2004f). So although access has significantly 2002). Médecins Sans Frontières (MSF) in 1999 initiated its improved in a number of countries, a large fraction of the advocacy program, Drugs for Neglected Diseases, which has world's population still has no effective access to modern medi- highlighted this gap (http://www.accessmed-msf.org/). cines or vaccines. The majority of these people are either Access to effective medicines and vaccines requires a com- extremely poor or are living in remote rural areas where the sup- plex and coordinated system. It must encompass production ply of drugs is limited or nonexistent--or both. that ensures good quality, selection, procurement, and distri- Many diseases can be effectively treated, managed, or pre- bution; correct prescription and dispensing and correct use by vented with pharmaceuticals and vaccines. The WHO figure of patients; adequate financing; and effective monitoring of the 30 percent of the world's population lacking access understates system. Multiple delivery systems involving public, private, and the reality; even within countries with apparently good serv- nongovernmental organization (NGO) sectors frequently ices, some populations lack access. Similarly, immunization coexist, and patients are very likely to use multiple systems to coverage globally has remained static for more than a decade at access these products. about 75 percent of children fully immunized, with about 27 million children born every year with no access to immu- nization services. Some effective vaccines, such as hepatitis B DRUG POLICIES (HepB), are still not in routine use in many countries. Medicines and vaccines are developed as a result of innova- In any country, many stakeholders are interested in the national tion by researchers and pharmaceutical companies. The global policy on pharmaceuticals. In broad terms, they can be charac- pharmaceutical market was worth more than US$400 billion terized as producers, importers, distributors, prescribers, in 2004, and more than 80 percent of this market is in North finance providers, and consumers. Each has a different set of 1323 interests, which in some cases are contradictory and in other vidual, because vaccination prevents transmission. It can cases congruent. To reconcile these disparate interests, many also be argued that the elimination or eradication of a dis- countries have developed a national drug policy. Managers of ease as a public health problem has public good characteris- disease control programs need to be involved in these tics: the benefits of the absence of disease are available to discussions at an early stage to prevent policy decisions from everyone, and all persons benefit at the same time. adversely affecting their programs. Any national drug policy Therefore, governments must take an active role in ensuring broadly relates to three key objectives: increasing access, that adequate vaccines of assured quality are available for improving and ensuring quality, and ensuring rational pre- comprehensive immunization programs within the country. scription and use by providers and patients. The recent flu vaccine shortage and resulting rationing The primary components of a drug policy are selection of problems in the United States have illustrated this principle essential medicines; assurance of affordability, which includes clearly. issues of pricing, taxation, generic competition, and policies · Countries should strive toward financial sustainability for related to the Agreement on Trade-Related Aspects of the national immunization program. Intellectual Property Rights (WHO 2001a); financing options; supply systems; regulation and quality assurance; rational use; A vaccine policy normally has six specific objectives: operational research and drug development; clinical research, including clinical trials; human resource development for · To provide a coordinated approach to national vaccines and pharmaceutical policy and program management; and moni- equipment needs, including national vaccine production toring and evaluation. WHO has developed manuals and has where applicable provided technical support to countries to develop such · To provide criteria for vaccine selection and introduction, national policies (WHO 2001b, 2003a). including burden-of-disease studies where relevant · To develop a financial sustainability plan that ensures avail- ability of vaccines in the longer term VACCINE POLICIES · To define guidelines for private-public partnerships, includ- ing vaccine research Every country has a national vaccine policy, usually laid down · To define national research priorities in a national health policy or through the establishment of · To support the implementation of the national immuniza- well-defined elements of such a policy. WHO defines global tion programs. frameworks and produces policy documents to advise develop- ing countries (WHO 2002b). At the national level these guide- Policy setting is a continuous process that must keep up lines may be adapted to fit national needs and capacities. with global developments and changing national needs. Standards and norms for vaccines are also set by WHO and are Countries will normally formulate policies that are based on generally adhered to worldwide (WHO 2003b). the technical work of a national committee of experts, who WHO's creation of the Expanded Program on Immunization meet regularly under the auspices of the ministry of health. (EPI) in the 1970s established a policy for selection and use Bilateral donors and the Global Alliance for Vaccines and of vaccines that the vast majority of developing countries Immunization (GAVI) may influence policy setting, such as the adopted. Only three vaccines--HepB, yellow fever (YF), and timing for introduction of new vaccines, where they contribute Haemophilus influenzae type B (Hib)--have been added since significantly to the national immunization programs. then, and the overall program directions remain largely intact. In recent years, WHO has published a set of policy guide- SELECTION OF DRUGS lines for vaccines not included in the global recommendations. These position papers are regularly updated. Three guiding Selection of a limited list of essential medicines that should principles provide the pillars for any national vaccine policy: always be available is necessary both for supply officials who work on procurement, storage, and distribution and for clini- · Every eligible child must have equal access to nationally cians who aim to use medicines most effectively. adopted vaccines regardless of religion, caste, or economic In 1977, WHO defined the first Model List of Essential status. Drugs (WHO 1977); since then it has updated the list 14 times. · Vaccines require active government financial participation The latest list defines essential medicine as follows: to ensure that they are provided and used in adequate quantities, thus ensuring the benefit of their considerable Essential medicines are those that satisfy the priority health externalities. For example, the benefits to society of an indi- care needs of the population. They are selected with due regard vidual's being vaccinated are greater than those to the indi- to public health relevance, evidence on efficacy and safety, and 1324 | Disease Control Priorities in Developing Countries | Susan Foster, Richard Laing, Bjørn Melgaard, and others comparative cost effectiveness. Essential medicines are selection of treatments would ideally have already been under- intended to be available within the context of functioning taken on the basis of available evidence or clinical trial data health systems at all times in adequate amounts, in the appro- from the country. The medicines identified within these guide- priate dosage forms, with assured quality and adequate infor- lines would thus become the medicines on the essential mation, and at a price the individual and the community can medicines list. This list would then serve as the basis for pro- afford. The implementation of the concept of essential medi- curement, storage, and distribution activities. The evidence- cines is intended to be flexible and adaptable to many different based treatment guidelines would define treatment practices situations; exactly which medicines are regarded as essential and be the basis of training (including examinations and remains a national responsibility. licensing) and supervision. In 2003, the WHO Expert Committee on Selection and Use of Essential Medicines decided to define the criteria for core and complementary lists, as follows: SELECTION OF VACCINES The core list presents a list of minimum medicine needs for a Developing countries select vaccines used in national immu- basic health care system, listing the most efficacious, safe, and nization programs primarily on the basis of WHO policy cost-effective medicine for priority conditions. Priority condi- guidelines. Most countries have adhered strictly to the six tions are selected on the basis of current and estimated future original vaccines--Bacillus Calmette-Guérin (BCG), oral public health relevance, and potential for safe and cost-effective polio, diphtheria, pertussis, tetanus, and measles. On the treatment. recommendation of the WHO Global Advisory Group on The complementary list presents essential medicines for Immunization, HepB vaccine was included in the global priority diseases, for which specialized diagnostic or monitor- guidelines in 1987, and YF vaccine was added in 1988; Hib ing facilities, or specialist medical care, or specialist training, or vaccine was added in 1994. These remain the only vaccines all three are needed. In case of doubt, medicines may also be listed as complementary because of their consistently higher recommended by the WHO for national use, and the recom- costs or less attractive cost effectiveness in a variety of settings mendation presupposes that a disease burden of public health (WHO 2003d, 28). importance is present (see table 72.1). A few vaccines, such as YF and Japanese encephalitis (JE), At its 2002 meeting, the WHO Expert Committee changed have regional importance in accordance with the prevalence of its attitude toward fixed-dose combinations (FDCs). The com- the disease. These vaccines are used in only a small number of mittee stated that most essential medicines should be formu- developing countries. WHO has not generally recommended lated as single compounds. Fixed dose combination products JE, although no evidence indicates that the disease burden of are selected only when the combination has a proven advantage over single compounds administered separately in therapeutic effect, safety, adherence or in delaying the development of drug Table 72.1 Current Vaccines Recommended by the World resistance in malaria, tuberculosis and HIV/AIDS (WHO Health Organization 2002c). This change reflected the interest in preventing the Age development of resistance and in promoting adherence. Although controversial, these FDC products will very likely be Vaccine Birth 6 weeks 10 weeks 14 weeks 9 months the main form of treatment for AIDS, tuberculosis (TB), and BCG X malaria. Oral polio Xa X X X The number of medicines on the WHO list has increased Diphtheria- X X X over time. The 2003 WHO list has 320 drugs in 559 formula- tetanus- tions (Laing and others 2003). At country level, the essential pertussis drugs list is used as a guide rather than as a template. A study Hepatitis Bb X X X of 17 national lists of essential drugs showed that 68 percent Haemophilus X X X had fewer than 300 drugs. The number of drugs per list ranged influenzae type Bb from 108 in Liberia to 389 in Karnataka state, India. Nine of the Yellow fever Xc drugs on the WHO list were not on any of the 17 national drug Measles Xd lists in the study (Laing and others 2003). Source: WHO 2002d, 88. At the first stage of identifying common diseases and com- a. In endemic countries. plaints, managers of disease control programs are in a strong b. Only a few African countries have been able to introduce the vaccines to date. position to provide epidemiological information about the c. In countries where yellow fever poses a risk. d. In addition, a second opportunity to receive a dose of measles vaccine should be provided for incidence or prevalence of a condition. At the second stage, the all children. Ensuring Supplies of Appropriate Drugs and Vaccines | 1325 YF is greater than the JE disease burden. However, a safety issue procurement were recommended under different circum- exists with mouse brain­derived vaccines such as JE. stances. These methods were open tender, restricted tender Oral polio vaccine (OPV) is the vaccine of choice in devel- with performance monitoring, negotiated procurement, and oping countries, because it is easy to administer and the pro- direct procurement. At that time, the World Bank was a major tective effect spreads to close contacts of vaccinees. It is suitable funder, and many countries favored the use of open tender. for mass campaigns, so the vaccine is used in poliomyelitis However, a major shift has taken place to restricted tender eradication programs. In 2005, the monovalent type 1 OPV, for based on prequalification and direct procurement from non- which seroconversion rates are substantially higher than for profit suppliers. The World Bank has produced a number of trivalent OPV, has started to be used in areas where only type 1 useful documents and resource materials that can be used for wild poliovirus remains in circulation. national procurement activities (World Bank 2000). The selection of the original EPI vaccines was made on pro- Another method of procurement that has been more grammatic criteria rather than on considerations of disease widely used has been procurement from nonprofit suppliers, burden. The need for consistent and standardized regimens such as the United Nations Children's Fund (UNICEF) or determined the strategies selected by EPI. Adaptations over the International Dispensary Association (International time, as new vaccines came along and local needs changed, Dispensary Association 2004; UNICEF 2004). These organiza- were slow, and the uptake of newer vaccines remains a major tions produce price lists twice yearly, and products can be constraint in most developing countries, although support ordered directly. Management Sciences for Health (2004) provided through GAVI has improved the situation. publishes an international drug price indicator guide annually The term vaccine gap is used to describe the phenomenon that reports these prices and other procurement prices. whereby children in developing countries receive at most eight Interestingly, the trend in drug prices generally has been down- vaccines, if they are reached by immunization programs at all, ward. Prices of some TB drugs fell by more than 90 percent whereas children in industrial countries normally receive 10 to when procurement managers opened intensive negotiations 12 antigens, depending on national schedules. Furthermore, with suppliers. And since 2000, prices of many important first- two vaccines in the routine schedule in the affluent world are line antiretroviral drugs have fallen considerably. This trend is less reactogenic than those given in developing countries: acel- attributable to "advocacy, corporate responsiveness, competi- lular pertussis vaccine (aP) and inactivated polio vaccine tion from generic manufacturers, sustained public pressure, (IPV). The combination measles-mumps-rubella (MMR) vac- and the growing political attention paid to the AIDS epidemic. cine is normally given twice, offering long-term protection In addition, originator companies began announcing discount against measles and rubella, important diseases in the develop- offers for the benefit of the poorest countries or those where ing world. The vaccine gap therefore consists mainly of three the HIV/AIDS prevalence is the highest" (WHO 2004d, 5). The vaccines--aP, IPV, and MMR. influence of economies of scale, in which unit costs have fallen The pneumococcal vaccine, which is currently used in some because of the increased consumption of needed pharmaceuti- countries, illustrates an additional aspect of the vaccine gap. Its cals, might also have played a role. composition is directed against the most prevalent strains, Generic drugs obtained in bulk are almost always 10 or which cause otitis media in children. At the same time, millions more times less expensive than brand-name drugs. Bulk pur- of children die in the developing world from pneumonia chase of generic drugs is the single best way to make a given caused by other strains of the bacteria, but no vaccine is cur- budget go farther to satisfy the drug needs of a population. For rently available against those strains. price comparison, international prices (ex-factory, free- on-board--that is, not including insurance, freight charges, or taxes) are available online from Management Sciences for PROCUREMENT OF DRUGS Health (2004). Local prices must, of course, include trans- portation and freight, as well as any applicable local taxes. After the drugs have been selected, the next step is to decide Organizations that procure drugs in bulk but then sell smaller how much to order. Usually this decision is based in part on quantities, usually to nonprofit organizations, can help smaller past consumption, but it is also based on treatment guidelines purchasers obtain the advantages of competitive tendering. and morbidity patterns. Concentrating on larger quantities of Examples include the International Dispensary Association, fewer drugs and dosage forms simplifies the process of order- which is based in the Netherlands, and the Copenhagen office ing and reduces the chances of running out of stock. Ordering of UNICEF, which is able to supply drugs at very low cost to tablets or capsules rather than syrups or injections saves a great government-supported institutions. deal of money. Recent studies have revealed just how much the local In 1997, when the second edition of Managing Drug Supply component of drug costs can be, particularly in the private was published (Quick and others 1997), four methods of sector. A survey of costs in nine countries found an average 1326 | Disease Control Priorities in Developing Countries | Susan Foster, Richard Laing, Bjørn Melgaard, and others markup of 68.6 percent, with retail markups of 16 to 50 percent ranging from 20 percent to 67 percent for chloroquine tablet comprising the largest single component in most cases. In (CQT) and 5 percent to 38 percent for sulphadoxine/ countries that charge a value added tax on drugs, the tax can pyrimethamine tablet (SPT) and dissolution failures ranging add 15 to 20 percent to the price of the drug (Levison and Laing from 5 percent to 29 percent for CQT, and 75 percent to 100 per- 2003). Many of these cost elements are within the control of cent for SPT (WHO 2003c). Good procurement practices of national policy makers. both brand-name and generic drugs require that suppliers be Finally, when considering a change of standard therapy, prequalified through the inspection of dossiers and factory managers of disease control programs need to take into inspections for good manufacturing practice (GMP) and that account the long lead time between ordering a particular drug their performance be monitored. and having it arrive in the country ready for use--which can be Counterfeit medicines are a particularly difficult problem. a year or more. Time will also be required to prepare, print, and Counterfeit medicines "are deliberately and fraudulently misla- disseminate new guidelines, to train prescribers and dispensers, beled with respect to identity and/or source. Counterfeiting can and to dispose of drugs used in the older therapy (Williams, apply to both branded and generic products, and counterfeit Durrheim, and Shretta 2004). medicines may include products with the correct ingredients but fake packaging, with the wrong ingredients, without active ingredients, or with insufficient active ingredients" (WHO PROCUREMENT OF VACCINES 2005b, 1). In industrial countries, the newer brand-name med- icines are counterfeited most often; Viagra is the frequent sub- Countries can be grouped into three categories according to ject of counterfeiters. In poorer developing countries, the most the way they procure vaccines: procurement through United commonly used antimalarials, antibiotics, and now antiretrovi- Nations (UN) agencies, direct procurement, and local produc- rals are the targets of the counterfeiters. The U.S. Food and tion. Some countries procure their vaccines from a range of Drug Administration estimates that up to 25 percent of medi- sources and may cut across all three categories. Over the past 25 cines in developing countries are either counterfeit or other- years, UNICEF has been the main bulk procurer of traditional wise substandard and that earnings from counterfeit drugs vaccines for most of the developing world, with the Pan are more than US$32 billion per year (WHO 2005b). American Health Organization Revolving Fund for Vaccine Procurement serving most Latin American countries. Because the fund takes advantage of large volume purchasing, it obtains THE WHO PREQUALIFICATION SCHEMES prices comparable to those of UNICEF, which are available to all participating countries--regardless of their income level or Because undertaking prequalification tasks may be beyond the size. The Gulf Cooperation Council also operates a purchasing capabilities of national authorities, WHO has, on behalf of all program for its member states. That program includes 43 dif- UN agencies, started a prequalification scheme (WHO 2004d) ferent vaccines and sera. covering AIDS, TB, and artemisinin-containing malaria drugs. Some countries, where governments take on an increasing The prequalification process is rigorous but efficient. WHO share of vaccine financing, purchase the vaccines directly from provides a positive list of prequalified products and manufac- the producers or their representatives. Unfortunately, procure- turers that have applied for and received favorable product ment is often being undertaken with little recognition that assessments and manufacturing site inspections. Since January stringent quality assurance procedures must be in place to 2005, the Global Fund to Fight AIDS, Tuberculosis, and oversee the entire process. WHO is organizing workshops Malaria has required recipients to use WHO-prequalified specifically targeted at vaccine procurement and has developed products. a vaccine procurement manual to guide such countries (WHO Ensuring quality is also an important aspect of any immu- 2005a). nization program. For countries receiving their vaccines through UN agencies, WHO advises on the quality, efficacy, and safety of vaccines on the market through a prequalification QUALITY ASSURANCE FOR PHARMACEUTICALS of vaccines that entails the following steps: (a) reliance on a AND VACCINES fully functional national regulatory authority (NRA) in the country of production and (b) verification of compliance with In an ideal world, all products to be imported into a country specifications through a thorough process of independent would be registered by a fully competent national drug regula- dossier reviews, testing of samples, site visits, ongoing moni- tory authority to ensure quality. Unfortunately, this situation is toring of quality, and follow-up of complaints. not always the case. A study of antimalarial samples from seven For a successful prequalification process, the NRA of the African countries found that failures in ingredient content country of production must be functional and empowered by Ensuring Supplies of Appropriate Drugs and Vaccines | 1327 the government. A set of laws and structures must be in vertically to some extent, and because they were concerned place that guarantee the NRA's authority and independence with a far more limited range of products, the task was some- and that the NRA exercises the following functions: licensing, what simpler. postmarketing surveillance, lot release, laboratory access, GMP A push has been made to integrate the distribution systems inspections, and evaluation of clinical performance. These for drugs, vaccines, and contraceptives, although in most places functions constitute the prerequisite for vaccines of assured separate systems are still operating, at least at the national level quality and are the focus in vaccine regulation. and often down to provincial levels. Vian and Bates (2003) noted a number of changes to the distribution systems in the LOCAL PRODUCTION OF PHARMACEUTICALS past few years. In many countries, health sector reform pro- grams included efforts to reform central medical stores to allow Large-scale production of pharmaceuticals in the developing more autonomy and to introduce commercial incentives and world is limited to a few larger countries, most of which export improved management methods. In some cases, this reform primarily to other developing countries. has led to higher staff productivity, better performance, and Whether local production of pharmaceuticals should be more enforcement of payment policies. However, disruption in encouraged in low- and middle-income countries is a contro- supply often occurs during central medical store transition versial issue. During the 1980s and early 1990s, the United phases. Increased integration of commodities, including con- Nations Industrial Development Organization encouraged the traceptives and vaccines, has also been noted. In some cases, it establishment of national production facilities. Recently, the has decreased the amount and reliability of data collected on World Bank and the executive board of WHO have reviewed logistics, creating problems for needs estimation and for track- this issue (Kaplan and others 2003; WHO 2004a). The more ing of consumption (Vian and Bates 2003). extensive World Bank report concluded: Another trend is the increasing use of private transporters and contracting out for transportation management; contract- In many parts of the world, there is no reason to produce med- ing transport can generate cost savings and improve services. icines domestically, since it makes little economic sense. In the Finally, a trend toward computerized systems exists, particu- local pharmaceutical manufacturing sector, local production is larly involving the use of donor-financed software for often not reliable and, even if reliable, it does not necessarily improved management of logistics as well as a number of com- mean that medicine prices are reduced for the end user. If local prehensive assessment tools and indicator sets for evaluating production is adopted by many countries, it may lead to less access to medicines, since there are no economies of scale drug supply systems. But the proliferation of software systems, in having a production facility in each country" (Kaplan and with little coordination and not enough support and mainte- others 2003). nance of complex and fragile computer systems, can be coun- terproductive, especially if paper-based systems that are Profit margins on bulk generic drugs are low so that public difficult to reintroduce upon failure of the computer system are production must be as efficient as private manufacturing. For abandoned (Vian and Bates 2003). many countries, technical expertise, raw materials, quality stan- dards, and production and laboratory equipment all need to be imported, so foreign exchange savings may be small or nonex- Storage and Stocks Management istent. Few developing countries have the capacity to produce Drugs require secure storage in controlled climatic conditions active ingredients for pharmaceutical manufacture. Industrial and a reliable method of stock rotation. The FEFO rule (first investment to promote local manufacture of pharmaceuticals expiry, first out) helps ensure that older stock is used up first. in most, but not all, developing countries could be better used Security is another major consideration: access to the store- to improve health infrastructure (Kaplan and others 2003). house must be carefully controlled so that theft and embezzle- In summary, a manager of a disease control program is ment are minimized, and the persons who control access must likely to obtain quality-assured products by procuring them themselves be trustworthy. Proper storage conditions, includ- from prequalified suppliers at the lowest prices without con- ing minimizing exposure to heat, light, and humidity, are sidering whether the products are locally produced. important for some drugs, but most drugs have proved remarkably resistant to poor conditions. Notable exceptions are STORAGE AND DISTRIBUTION OF ESSENTIAL tetracycline products, which become toxic when exposed to DRUGS AND VACCINES heat, and oxytocin and ergometrine, which lose their potency when exposed to light and heat; all should thus be stored in the In the past, essential drugs, vaccines, and contraceptives were refrigerator. The same applies to insulin and, of course, most for the most part distributed using separate logistics systems. vaccines. Correct FEFO stock rotation will ensure that expo- For vaccines and contraceptives, such systems were organized sure to harsh conditions is minimized and that potency is 1328 | Disease Control Priorities in Developing Countries | Susan Foster, Richard Laing, Bjørn Melgaard, and others preserved as much as possible. Ensuring good air circulation At the country level, emphasis is being put on the use of new and preventing direct water contact are most important. tools, such as the vaccine vial monitor. This heat-sensitive label is a time-temperature indicator used to ensure that the vaccines have not been damaged by excessive exposure to heat, to iden- Management of Donated Drugs tify weaknesses in the cold chain, and to take vaccines beyond Management of donated drugs is a major problem in some the cold chain to children who have no access to fixed health areas, particularly if an emergency has precipitated an influx of facilities. drug donations. The best strategy is to accept only invited Together with the increased use of vaccine vial monitors, the donations of drugs that the facility has specifically asked for gradual adoption of the multidose vial policy contributes to the (WHO 1999a). Any drug that is neither vital nor essential, that reduction of wastage. This policy of using opened multidose is not labeled clearly with its generic name, that is expired, that vials of vaccine in subsequent immunization sessions applies to is in a package that contains only a few days' dosage, or that is all multidose vials of liquid vaccine containing thimerosal not on the national essential drugs list or on the facility's for- (WHO 2000). The policy was formulated in 1996 but its adop- mulary should be discarded--and the pharmacist should feel tion remains limited. no guilt and fear no sanctions about disposing of such materi- als. They take up space, require tracking like other drugs, and present a risk of being accidentally dispensed to a patient and PRESCRIPTION AND RATIONAL USE OF DRUGS causing the patient harm--a factor that must also be taken into account. Proper disposal can be a problem. These drugs consti- Rational drug use involves the correct drug being given to the tute potential toxic waste, and they should be treated as such correct patient, for the correct indication, in the correct dosage, and disposed of so that they cannot be retrieved and sold by the correct route of administration, for the correct duration (WHO 1999b). of treatment. The dispenser must also correctly dispense and label the drug and counsel the patient, and the patient must take the drug correctly or comply with or adhere to treatment. Vaccine Management An error at any stage of this complex process can prevent the Vaccines are delicate products that are destroyed if handled drug from being effective. Usually, at least half of these errors incorrectly. Vaccine management involves the use and distribu- are attributable to the failure of patients to adhere to treatment, tion of vaccines, from the manufacturers to the end users. but the other half of the errors occur before the patient actually Aspects of vaccine management include inventory and fore- begins taking the drug. Few of the recommended treatments casting, stock control, in-country distribution, storage and for common diseases involve more than one or two drugs, yet handling, equipment replacement plans, procedures for the use in actual practice, multiple drugs are often prescribed, even for of the vaccine, monitoring of vaccine storage, transport man- uncomplicated cases. Such overuse of drugs rapidly consumes agement, and operational management. stocks, does not add to the quality of care (although patients Forecasting of vaccine needs is the first building block of an may believe that more drugs are better), and allows stockpiling adequate management system. In 2002, 22 of 82 countries sur- by patients. veyed by UNICEF indicated that they had experienced a vac- The use of injections instead of equally effective oral prepa- cine stockout. In addition to lack of resources, the main reasons rations is also common. Not only are risks associated with the cited included poor or late forecasting. injections themselves, but also the cost of these injections is far In recent years, attention has focused on avoiding heat greater than for the equivalent oral preparations. If all the pre- exposure. The introduction of costly vaccines that are sensitive scribers at a facility can agree on and adhere to standard treat- to freezing has drawn attention to the need to protect vaccines ment guidelines that can be used as the basis for procurement from excessive exposure to cold as well as heat. WHO guide- and storage, the problem of overprescription and stockouts can lines for the international transport of vaccines now include gradually be eliminated. Uncertainties about dosages, particu- specific recommendations for each category of vaccine, includ- larly pediatric dosages, can also be reduced by the use of stan- ing freeze sensitivity. National cold stores are the next critical dard guidelines by age or weight. Doctors often cite their mis- level of the vaccine management system. A failure there-- trust or delay of laboratory results as a reason to "cover" the where vaccines are received, stored, and distributed in bulk-- patient for a variety of conditions. Dealing with laboratory effi- can result in extensive losses. The WHO-UNICEF Effective ciency or accuracy issues may be a worthwhile way to improve Cold Store Management Initiative encourages countries to pro- prescription practices that would also yield great benefits in cure equipment and adopt management and training practices terms of quality of care. Regularly reviewing a sample of pre- that fully protect vaccines in national and intermediate vaccine scriptions or case records and comparing treatments given to stores. the standard treatment guidelines is likely to have a dramatic Ensuring Supplies of Appropriate Drugs and Vaccines | 1329 effect on the improvement of treatment practices (Laing, exists on the advantages and disadvantages of user fees. On the Hogerzeil, and Ross-Degnan 2001). one hand, they do raise some revenue, but administrative costs have often taken a large proportion of it. Their net contribution DISPENSING has rarely exceeded 5 percent of a government's recurrent expenditure. On the other hand, they often accounted for as Finally, the last step in the chain of the drug supply system is high as 100 percent of nonsalary recurrent expenditures. delivery to the patient. Often, dispensing is done by untrained Moreover, where they have been retained at the facility level, staff members who know little about the drugs they are dis- they have allowed for improvements in infrastructure and staff pensing and are unable to communicate effectively with the income, as well as ensuring a more regular supply of drugs (Xu patient. Anecdotal stories about patients receiving a handful of and others forthcoming). white pills and throwing them on the ground are discouraging However, from a public health perspective, the disadvan- to staff members but demonstrate that patients need explana- tages are numerous. They are often applied inequitably, with tions about the drugs they are getting. Increasing the use of exemptions provided to richer people--such as government dispensing materials--paper or plastic bags--may be worth- workers, the military, and the police--while poorer people while if it improves adherence to treatments. Brief training must pay. But the main problem is that user fees discourage courses for dispensers can substantially improve the quality of some people, particularly the poor, from seeking care at all. dispensing. And among those who do seek care, the resulting costs can be Another major problem is the presence of dispensing doc- financially crippling, to the extent that households may sacri- tors. A number of studies in both developed and developing fice food, education, or other important purchases to pay for countries have demonstrated that dispensing doctors prescribe drugs. Some are forced into poverty as a direct result of user more by value and not according to national or accepted guide- fees. A related issue of increasing importance with the advent of lines (Trap, Hansen, and Hogerzeil 2002). The higher number effective antiretroviral therapy for AIDS is that user fees dis- of prescriptions is strongly associated with symptomatic treat- courage adherence to long-term treatment, resulting in treat- ment (that is, a drug was prescribed for every symptom); gen- ment failure, increased disease transmission, and the develop- eral overprescribing of antibiotics; overuse of injections; and ment of drug resistance. Fees are, therefore, particularly prescription of medicines with lower clinical value. From a pol- problematic for transmissible diseases. icy and safety perspective, the functions of prescribing and dis- On the basis of similar evidence, Creese (1997, 203) con- pensing should be separated whenever possible (Nizami, Khan, cluded, "A range of policy options other than user fees exists for and Bhutta 1996). dealing with situations of both under financing and rapid growth in expenditure. As an instrument of health policy, user fees have proved to be blunt and of limited success and to have ADHERENCE potentially serious side effects in terms of equity. They should be prescribed only after alternative interventions have been Delivering the drug to the patient is not the end of the story: considered." In this respect, WHO is now advocating that fees the patient must adhere to the therapy. Failure to comply or should be minimized and that countries should be supported adhere will result in poorer health outcomes; it may compro- in attempts to channel a high proportion of health expenses mise the effectiveness of treatment, decrease the quality of life, through taxes or prepayment mechanisms such as forms of increase preventable disability, and lead to premature death. It insurance. may also result in increased health care costs, more use of emergency rooms, more and longer hospitalizations, and Table 72.2 Measured World Pharmaceutical Spending, by potentially more use of intensive care units (Sabate 2003). per Capita Income Clusters, 1990­2000 (percent) Share of FINANCING ISSUES Share of expenditure world total on health The share of expenditures on health that goes to pharmaceuti- Income group 1990 2000 1990 2000 cals is presented in table 72.2. Asking patients to pay part or all of the cost of their drugs High income 80.2 78.7 13.0 13.8 can aid in holding down costs, reducing overuse, and replen- Middle income 17.1 18.8 22.5 24.8 ishing the funds for drugs in the system. Drugs are often Low income 2.7 2.4 20.8 19.2 targeted for such fees because it is felt that patients will pay for Source: WHO 2004f. them if they have no other choice. A substantial literature now Note: Income groups refer to World Bank classifications as of July 2000. 1330 | Disease Control Priorities in Developing Countries | Susan Foster, Richard Laing, Bjørn Melgaard, and others SUSTAINABLE FINANCING OF VACCINES the cost per fully immunized child now reaching US$30 if AND IMMUNIZATIONS HepB and Hib are included. This increased cost adds to the challenge. Governments in low-income countries and interna- Immunization is now generally accepted as representing one of tional development partners need to develop long-term strate- the "best buys" for the health sector that governments must gies to ensure adequate financing for key health programs and play a lead role in financing, but sustainable financing mecha- interventions, including vaccines. nisms have been largely absent in poor countries (WHO Since GAVI and the Vaccine Fund were established, renewed 2004a). The cost of immunizing a child against the six basic attention has been paid to financing issues as they relate to vac- diseases hovers between US$15 and US$20 at current levels of cine and immunization financing. GAVI has worked with coverage, representing no more than US$0.50 per capita, and WHO and countries to consider how much it would require to on average 0.2 percent of the gross domestic product in most maintain existing levels of coverage after GAVI funding ends, low-income countries. These costs suggest that immunizations whereas among the prerequisites for countries to obtain assis- are affordable for most developing countries from national tance from the Vaccine Fund is preparation of long-term budgets. However, immunization programs account for only financing plans for immunization programs. However, most 5 to 10 percent of total government health expenditures in low-income countries clearly will be unable to fund even a many countries, which often rely heavily on donor funds. minimum set of essential interventions in the short to medium Although the international community has recognized the term without the assistance of international partners, thereby important reasons that financing of vaccines cannot be left to increasing the need to develop the long-term financing strate- individuals or households, donor support has often been quite gies described above. A mix of such strategies would include erratic. The result is volatile financing that is vulnerable to raising additional domestic funds, ensuring that funds are used shifts in donor priorities. In addition, recipient governments effectively and efficiently, moving to greater reliance on pre- recognize that donors are more likely to fund vaccines than payment mechanisms, and ensuring increased and stable flows many other services, so they have taken the opportunity to of external funds. Table 72.3 summarizes and compares some spend their own resources on activities that are important to recent trends in the financing of essential drugs, vaccines, and them but are less attractive to donors. This phenomenon can be contraceptives. seen in the apparent mismatch between data on disease bur- den, stated government priorities for health, and the allocation of government funds. ISSUES FOR THE FUTURE The challenge facing governments in poor countries is how best to finance vaccines, taking into account the variety of As the world's population ages, health systems that formerly other health problems and the possible sources of funds. More focused primarily on infectious disease are being asked to funds could probably be raised from firms and households for deliver new types of care, mostly for chronic illnesses and health in general, but user fees for immunization, as for drugs, increasingly for mental illness. By 2020, the major causes of the discourage people from seeking vaccination for their children. burden of disease will shift from pneumonia, diarrhea, and However, helping countries move to a system in which more perinatal conditions to heart disease, mental illness (particu- prepayment exists for health services in general--either larly depression), and road traffic accidents. Tobacco will kill through taxes or the various possible forms of health more people than any other cause of disease, including HIV. insurance--would provide a pool of domestic funds that could Unlike the United States and the countries of Western Europe, be used for vaccines. If these funds were raised progressively, China and India will face the challenges of an aging population the rich could subsidize the poor. before they become high-income countries. Most health sys- A number of new issues relevant to immunization financing tems in the developing world are now prepared to deliver acute have arisen recently, including the evolving nature of the world care, particularly for infectious disease, rather than chronic market for vaccines, the growing divergence in vaccination care. They are ill suited to long-term chronic care and follow- schedules between developed and developing countries, the up; in general they lack recordkeeping, demonstrate little devel- increasing diversity of products and presentations available to opment of personal relationships with caregivers, and have countries, the emergence of developing country manufactur- little provision for enhancing patient adherence with medica- ers, and the importance of new global initiatives such as GAVI tion. In many situations, the irregular and intermittent supply and the Vaccine Fund. The Vaccine Fund focuses on helping of medications for chronic disease means that the chronically low-income countries introduce newer vaccines, such as HepB ill suffer many interruptions of their treatment. The changing and Hib, which are generally more expensive than the older nature of health care will require changes in drug supply, which vaccines. In addition, the technology associated with the pro- are only beginning to become visible. The (perceived) differ- duction of new combination vaccines has increased prices, with ence between "good" and "bad" care is often the availability of Ensuring Supplies of Appropriate Drugs and Vaccines | 1331 Table 72.3 Trends and Developments in Financing and Procurement of Essential Drugs, Vaccines, and Contraceptives Commodity Trends and developments Implications for logistics Essential drugs Use of loans and pooled or "basket" financing arrangements is In the short term, procurement delays, shortages, emergency pro- increasing, leading to increased government involvement in curement requests, higher prices, and greater waste will result procurement. while governments develop internal capacity to procure. Also, increasing government involvement can mean less predictable results because of politics and governance issues. Procurement models adapted for health reforms such as Donors are more concerned about how procurement is done, trans- . decentralization and privatization proliferate. lating into more technical assistance and emphasis on performance benchmarking. Difficulties in evaluating procurement systems are caused by a proliferation of models. Countries are moving toward restrictive tender and prequalifica- In the longer term, prequalification may shorten the procurement tion of suppliers. cycle and lower costs. Similar effects from use of NGO suppliers are possible. Role of NGO suppliers continues in some countries as well as role of international NGO suppliers. User fees represent a major trend for essential drugs, with many Fee systems can decrease demand unless mechanisms exist to health facilities operating on a cash-and-carry basis. ensure service for those unable to pay. Private sector role is increasing as it becomes more apparent Policy makers and managers will need to design and implement that public sector and NGO services cannot meet all needs. programs that promote appropriate use of the private sector Vaccines Donor contributions have been decreasing starting in 1990s. Concerns are similar to those with essential drugs. Government financing and procurement of vaccines is increas- GAVI supplies may require logistics changes because of new ing; a dependence on external resources persists. vaccines and injection equipment. Some shift in financing from grants to loans, and more use of Pressure on governments to finance vaccine purchases may lead to basket financing. less government funding for EPI operating expenses and other Ministry of Health programs. Use of pooled procurement mechanisms and revolving funds or Outcome-based support requires information systems resistant to other international financial mechanisms, some based on manipulation. achievement of outcomes, is increasing. New vaccines and vaccine combinations (new with old) are supplied through GAVI. Contraceptives Donor contributions have been flat or have decreased, starting Concerns are similar to those with essential drugs. in 1990s. Despite efforts to increase government contributions, there is Demand is created without supply keeping up. still a major dependence on external resources. Many governments continue to give contraceptives (as Constraints of Mexico City policy ("global gag rule") have limited compared with drugs) low priority for procurement with their funding for contraceptives. own funds. Financing through loans and basket financing are increasing, and governments are increasingly involved in procurement. Source: Adapted from Vian and Bates 2003. drugs and supplies. Programs and funding agencies that are tory infections (mainly pneumonia), diarrheal disease, HIV planning improvements in health care--for example, increas- and AIDS, tuberculosis, malaria, and measles. Drug resistance ing coverage or case detection rates--often overlook the fact complicates the effective treatment for nearly all of these acute that such improvements will increase drug needs and costs. infections. Furthermore, this trend is expected to accelerate in the coming decades. In the treatment of HIV and AIDS, the Drug Resistance increase of retroviral drug resistance is becoming a serious Although the burden of chronic and noninfectious disease is problem, especially in view of the limited number of treatment increasing rapidly in the developing world, infectious diseases regimens available to date. still account for nearly half of deaths in low-income countries. Drug-resistant malaria is now widespread. Chloroquine-- Most of these deaths are caused by six diseases: acute respira- once a cheap and reliable first-line treatment for malaria--is 1332 | Disease Control Priorities in Developing Countries | Susan Foster, Richard Laing, Bjørn Melgaard, and others no longer effective in most countries. Newer drugs are signifi- HAART; and President George W. Bush's Emergency Plan for cantly more expensive. Most recently, the trend has been AIDS Relief. A number of issues are raised by the delivery of a toward multidrug combinations of products, and the addition complex, lifelong, costly treatment to poorer communities, of more than one drug is often to "protect" the component especially in rural areas, one of which will be how to ensure drugs from developing resistance as well as to improve the ther- adequate adherence to treatment in different clinical settings, apeutic effect (WHO 2002a). ranging from district hospitals to health centers or even home Drug resistance in tuberculosis control--in particular, mul- settings, for HAART delivery. tidrug resistance--is a growing problem. Multidrug-resistant TB has now appeared around the world, and in many places more than 20 percent of resistant new tuberculosis cases are Aging and Chronic Diseases resistant to several drugs. Furthermore, the emergence of mul- One of the important results of the exercise to estimate the tidrug-resistant bacilli means that medication that once cost global burden of disease was to highlight the growing impor- US$20 must now be replaced with drugs that are significantly tance of chronic disease, particularly in the developing world. more expensive and more difficult to use (WHO 2002a). A large percentage of chronic illnesses are related to smoking Another major concern is the use of antimicrobials in farming, and lifestyle, and thus attempts to reduce smoking--or the because about half of the antimicrobials produced each year lethality of smoking--would have an important effect on the are used in farm animals. Some of the new resistant bacteria are need for medication for chronic disease. transmitted from food of animal origin or through direct con- Although many cancers are not yet curable, many are treat- tact with farm animals. Some reports indicate that as much as able with the goal either of slowing the spread or of palliating 50 percent of human antimicrobial resistance is caused by the symptoms of the disease. As the burden of cancer increases, growth promoters in livestock, which are added to feed in sub- palliative care, which involves the treatment of the symptoms therapeutic antibiotic doses (WHO 2002a). and especially the pain that accompanies most cancers, needs to be given much higher priority. At present, the vast majority of the millions of cancer patients in the developing world HIV and AIDS receive totally inadequate pain control and suffer needless The HIV epidemic has had a tremendous impact on the phar- agony, in part because of antiquated laws governing the use of maceutical supply situation. First, it has highlighted weaknesses opioid analgesics (particularly morphine) and attitudes of of drug supply and access around the world; the arrival of medical and nursing personnel toward pain control (as well as highly active antiretroviral therapy for the treatment of attitudes of family members in some settings). The myths HIV/AIDS (HAART) means that HIV is to a large extent now about morphine need to be dispelled. When used appropri- a treatable condition, yet treatment is not available to the ately, especially in oral form, morphine does not lead to addic- majority of those who suffer from HIV. Second, it has drawn tion, tolerance, respiratory depression, cognitive impairment, the world's attention to the growing gap between rich and poor or premature death. In fact, people live longer when their pain in terms of pharmaceutical provision. Unlike many other is controlled, and they can eat, sleep, and live normal lives highly prevalent illnesses in the developing world, HIV and (Merriman and others 2002). AIDS are also of major concern in the wealthier countries, and In countries where palliative care is fairly well developed thus significant research has been undertaken and has yielded and available, the consumption of morphine per capita aver- effective new medications (HAART, in particular). ages over 20 milligrams, but in most developing countries it is A recent WHO report highlights the issue of the affordabil- negligible, and most of the needs for pain relief are unmet ity of medications, pointing out that of the 23 countries that (Joranson, Rajagopal, and Gilson 2002). The World Bank rec- are estimated to make up 80 percent of the 2003 global need for ognized the importance of alleviating pain, which it included in HIV and AIDS treatment--estimated at about US$300 per its package of "essential clinical services" (World Bank 1993). annum per patient--only 8 have pharmaceutical expenditure As the population ages, the ability of the health care system to levels above US$5 per capita, far short of the level of expendi- provide palliative care must grow along with it. ture needed (WHO 2004c). Prices have fallen dramatically; The trend toward more sedentary lifestyles and toward con- WHO has continued to monitor the quality of AIDS drugs sumption of diets with higher fat and sugar content is leading available on the world market for sale in developing countries to a steep increase in the burden of diabetes, with 150 percent and has removed substandard drugs from its list when neces- increases in prevalence predicted for many countries by 2030; sary (WHO 2004e). Many high-profile initiatives to solve this the absolute numbers will grow from 171 million in 2000 to problem have been started, most notably the WHO's "3 5" about 366 million in 2030. The greatest increases in diabetes program; the Global Fund to Fight AIDS, Tuberculosis, and prevalence are predicted for the Middle East, Sub-Saharan Malaria; the Clinton Foundation's efforts to lower prices for Africa, and India (Wild and others 2004). Most of these Ensuring Supplies of Appropriate Drugs and Vaccines | 1333 Table 72.4 New Vaccines Needed Priority Close to Vaccines for Other vaccines Vaccines for vaccines licensure vaccines neglected diseases of importance new threats HIV, TB, malaria Meningococcus, Shigella, dengue, Human papilloma virus, SARS, anthrax, streptococcus Japanese encephalitis, respiratory syncytial virus, smallpox, pandemic pneumonia, rotavirus leishmaniasis, schistosomiasis, herpes simplex, enterotoxigenic, influenza cholera Escherichia coli Source: Authors. SARS Severe acute respiratory syndrome. new cases will be type 2 and, thus, most will not be insulin more than 5 million deaths per year, or about 50 percent of all dependent, but they will require oral diabetic medications. For infectious disease deaths (see table 72.4). The total investment those who do require insulin, given the current state of tech- in vaccine against these diseases is not impressive, and it will nology, the main barrier (other than cost) is the need for stor- probably take at least 5 to 10 years before a vaccine against any age of the insulin in a cold or cool location and for sterile injec- of them is available. tion equipment. In either case, to meet the predicted rise in GAVI has selected three vaccines for accelerated develop- cases and to treat them with current drugs, a major expansion ment: meningococcal meningitis, rotavirus, and pneumococcal of drug supply for diabetes must be anticipated. Many diabet- vaccines. They have been selected because they are considered ics currently do not receive adequate treatment. The pressure to close to licensure, or "near term." Other important diseases are provide adequate treatment will increase as the population ages considered neglected in terms of vaccine development, among and begins to demand treatment of its chronic afflictions--and them shigella dysentery and dengue fever. in that case the increase in demand for diabetes medications New diseases emerge and old ones reemerge, influencing would potentially be much more than 150 percent. priorities in vaccine research. The severe acute respiratory syn- Another important finding of the global burden-of-disease drome (SARS) epidemic, the outbreak of avian influenza, and exercise was the high number of DALYs lost to mental illness, the emergence of bioterrorism threats such as anthrax have led depression in particular. In 2020, unipolar depression is pro- to a new search for vaccines against these infections. The threat jected to be the leading cause of morbidity and disability of a pandemic of a reassorted influenza virus strain has recently among females worldwide and in developing countries. highlighted the need for much greater resources and attention Whereas in the industrial countries a pharmacological solution to be devoted to the development and distribution of effective is often used, this approach may not be feasible in the develop- flu vaccines. ing world, at least not at present price levels. Recent research in the developing world has shown good results with weekly group interpersonal therapy, without the use of antidepres- New Vaccine Technologies sants. Trained laypersons ran the therapy sessions, not psy- Alternative routes of administration would improve program chiatrists or medical personnel (Bolton and others 2003). safety, avoiding needle transmission of bloodborne pathogens. The ability of nonprofessionals to administer vaccines would also ease vaccine delivery strategies. New administration VACCINE RESEARCH PRIORITIES routes, such as oral, nasal, and transcutaneous routes, are being explored. An interesting project concentrates on the develop- In the past two decades new advances in biotechnology have ment of a nasal measles vaccine that would greatly enhance the resulted in the licensure of new vaccines, such as Hib, acellular feasibility of eliminating this disease by facilitating the admin- pertussis, HepB, and attenuated varicella. Research institutions istration during mass campaigns. in the public sector have generated most of the basic scientific The concept of using plant-derived or edible vaccines breakthroughs, whereas the large pharmaceutical companies involves encoding protective antigens from pathogens into have borne the cost for clinical development. Because such transgenic plants. The plants are processed so that they can development requires heavy investments that need to be deliver a uniform dose of vaccines. Human clinical trials have recouped from profits, new vaccines are expensive and there- been conducted with, for example, bananas and raw potatoes, fore out of reach for poor populations. which have shown encouraging antibody responses. The Of all the vaccines currently under development, the three potential advantages of this technology could include ther- most needed today are vaccines to prevent the three big mostability, low investment needs, multivalency, and oral diseases--AIDS, TB, and malaria--which jointly account for administration. 1334 | Disease Control Priorities in Developing Countries | Susan Foster, Richard Laing, Bjørn Melgaard, and others New Immunization Technologies world. Only a limited number of research centers have the Priority is given to new delivery technologies that will expand capacity and experience required to conduct phase 2 and access, improve safety, and cut the cost of immunization pro- especially phase 3 trials of new vaccines, and they are mainly grams. They include the following technologies: located in industrial countries. The capacity to conduct phase 3 trials in developing countries needs to be strengthened; the · The reuse of disposable syringes and needles is widespread current situation impedes further development of vaccines and contributes significantly to the transmission of hepati- needed in those countries. tis B and C and HIV. The autodisabled syringe prevents Pilot lot production of vaccines is required for all phases of reuse, and disposal in safety boxes reduces the risk to health clinical trials. The global capacity to produce pilot lots is, how- staff and the general public from contaminated syringes and ever, inadequate to meet demand. Close public-private part- needles. nerships are necessary to ensure that the production capacity is · Four different technologies are being explored to minimize available. the risk of infection from accidental exposure to sharps: cor- Manufacturers need markets to provide some assurance that rosive disinfectants, thermoprocessing, needle destruction, the development cost for new products can be recouped. Such and plastic melting. However, none of these options will incentives require realistic forecasts of demands. Various mech- soon be put into use in the field. anisms have recently been put in place to try to guarantee · Although the adoption of the multidose vial policy will future markets, most notably the Vaccine Fund. contribute to the reduction of wastage, the ultimate aim is Last, disease burden data are needed for both selections of to provide all immunizations as monodose preparations. vaccines for national programs and for estimations of vaccine Injection devices prefilled with a monodose increase quality requirements, including market projections. However, such and safety at the point of use. UniJect is one such device that data are lacking in many countries and regions. Existing data has been tested with HepB and tetanus toxoid. Village health are especially weak for respiratory disease of both bacterial and workers or traditional birth attendants can use such devices. viral origin. Currently, the cost of the device and the need for additional cold storage space when multidose presentation is ex- changed for monodose pose obstacles to implementation. PRIORITIES FOR PHARMACEUTICAL RESEARCH · Needle-free injection devices deliver vaccine at high velocity into the skin without penetration of a needle, reducing the The WHO Priority Medicines Project, a recent exercise that risk of transmission of bloodborne pathogens. Technologies used evidence-based methods to outline the priorities for are being developed for both mono- and multidose presen- public funding of pharmaceutical research, has recently been tations. Available multidose injectors have not been found published (WHO 2004b). It incorporated data from the safe, and new models are under development. However, reg- burden-of-disease rankings and from the Cochrane Database ulatory obstacles and high cost have rendered the monodose of Systematic Reviews of data on clinical efficacy. It also incor- injector models that are currently available infeasible for porated the use of criteria of social justice, social solidarity, and large-scale programs. equity, so that neglected diseases and the needs of special · Vaccine distribution and storage without a cold chain would patient groups (the elderly, women, and children) were also considerably simplify the delivery system, reduce cost, and taken into account. The research identified 20 major diseases allow for an integrated supplies mechanism. Development that account for 60 percent of the total DALY burden both in of vaccines that do not need a cold chain should be the high- Europe and in the rest of the world--diseases that are common est priority for technology research. Sugar glass drying is to both groups included unipolar depression, ischemic heart one such technology that has shown great promise. It can be disease, cerebrovascular disease, chronic obstructive pul- used to produce multivalent vaccines that are completely monary disease, and digestive diseases (excluding diarrheal heat stable except under extreme climatic conditions. The diseases). high cost of regulation and licensing and the uncertainty The authors also mention the important contributions of about market prospects in industrial countries have so far various cancers, lower respiratory tract infections, and diabetes impeded the development and use of this technology. to the burden of disease, which is common to both developed European countries and to the developing world (WHO 2004b). Smoking is clearly a major contributing risk factor, and Obstacles to Vaccine Research the authors caution that expenditure on pharmaceuticals for A host of obstacles confront vaccine research, the most impor- smoking cessation must not divert resources from other efforts tant being the low level of investment for vaccine development to reduce smoking. The priority areas identified by this exercise when there are limited market prospects in the industrial are presented in table 72.5. Ensuring Supplies of Appropriate Drugs and Vaccines | 1335 Table 72.5 Preliminary List of 16 Priority Areas for IMS Global Learning Consortium. 2003. "2002 World Pharma Sales Pharmaceutical Research Growth: Slower, but Still Healthy." http://www.ims-global.com/ insight/news_story/0302/news_story_030228.htm. Rank Condition International Dispensary Association. 2004. E-Catalogue of Products. http://www.ida.nl/en-us/content.aspx?cid=42. 1 Infections caused by antibacterial resistance Joranson, D. E., M. R. Rajagopal, and A. M. Gilson. 2002. "Improving 2 Pandemic influenza Access to Opioid Analgesics for Palliative Care in India." Journal of Pain 3 Cardiovascular disease and Symptom Management 24 (2): 152­59. 4 Diabetes (types 1 and 2) Kaplan, W. A., R. O. Laing, B. Waning, L. Levison, and S. D. Foster. 2003."Is Local Production of Pharmaceuticals a Way to Improve Pharmaceutical 5 Cancer Access in Developing and Transitional Countries? Setting a Research 6 Acute stroke Agenda." World Bank, Washington, DC. http://www1.worldbank. org/hnp/hsd/documents/LOCALPRODUCTION.pdf. 7 HIV/AIDS Laing, R. O., H. V. Hogerzeil, and D. Ross-Degnan. 2001. "Ten 8 Tuberculosis Recommendations to Improve Use of Medicines in Developing 9 Neglected diseasesa Countries." Health Policy and Planning 16 (1): 13­20. 10 Malaria Laing, R. O., B. Waning, A. Gray, N. Ford, and E. 't Hoen. 2003. "Twenty- 11 Alzheimer's disease Five Years of the WHO Essential Medicines Lists: Progress and Challenges." Lancet 361 (9370): 1723. 12 Osteoarthritis Levison, L., and R. O. Laing. 2003. "The Hidden Costs of Essential 13 Chronic obstructive pulmonary disease Medicines." Essential Drugs Monitor 33: 20­21. 14 Alcohol use disorders: alcoholic liver disease and alcohol Management Sciences for Health. 2004. International Drug Price Indica- dependency tor Guide. http://erc.msh.org/mainpage.cfm?file=1.0.htm&module= 15 Depression in the elderly and adolescents DMP&language=English. Merriman, A., adapted from D. Doyle, and T. F. Benson. 2002. Palliative 16 Postpartum hemorrhage Medicine: Pain and Symptom Control in the Cancer and/or AIDS Patient Source: WHO 2004b. in Uganda and Other African Countries. 3rd ed. Kampala: Hospice a. Neglected diseases include, but are not limited to, trypanosomiasis (sleeping sickness), Buruli Africa Uganda. ulcer, leishmaniasis, and Chagas disease. Nizami, S. Q., I. A. Khan, and Z. A. Bhutta. 1996. "Drug Prescribing Practices for General Doctors and Paediatricians for Childhood Diarrhoea in Karachi, Pakistan." Social Science and Medicine 42 (8): CONCLUSION 1133­39. Quick, J. D., J. R. Rankin, R. O. Laing, R. W. O'Connor, H. V. Hogerzeil, Ensuring that needed essential medicines and vaccines are M. N. G. Dukes, and A. Garnett, eds. 1997. Managing Drug Supply. available is critical for the success of any disease control pro- 2nd ed. Management Sciences for Health in collaboration with the gram. A great deal is known about what works and what does World Health Organization. West Hartford, CT: Kumarian Press. not work. Careful selection, procurement from prequalified Sabate, E., ed. 2003. Adherence to Long-Term Therapies: Evidence for Action. Geneva: World Health Organization. suppliers, proper storage and distribution using secure reliable Trap, B., E. H. Hansen, and H. V. Hogerzeil. 2002. "Prescription Habits of channels, and assurance of rational use and correct dispensing Dispensing and Non-Dispensing Doctors in Zimbabwe." Health Policy are all critical components of any drug and vaccine supply and Planning 17 (3): 288­95. system. Ensuring that adequate funds are available to pay for Trouiller, P., P. Olliaro, E. Torreele, J. Orbinski, R. O. Laing, and N. Ford. the procurement, distribution, and quality assurance of all 2002. "Drug Development for Neglected Diseases: A Deficient Market and a Public-Health Policy Failure." Lancet 359 (9324): 2188­94. medicines and vaccines is equally critical. Depending on the UNICEF (United Nations Children's Fund). 2004. "Supply Catalogue." circumstances, either the public or the private sector or a com- http://www.supply.unicef.dk/catalogue/index.htm. bination of both can efficiently deliver quality-assured Vian, T., and J. Bates. 2003. Implications and Recommendations for medicines and vaccines. The experience of a number of coun- Contraceptive Security. Vol. 1 of Commodity Security and Product tries and programs has demonstrated that essential medicines Availability Issues for Essential Medicines, Vaccines, and Contraceptives. and vaccines can be reliably delivered to poor people using the Arlington, VA: John Snow/U.S. Agency for International Development. approaches described in this chapter. WHO (World Health Organization). 1977. The Selection of Essential Drugs: Report of a WHO Expert Committee. WHO Technical Report Series 615. Geneva: World Health Organization. REFERENCES ------. 1999a. Guidelines for Drug Donations. 2nd ed. WHO/ Bolton, P., J. Bass, R. Neugebauer, H. Verdeli, K. F. Clougherty, P. EDM/PAR/99.4 1999. Geneva: WHO. Wickramaratne, and others. 2003."Group Interpersonal Psychotherapy ------. 1999b. Guidelines for Safe Disposal of Unwanted Pharmaceuticals for Depression in Rural Uganda: A Randomized Controlled Trial." in and after Emergencies. WHO/EDM/PAR/99.2 1999. Geneva: WHO. Journal of the American Medical Association 289: 3117­24. ------. 2000. The Use of Opened Multi-dose Vials of Vaccine in Subsequent Creese, A. 1997. "User Fees" (editorial). British Medical Journal 315: Immunization Sessions. WHO/V&B/00.09. Geneva: WHO. http://www. 202­3. who.int/vaccines-documents/DocsPDF99/www9924.pdf. 1336 | Disease Control Priorities in Developing Countries | Susan Foster, Richard Laing, Bjørn Melgaard, and others ------. 2001a. Globalization, TRIPS and Access to Pharmaceuticals. WHO ------. 2004b. Priority Medicines for Europe and the World. WHO/ Policy Perspectives on Medicine 3. WHO/EDM/2001.2. Geneva: EDM/PAR/2004.7. Geneva: WHO. WHO. ------. 2004c. Sources and Prices of Selected Medicines and Diagnostics for ------. 2001b. How to Develop and Implement a National Drug Policy. People Living with HIV/AIDS. A joint UNICEF-UNAIDS Secretariat- 2nd ed. Geneva: WHO. http://whqlibdoc.who.int/publications/ WHO-Médecins Sans Frontières project. Geneva: WHO. 924154547X.pdf. ------. 2004d. The WHO Prequalification Project. Geneva: WHO. http:// ------. 2002a. "Antimicrobial Resistance and Rational Use of mednet3.who.int/prequal/. Antimicrobial Agents." EM/RC49/8. Geneva: WHO Regional ------. 2004e. "WHO Statement on Removal of Two AIDS Medicines Committee for the Eastern Mediterranean. from List of Prequalified Products." WHO, Geneva. http://www. ------. 2002b. "Core Information for the Development of Immunization who.int/mediacentre/statements/2004/statement_aidsprequal/en/. Policy: 2002 Update." WHO/V&B/02.28. Department of Vaccines and ------. 2004f. World Pharmaceutical Situation Report 2004. Geneva: Biologicals. WHO, Geneva. http://www.who.int/vaccines-documents/ WHO. DocsPDF02/www557.pdf. ------. 2005a. Procurement of Vaccines for Public-Sector Programmes. ------. 2002c. "Fixed Dose Combinations for HIV/AIDS, Tuberculosis, WHO/IVB/03.16. Geneva: WHO. and Malaria. Current Status and Future Challenges from Clinical, ------. 2005b. "Substandard and Counterfeit Medicines." Fact Sheet 275. Regulatory, Intellectual Property, and Production Perspectives." WHO, Geneva. http://www.who.int/mediacentre/factsheets/fs275/en/. WHO, Geneva. Wild, S., G. Roglic, A. Green, R. Sicree, and H. King. 2004. "Global ------. 2002d. "State of the World's Vaccines and Immunization." Prevalence of Diabetes: Estimates for the Year 2000 and Projections for ------. 2003a. How to Develop and Implement a National Drug Policy. 2030." Diabetes Care 27: 1047­53. Policy Perspectives on Medicines. Geneva: WHO. http://www.who. Williams, H. A., D. Durrheim, and R. Shretta. 2004. "The Process of int/medicines/library/general/PPMedicines/PPM_No6-6pg-en.pdf. Changing National Malaria Treatment Policy: Lessons from Country- ------. 2003b. Vaccines and Biologicals Catalogue 2003. WHO/V&B/ Level Studies." Health Policy and Planning 19 (6): 356­70. 02.06. Geneva: WHO. World Bank 1993. World Development Report 1993: Investing in Health. ------. 2003c. The Quality of Antimalarials--A Study in Selected African New York: Oxford University Press. Countries. Geneva: WHO. ------. 2000. "Procurement of Health Sector Goods: Technical Note." ------. 2003d. Report of the 13th Expert Committee on the Selection World Bank, Washington, DC. http://siteresources.worldbank.org/ and Use of Essential Medicines. Geneva: WHO. http://whqlibdoc. PROCUREMENT/Resources/health-tn-ev2.doc. who.int/trs/WHO_TRS_920.pdf. Xu, K., D. B. Evans, P. Kadama, J. Nabyonga, P. Ogwang Ogwal, ------. 2004a. Economics of Immunization: A Guide to the Literature and P. Nabukhonzo, and A. M. Aguilar. Forthcoming. "Understanding the Other Resources. WHO/V&B/04.02. Geneva: WHO. http://www.who. Impact of Eliminating User Fees: Utilization and Catastrophic Health int/vaccines-documents/DocsPDF04/www769.pdf. Expenditures in Uganda." Social Science and Medicine. Ensuring Supplies of Appropriate Drugs and Vaccines | 1337 Chapter 73 Strategic Management of Clinical Services Alexander S. Preker, Martin McKee, Andrew Mitchell, and Suwit Wilbulpolprasert Financial resources alone are insufficient for individuals to mortality rates. What such a system could not do, however, was benefit from the opportunities presented by modern health respond effectively to the possibilities opened up by the explo- care systems. Some countries have achieved much better levels sion in diagnostic and therapeutic knowledge that began in the of health than would be expected given their financial re- mid 1960s with the availability of new and easily tolerated sources (Mehotra 2000); many examples of poor-quality care treatments for many common chronic disorders. As individu- in countries at all levels of development reflect not only scarce als became able to survive with their chronic diseases, they aged resources but also inadequate management of what resources and acquired other conditions, many of which could now be are available (see chapter 70). treated effectively, constantly increasing the complexity of the Many inputs must come together at the appropriate time and health care required. The inability to manage this increased in the appropriate place to achieve maximum health gain. These complexity resulted in persistently high mortality rates from inputs include human resources (in particular, trained staff); treatable conditions at a time when corresponding mortality physical resources (such as pharmaceuticals and technology); rates were falling in Western countries (Andreev and others and intellectual resources (in the form of evidence and the abil- 2003). ity to apply it appropriately). This congruence requires that the This situation has certain parallels with that faced by many production, distribution, and combination of these resources be low- and middle-income countries today. Through a variety of actively managed and that the relations among the various ele- mechanisms, a political commitment to the health sector is ments that contribute to health care be optimized. manifest in the increased availability of funding, such as The challenge of bringing these diverse inputs together is through the Global Fund for AIDS, Tuberculosis, and Malaria becoming increasingly complex. Until the 1950s, providing (http://www.theglobalfund.org/en/). Much discussion has basic care at low cost to large populations was relatively focused on one of the elements of health care that these initia- straightforward, given political will and sufficient resources. tives will support: the supply of drugs that target the microbi- Relatively few effective drugs were available; even fewer drugs ological agents responsible for these three diseases. Yet were effective in managing chronic disease. The available tech- improved outcomes will be achieved only if such agents are nology was limited to simple x-ray machines and chemistry linked to the many other elements required to diagnose and tests that required few skills to administer. Consequently, scal- treat these patients. Most obviously, the supply of drugs such as ing up the delivery of basic health care was relatively easy. antiretroviral agents must be coupled with those used to treat The situation in the former Soviet Union illustrates this the opportunistic infections that exacerbate AIDS. Care for the state of affairs. Beginning in the 1930s, the Soviet Union imple- acute episode of illness should be linked to general support for mented a vast system to provide basic health care where almost patients and family members as well as to activities designed to none had existed. The very simple care available was sufficient prevent further spread of the disease. Furthermore, as drugs to produce significant reductions in maternal and childhood to combat infections become more widely available, it is 1339 probable--unless highly developed prevention systems have as poor communication and transportation links; or when been put in place--that drug resistance will increase; this out- there are weaknesses in the banking system. come has been evidenced with tuberculosis in those parts of · The fourth level refers to the environmental and contextual the world where treatment has been available but poorly man- constraints on effective policies. The delivery of effective aged, such as the former Soviet Union and Peru (Farmer, care may be affected by the physical environment, including Reichman, and Iseman 1999). The resulting resistant infections climate and population dispersion. However, an equally are much more complicated and expensive to treat. The rise in important constraint is weak governance working within antibiotic resistance provides one of the most graphic examples unsupportive policy frameworks, which may be compro- of the consequence of the failure to manage the delivery of mised further by corruption, weak rule of law, political health care (see chapter 55). instability, weak public accountability, and lack of a free Yet even where the financial resources and the political will press. For example, de Soto (2000) has shown that, in many exist to deliver effective health care, many health care systems middle-income countries, it is almost impossible even to contain numerous constraints to success (Hanson and others create a simple garment repair business because of a failure 2003): of legislative reform, in particular a lack of clearly defined property rights. As a result, much of the economic activity · At the first level, that of the community or household, there in those countries is informal or even marginally illegal, a may be inadequate demand for services or physical, finan- response that is of particular concern in health care, given cial, or social obstacles to their use. This situation calls for the scope for unlicensed and incompetent providers to action to increase access and affordability, including health endanger the public. care financing reforms (see chapter 12). It also requires poli- cies to ensure that services are culturally appropriate, that This framework underscores the importance of coordinat- they address the particular needs of underserved popula- ing action at multiple levels. Health services can operate effec- tions, and that they provide dignity and privacy. Moreover, tively only if policies are in place at the community level to services should be physically accessible, both in terms of dis- ensure that those in need have access to services, and only if tance from population settlements and in terms of their policies are in place at higher levels to ensure that the resources construction--that is, facilities must be responsive to the are available to provide those services. needs of persons with disabilities. This analysis of different levels demonstrates the impor- · At the second level, the delivery of health care, there may be tance of taking a systemwide approach to the management of a shortage of resources, such as staff members, drugs, and health services. However, because of limited space, this chapter equipment. However, to bring these resources together focuses primarily on the third level, that of strategic manage- would require actions at the third level that anticipate future ment. It first examines the nature of management in general needs, as well as actions that ensure that the needed drugs and the specificities of management in the health system. It and equipment are purchased at the best price possible, are then explores some issues that arise when managing health subject to appropriate quality controls, and are distributed services in different settings. It concludes with an exploration where needed. of some of the strategies used in low- and middle-income · The third level includes health sector policy and strategic countries to optimize the delivery of care, using a framework management. Effective action may be constrained by weak developed by Oliveira-Cruz, Hanson, and Mills (2003). systems of management that are unable to take into account the changing health needs of the population and the chang- ing demands on health care providers. Management weak- WHAT IS MANAGEMENT? nesses include inadequate pharmaceutical regulation and supply, ineffective training of health professionals, inability One of the earliest definitions of management was that of to engage with civil society, and a failure to put in place French mining engineer Henri Fayol. Writing at the beginning incentive systems to facilitate effective health care. of the 20th century, he stated, "To manage is to forecast and Constraints at this level may originate outside the country, plan, to organise, to command, to coordinate, and to control" as governments are faced with demands by donors to follow (Fayol 1949). Put simply, managing is about assessing probable paths that either undermine their policy goals or remove the future scenarios, deciding how best to respond to them, bring- flexibility needed to achieve them. Constraints acting at this ing together the resources needed for that response, and deploy- level also arise when policies in other areas affect the health ing them as effectively as possible. Until relatively recently, sector, such as when a weak, overly bureaucratic, and most management research was concerned with industrial unreformed civil service system implements obsolete regu- production, for which outputs could be measured relatively lations; when there are inadequacies in infrastructure, such easily. Relatively less attention was given to management of 1340 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others Box 73.1 Specificities of Health Care Organizations Health care services differ from many other organizations · The work involves a high degree of specialization. in many ways: · Workers are highly professional, with a primary loyalty to the profession rather than to the organization. · Defining and measuring outputs is difficult. · There is limited scope for effective organizational or · The work involved is more variable and more complex managerial control over clinicians, the group most than in many other organizations. responsible for generating work and expenditure. · Much of the work is of an emergency nature and cannot · Dual lines of responsibility often create problems of easily be deferred. coordination, accountability, and confusion of roles. · The consequences of error can be severe. · Activities by different groups of staff members are highly interdependent, requiring a high level of coordination. Source: Shortell and Kaluzny 1983. service industries in general and health care services in partic- concepts that arise from new institutional economics. Included ular. As Shortell and Kaluzny (1983) noted, health care services in these is contestability (Baumol, Panzar, and Willig 1982), in are different from many other organizations. Of course, many which the benefits that competition is thought to bring can of the specificities shown in box 73.1 are differences of degree, arise even when competition is absent, thus ensuring that the with health services sharing many features with other service barriers to market entry are sufficiently low to allow other organizations. Yet important differences exist. providers to emerge. User choice is given priority over most other goals, including equity. The enthusiasm for new public management was largely Managing Health Care Services ideological, reflecting the contemporary rejection of an During the 1970s, health care services in many countries faced expanded role for the state; the extent to which this model was growing criticism for their perceived failure to articulate ex- actually able to achieve what was claimed for it remains highly plicit goals or to develop the means to achieve them (Enthoven contested (Le Grand and Bartlett 1993; Stewart 1998). In par- 1985; Griffiths 1984). This failure was contrasted with the per- ticular, critics drew attention to the high transaction costs ceived success of the private sector, which was seen as more involved (Evans 1997) and the lack of evidence that competi- capable of innovation and more responsive to demand. tion can actually bring about the intended improvements in These developments gave rise to what has been termed new quality of care (Maynard 1998). public management (Hood 1991), which is characterized by the One feature of the new public management is its emphasis following: on general management, with managers possessing skills and expertise that can be applied to any sector. These managerial · greater role for professional management in the public attributes are considered to be of greater importance than tech- sector nical or professional knowledge. As a consequence, in some · closer scrutiny of the work of professionals, involving per- countries, the balance of power has begun to shift away from formance measurement and target setting health professionals and toward general managers. In many · link between resource allocation and measurable outputs places, the initial enthusiasm has given way to disillusionment · "unbundling" of previously integrated units, with contract- and subsequently to a more balanced view that, though the pre- ing for previously integrated services cise solution will reflect the particular circumstances of the · shift to competition as a key to reducing costs and an health care system, what is needed is a partnership between emphasis on a private-sector management style these two groups. · careful use of resources to drive down the cost of labor and In some countries, this development will mean that man- other inputs, where possible. agers must assume a greater role in relation to the delivery of clinical care. Such an expanded role will extend from their Recognizing the many reasons for market failure in health traditional responsibilities, such as financial controls, hotel care (Arrow 1963), new public management builds on several services, and payroll management, to active participation in Strategic Management of Clinical Services | 1341 setting and monitoring standards for care delivery, linked to a involve direct contact between a patient and a health care pro- responsibility for ensuring that the resources needed for care fessional, as well as indirect contact, such as when a pathologist delivery are available. provides a diagnosis on a biopsy or blood sample. Reflecting the In other countries, this role may involve stepping back a focus of much existing research, this section is structured in little. In an analysis of the British National Health Service, in terms of different settings of care: ambulatory care, hospitals, which the degree of managerial control over the delivery of and community care. Unfortunately, rather less research tran- health care has proceeded further than in many other industrial scends these often artificial and arbitrary divisions to look at the countries, Harrison and Pollitt (1994) note how clinical deci- more important issue of the patient's journey through the health sion making is increasingly driven by diagnostic and treatment care system, given that one of the greatest managerial challenges protocols. Although often developed locally, working arrange- facing those delivering clinical services is how to ensure that the ments are increasingly specified, with the introduction of journey is efficiently navigated (McKee and Nolte 2004). timetabled job plans for medical specialists and much greater Furthermore, available research that focuses on health facilities measurement of outcomes. Yet Harrison and Pollitt argue that is often difficult to generalize because of the different meanings the growth of managerial control over professional activity is attributed to common terms such as hospital, health center, likely to be constrained by the increasing involvement of pro- or more prosaically, hospital bed. For example, a major fessionals in management, even if they do not fully adopt the teaching hospital in a capital city, such as the Kenyatta National managerial agenda. A further constraint is the persisting ability Hospital in Nairobi (http://www.kenyattanationalhospital.org/ of professionals, because of their specialized knowledge, to services.html), which offers invasive cardiology, renal trans- resist managerial control and the related unwillingness of lay plants, and radiotherapy, is very different from a rural hospital managers to extend their control into certain areas in which with perhaps 100 beds and a single operating theater that they do not feel competent. provides only the most basic surgical and obstetric care. Managing for Improved Quality of Care Hospitals An increasing volume of research in industrial countries has Although hospitals are rarely the first places of contact between focused specifically on managerial and organizational responses patients and health systems, and although hospitals do not pro- to evidence that health systems often deliver suboptimal care vide the greatest share of health care, it is appropriate to begin (Institute of Medicine 2001). Quality of care is addressed in with them because they often account for the largest share of more detail in chapter 70. However, some of the key messages public health sector expenditure (OECD 2003). They are also from this research are relevant here. particularly difficult to manage for several reasons: One message is that change must take place at all levels of the system. In this context, Ferlie and Shortell (2001) identify · One reason is the diversity of tasks that a hospital must four such levels: the individual, the group or team, the organi- undertake (Healy and McKee 2002b). Many hospitals ful- zation, and the larger system or environment. They note the fill roles that go beyond the delivery of patient care to pro- growing evidence that strategies that focus on individuals alone vide training and research, support to community-based are unlikely to be successful, whereas those that are embedded facilities, and even local employment or civic identity within broader organizational change are more likely to be symbols. effective (Davis and others 1995). A second key message is the · A second reason is the blurring of boundaries between importance of teamwork, with evidence that well-functioning, hospitals and the rest of the health care system, which has multiprofessional teams provide better quality care (Aiken, occurred as a result of the emergence of many innovative Sochalski, and Lake 1997). However, change may be inhibited models of care that cross the boundary between secondary by barriers at the level of the organization, including lack of a and either primary or social care. A related issue is the shift consistent focus on quality, inadequate information, lack of taking place in many countries to managing patients physician involvement, and inadequate managerial support through a complex combination of short inpatient stays (Shortell, Bennett, and Byck 1998). and visits as an outpatient to specialist clinics and diagnos- tic facilities (McKee and Healy 2001). This approach is MANAGING CLINICAL SERVICES vastly more complicated to manage than the traditional IN DIFFERENT SETTINGS model in which patients were admitted to wards from which they were taken to undergo investigations and treat- Clinical services are provided in a variety of settings, from the ment at a time convenient for the specialist concerned, a patient's home to ambulatory care facilities and hospitals process managed by senior nurses. The new model requires providing inpatient care. They include those services that new health worker roles, which might be termed case 1342 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others managers. These case managers help patients to navigate with the local staff, leading to more efficient use of resources the system. (Levy-Bruhl and others 1997). In Zambia and Tanzania, · A third reason is the contrast between the rapidly changing strengthening of management capacities in primary care facil- demands on hospitals and the structural rigidities of hospi- ities through a team-based approach to decision making that tals themselves (McKee and Healy 2000). The original justi- linked planning to budgeted action plans led to improved client fication for creating hospitals as institutions was the need perceptions of facilities and to a marked increase in utilization to concentrate equipment such as operating theaters, x-ray (Few and Harpham 2003). However, the challenges of manage- machines, and laboratories, and expertise such as medical ment in the ambulatory care sector are great in many countries; specialists. Yet changes in the nature of health care are rais- this sector is often highly fragmented, with extensive and ing questions about how hospitals of the future should be largely unregulated private provision and few levers to exert configured. Many laboratory functions are being replaced pressure for change. by testing kits that can be used at the bedside, diagnostic equipment such as ultrasound scanners is being used in pri- mary care, and a new generation of primary care workers Community and Social Care are acquiring greatly augmented skills. In this rapidly chang- A particular challenge is how best to link long-term manage- ing environment, managers may be faced with aging hospi- ment of medical conditions with community and social care in tal buildings that may lack sufficient electrical sockets for those cases in which an effective response to health needs spans the greatly increased amount of equipment now available, the interface. Management of chronic physical or mental illness or managers may have staff members with deeply ingrained in the elderly, for example, can fall under the responsibility of ways of working who pose a particularly acute managerial home care and volunteer agencies, day centers, day hospitals, problem. rehabilitation hospitals, and long-term care institutions, as well as community-based health teams (Bergman, Beland, and These issues can be seen in the Kenyatta National Hospital Perrault 2002). A systematic review of community-based care in Nairobi, where a new managerial approach was developed for elderly people in industrial countries concluded that such but faced problems because of an unclear understanding of the schemes can favorably affect rates of institutionalization and kind of services to be provided, weak managerial capacity, and costs. However, comprehensive approaches involving program a lack of focus in targeting services (Collins and others 1999). restructuring are often necessary, and cost-effectiveness In Zambia, financial management and accountability improved depends on the characteristics of the health and social care sys- when the hiring of hospital staff was delinked from the national tems. The review's authors identified as a critical challenge the civil service, yet the process has been derailed on a wider scale expansion of those programs considered to be successful because of trade unions' resistance to the changes (Hanson and (Johri, Beland, and Bergman 2003). others 2002). Low-income countries face particular obstacles because they often lack effective alternatives to hospital care. As a result, patients are frequently cared for by their families but with little Ambulatory Care support, or they are consigned to large, poorly equipped, and Ambulatory care, delivered on an outpatient basis, is the com- poorly staffed institutions. This situation has stimulated the monest form of contact between patients and health care development of models of "community care," in which health providers. Although it often receives relatively little attention care providers work with communities to deliver services. In from policy makers compared with the more resource- the area of mental health, for example, the World Health intensive inpatient hospital care, ambulatory care contributes Organization (WHO) has developed models of care that cover substantially to health care system performance (Berman a range of care settings, including community centers and out- 2000). Good management of ambulatory services is essential reach services and residential homes, backed up by access because these services are often the entry and exit points for to hospital outpatient and emergency care (WHO 2001). consumers; however, these services can be difficult to manage Similarly, the complexities of caring for people with disabilities effectively (Waghorn and McKee 2000). in low-income countries have led to internationally developed Effective coordination of ambulatory and inpatient services guidelines that advocate a shared role for heath care providers is needed to ensure that patients are cared for in the most and local communities (Helander 2000). Accordingly, effective appropriate settings, thereby reducing inefficiencies such as the coordination of those services clearly depends in large part on overuse of hospitals for nonemergency care. Such coordination effective management. Shifting from hospital-focused care to often involves developing shared protocols for referral and community care introduces many managerial challenges. One management. In Benin and Guinea, for example, diagnostic element of an effective response should be to heighten the and treatment decision trees were developed collaboratively autonomy of patients in managing their diseases, but this Strategic Management of Clinical Services | 1343 response requires attitudinal changes among providers, who mixed, training programs were overall more likely to have pos- must commit to a real shift of power to patients, supported by itive rather than negative effects. Several studies focused on effective information systems and safeguards for vulnerable communication and counseling skills, which often lead to patients (Litwin and Lightman 1996). improved client satisfaction. A study from Zambia showed that Health care systems are generally poor at addressing long- training must be linked to other resources; although training term illnesses, especially when those illnesses require integrated was associated with improved transmission of information, care spanning primary, secondary, and community providers there was no decline in the number of complaints from clients (McKee and Nolte 2004; WHO 2001). The often low status who remained unhappy about long waits and short contact accorded to these conditions in the hierarchy of priorities, cou- time (Faxelid and others 1997). pled with fragmentation between health and social sectors (WHO 2002), will require greater commitment to managerial Changing Skill Mix. The division of tasks among different reforms that can improve the delivery of appropriate services. health care workers reflects many considerations, but evidence about who would be best at doing these tasks is rarely considered. There may be regulations restricting tasks to one THE SPECIFICS: WHAT WORKS? professional group, such as the right to prescribe, or there may be cultural norms, which while unwritten have just as great an This section turns to those policies that are designed to effect. Underlying these factors is a set of issues that includes a enhance the resources available to deliver health care and to difference in the power of different professions, itself often a combine them in ways that optimize the potential benefits. It reflection of gender relationships in society, with a predomi- looks, in turn, at the different elements required to deliver nantly male medical profession controlling a predominantly effective care: human resources, physical resources, intellectual female nursing profession. However, increasing evidence sug- resources, and the organizational or social resources that bind gests that traditional demarcations do not support the optimal them together. The section begins with the most important ways to provide care, and there is considerable scope for chang- resources for health care systems: the people who provide care. ing the mix of skills involved in delivering many aspects of health care. This topic has recently been reviewed systematically by Developing Human Resources Sibbald, Shen, and McBride (2004), who have developed a tax- A key element in the delivery of effective health care is how to onomy of the types of change in skill mixes that are possible provide staff members with the appropriate combination of (table 73.1). Their review shows that many tasks undertaken by skills to do their jobs effectively. one professional group can yield comparable and often better results when performed by another group. In particular, they Increasing Skills. In their review undertaken to inform the show how nurse-led clinics often achieve better outcomes than Commission on Macroeconomics and Health, Oliveira-Cruz, traditional doctor-led service (Connor, Wright, and Fegan Hanson, and Mills (2003) identified 13 studies that assessed the 2002; Stromberg and others 2003; Vrijhoef, Diederiks, and effects of training to enhance skills. Though the results were Spreeuwenberg 2000; Vrijhoef and others 2001, 2003). Table 73.1 A Taxonomy of Changes in Skill Mix in Health Care Changing roles Enhancement Increasing the depth of a job by extending the role or skills of a particular group of workers Substitution Expanding the breadth of a job, in particular by working across professional divides or exchanging one type of worker for another Delegation Moving a task down a traditional unidisciplinary ladder Innovation Creating new jobs by introducing a new type of worker Changing the interface between services Transfer Moving the provision of a service from one health care setting to another (for example, substituting community for hospital care) Relocation Shifting the venue from which a service is provided from one health care sector to another without changing the people who provide it (for example, running a hospital clinic in a primary care facility) Liaison Using specialists in one health care sector to educate and support staff members working in another (for example, hospital outreach facilitators in primary care) Source: Sibbald, Shen, and McBride 2004. 1344 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others Although Sibbald, Shen, and McBride focus their review on needs (England 2000; Preker, Harding, and Travis 2000). experience in industrial countries, by challenging many deeply Capital charging--requiring managers to explicitly account for held beliefs they indicate what could be done in other settings the value of physical assets out of funding allocation or con- around the world, after taking into account local circumstances tract revenues--has been developed as a response, successfully such as the skills and expertise of those involved, as well as any heightening public sector management of capital investments salient regulatory or training issues. in the United Kingdom and New Zealand (Heald and Scott 1996). Capital charging has been proposed as a strategy to Strengthening Management stimulate better capital management in developing countries as well. For example, in Malaysia a corporatized hospital has been In their review of constraints to health service delivery, required to reimburse invested capital through dividends, with Oliveira-Cruz, Hanson, and Mills (2003) identified 10 studies the Malaysian government recouping one-third of its original that evaluated the effect of management strengthening. The investment within five years (Hussein and Al-Junid 2003). activities in those studies included the following: Similarly, the Kenyatta National Hospital in Nairobi was obliged to account for all accruals (for example, property and · workshops for identifying and prioritizing managerial depreciation) when it was given greater autonomy. Though programs changes in accounting management have experienced some · introduction of regular planning and evaluation cycles shortcomings, improvements have been seen in financial trans- · quality assurance methods parency, timeliness of reporting, donor satisfaction, and rev- · establishment of routine communication systems enue collection (Collins and others 1999). · training activities. Within the public sector, changes in line management have facilitated the incorporation of more explicit infrastructure They concluded that the results were generally positive, with concerns into the planning process. The central authority in more rational use of funds; greater availability of funds as a Hong Kong (China) has made capital acquisition decisions consequence of better planning; improved coordination and jointly with hospitals during annual planning processes (Yip integration of programs; improved methods of working; better and Hsiao 2003). The introduction of business planning to dis- staff morale; enhanced data collection, reporting, and use; and trict-level planning in Turkmenistan heightened accountability increased community participation. WHO has developed an for maintaining physical infrastructure: use of a global budget- approach to strengthening management that has been success- ing model (that is, increased autonomy in line management as ful in a variety of settings (Cassels and Janovsky 1995). well as performance monitoring) led to reduced resource allo- It is important to identify where specific managerial skills cation to personnel and a greater than fivefold increase in are lacking and to explore different ways of obtaining them, maintenance expenditures (Ensor and Amannyazova 2000). whether through training, recruitment, or links with related Explicitly managing capital investments in both the short and organizations. For example, improved financial management the long term may facilitate efficient resource allocation. in district health teams in Ghana was made possible by inte- Although capital charging is a relatively straightforward grating staff members from local government accounts offices technical solution, capital investment can be particularly sus- (Kanlisi 1991); a similar initiative was successful in The ceptible to political derailment (Anell and Barnum 1998). In Gambia (Conn, Jenkins, and Touray 1996). However, a word of the hospital sector, for instance, many transition economy caution is required. Although a management strengthening countries have had difficulty downsizing infrastructure exercise undertaken in Tanzania was successful when imple- because those with decision rights to manage capital (that is, mented at the local level, it failed when scaled up because the local governments) are different from those who have incen- same degree of involvement by the originating team was no tives to do so, such as hospital managers (Jakab and Preker longer possible (Barnett and Ndeki 1992). 2003). Managing Physical Resources Managing infrastructure and other capital assets such as hospi- Strengthening Drug Procurement, Regulation, tals and health centers requires investment planning in both and Distribution the short term (for example, maintenance) and the long term Managing pharmaceutical resources is crucial for ensuring (for example, new acquisitions). Historically, however, costs access to essential drugs and promoting their rational use associated with capital consumption and maintenance have not (Mossialos, Mrazek, and Walley 2004). WHO defines the goals been met through operating budgets, resulting in few incen- of rational use of drugs as delivering medications effectively-- tives for public sector health planners to efficiently manage appropriate to patients' clinical needs and at dose levels and infrastructure or to respond to market demand and consumer durations appropriate to their individual requirements--and Strategic Management of Clinical Services | 1345 at an affordable cost (WHO 1985). The public sector plays a from those same stakeholders (Kwon 2003). Management of key role in providing the framework for rational use of drugs pharmaceuticals thus presents a complicated challenge, requir- (Quick 1997) through measures ranging from drug regulation ing significant investment and flexible responses. to clinical practice guidelines. National drug policies (NDPs) can be effective in regulating private and public sector provision of essential medicines. The Using Intellectual Resources Lao People's Democratic Republic's NDP has been important The process of generating, disseminating, and using knowledge in improving private pharmacy service quality (Stenson, is frequently imperfect. Pang and others (2003) have argued Tomson, and Syhakhang 1997). In Burkina Faso, an NDP has that a well-functioning health care system must have in place enhanced the performance of rural pharmacies (Krause and mechanisms that allow it to access and use research and the others 1998). At the local and facility levels, increasing products of research. They highlight the weaknesses of much of accountability can also lead to a more rational use of drugs. A the existing health care research, including fragmentation, simulation exercise in Tunisia that required physicians to relate overspecialization, and damaging competition among re- pharmaceutical budgeting to involvement in the procurement searchers, who are frequently isolated from other researchers process improved prescribing practices by containing costs and from the policy-making community. Drawing on concepts while increasing the use of essential drugs (Garraoui, Le of the functions of a health system, they identify a series of four Feuvre, and Ledoux 1999). Enhanced management informa- roles for a health research system: tion systems, with corresponding supervision, monitoring, and top-level support, have improved contraceptive management · stewardship, which includes defining and articulating a in several countries (Kinzett and Bates 2000). The introduction vision for a national research system, identifying appropri- of standard treatment guidelines and formularies has reduced ate priorities, and setting and monitoring ethical standards overprescribing in several countries, and educational materials · financing, which includes obtaining research funds and allo- for consumers in Cameroon increased compliance with antibi- cating them accountably otic regimens (Nabiswa, Makokha, and Godfrey 1993). · creating and sustaining resources, which includes the physical A comprehensive review of interventions used in Sub- and human capacity to conduct, absorb, and use research Saharan Africa, where health systems are plagued by shortages · producing and using research, which includes generating of supplies, high costs, large-scale use of proprietary drugs, valid research outputs; translating research into formats that waste, and theft, provided considerable evidence to suggest inform health policy, practices, and public opinion; and what works in those countries (Foster 1991). Successful inter- promoting the use of research to support innovation. ventions included the following: Such a system must be able to answer the many different · selection and precise quantification of drug needs--in questions requiring research, from basic laboratory science, particular, the creation of essential drug lists such as new drug development, through health services · improved procurement, with greater use of generics, com- research, such as comparisons of the cost-effectiveness of dif- petitive bidding, and international procurement agencies ferent drug regimens, to organizational research, such as the · improved storage and distribution, with better storage con- best way of delivering the most cost-effective drug regimen. ditions, inventory controls, security systems, and use of Although the majority of health systems and services research prepacked kits. continues to be undertaken in the industrial countries, a grow- ing volume of research addresses the needs of low- and middle- At the same time, several factors constrain better manage- income countries, such as that by the participants in the ment of pharmaceuticals. Considerable resources are needed to Effective Health Care Alliance Programme (EHCAP), an inter- adequately monitor NDPs, and implementation can be difficult national research network that is undertaking systematic (Petrova 2002). Furthermore, much of the pharmaceutical use research within the framework of the Cochrane Collaboration is outside the control of the public sector: two-thirds or more of (http://www.liv.ac.uk/lstm/ehcap/introduction.htm). health problems are self-medicated. Though the public sector may strive to inform consumers, patients' nonadherence remains high (Le Grand, Hogerzeil, and Haaijer-Ruskamp Establishing Relationships 1999). As in management of other inputs, political considera- The debate about the relative benefits of vertical (in which a tions can thwart managerial responses. The Republic of Korea single disorder is tackled by a program managed across levels decided to divide its prescribing and dispensing functions from the Ministry of Health to the health care provider) and precisely to address high levels of pharmaceutical overuse and horizontal (in which health care for a wide range of disorders misuse, but it subsequently faced strikes and stiff opposition is delivered through a system that is integrated at each level) 1346 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others systems of health care delivery has been examined in detail in a and subsequent attrition of volunteers. The programs' develop- major review of relevant literature by Oliveira-Cruz, Kurowski, ers will have learned lessons from other similar programs, in and Mills (2003). They note how many activities lie on a con- relation to both organizational and technical issues. Where tinuum between the two extremes, with the Global Polio several vertical programs coexist, the programs' developers Eradication Initiative more vertical than the Expanded should explore how they can share common elements. Programme on Immunization, which in turn is more vertical than the integrated management of childhood illness approach. They identify certain features that are often associ- Contracting for Services ated with vertical programs and that promote success: specific The setting of contracts by public agencies to purchase health objectives, clear work schedules, well-defined techniques, and care services is increasingly common in a number of low- and frequent supervision. They also identify characteristics that are middle-income countries. The theoretical case for contracting often associated with horizontal programs and that can ham- out identifies potential advantages from combining public per effectiveness: shortage of essential drugs, lack of adequate finance with private provision. However, there may also be dif- staff training, intermittent supervision, and limited backup. ficulties, such as ensuring that competition takes place among However, they note that horizontal programs have considerable potential contractors, that competition leads to efficiency, and potential to deliver effective services if they are adequately that contracts and the process of contracting are effectively funded, staffed, and managed, largely because of their managed; consequently, these advantages may not always be economies of scale and scope. realized (McPake and Banda 1994). To some extent, the approach is determined by the nature of Unfortunately, the question of whether the advantages out- the program. Vertical programs are most effective when the weigh the disadvantages has been the subject of relatively little technology involved is very sophisticated or when it includes empirical study in low- and middle-income countries, and procedures different from the usual tasks and thus requires spe- what exists is often highly context specific. For example, in cialist skills. Vertical programs may be more appropriate when Zimbabwe, a comparison of a hospital owned by a colliery, there is a need to rapidly achieve major reductions in the bur- from which services were purchased by the government, and a den of a disease, although this situation does not preclude nearby government hospital found that the colliery hospital embedding the management of the program within existing offered services of at least comparable quality at prices lower organizations. These programs are often a response to weak than the unit costs of the government hospital after capital management capacity in the existing system, although it is costs were included (McPake and Hongoro 1995). However, argued that they can perpetuate this problem or even under- failure to establish policies on thresholds for use meant that mine what does exist, diverting the attention of staff members growth in expenditure on the colliery hospital was not con- from their usual tasks. Such programs often have a short time trolled. The authors argue that contracted facilities can achieve horizon, either being absorbed into existing systems or brought powerful bargaining positions if there are no viable competi- to an end. In part, their duration is linked to the source of their tors and the government does not retain the ability to offer an funding, which is often from donors who themselves have a alternative service. They also identify a need for specific skills to short time horizon. manage contracts at all levels. Where a policy of contracting is Integrating previously vertical programs into mainstream a response to crises arising from civil service retrenchment systems can be successful, as with schistosomiasis programs in and public expenditure cuts, these skills are unlikely to be Saudi Arabia (Ageel and Amin 1997) and Brazil (Coura Filho developed. and others 1992). However, a systematic review of integration Another study examined the economic arguments for con- failed to identify consistent benefits, largely because of the very tracting for district hospital care in South Africa, by using pri- limited extent of the evidence available and the context-specific vate for-profit providers, and in Zimbabwe, by using non- nature of this process (Briggs, Capdegelle, and Garner 2001). governmental (mission) providers (Mills, Hongoro, and The authors of that review concluded that the question facing Broomberg 1997). In the South African setting, there were no policy makers is not whether one approach is invariably better significant differences in quality among three contractor hospi- than the other; rather, it is how best to build on the synergies tals and three government-run hospitals, but the contractor among them to maximize overall benefits. They note, for hospitals provided care at significantly lower unit costs. example, how the many successes of the Malaria Eradication However, the overall cost to the government was similar for the Programme in the 1950s and 1960s were not sustained because two options because of the additional cost of contracting, with active case surveillance was not integrated into routine health the efficiency gains captured almost entirely by the contractor. services (see also Bradley 1998). In Zimbabwe, two district-designated mission hospitals deliv- Successful vertical programs are likely to involve community ered similar quality care at lower cost than did two government participation, but not to the extent that there is overdependence hospitals. However, the contract between the government and Strategic Management of Clinical Services | 1347 the missions was implicit, rather than explicit, and was of long somewhat mixed. Thus, a study of dispensaries run by the gov- standing. As in the other Zimbabwean example, the authors ernment and by nongovernmental organizations in Tanzania identified the importance of developing the government's found considerable variation in both sectors (Gilson 1995). capacity to design and negotiate contracts that allow the gov- This finding was consistent with another study in Tanzania of ernment to derive significant efficiency gains from contractual primary care providers in Dar es Salaam. In the latter, although arrangements. the quality of care offered by private providers was, on average, better, much low-quality care was found in both types of facil- ities (Kanji and others 1995). Considerable variation in Increasing Provider Autonomy providers of both types, although with overall better quality A review of cross-country experiences with enhanced auton- in the private sector, was also reported in a study in Senegal omy of hospitals found improvements in service delivery. The (Bitran 1995). In summary, there is very little evidence to sup- most successful cases--in Hong Kong (China) and Tunisia-- port the contention that private provision is better than public applied private sector management techniques and training, provision, and what evidence exists indicates considerable vari- with appropriate performance assessment systems for staff. In ations in both. countries where reforms were considered less successful, man- agers had been granted greater autonomy without suitable performance-oriented incentives (New Zealand) or vice versa Strategic Purchasing (Indonesia) (Hawkins and Ham 2003). In the Kenyatta The quest to deliver effective health care is a dynamic process, National Hospital, greater autonomy led to the introduction of adapting continually to changing health needs and the oppor- performance appraisal linked to incentives, enabling the dis- tunities that arise that make it possible to respond in new and missal of poor performers and increased benefits and greater better ways. However, health systems that have failed in the past responsibilities for good performers. This change was coupled to respond to these changing circumstances face even greater with clarification of clinical management roles. Comple- problems. The pace of change is constantly increasing, with fac- menting increased salaries for staff nurses, these changes tors such as greater population mobility contributing to the helped improve the hospital's strategic management, donor reemergence of infectious diseases and with demographic accountability, and performance reporting (Collins and others changes and lifestyle changes giving rise to a new burden of 1999). chronic diseases. Implementing such management strategies in a coherent Health care providers have faced difficulties in responding fashion is not an easy task. Hospital governance in several to this challenge on their own. Although they may possess a Eastern European countries, which has been transferred to great deal of information about the patient sitting in front of local governments to improve responsiveness, has included them and, on the basis of their training and accumulated expe- measures such as performance-based payment mechanisms. rience, about what might be done to help that patient, health Performance did not improve as expected because of an care providers confront several important information gaps: "inconsistent incentive environment"; rewards and sanctions were not linked to performance. Important factors in that fail- · First, they may know little about those who, despite being in ure to improve were weak stewardship functions and an need of health care, do not seek help. These people will often absence of effective governance at the regional level, which be the most disadvantaged in a society, with few means of made it difficult to change the initial configuration of the hos- making their voices heard. pital system. Instead, increased hospital autonomy was used to · Second, they may have inadequate knowledge about newly ensure the survival of the institution rather than to meet the emerging treatments or more effective ways of providing needs of the population. Thus, a continuing excess of capacity, those treatments, especially if the treatments involve creat- inefficiency, and poor responsiveness to patient expectations ing multidisciplinary teams with new sets of skills, working remains (Healy and McKee 2002a). A review of experience with in ways outside their experience. programs that increased autonomy in Sub-Saharan Africa also · Third, even if providers introduce changes, they may have identified only modest success in achieving the stated goals inadequate knowledge of whether such changes have been (McPake 1996). effective. These knowledge gaps provide the justification for action to Public or Private Provision? improve the delivery of health care at several levels above that Although there has been considerable enthusiasm for privatiz- of the individual encounter between the patient and the health ing state facilities because of the supposed efficiency gains professional. Strategic purchasing brings together a series of achieved in the private sector, in reality the evidence is interlinked activities: assessing health needs, using appropriate 1348 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others Health strategy SUMMARY Health systems worldwide face unprecedented challenges in Assess needs responding to the increasing complexity of health care. Systems that were capable of providing basic care to populations for whom diseases were either simple or complex but self-limiting confront a fatal struggle to keep up with the increasing oppor- tunities that modern science has provided. The challenge is especially great for health systems in low- and middle-income Monitor Specify care countries because the global community is no longer willing to outcome model sit back while millions of people die from treatable diseases such as malaria and tuberculosis and fail to receive life- prolonging therapies for AIDS. As a consequence, some of the resources, primarily pharmaceuticals, are being made available to those who need them. However, the challenge that health Purchase systems face is not simply a lack of money to purchase phar- care maceuticals; effective management systems are requisite as well to create the infrastructure to identify those in need, establish Source: McKee and Brand 2005. appropriate treatments, and ensure provision of these treat- ments as long as necessary. Emerging challenges must be iden- Figure 73.1 A Framework for Strategic Purchasing tified, and the necessary resources to deal with them must be brought together and applied effectively. evidence to develop models of care that meet priority needs, Many countries have a clear need to invest in the develop- creating the appropriate combination of regulations and incen- ment of human resources. Although in many cases this invest- tives to implement those models, and then evaluating the ment will require new and wide-ranging human resource response and reassessing whether the need remains (Figure strategies involving training, career progression, and retention, 73.1). All of these activities should take place within an overall there seems to be scope for rapid gains from some shorter health strategy that takes into account the goals of a health care forms of training, particularly, in communications skills. system, such as those defined by WHO (2000), of increasing Although the evidence for the effectiveness of current models health attainment, providing services responsive to the popula- of management strengthening is somewhat mixed, gains may tion's needs and expectations, and financing those services be realized by identifying and filling key gaps, such as those in equitably. financial expertise. Changing the skill mix can do much to The development of a strategic purchasing function is com- match available skills to tasks. plicated, requiring high levels of information resources, both Much also can be done to manage the capital stock better or, on health needs and on effectiveness. Strategic purchasing in most cases, to manage it at all. For example, mechanisms involves using technical and political skills, determining the such as capital charging can focus greater management atten- needs of the population, identifying evidence of the effective- tion on this issue, although this will work only if sufficient ness of different care packages, and setting priorities within capacity can be focused. Important gains can be made from limited resources. The last of these components is arguably the better management of pharmaceuticals, an issue of increasing most difficult, given the high level of need and the scarcity of importance because of the new funds made available for their resources in many places. This list of components highlights purchase. why, in addition to having skills in financial and personnel Modern health care is based on the growth of knowledge, management, the effective health service manager needs at least and it is as important to manage intellectual resources as it is to a working knowledge of clinical epidemiology and economic manage people and equipment. Doing so means investing in a evaluation. health research strategy that includes the generation, synthesis, Even in industrial countries, the strategic purchasing func- and adoption of knowledge. tion is often poorly developed. Given its many interlinked com- Finally, it is necessary to bring these resources together opti- ponents and the problem of isolating any benefits from wider mally, which raises issues of relationships between different changes in the health system, this function is very difficult to levels of the system, between the public and private sectors, and evaluate. Nonetheless, it is included here as a model from between vertical and horizontal programs. Unfortunately, which concepts may be adopted in low- and middle-income despite the large amount of rhetoric on these often highly ide- settings. ological issues, there is surprisingly little research to inform Strategic Management of Clinical Services | 1349 policy. More than ever, the issue of context specificity Control in Peri-Peri, MG, Brazil." Revisto do Instituto de Medicina reemerges, leading once more to the answer "it depends." Tropical de São Paulo 34: 543­48. The delivery of optimal health care requires well-developed Davis, D., M. Thomson, A. Oxman, and R. Haynes. 1995. "Changing Physician Performance. A Systematic Review of the Effect of managerial skills to apply methods that are appropriate for Continuing Medical Education Strategies." Journal of the American the setting in which they are being applied. However, it also Medical Association 274: 700­05. requires governments to provide oversight of their health sys- de Soto, H. 2000. The Mystery of Capital: Why Capitalism Succeeds in the tems and to anticipate changes and give managers the tools West and Fails Everywhere Else. Boulder CO: Basic Books. with which to respond to those changes. England, R. 2000. Contracting and Performance Management in the Health Sector: A Guide for Low and Middle Income Countries. London: Depart- ment for International Development Health Systems Resource Centre. REFERENCES Ensor, T., and B. Amannyazova. 2000. "Use of Business Planning Methods to Monitor Global Health Budgets in Turkmenistan." Bulletin of the World Health Organization 78: 1045­53. Ageel, A. R., and M. A. Amin. 1997. "Integration of Schistosomiasis- Control Activities into the Primary-Health-Care System in the Gizan Enthoven, A. C. 1985. Reflections on the Management of the National Health Region, Saudi Arabia." Annals of Tropical Medicine and Parasitology 91: Service. London: Nuffield Provincial Hospitals Trust. 907­15. Evans, R. G. 1997. "Health Care Reform: Who's Selling the Market, and Aiken, L., J. Sochalski, and E. Lake. 1997. "Studying Outcomes of Why?" Journal of Public Health Medicine 19: 45­49. Organizational Change in Health Services." Medical Care 35 (11 Farmer, P., L. Reichman, and M. Iseman. 1999. The Global Impact of Drug Suppl.): NS6­18. Resistant Tuberculosis. Boston, MA: Harvard Medical School and Open Andreev, E. M., E. Nolte, V. M. Shkolnikov, E. Varavikova, and M. McKee. Society Institute. 2003. "The Evolving Pattern of Avoidable Mortality in Russia." Faxelid, E., B. M. Ahlberg, S. Freudenthal, J. Ndulo, and I. Krantz. 1997. International Journal of Epidemiology 32: 437­46. "Quality of STD Care in Zambia: Impact of Training in STD Anell, A., and H. Barnum. 1998. "The Allocation of Capital and Health Management." International Journal for Quality in Health Care 9: Sector Reform." In Critical Challenges for Health Care Reform in 361­66. Europe, ed. R. B. Saltman, J. Figueras, and C. Sakellarides, 179­96. Fayol, H. 1949. General and Industrial Management. Translated from the Buckingham, U.K.: Open University Press. original, Administration Industriele et Générale, 1916. London: Pitman. Arrow, K. 1963."Uncertainty and the Welfare Economics of Medical Care." Ferlie, E., and S. Shortell. 2001. "Improving the Quality of Health Care in American Economic Review 53: 941­73. the United Kingdom and the United States: A Framework for Change." Barnett, E., and S. Ndeki. 1992. "Action-Based Learning to Improve Milbank Quarterly 79: 281­315. District Management: A Case Study from Tanzania." International Few, R., and T. Harpham. 2003. "Urban Primary Health Care in Africa: A Journal of Health Plan Management 7: 299­308. Comparative Analysis of City-Wide Public Sector Projects in Lusaka Baumol, W., J. Panzar, and R. Willig. 1982. Contestable Markets and the and Dar es Salaam." Health and Place 9: 45­53. Theory of Industrial Structure. New York: Harcourt Brace Jovanovich. Foster, S. 1991. "Supply and Use of Essential Drugs in Sub-Saharan Africa: Bergman, H., F. Beland, and A. Perrault. 2002. "The Global Challenge of Some Issues and Possible Solutions." Social Science and Medicine 32: Understanding and Meeting the Needs of the Frail Older Population." 1201­18. Aging Clinical and Experimental Research 14: 223­25. Garraoui, A., P. Le Feuvre, and M. Ledoux. 1999. "Introducing Berman, P. 2000. "Organization of Ambulatory Care Provision: A Critical Management Principles into the Supply and Distribution of Medicines Determinant of Health System Performance in Developing Countries." in Tunisia." Bulletin of the World Health Organization 77: 525­29. Bulletin of the World Health Organization 78: 791­802. Gilson, L. 1995. "Management and Health Care Reform in Sub-Saharan Bitran, R. 1995. "Efficiency and Quality in the Public and Private Sectors Africa." Social Science and Medicine 40: 695­710. in Senegal." Health Policy and Planning 10: 271­83. Griffiths, R. 1984. National Health Service Management Inquiry Report. Bradley, D. J. 1998."The Particular and the General: Issues of Specificity and London: Department of Health and Social Security. Verticality in the History of Malaria Control." Parassitologia 40: 5­10. Hanson, K., L. Atuyambe, J. Kamwanga, B. McPake, O. Mungule, and F. Briggs, C. J., P. Capdegelle, and P. Garner. 2001. "Strategies for Integrating Ssengooba. 2002. "Towards Improving Hospital Performance in Primary Health Services in Middle- and Low-Income Countries: Uganda and Zambia: Reflections and Opportunities for Autonomy." Effects on Performance, Costs, and Patient Outcomes." Cochrane Health Policy 61: 73­94. Database of Systematic Reviews (4): CD003318. Hanson, K., M. K. Ranson, V. Oliveira-Cruz, and A. Mills. 2003. Cassels, A., and K. Janovsky. 1995. Strengthening Health Management in "Expanding Access to Priority Health Interventions: A Framework for Districts and Provinces. Geneva: World Health Organization. Understanding the Constraints to Scaling Up." Journal of International Collins, D., G. Njeru, J. Meme, and W. Newbrander. 1999. "Hospital Development 15: 1­14. Autonomy: The Experience of Kenyatta National Hospital." Harrison, S., and C. Pollitt. 1994. Controlling Health Professionals. International Journal of Health Planning and Management 14: 129­53. Buckingham, U.K.: Open University Press. Conn, C. P., P. Jenkins, and S. O. Touray. 1996. "Strengthening Health Hawkins, L., and C. Ham. 2003. "Reviewing the Case Studies: Tentative Management: Experience of District Teams in The Gambia." Health Lessons and Hypotheses for Future Testing." In Innovations in Health Policy and Planning 11: 64­71. Service Delivery: The Corporatization of Public Hospitals, ed. A. S. Connor, C. A., C. C. Wright, and C. D. Fegan. 2002. "The Safety and Preker and A. L. Harding. Washington, DC: World Bank. Effectiveness of a Nurse-Led Anticoagulant Service." Journal of Heald, D., and D. Scott. 1996. "Assessing Capital Charging in the National Advanced Nursing 38: 407­15. Health Service." Financial Accountability and Management 12: 225­44. Coura Filho, P., R. S. Rocha, E. De Lima, M. F. Costa, and N. Katz. 1992. "A Healy, J., and M. McKee. 2002a. "Implementing Hospital Reform in Municipal Level Approach to the Management of Schistosomiasis Central and Eastern Europe." Health Policy 61: 1­19. 1350 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others ------. "The Role and Function of Hospitals." In Hospitals in a Changing McPake, B., and E. E. Banda. 1994. "Contracting Out of Health Services in Europe, ed. M. McKee and J. Healy, 59­80. Buckingham, U.K.: Open Developing Countries." Health Policy and Planning 9: 25­30. University Press. McPake, B., and C. Hongoro. 1995. "Contracting Out of Clinical Services Helander, E. 2000. "25 Years of Community-Based Rehabilitation." Asia in Zimbabwe." Social Science and Medicine 41: 13­24. Pacific Disability Rehabilitation Journal 11: 4­9. Mehotra, S. 2000."Integrating Economic and Social Policy: Good Practices Hood, C. 1991. "A Public Management for All Seasons." Public from High Achieving Countries." Innocenti Working Paper 80, Administration 69: 3­19. UNICEF Innocenti Research Centre, Florence, Italy. Hussein, R., and S. Al-Junid. 2003. "Corporatization of a Single Facility: Mills, A., C. Hongoro, and J. Broomberg. 1997. "Improving the Efficiency Reforming the Malaysian National Heart Institute." In Innovations in of District Hospitals: Is Contracting an Option?" Tropical Medicine and Health Service Delivery: The Corporatization of Public Hospitals, ed. International Health 2: 116­26. A. S. Preker and A. L. Harding. Washington, DC: World Bank. Mossialos, E., M. Mrazek, and T. Walley. 2004. Regulating Pharmaceuticals Institute of Medicine. 2001. Crossing the Quality Chasm: A New Health in Europe: Striving for Efficiency, Equity, and Quality. Buckingham, System for the 21st Century. Washington, DC: Institute of Medicine. U.K.: Open University Press. Jakab, M., and A. S. Preker. 2003. "The Missing Link? Hospital Reform in Nabiswa, A. K., J. D. Makokha, and R. C. Godfrey. 1993. "Malaria: Impact Transition Economies." In Innovations in Health Service Delivery: The of a Standardized Protocol on Inpatient Management." Tropical Doctor Corporatization of Public Hospitals, ed. A. S. Preker and A. Harding. 23: 25­26. Washington, DC: World Bank. OECD (Organisation for Economic Co-Operation and Development). Johri, M., F. Beland, and H. Bergman. 2003. "International Experiments in 2003. OECD Health Data. Paris: OECD. Integrated Care for the Elderly: A Synthesis of the Evidence." Oliveira-Cruz, V., K. Hanson, and A. Mills. 2003. "Approaches to International Journal of Geriatric Psychiatry 18: 222­35. Overcoming Constraints to Effective Health Service Delivery: A Review Kanji, N., P. Kilima, N. Lorenz, and P. Garner. 1995. "Quality of Primary of the Evidence." Journal of International Development 15: 41­65. Outpatient Services in Dar-es-Salaam: A Comparison of Government Oliveira-Cruz, V., C. Kurowski, and A. Mills. 2003. "Delivery of Priority and Voluntary Providers." Health Policy and Planning 10: 186­90. Health Services: Searching for Synergies with the Vertical versus Kanlisi, N. 1991. "Strengthening District Health Teams in Ghana: The Horizontal Debate." Journal of International Development 15: 67­86. Experience of Ejisu District." Tropical Doctor 21: 98­100. Pang, T., R. Sadana, S. Hanney, Z. A. Bhutta, A. A. Hyder, and J. Simon. Kinzett, S., and J. Bates. 2000. Bangladesh: Contraceptive Logistics System, 2003. "Knowledge for Better Health: A Conceptual Framework and Review of Accomplishments and Lessons Learned. Arlington, VA: Family Foundation for Health Research Systems." Bulletin of the World Health Planning Logistics Management and John Snow Inc. Organization 81: 815­20. Krause, G., J. Benzler, R. Heinmuller, M. Borchert, E. Koob, K. Ouattara, and Petrova, G. I. 2002. "Prescription Patterns Analysis--Variations among H. J. Diesfeld. 1998. "Performance of Village Pharmacies and Patient Bulgaria, Romania, Macedonia, and Bosnia Herzegovina." Central Compliance after Implementation of Essential Drug Programme in European Journal of Public Health 10 (3): 100­03. Rural Burkina Faso." Health Policy and Planning 13: 159­66. Preker, A. S., A. Harding, and P. Travis. 2000. "`Make or Buy' Decisions in Kwon, S. 2003. "Pharmaceutical Reform and Physician Strikes in Korea: the Production of Health Care Goods and Services: New Insights from Separation of Drug Prescribing and Dispensing." Social Science and Institutional Economics and Organizational Theory." Bulletin of the Medicine 57: 529­38. World Health Organization 78: 779­90. Le Grand, A., H. V. Hogerzeil, and F. M. Haaijer-Ruskamp. 1999. Quick, J. D. 1997. Managing Drug Supply: The Selection, Procurement, "Intervention Research in Rational Use of Drugs: A Review." Health Distribution, and Use of Pharmaceuticals. West Hartford, CT: Kumarian Policy and Planning 14 (2): 89­102. Press. Le Grand, J., and W. Bartlett. 1993. Quasi-Markets and Social Policy. Shortell, S., C. Bennett, and G. Byck. 1998. "Assessing the Impact of London: Palgrave Macmillan. Continuous Quality Improvement on Clinical Practice: What It Will Levy-Bruhl, D., A. Soucat, R. Osseni, J. M. Ndiaye, B. Dieng, X. De Bethune, Take to Accelerate Progress." Milbank Quarterly 76: 593­624. and others. 1997. "The Bamako Initiative in Benin and Guinea: Shortell, S. M., and A. D. Kaluzny. 1983. Health Care Management: A Text Improving the Effectiveness of Primary Health Care." International in Organisation Theory and Behaviour. New York: Wiley. Journal of Health Planning and Management 12 (Suppl. 1): S49­79. Sibbald, B., J. Shen, and A. McBride. 2004. "Changing the Skill-Mix of the Litwin, H., and E. Lightman. 1996. "The Development of Community Health Care Workforce." Journal of Health Services Research and Policy Care Policy for the Elderly: A Comparative Perspective." International 2004 9: 28­38. Journal of Health Services 26: 691­708. Stenson, B., G. Tomson, and L. Syhakhang. 1997. "Pharmaceutical Maynard, A. 1998. "Competition and Quality: Rhetoric and Reality." Regulation in Context: The Case of Lao People's Democratic International Journal for Quality in Health Care 10: 379­84. Republic." Health Policy and Planning 12: 329­40. McKee, M., and H. Brand. 2005. "Purchasing to Promote Population Stewart, J. 1998. "Advance or Retreat: From the Traditions of Public Health." In Effective Purchasing for Health Gain, ed. J. Figueras, R. Administration to the New Public Management and Beyond." Public Robinson, and E. Jakubowski, 140­63. Buckingham, U.K.: Open Policy and Administration 13: 27. University Press. Stromberg, A., J. Martensson, B. Fridlund, L. A. Levin, J. E. Karlsson, and McKee, M., and J. Healy. 2000. "The Role of the Hospital in a Changing U. Dahlstrom. 2003. "Nurse-Led Heart Failure Clinics Improve Environment." Bulletin of the World Health Organization 78: 803­10. Survival and Self-Care Behaviour in Patients with Heart Failure: ------. 2001. "The Changing Role of the Hospital in Europe: Causes and Results from a Prospective, Randomised Trial." European Heart Journal Consequences." Clinical Medicine, 1: 299­304. 24: 1014­23. McKee, M., and E. Nolte. 2004. "Responding to the Challenge of Chronic Vrijhoef, H. J., J. P. Diederiks, and C. Spreeuwenberg. 2000. "Effects on Disease: Ideas from Europe." Clinical Medicine 4: 336­42. Quality of Care for Patients with NIDDM or COPD When the McPake, B. 1996. "Public Autonomous Hospitals in Sub-Saharan Africa: Specialised Nurse Has a Central Role: A Literature Review." Patient Trends and Issues." Health Policy 35: 155­77. Education and Counseling 41: 243­50. Strategic Management of Clinical Services | 1351 Vrijhoef, H. J., J. P. Diederiks, C. Spreeuwenberg, and B. H. Wolffenbuttel. Review of Major Issues." Paper prepared for the Conference of Experts 2001. "Substitution Model with Central Role for Nurse Specialist Is on the Rational Use of Drugs, Nairobi, Kenya, November 25­29. Justified in the Care for Stable Type 2 Diabetic Outpatients." Journal of ------. 2000. World Health Report 2000: Health Systems: Improving Advanced Nursing 36 (4): 546­55. Performance. Geneva: WHO. Vrijhoef, H. J., J. P. Diederiks, G. J. Wesseling, C. P. Van Schayck, and C. ------. 2001. Innovative Care for Chronic Conditions. Geneva: WHO. Spreeuwenberg. 2003. "Undiagnosed Patients and Patients at Risk for ------. 2002. Community Home-Based Care in Resource-Limited Settings: COPD in Primary Health Care: Early Detection with the Support of A Framework for Action. Geneva: WHO. Non-Physicians." Journal of Clinical Nursing 12: 366­73. Yip, W., and W. Hsiao. 2003. "Autonomizing a Hospital System: Corporate Waghorn, A., and M. McKee. 2000. "Why Is It So Difficult to Organise an Control by Central Authorities in Hong Kong." In Innovations in Outpatient Clinic?" Journal of Health Services Research and Policy 5: Health Service Delivery: The Corporatization of Public Hospitals, ed. 140­47. A. S. Preker and A. L. Harding. Washington, DC: World Bank. WHO (World Health Organization). 1985. "The Rational Use of Drugs: 1352 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others Glossary Age-standardized rate An age-standardized rate is a weighted bronchial airflow is usually reversible and between asthma average of the age-specific rates, where the weights are the pro- episodes the flow of air through the airways is usually good. portions of a standard population in the corresponding age CVD: Cardiovascular disease Cardiovascular disease covers a groups (q.v.). The potential confounding effect of age is wide array of disorders, including diseases of the cardiac mus- removed when comparing age-standardized rates computed cle and of the vascular system supplying the heart, brain, and using the same standard population. other vital organs. The most common manifestations of CVD Age weights Factor specifying the relative value of a year of are ischemic heart disease, congestive heart failure, and stroke. healthy life lived at different ages. The DALY can incorporate CVD is used here as an abbreviation for cardiovascular disease, non-uniform age weights which give less weight to years of life not cerebrovascular disease (q.v.) lived in early childhood and at older ages (see Chapter 5). DALY: Disability Adjusted Life Year A measure of the gap in AIDS: Acquired Immunodeficiency Syndrome Disease due to healthy years of life lived by a population as compared with a infection with the human immunodeficiency virus (HIV). normative standard. More formally, DALYs are a time based measure which adds together years of life lost due to premature BMI: Body mass index A measure of underweight and over- mortality with the equivalent number of years of life lived with weight calculated as weight (kg) divided by height squared (m2). disability or illness. DFLE: Disability-free life expectancy A form of HE (q.v.) Case Fatality Rate The proportion of cases of a disease or which gives a weight of 1 to states of health with no disability injury that die as a result of their disease or injury over a above an explicit or implicit threshold and a weight of 0 to specified time period. states of health with any level of disability above that threshold. CHD: Coronary heart disease Synonymous with ischemic DBP: Diastolic blood pressure heart disease (q.v.). Demography The study of population size, growth and age Childhood-cluster diseases GBD (q.v.) cause group including structure, and of the forces (fertility, mortality, migration) that the following vaccine-preventable diseases of childhood: pertussis, lead to population change. poliomyelitis, diphtheria, measles and tetanus. Disability Restriction or lack of ability (resulting from an CODMOD: Cause of death model A statistical model for the impairment or health condition) to perform an activity in the prediction of the broad distribution of causes of death based manner or within the range considered normal. Although the on observed historical data on the relationships between cause word "disability" is widely used, the ICF (q.v.) uses this term distributions, and overall levels of mortality and per-capita only as a broad umbrella term for capacity and performance in income. activity/participation domains. The GBD (q.v.) used the term Comorbidity Presence of more than one disease or health disability, as in the DALY (q.v.), as a synonym for health states condition in an individual at a given time. (q.v.) less than full health (q.v.). Disability is also commonly used to refer only to long-standing limitations in carrying out COPD: Chronic obstructive pulmonary disease Lung activities of daily living. diseases that persistently obstruct bronchial airflow. COPD mainly involves two related diseases--chronic bronchitis Disability weight Measure of the relative valuations of a and emphysema. COPD is also called chronic obstructive lung health state on an interval scale. In the GBD (q.v.), health disease. Asthma is not included in COPD, as the obstruction to state valuations lie between 0 (full health q.v.) and 1 (states 1353 equivalent to death). The disability weight quantifies judg- Group II causes Major disease and injury cause group used in ments about overall levels of health associated with different GBD (q.v.). Comprises non-communicable diseases, including health states (q.v.), not judgments on the relative values of lives malignant neoplasms, cardiovascular diseases, chronic respira- lived, persons, or of overall well-being, quality of life or utility. tory diseases, digestive, musculoskeletal and genitourinary The GBD disability weights are intended to reflect average conditions, as well as mental disorders and neurological global valuations. conditions. Discounting Process applied to costs, benefits, and outcomes Group III causes Major disease and injury cause group used based on the concept that there is preference for money or in GBD (q.v.). Includes unintentional and intentional injuries. health in the present relative to the future. HALE: Health-adjusted life expectancy Any of a number of DisMod An epidemiological disease model linking popula- summary measures which use explicit weights to combine tions exposed to risk of disease with incident cases, prevalent health expectancies for a set of discrete health states into a sin- cases, case fatality and the duration of time lived with a disease gle indicator of the expectation of equivalent years of good or injury, including its sequelae. health. Also referred to as `Healthy life expectancy'. DSP Disease Surveillance Points System run by the Chinese HE: Health expectancy Generic term for summary measures Centre for Disease Control and Prevention for the surveillance of population health which estimate the expectation of years of of mortality and morbidity. life lived in various health states. Epidemiological transition The process whereby major com- Healthy life expectancy Synonym for HALE (q.v.) or Health- municable diseases and conditions of poverty (e.g. malnutri- adjusted life expectancy. tion) are progressively replaced by non-communicable diseases such as cancers and CVD. Health state Health state refers to an individual's levels of functioning within a set of health domains such as mobility, Epidemiology The study of the occurrence and causes of dis- cognition, pain, emotional functioning, self-care, etc. More ease and injury in populations. specifically, in terms of ICF (q.v.) concepts, health state is Full health Health state (q.v.) characterized by optimal levels defined as the capacities of an individual in all important of functioning or capacity in all the important domains of domains of health, where such domains may include domains health, and freedom from any type of illness or disease. The of body structure and function, and domains of activities/par- "optimal" levels of functioning are defined as those levels above ticipation. Health states do not include risk factors, diseases, which further gains would not (in general) be regarded as prognosis or the impact of health states on overall quality of improvements in health. States of exceptional functioning life, well-being or satisfaction. above these levels are thus considered to be talents or excep- Health status A general term referring to all aspects of the tional abilities, not higher states of health. health of individuals or populations. Usually understood to Garbage codes ICD codes (q.v.) for ill-defined or residual cat- include mortality risks, diseases, health states (q.v.), impair- egories of major disease groups (e.g. cardiovascular diseases) ments and disability. May also include some risk factors or that do not provide meaningful information on underlying prognosis information. disease or injury causes of death. Examples include ill-defined High income Category in the World Bank income grouping of primary site of cancer and atherosclerosis. countries used for countries with Gross National Income (GNI) GBD: Global burden of disease A comprehensive demographic per capita of US$9,206 or more (exchange rate adjusted curren- and epidemiological framework to estimate health gaps (q.v.) for cies) in 2001. See Table 3A-3 for list of countries included. an extensive set of disease and injury causes, and for major risk HIV Acronym for the Human Immunodeficiency Virus, the factors, using all available mortality and health data and meth- cause of AIDS (acquired immunodeficiency syndrome). ods to ensure internal consistency and comparability of esti- mates. In the first global burden of disease study, Murray and Ideal health Synonymous with full health (q.v.). Lopez estimated health gaps using DALYs (q.v.) for eight regions Incidence New cases of disease or injury occurring in a speci- of the world in 1990. This book presents updated estimates for fied population in a given time period. the year 2001 for the world and for World Bank regions. Incidence rate New cases of disease or injury occurring per Group I causes Major disease and injury cause group used in unit of population, per unit time. GBD (q.v.). Includes communicable, maternal, perinatal and nutritional conditions. These are causes which are ICD: International Statistical Classification of Diseases and characteristically common in populations who have not yet Related Health Problems A classification of diseases and completed the epidemiological transition (q.v.). other causes of mortality prepared by the World Health 1354 | Glossary Organization since 1948, periodically revised as necessary. The Perinatal causes or conditions The cause category Perinatal current tenth revision was issued in 1992 to come into effect on causes refers to the ICD cause group "Conditions arising in the 1 January 1993. The ICD is a member of the WHO family of perinatal period". Deaths from these causes (primarily low birth international classifications. weight and birth trauma/ asphyxia) may occur at any age, but are largely confined to the perinatal period. ICF: International Classification of Functioning, Disability and Health A classification of body structures and functions Period life expectancy A summary measure of a population's (impairments) and activities/participation domains (perform- mortality that measures the expectation of years of life lived by ance and capacity). The ICF was endorsed by the WHO World a fictitious birth cohort assuming that at each age the cohort Health Assembly in 2001 as a successor to the 1980 experiences the age-specific mortality rates observed in the real International Classification of Impairment, Disability and population during a specified time period (such as a given cal- Handicap (ICIDH). The ICF is a member of the WHO family of endar year). See also: life expectancy. international classifications. Postneonatal period Persons between the age of 28 days and 1 IHD: Ischemic heart disease Any of a number of heart condi- year are in the postneonatal period. tions in which heart muscle is damaged or works inefficiently Prevalence Actual number of cases of disease or injury present because of an absence or relative deficiency of its blood supply; in a population at any particular moment in time. most often caused by atherosclerosis, it includes angina pec- toris, acute myocardial infarction (heart attack), chronic Probability of death The chance that an individual, alive at ischemic heart disease and sudden death. The term coronary age x, will be dead before his or her (x n)th birthday, usually heart disease is synonymous with IHD. written as n qx . q0 denotes the probability that a newborn 5 infant will die before his or her fifth birthday. Life expectancy The average number of years of life expected to be lived by individuals who survive to a specific age. See also: PTO: person trade-off A method for valuation of health states Period life expectancy. that asks respondents to choose between hypothetical interven- tions that offer health benefits to groups of individuals in dif- Logit transformation A mathematical function that trans- ferent health states. forms a variable such as probability of death into another func- tional form, characterized by asymptotic values. QALY: Quality-adjusted life year A measure of years of life lived (or gained through an intervention) adjusted for quality Low- and middle-income Category in the World Bank of life using health state preferences ranging between 0 (states income grouping of countries used for countries with Gross equivalent to death) through to 1 (full health). QALYs National Income (GNI) per capita of less than US$9,206 in were developed for the assessment of the cost-effectiveness of 2001 (exchange rate adjusted currencies). See Table 3A-3 for list interventions in health economics. QALYs gained and DALYs of countries included. averted through an intervention are calculated in very similar MONICA Study The MONICA (MONItoring CArdiovascular ways, and the main differences relate to the interpretation of disease) Study was an international research project coordinat- the weights. Whereas the disability weights in the DALY ed by the World Health Organization from the mid-1980s to quantify loss of health, the corresponding QALY weights the mid-1990s in which teams from 38 populations in 21 coun- are often interpreted in terms of well-being, quality of life, tries studied heart disease, stroke and risk factors in their or utility. populations. Risk Factor A risk factor is an attribute or exposure which is causally associated with an increased probability of a disease Neonatal period Persons under the age of 28 days are in the or injury. neonatal period. The neonatal period is itself divided into the early neonatal period, age less than 7 days, and the remaining RR: Relative risk Relative risk is a measure of the strength of late neonatal period. an association. It is calculated as a ratio of the risk of occur- rence of a disease or death among two population groups, such PAF: population attributable fraction Proportional reduc- as those exposed to a risk factor and those not exposed. tion in disease or injury that would occur if population expo- sure to a risk factor or group of risk factors were reduced to an SD: Standard deviation A measure of the dispersion or spread alternative distribution. of values of a variable (e.g. body weight) around a population mean value. Perinatal deaths Includes stillbirths and neonatal deaths from any cause, including tetanus and congenital malformations. Sensitivity analysis Systematic investigation of the effects on The perinatal period includes the period from 27 weeks of ges- estimates or outcomes of changes in data or parameter inputs tation to 28 days of life. or assumptions. Glossary | 1355 Sequelae The medical conditions that can occur among peo- Sullivan's method A method of calculating health expectan- ple who contract a disease or suffer an injury. The GBD (q.v.) cies using data on the current prevalence of health states in a focuses on disabling sequelae of diseases and injuries; these population together with a period life table for the population. may remain present long after the initiating disease episode or Theoretical-minimum-risk exposure distribution The pop- injury event. ulation distribution of exposure to a risk factor that would Standard gamble (SG) A method for valuation of health states result in the lowest population disease burden. based on the axioms of expected utility theory. The standard TTO: time trade-off A method for valuation of health states gamble asks respondents to make choices that weigh improve- that asks respondents to make hypothetical choices that weigh ments in health against mortality risks. improvements in health against reduced longevity. Standard Population A population structure that is used to Uncertainty analysis Estimation of range or distribution of provide a constant age or covariate distribution, so that the uncertainty in estimates based on an assessment of the uncer- age- and sex-specific rates within different populations can be tainty or confidence intervals for all data and parameter inputs. applied to it and can be compared without confounding by the Uncertainty intervals should ideally include all sources of different age or covariate distributions of the populations. uncertainty, including those arising from systematic biases and STD: Sexually transmitted disease See: STI. measurement error. In contrast, generally reported confidence intervals are based solely on the variation observed in sample STI: Sexually transmitted infection An infection that can be data. transferred from one person to another through sexual contact. Among the sexually transmitted infections (STIs) are Visual analogue scale A method for valuation of health states HIV/AIDS, chlamydia, genital herpes, gonorrhea and syphilis. in which respondents are asked to directly assess health levels The term "sexually transmitted infection (STI)" corresponds to associated with different health states. Individuals place these the older term "sexually transmitted disease (STD)". on a 0 to 1 scale representing a continuum from health states considered equivalent to death through to full health (q.v.) SMPH: Summary measures of population health Indicators Verbal autopsy A method of inquiry to ascertain the likely that summarize the health of a population into a single number. cause of death in populations where vital registration of deaths SMPH combine information about mortality and population is incomplete and unreliable. Relatives of the deceased are health states.They may summarize either the average health level interviewed about symptoms and signs experienced by the or health inequality for a population. The two main classes of deceased prior to death, from which a diagnosis of the proba- summary measures are health expectancies (q.v.) and health gap ble cause of death is made. measures, of which the DALY (q.v.) is the best-known example. Vital registration A system for the registration of vital events Stillbirth Stillbirth refers to the birth of a dead fetus weighing in a population, including births and deaths, with medical cer- more than 1,000 grams up to 0.25 years (13 weeks) prior to the tification of the cause of death according to the rules and pro- expected time of birth (corresponding to 27 weeks of gesta- cedures of the ICD. tional age). YLD: Years Lived With Disability The component of the Stroke Stroke is defined as a condition that results in a disrup- DALY (q.v.) that measures lost years of healthy life through tion of blood flow to a region of the brain causing irreversible living in health states of less than full health (q.v.). "death" of brain tissue. There are two main types of stroke: hemorrhagic and ischemic stroke. Stroke is the main cause of YLL: Years of Life Lost The component of the DALY (q.v.) that mortality and burden for cerebrovascular disease (q.v.). measures years of life lost due to premature mortality. 1356 | Glossary Index Boxes, figures, notes, and tables are indicated by b, f, n, and t following the page number. abortions, 499, 510, 1077­1078, 1078t DALYs and, 1109­1111, 1110t Africa Down syndrome and, 942­945, 944t, 948 deworming and, 1116 See also specific countries and regions illegal and unsafe, 1078 drug use by, 1111 abortions in, 1078 integration into contraception economic benefits of interventions, African trypanosomiasis in, 456 programs, 1085 1117, 1120t asthma in, 686 postabortion care, 512­513 health challenges of, 1109 burn-related injuries in, 741 Abuja Declaration on HIV/AIDS, health risk behaviors of, 1111­1112 cleft lip and palate in, 731 Tuberculosis, and Other Related improving health systems to meet community-based health in, 1070­1071 Infectious Diseases, 229 needs of, 1122 contraceptive use in, 1082, 1086 accountability of health provider injuries and, 1111 coronary artery disease in, 833 organizations, 188 interventions, 1114­1117 dental treatment in, 729 accreditation lessons learned, 1120­1122 development assistance for health to, 245 hospitals, 1223­1224, 1224f, 1226, 1296 mass media as influential on, 1116 diarrheal diseases in, 778 traditional or alternative medicine, 1287 mental disorders and, 621, 1111 epilepsy in, 630 acne, 708 nutrition and, 1111­1112, 1116 excreta disposal in, 779, 780 ACT. See artemisinin combination therapy poverty as factor in health of, 1112 family planning in, 1079 acupuncture, 975, 1282, 1285, 1288 principles of health programming hemoglobinopathies in, 668, acute management, defined, 59 for, 1121 676­677, 678 acute myocardial infarction (AMI), promising but not proven interventions, hepatitis B in, 396 646, 650­651, 653­654 1116­1117 HIV and TB in, 299 See also cardiovascular disease (CVD) research and development agenda, hygiene promotion in, 784­785 acute pharyngitis in children, 483­484 1122­1123 insecticide-treated nets (ITNs) for acute renal failure, 698 scaling up of programs for, 1121 malaria prevention, 217 See also kidney and urinary sexual and reproductive behaviors of, interpersonal violence in, 757 system diseases 1112, 1113t, 1114­1116 kidney and urinary system diseases in, acute respiratory infections. See respiratory sexually transmitted infections 695, 697 diseases of adults; respiratory and, 1112 lymphatic filariasis in, 439, 440 diseases of children smoking and, 875, 1111, 1116 malaria in, 413, 414­415, 416t, 417­418, addiction. See alcohol use and control; drug social marketing and, 1116 420­421, 426 dependence; tobacco use and control in workplace, 1116 measles elimination campaign in, 98­99, ADHD (attention deficit hyperactivity youth development programs, 1116 168, 170, 173 disorder), 935b advertising mental disorders in, 620­621 adolescents and young adults, 1109­1125 See also social marketing neurological diseases, 641 alcohol use by, 1111 alcohol products, 894 onchocerciasis in, 957 burden of disease and, 1109­1112, 1110t food marketing aimed at Parkinson's disease in, 632 cannabis and, 918­919 children, 842 polio in, 1170 clinical health services for, 1116 tobacco products, 876 schistosomiasis in, 471 contraception and, 1112 Afghanistan school health programs in, 1100 cost-benefit analysis and, 1117­1120 leishmaniasis in, 453 skin diseases in, 714 costs and cost-effectiveness of as opium producer, 908 social marketing to reach adolescents interventions, 1117, 1118­1119t aflatoxin, 573­574 in, 1122 1357 Africa (Continued) exposure to air pollutants, 819 sales hours and places of purchase, traditional medicine in, 1287 exposure to chemical water pollution, 822 regulation of, 893­894, 900­901 vital events registration in, 1021 health impacts social determinants of exposure and water supply in, 771, 772, 773, 774, 776 air pollution, 819­820 risk, 890 WHO­ILO Joint Effort on Occupational water pollution, 822 taxation to discourage use, 218, 893, Health and Safety in Africa, 1134 industrial sectors contributing to, 818, 895t, 900 yellow fever in, 396­397 819t, 821 Alcoholics Anonymous, 911 Africa Measles Partnership, 404 interventions, 823­824, 823f alfacalcidol, 972 African Americans air pollution, 823­824, 824b allergies and homeopathy, 1288 coronary artery disease of, 833 water pollution, 824 Alma Ata Declaration on Primary Health diabetes of, 598 lessons learned Care (1978), 88, 1193 African Programme for Onchocerciasis air pollution, 828­829 alternative medicine. See complementary Control (APOC), 168, 172, 437, water pollution, 829 and alternative medicine (CAM) 444, 957 research and development agenda, Alzheimer's disease (AD) and other African trypanosomiasis 829­830 dementias, 124, 131, 627­629 burden of, 456 sources of air pollution, 818­819 See also neurological disorders characteristics and transmission, 453­454 sources of chemical water pollution, burden of, 628, 629t costs and cost-effectiveness of 820­822 cost-effectiveness of interventions in interventions, 46, 459­460 water pollutants, 820­822 developing countries, 637 economic impact of, 456 albinism, 717­718 interventions, 628­629 management and control strategies of, alcohol use and control, 887­906 multi-infarct dementia, 628, 629 458, 462­463 addiction as clinical disorder, 887­888 personal interventions, 629 social impact of, 456 adolescents and, 1111 population-based interventions, African Union, 462 advertising bans, 894 628­629 See also Pan African Initiative beneficial effects of moderate alcohol prevalence and incidence rate, 628 aging use, 887 recommendations, 640 arthritis and, 963 costs and cost-effectiveness of research and development agenda, chronic diseases and, 107, 1333­1334 interventions, 49, 896­899, 898t 640, 641 dementia and, 124 drowning and, 741 risk and protective factors and See also Alzheimer's disease (AD) and drunk driving and, 739­740, 893, 894t, survivorship, 628 other dementias 901­902 women and, 201, 205­206 hearing loss and, 958 economic benefits of interventions, ambulances, 1267­1271, 1269t, 1270t pregnancy of women over 40, 1079 899­900 ambulatory care. See primary care shift of burden of disease due to, 1331 epidemiology of, 887­890, 888f American Academy of Pediatrics understanding biological basis of, 124 fetal alcohol syndrome, 936b recommendation on television Agreement on Agriculture (WTO), 220 high-risk drinking watching, 835 Agreement on Trade-Related Aspects of burden of disease related to, American Heart Association, 651, 658 Intellectual Property Rights. See 890­892, 891t American Psychiatric Association, 908 TRIPS (Agreement on Trade-Related DALYs attributable to, 892t American trypanosomiasis. See Chagas Aspects of Intellectual Property epidemiology of, 888­889, 889t disease Rights) interventions for reducing, 892­896 AMI. See acute myocardial infarction agricultural productivity, effect of, relationship to alcohol use amphetamines. See drug dependence 190, 220 disorders, 889 anemia agricultural runoff, 821 interpersonal violence and, 760 in adolescents, 1111 AIDS. See HIV/AIDS interventions in AIDS patients, 359­360 AIDSVAX, 360b brief interventions, 894, 902 in children, 553­554, 1094 air and water pollution, 817­832 population-level effects and costs, costs and cost-effectiveness of See also indoor air pollution 896­899, 897t interventions, 561 air pollutants, 818­820 self-help, 911 developmental delays and, 562 causes and burden of, 817­822 kidney disease and, 699 helminth infections and, 467­468, costs and cost-effectiveness of lessons learned, 900­902 471, 553 interventions, 825­827, 826t liver cancer and, 573 interventions, 558 air pollution, 825­826 musculoskeletal disorders and, 966 lack of data on interventions for, 563­564 comparison of interventions, 825, 827t relationship between alcohol use and in pregnant women with malaria, 420 water pollution, 826­827 disease categories, 889­890 in school-age children, 1094 economic benefits of interventions, research and development agenda, ankylosing spondylitis, 976 827­828 902­903 anthrax, 107 1358 | Index antibiotics, indiscriminate use of, drug dependence in, 908, 916 excreta disposal in, 780, 782 122, 1340 excreta disposal in, 780 family planning program in, 167 See also drug resistance hepatitis B in, 396 helminth disease control in, 784 in animals, 1032­1033, 1041 kidney and urinary system diseases hospitals in, 1220b, 1257b in humans, 314, 378, 1031­1032 and, 697 IMCI program in, 1180­1181, 1187 antiretroviral drugs (ART) leprosy in, 445 interpersonal violence in, 756, 757 costs and cost-effectiveness of, 43, malaria in, 413, 417, 421 leishmaniasis in, 452 355­358 rheumatic heart disease in, 647 leprosy in, 441 HIV/AIDS and, 14, 110, 294, 302, 305, road traffic injuries in, 740 medical workers in, 1312 331, 345, 349­350, 351, 354­358, water supply in, 772, 773­774, 774, 776 neonatal deaths in, 531, 533, 1079 357b, 359, 1070, 1204, 1333 Asian Development Bank, 1070 nutrition investment in, 248b low levels of use, 184 Asian-Pacific International Molecular respiratory diseases of children in, 492 success in high-income countries, 27 Biology Network, 133 road traffic injuries in, 746 anxiety disorders, 611­613 aspirin TB in, 291, 295 See also mental disorders cancer and, 574 tidal wave in, 1149 cost-effectiveness methods and results, as inexpensive intervention for CVD, BCG. See bacillus Calmette-Guérin 613­619, 614­618t 109, 123f behavioral change interventions generalized anxiety disorder, 611 asthma. See chronic obstructive pulmonary successful programs, 175­177 interventions for panic disorder, 612­613 disease (COPD) and asthma Belarus natural history and course, 612 attention deficit hyperactivity disorder cardiovascular disease in, 649, 649t obsessive-compulsive disorders, 612 (ADHD), 935b Belize phobias, 612 Australia interpersonal violence in, 766 posttraumatic stress disorder (PTSD), COPD and asthma in, 686 Bellagio Study Group on Child Survival, 605, 610, 621 dengue in, 452 539, 541, 1196, 1202 social anxiety disorder, 611 detoxification methods in, 912 Belmont Report of National Commission APOC. See African Programme for ecstasy (MDMA) use in, 924 for the Protection of Human Onchocerciasis heroin use in, 908, 909 Subjects, 267 "appropriate science" for developing illicit drug use in, 909, 922 benign prostatic hypertrophy, 698 world, 109­111 interpersonal violence in, 763 See also kidney and urinary system Argentina kidney and urinary system diseases diseases arsenic contamination of water in, 821 in, 700 Benin Chagas disease in, 438, 442 skin diseases in, 708, 710 excreta disposal in, 780 fluoridation of water in, 725 tobacco taxes in, 882 hemoglobinopathies in, 675 kidney and urinary system diseases traditional medicine in, 1289 road traffic injuries in, 749b in, 703 autism, 935b Bernard, Claude, 119 array technology, 128 avian influenza threat, 112b, 117, Better Health in Africa (World Bank), 1194 arsenic contamination of water, 821, 682­683b, 1005­1006 Bhopal catastrophe (India 1984), 819, 819b, 822, 822b ayurveda, 1286 1133, 1149 ART. See antiretroviral drugs bicycles. See transportation artemisinin, use of, 111, 1288 bacillus Calmette-Guérin (BCG), 40, 121, Bill & Melinda Gates Foundation artemisinin combination therapy (ACT), 292, 389, 441, 1325 AIDS prevention efforts and, 254 419, 420­421, 423, 424, 425, back pain. See musculoskeletal disorders development assistance for health from, 427, 1043 bacterial skin infections, 708, 709, 710­713, 165, 240n5, 243, 244­245, 244t, 397 arthritis. See musculoskeletal disorders 712­713t disease control priorities and, 111, 115 asbestosis. See occupational lung diseases Bamako Initiative, 227 Diseases of the Most Impoverished ascariasis Bangladesh initiative and, 378 excreta disposal and, 783­784 arsenic contamination of water in, economic issues for future medical water supply and, 779 821, 822b research and, 134 Asia child mortality of under age five in, 246 food fortification efforts and, 254 See also specific countries and regions community-based programs in, 1067 funding commitment for disease control abortions in, 1078 contraceptive use in, 1083 in developing countries, 147 asthma in, 686 contracting for health care services in, MCE and, 1190 avian influenza in, 682­683b 188, 1320 MDGs and, 182 burn-related injuries in, 741 development assistance and health polio eradication efforts and, 24, contraceptive use in, 1076, 1086 policy, 248b 254b, 1171 dengue in, 461 diarrheal diseases in, 170, 176, 372, 374 stimulating growth of public-private diarrheal diseases in, 373, 778 drowning in, 741 partnerships, 147 Index | 1359 biologic terrorism dengue in, 461 burden of disease, 105­106 surveillance for, 998, 1008 diarrheal diseases in, 374, 375, 376, See also specific diseases and categories biologically based practices, 1282 378, 778 of diseases (e.g., infectious diseases, See also complementary and alternative excreta disposal in, 783 tropical diseases) medicine (CAM) fluoridation of water in, 725 broad patterns of (by region), biomedical research priorities, 126, 132 health knowledge, dissemination 105t, 1335 bipolar disorder, 608­610 to mothers, 187 children under age five, 105­106 See also mental disorders HIV/AIDS in, 13, 106 estimates for different diseases and costs and cost-effectiveness of IMCI program in, 1181, 1183, injuries, 28­32 interventions, 47, 613­619, 1185b, 1186 as factor in cost-effectiveness analysis, 614­618t leishmaniasis in, 452 272­273 epidemiology and burden, 609 leprosy in, 437 leading causes of (by region), 105t genomics and, 131 lifestyle changes in, 839b, 841b in low- and middle-income interventions, 609­610 malaria in, 418 countries, 31t natural history and course, 609 Multi-Country Evaluation of IMCI purpose of measures of, 29­32 policy and service implications, 619­621 Effectiveness, Cost, and Impact shift from infectious to noncommunicable bird flu. See avian influenza threat (MCE) in, 1183 diseases as major causes, 1331, 1333 blackfly control, 178, 435 occupational health in, 1140 women's excess burden, 196, 198­199, blindness, 953­957 physical activity initiative in, 841b 199t, 200t, 201t burden of, 954­956, 955t road traffic injuries in, 744, 745­746, Burkina Faso cataract surgery, 956, 1258b 900­901 African trypanosomiasis in, 458 causes and epidemiology of, sexually transmitted infections in, 323 drug policy in, 1346 953­957, 954t skin diseases in, 715 health care worker safety for childhood blindness, 957 smallpox in, 1163 immunizations and injections cost-effectiveness of interventions, 954t surveillance project in, 1021 in, 1132 definitions of visual impairment transportation policy in, 839b hygiene promotion in, 785 levels, 954t water pollution in, 829 malaria in, 421 interventions, 956­957 breast cancer, 205 sexually transmitted infections in, 325 river blindness. See Onchocerciasis age-specific incidence of, 582f burn-related injuries. See unintentional blood pressure. See cardiovascular chemotherapy for, 268, 580­581 injuries disease (CVD) mind-body interventions and, blood-related disorders. 1288­1289 calcitonin, 972 See hemoglobinopathies radiation for, 581 calcium supplementation, 967, 971­972 blood safety measures, 346 rates of, 570 CAM. See complementary and blood transfusions screening for, 577­579, 578t, 579t, 580t alternative medicine Chagas disease and, 434, 438­439 tamoxifen and, 574 Cambodia HIV/AIDS and, 346­347 treatment interventions for, 580­581 avian influenza in, 682b Bolivia breastfeeding dengue in, 454 Chagas disease in, 436, 438, 442 community health and nutrition health sector reform in, 1315 newborn project in, 535 programs, 1058 IMCI program in, 1186, 1187 Bone and Joint Decade (2000­10), 963, 978 diarrheal disease prevention due to, medical workers' New Deal experiment bone marrow transplantation. See 375­376, 379 in, 1317­1318, 1319b hemoglobinopathies duration of, 555 oral precancer and cancer in, 730 Botswana infant mortality and, 539 outsourcing of primary health care African trypanosomiasis in, 458 programs to promote, 559 services in, 96b HIV/AIDS in, 349, 360b supplemental or substitute feeding, performance-based contracts with NGOs brain drain of medical personnel, 94, 95b, 345, 375­376, 555, 557b, 1058 in, 253 1309­1310 transmission of HIV and, 334, refugee populations and Brazil 336, 375 surveillance, 1009 alcohol consumption regulation in, bronchiolitis, 485 sexual risk behaviors in, 312 900­901 bronchitis. See chronic obstructive water pollution in, 829 breast cancer management in, 205 pulmonary disease (COPD) and Cameroon cardiovascular disease in, 650, 659 asthma drug treatment in, 1346 cervical cancer screening in, 576, 577t Brunei malaria in, 426 Chagas disease in, 436, 438, 442 fluoridation of water in, 725 sexually transmitted infections combination of high blood pressure, Bulgaria in, 323 cholesterol, and obesity in, 853­854 milk fluoridation in, 726 user fees and improved quality in, 89 1360 | Index Canada costs and cost-effectiveness of pharmacological interventions, 15­16, cigarette smoking in, 875, 877 interventions, 47­48, 650­658, 650­654, 656, 658, 856­858, 859t, development assistance for health from, 655t, 656t, 657t, 858­863, 861t 860­863, 862t, 863t 244, 254 demographic changes and, 105 population-based primary prevention of, fluoridation of water in, 725 diabetes and, 597 47­48 hemoglobinopathies in, 673, 675 diet contributing to, 652 predominant diseases, 645­649 illicit drug use costs in, 909 See also obesity primary prevention, 658 interpersonal violence in, 760, 765 epidemiology of, 645­649, 648t regional burdens, 649­650 pollution control in, 828 equity issues and prevention research and development agenda, SARS in, 1005 strategies, 863 658­659, 864­865 cancer, 569­589 folic acid and, 836 rheumatic heart disease (RHD), See also specific types glossary, 646b 647, 656 burden of, 107, 569­572, 570f hypertension burden of, 650 DALYs lost due to, 570, 571t CHF and, 646 cost-effectiveness of interventions, 653, failure of U.S. "war on cancer," 112 control of, 646 656, 657t, 658 interventions IHD and, 646 epidemiological transition and, costs and cost-effectiveness of, 46, kidney disease and, 698, 699, 647, 648t 572­582, 582­583 700, 701 strokes and, 634 diagnosis and treatment, 572, as risk factor, 649, 652 risk factors, 649, 851­862, 854t 579­582 strokes and, 633, 634, 641 social and economic impact of, 650, 864 dietary and related interventions, hypertension in combination with strokes and, 634, 645­646, 646, 647, 648t, 573­574 cholesterol and obesity, 851­868 649­650, 657t, 658 early detection and secondary burden of, 852­854, 853f, 853t types of inexpensive interventions for, prevention, 572, 574­579, 582 cost-effectiveness of interventions, 109, 123f genomics and, 128 858­863, 861t women and, 201 immunization, 572­573 economic benefits of Caribbean. See Latin America medical advances in, 123 interventions, 864 and Caribbean palliative care, 572, 581­582 effectiveness of interventions, Carter Center, 167 pharmacological interventions, 574 855­858 cash-transfer programs to increase school primary prevention, 572­574 epidemiology of, 851­852, 852f enrollment, 1101 screening, 123, 573, 574­575, 582 equity issues, 863 cataracts. See blindness starting small and scaling up, 585 financial burden of, 853­854 cattle time horizon and approach to control interventions, 854­858 African trypanosomiasis and, 458 programs, 585 personal interventions, 856­858 CBHI. See community-based tobacco and alcohol control pharmacological interventions, health insurance programs, 573 856­858, 859t CDC. See U.S. Centers for Disease Control types of, 572 population-based interventions, and Prevention pain control, 572, 581­582 855­856, 859 cell theory, 119 See also palliation research and development agenda, census data, 1020 research and development priority, 864­865 Center for Global Development's 583­584, 1333 surgical interventions, 856­858 Global Health Policy Research women and, 203, 205, 570 surveillance and, 1009 Network, 165 cannabis. See drug dependence ischemic heart disease. See ischemic Central America. See Latin America capital charging and management of heart disease (IHD) and Caribbean physical resources, 1345 lifestyle changes in response to, 652 Central Asia cardiovascular disease (CVD), 15­16, linking costs and effectiveness cardiovascular disease in, 649, 654, 645­662 in developing countries, 653­656, 655t, 656t acute management of, 48 655t, 656t contraception demand in, 1076 acute myocardial infarction (AMI), 646, in high-income countries, diarrheal diseases in, 378 650­651, 653­654 656­658, 657t mental disorders in, 618, 619 advances in treatment of, 122­123, 123f long-term management of, 651­652 proposed guidelines for HIV/AIDS in, atrial fibrillation, 633, 646 invasive interventions, 651 348­349 burden of, 107, 649­650, 649f nonpharmacological road traffic injuries in, 746 as cause of death, 107 interventions, 652 strokes in, 634 clinical research, 659 pharmacological interventions, TB, spread of, 291 congestive heart failure. See congestive 651­652 vaccine-preventable diseases in, 394, 399, heart failure (CHF) personal interventions for, 48, 856­858 400, 402, 408 Index | 1361 cerebral palsy, 947, 948 indirect causes of, 534 integrated services. See integrated cerebrovascular disease. See Alzheimer's interventions for, 4b, 43­44, 535­541, management of childhood illness disease (AD) and other dementias; 536­538t (IMCI) stroke lessons learned, 543­547 interpersonal violence. See interpersonal cervical cancer low birthweight and, 534, 540 violence age-specific incidence of, 582f MDGs on, 7, 531, 534, 547­548 learning disabilities. See learning and HBV vaccine and, 573 monitoring of coverage and measuring developmental disabilities (LDDs) HPV, association with, 569 for cost-effectiveness, 546­547 malnutrition. See stunting, wasting, mortality rates and, 570 neural tube defects and, 939 and micronutrient deficiency radiation for, 581 poverty and, 534 disorders screening for, 203­205, 575­577, 577t research priorities, 547 mental disorders and, 621 treatment interventions for, 312, 581 resuscitation program and, 539­540, out-of-school children and youth, 1098, Chagas disease, 433­434 539b, 545 1121­1122 burden of, 436 systematic scaling-up of newborn perinatal conditions, 499­529 cardiovascular disease and, 650 care, 546 See also perinatal conditions costs and cost-effectiveness of time trends and, 532­533, 532f research priority of, 126 interventions, 45­46, 442 malaria and, 414, 416t, 417, 418, 421, respiratory disease. See respiratory interventions and effectiveness of, 423, 556 diseases of children 438­439 child sexual abuse, 764 rheumatic heart disease and, 647 research priorities, 446 See also interpersonal violence school-based health. See school Southern Cone countries initiative on, Child Survival and Development Projects health programs 167, 170, 177, 438 (World Bank), 1068 stunting, wasting, and micronutrient defi- chemotherapy childbirth conditions ciency disorders, 551­567 for cancer, 268, 574, 580­581, 583 See also maternal conditions TB and, 293 for schistosomiasis, 474, 479 access to skilled attendance, 44, Tinea capitis, 714­715, 716t, 718, 719t for TB, 1045 512, 547b unintentional injuries. See unintentional Chernobyl nuclear disaster (1986), birth preparedness, 511 injuries 819, 1133 components of comprehensive safe moth- childhood illnesses of children under age 5 CHF. See congestive heart failure erhood strategy, burden of, 105­106 child abuse, 756, 758, 765 508­509, 509b deaths. See child mortality rates See also interpersonal violence cost-effective interventions, 44 interventions, 43­45 Child Health Epidemiology Reference fetal monitoring and, 948 malnutrition and, 552, 553t Group, 533 male involvement, 511 Chile child mortality rates maternal death. See maternal mortality arsenic contamination of water in, 821 under age 5 mortality, 11­13, 12f midwives, role of, 174, 511 Chagas disease in, 436, 438, 442 diarrheal disease and, 374 neonatal death. See child mortality rates combination of high blood pressure, interventions for, 11­13 preterm birth, 534, 547 cholesterol, and obesity in, 856 malaria and, 416t prevention of, 947­948 fluoridation of water in, 725 malnutrition and, 552, 552t retinopathy of prematurity, 954­955 food fortification in, 558 MDGs for, 7, 11 quality of evidence on, 509­510, 512 Hib disease prevention in, 167, 172 rate of progress in reducing, 12f, 246 risks associated with, 1078­1079 interpersonal violence in, 760 infant and neonatal mortality, 4b, sexually transmitted infections of mother life expectancy, increases in, 3 505­508, 507t, 531­549 and, 315 noncommunicable diseases and causes achieving optimal newborn care strokes and, 634 of death in, 6, 7f within system constraints, technological improvements in, 131 occupational health in, 1133 544­546 childhood health, 10­13 pain control in, 987b advocating for newborn health, 544 See also specific diseases and conditions public health improvements in, 3 cost of scaling up universal neonatal under age five. See childhood illnesses of respiratory diseases of adults in, 691 packages, 541­543, 542t, 544b children under age 5 universal coverage approach of, 237 delays in access to care and, 534 community health and nutrition China direct causes of, 533­534 programs. See community health abortions in, 1078 emergency care and, 540­541 and nutrition programs (CHNPs) arsenic contamination of water in, 821 essential newborn care and, 535­539 dental health. See dental caries artemisinin use in, 1288 family-community care and, 535 helminth infections and, 471, 473, 474 asbestos exposure in, 1128 in first week after birth, 532 improved quality of treatment for burn-related injuries in, 741 income level of countries and, 532, children with childhood diseases, cardiovascular disease in, 647, 649, 650, 533f, 533t 1303­1304 658, 659 1362 | Index combination of high blood pressure, cholesterol levels climate cholesterol, and obesity in, in combination with high blood pressure See also global warming 853­854, 856 and obesity, 851­868 health systems' efficacy and, 1340 contraceptive use in, 1080 See also cardiovascular disease (CVD) helminth infections and, 470 dengue in, 452 CVD and, 15 Clinton Foundation, 359, 1333 diabetes in, 593 diabetes and, 598 cocaine. See drug dependence diarrheal diseases in, 374, 377 drugs to control, 857, 1323 cognitive function drug addiction in, 908, 916 modeling of interventions for lifestyle dis- cannabis and, 918 fall-related injuries in, 748­749 eases, 844­845 malnutrition's effect on, 562 folic acid deficiency in, 836 chronic diseases, 833­850 of school-age children. See school food manufacture in, 845 See also specific diseases health programs food taxes in, 219 advances in treatment of, 123 COHRED. See Council on Health Research hemoglobinopathies in, 668, 670 alcohol use and, 890 for Development hypertension in, 835 burden of, 107, 1333 collaboration for research. See research incentive pay to doctors in, 94­95, community-based interventions, 842, 846 and development 95b, 170 cost-effectiveness of interventions, collective efficacy, 1067 indoor air quality and energy programs 844­846 Colombia in, 190, 795, 809, 810b diet and lifestyle changes to prevent, ambulatory surgery in, 1256 influenza in, 684 833­836, 847 contraception, access to, 1076 kidney and urinary system diseases in, evidence of effectiveness, 836­837 interpersonal violence in, 756, 763 702, 703 educational interventions, 837­838 stomach cancer in, 573 lung cancer in, 795 health care providers modifying universal coverage approach of, 237 lymphatic filariasis in, 436 unhealthy behaviors, 838 volcanic eruptions in, 1150 malaria in, 421, 425 interventions, 837­844 colorectal cancer, 570, 574, 575, 575t, mental disorders in, 620 recommended priority interventions, 581, 836 milk fluoridation in, 726 846­848 Commission on Health Research in neonatal deaths in, 531 malnutrition in childhood associated Development, 106 neonatal resuscitation program in, 539b with, 563 Commission on Macroeconomics and neurological disorders in, 628 medications for, 1331 Health (WHO) oral and craniofacial diseases and modeling of likely interventions, on close-to-client services, 521 disorders in, 723 844­846 on constraints of low-income countries to oral health and education program physical activity interventions, 835, 838, improve health care, 98 in, 726 841b, 842, 846, 847 on cost-effective analyses, 234, 261 road traffic injuries in, 740, 747, 839 research and development agenda, 846 on district hospitals, 1214 rotavirus vaccine in, 377 surveillance and, 1009­1010 on external financing for health, 245 salt iodination program in, 168, 173 transportation policy and environmental on HIV/AIDS, 9 SARS in, 997, 1005 design as intervention, 839­840 on mobilizing skills and resources, 134 school-based physical activity in, 838 worksite interventions, 838, 838b on primary care interventions, 246, stomach cancer in, 573 chronic kidney disease (CKD). See kidney 1194, 1200 strokes in, 633 and urinary system diseases on scaling up health coverage, 190 TB control program in, 36b, 94­95, chronic obstructive pulmonary disease on training to enhance skills of medical 168­169, 174, 247 (COPD) and asthma, 684­689, workers, 1344 tobacco use in, 871, 1128 685f, 687t Common Rule of Conduct of U.S. Code of transportation choices in, 839­840 See also respiratory diseases of adults Federal Regulations, 266 vital events registration in, 1021 air pollution and asthma in communicable diseases. See infectious and water pollution in, 829 children, 820 communicable diseases women in, 195 cigarette smoking. See tobacco use and communication for emergency medical chiropractic medicine, 1282, 1286, 1288 control services, 1267 chlamydia, 203, 312, 314­315, 316 circumcision to protect against sexually community-based health insurance See also sexually transmitted transmitted infections, 320, (CBHI), 231, 236 infections (STIs) 322, 360b community-based programs chlorination, 821 civil service reform, 96 and treatment chloroquine, 414, 418, 420f, 423, 1035f, CKD (chronic kidney disease). See kidney breastfeeding, programs to promote, 559 1043f, 1327 and urinary system diseases central government transfers to local CHNPs. See community health and Clean Air Act of 1970 (U.S.), 828, 829t authorities for, 253 nutrition programs clean drinking water. See water supply community health and nutrition cholera, 120, 377, 379, 774, 777, 778 cleft lip and palate, 731 programs, 1053­1074 Index | 1363 community-based programs and micronutrient supplementation, congestive heart failure (CHF), 646­647 treatment (Continued) 1058­1059 See also cardiovascular disease (CVD) See also community health and oral rehydration, 1059 cost-effectiveness of interventions, 653 nutrition programs (CHNPs) organization of, 1064 in developing countries, 654, 656 decentralizing health functions to, 253 political commitment and, 1066 in high-income countries, 656, 657t, 658 depression treatment, 110 prenatal care, 1058 Congo, Democratic Republic of epilepsy treatment for children, 110 process indicators and, 1067 African trypanosomiasis in, 454 HIV/AIDS, 1070­1071 programmatic factors for, 1060­1064 Ebola in, 1006 home-based care, 1204 research and development agenda, consumer subsidies, 212­218 IMCI and, 1178, 1180 1071­1072 consumer taxes, 218­220, 219b infant and neonatal mortality and, 535 success factors for, 1057­1058 contextual knowledge, 110 learning and developmental disabilities, supplementary feeding, 1059 contingent valuation, 280 946­947 technology and, 1063 continuous quality improvement in health management of health services in, underweight children and, 1055, 1056f care, 1298 1343­1344 women's health and nutrition, 1058 contraception, 1075­1090 mental disorders, 608, 613, 615­616t, workers in, 1063­1064 See also family planning 615t, 619, 621 community violence, 756 adolescents and, 1112 neurological disorders, 631, 641 See also interpersonal violence condom use, 1080, 1082, 1086 nutrition programs, 559, 561, 842, 846 complementary and alternative medicine costs and cost-effectiveness of See also community health and (CAM), 1281­1291 interventions, 1081­1085, 1082t nutrition programs (CHNPs) See also specific types of treatment costs per births averted, 1083­1084 onchocerciasis control program, 440 (e.g., acupuncture, homeopathy) costs per DALY, 1084 oral health and, 732 ancillary benefits and costs of, 1286 costs per death averted, 1084 physical activity programs, 842, 846 defined, 1281 outside of programs, 1083­1084, 1084t primary care, 1202, 1204 demography and use of, 1282­1283, within programs, 1084­1085 scabies and, 710 1282t, 1283t demand for, 1076, 1077t skin diseases, 710, 715, 718, 720 domains of, 1281­1282 diaphragms, 1080 Tanzania Essential Health Interventions economic evidence on use of, 1285­1286 HIV/AIDS and, 360b Program approach to health economic factors influencing use of, economic benefits of intervention, 1086 planning for. See Tanzania 1283­1285 emergency contraception, 1081 Essential Health Intervention efficacy of, 1283, 1284t equity issues and, 1086 Program (TEHIP) epilepsy and, 630 female sterilization, 1080, 1082 TB, quality of care and treatment of, 295 expanding beneficial use of, 1286­1287 fertility and population growth, community clinics and surgery, 1249, lessons learned, 1287­1289 1075­1076, 1086 1253­1255, 1253t, 1254t, 1256f for mental disorders, 620 financing of, 1332t community health and nutrition programs for neurological disorders, 629, 641 health consequences and, 1076­1079 (CHNPs), 1053­1074 for Parkinson's disease, 632, 640, 641 interventions, 1080­1081 breastfeeding and, 1058 research and development agenda, 1289 IUDs, 1080 community- and facility-based programs, for skin diseases, 715 methods, 1080­1081 1056­1057, 1059f, 1060b toxicity of herbal products, 1283 "the pill," 1080, 1082 complementary feeding of children conditionality and development research and development agenda, 1087 and, 1058 assistance, 247 risks associated with pregnancy and birth, contextual factors for, 1065­1067, 1065t condoms 1078­1079 cost-effectiveness of, 1068­1070 See also HIV/AIDS risks associated with unwanted coverage, targeting, and resource import taxes not charged on, 220 pregnancies, 1077­1078, 1078t intensity, 1060­1063, 1061­1062t subsidy programs, 216 social marketing of, 1081 description of, 1058­1059 Thailand distribution program. tubal ligation, 1080 effectiveness of, 1054­1056, 1055t See Thailand vasectomy, 1080 estimated contributions to disease use as contraceptive, 1080, 1082, 1086 contracting for health services. burden, 1054t See also contraception See outsourcing of health care future applications of, 1070­1071 congenital and developmental disorders services growth monitoring, 1058 cleft lip and palate, 731 Convention on the Rights of the HIV/AIDS and, 1070­1071 hyperthyroidism, 945­946, 949 Child, 1095 IMCIs and, 1056 medical treatment of, 123 cooking fuels and air quality. See indoor immunizations, 1059 oral and craniofacial diseases and air pollution impact of, 1067­1068, 1069t disorders, 731 COPD. See chronic obstructive pulmonary micronutrient fortification, 1059 research priority of, 126 disease (COPD) and asthma 1364 | Index cost-effective interventions, 24­27, 25t, unit of measurement of health, 278­280 data collection. See information systems 35­86, 36f See also DALYs (disability-adjusted Daubert v. Merrell Dow Pharmaceuticals See also specific diseases and conditions life years) (1993), 153 analysis of. See cost-effectiveness analysis weighting of life years in, 261 day care at work, 221 assessing evidence of, 40­50, 41­42f cost-effectiveness ratio DCPP. See Disease Control Priorities building health systems, 27, 53­58 average and incremental, use of, 275, Project communicable diseases, 53 276­277b, 283 deafness. See hearing loss lack of reliable data on, 58 defined, 271­272 death. See child mortality; mortality rates; methodology of study, 38­40, 39t use of, 272, 282 specific diseases noncommunicable diseases, 53 Costa Rica decentralizing health functions to local personal. See personal interventions child mortality in, 1068 governments, 253 policy setting and, 36b community-based programs in, 1065, decision making, information for, population-based. See population-based 1066, 1067, 1068 1017­1030 interventions diet in, 836 See also information systems priority setting, 35­38, 37­38b, 271­273 disaster damage to hospitals in, 1156 Declaration of Helsinki, 266, 267, 268, private sector and, 38 health sector reform in, 1298 269n3 quality as determinant of interpersonal violence in, 766 dehydration. See oral rehydration cost-effectiveness, 50 universal coverage approach of, 237 therapy (ORT) regional variations, 39 Côte d'Ivoire dementia. See Alzheimer's disease (AD) and target audiences and, 38 African trypanosomiasis in, 459, 460 other dementias technological progress and, 53 malaria in, 426 demographic changes of 20th century, 124 women and, 203­206, 204­205t road traffic injuries in, 749b Demonstration in Small Industries for cost-effectiveness analysis, 26b, 271­285 Council for International Organizations of Reducing Wastes Project, 825b aggregation and cost differences in, Medical Sciences, 266 dengue 263­264 Council on Health Research for burden of, 454­455 amount of health (by service or Development (COHRED), 161, 1297 characteristics and transmission, 451­452 intervention) US$1 million will counterfeit medicines, 1327 costs and cost-effectiveness of buy, 25t, 284 craniofacial diseases. See oral and interventions, 459 burden of disease and, 272­273 craniofacial diseases and disorders economic impact of, 455 CER, use of. See cost-effectiveness ratio criminal enforcement for dependent illicit management and control strategies costs of producing intervention used in, opioid use, 911­912 of, 457 261­262 Croatia risk factors of, 461 currency units used in, 281 cardiovascular disease in, 649, 649t social impact of, 455 data for, 281­282 cross-level services and inputs vaccine development for, 460­461 defined, 271 district hospitals and, 1215, 1222­1226 dental caries, 723­729, 724t, 725f discounting in, 262, 278 Crossing the Quality Chasm (IOM), 1295 common-risk-factors intervention discrimination against persons with Cuba programs, 731­732, 732t disabilities, 264­265 dengue in, 454, 455, 461 cost-effectiveness of oral health care, estimating effectiveness in health, public health improvements in, 88 733­734, 733f 277­280 cure effectiveness of oral health programs, 726, ethical considerations for, 259­260, 272 defined, 59 727­728t, 729 evaluation of health benefits in, CVD. See cardiovascular disease fluoridation and, 168 260­261 Cyprus intervention programs, 725­726 evaluation of nonhealth benefits in, 279 hemoglobinopathies in, 665, 673 oral health education and promotion pro- fair chances and best outcomes in, 264 malaria in, 421 grams, 726 improving, 284 Czech Republic research and future action, 733 modeling in, 281 cardiovascular disease in, 649 treatment of, 729 population impact and, 283­284 depression quality-adjusted life years in. DALYs (disability-adjusted life years) See also major depressive disorders See quality-adjusted life See also specific diseases and conditions burden of, 107, 1334 years (QALYs) gender differences, 197, 198t, 199t community-based treatment of, 110 reasons to develop, 273­277 in low- and middle-income countries, 31t costs and cost-effectiveness of reliability and uncertainty of, 282­283 for tropical diseases, 437­438, 438t interventions, 47 resource allocation based on, problems use to assess burden of disease, 29­32 women and, 107, 196, 201, 202, 206 from, 262­265 use to measure health, 278­280 detoxification. See drug dependence for South Asia and Sub-Saharan Africa, use to value health, 158, 161 developed countries. See high-income 50, 51­52t Darwin, Charles, 119 countries Index | 1365 developing countries. See low-income increasing burden of, 107, 591­592, poverty and, 373 countries 600, 1333 public health significance of, 373, 373f, development assistance and health policy, insulin resistance syndrome and, 834b 374­375, 374f 22­25, 28, 243­257 interventions, 593­595 research and development agenda, See also Millennium Development kidney disease and, 695, 698, 699, 383­384 Goals (MDGs) 700, 702 rotavirus immunization, 376­377 absorptive capacity of countries and, lessons learned, 599­600 successful programs in combating, 167, 232­233, 232t lifestyle changes in response to, 593 176, 375­377, 1303­1304 aid effectiveness and, 233, 250­256 low birthweight and, 121 syndromic diagnosis of, 372 reaching local communities, management of, 594­595 transmission of, 371­372 255­256 nature and distribution of, 591­593 water and sanitary facilities and, budget support and, 250­251 periodontal disease and, 729­730 187, 189­190, 377, 776, 777, 777t, central government transfers to local prevention, 599, 600 782, 785, 786­791, 786t, 787f, 789t, authorities, 253 quality of treatment and care, 790t, 791t conditionality and, 247 599­600 See also sanitation improvements; water donor disbursements to national research and development agenda, 600 supply governments, 253 risk factors for, 593 zinc supplementation and, 378 fungibility vs. earmarking of funding, screening for, 594, 598 diet. See nutrition 247­249 self-management of, 598 Difference Principle, 262 harmonizing procedures in, 249, 251 stokes and, 633­634 diphtheria, pertussis, tetanus lessons learned, 246­250 type 1 diabetes, 593 See also Expanded Program on performance-based financing, 251­253 type 2 diabetes, 593 Immunization (EPI); policy environment and, 246­247 Diabetes Prevention Program, 593 vaccine-preventable diseases pooling of donor funding, 251 Diagnostic and Statistical Manual of Mental control of, 394­395 private sector and, 253­255 Disorders, 4th ed.(DSM-IVTR) neonatal deaths caused by tetanus, sources and amounts, 243­244, 244t (American Psychiatric Association), 533, 534 transaction costs of, 249 607, 908 vaccine, combined with Hib trends and gaps in, 243­246 diagnostics, development of, 144­146, vaccine, 405b unpredictability of, 249­250, 250t 145, 145t vaccine for, 121, 389, 1325 developmental disabilities, 933­951 diaphragms. See contraception directly observed therapy short course See also learning and developmental diarrheal diseases, 371­387 (DOTS), 36b, 110, 170, 174, 289, disabilities (LDDs) See also oral rehydration therapy (ORT) 293­294, 293­295, 303­304, 1041, developmental disorders at birth. bloody diarrhea 1044­1045 See congenital and developmental diagnosis of, 372 agenda for action, 1045­1046 disorders management of, 378 disabled persons deworming. See helminth infections breastfeeding preventing, 375­376, 379 discrimination against, 264­265 Diabcare-Asia project, 599 burden of, 373­375, 373f disaster relief diabetes, 591­603 case management of, 378 cash assistance, 1157­1158 combination of hypertension, cholesterol, cholera immunization and, 377 coordination of humanitarian and obesity and, 851 complementary feeding practices effort, 1154 cost-effectiveness of interventions, 594t, and, 376 costs and cost-effectiveness of 595­598, 600­602, 601t cost-effectiveness of interventions, interventions, 1156­1158 estimating in developing countries, 378­383, 380­382t, 383f, 383t disease prevention and control, 1157 595, 596t drug therapy for, 378 donations and supplies following, ranking implementation priorities, environment and, 373 1154, 1157 595­598 hand washing to prevent. See hand emergency preparedness of health sector, death and disability rate of, 591­592 washing as intervention 1155, 1158 diet contributing to, 836 interventions, 45, 375­377 field hospitals for, 1157 economic burden of, 592­593 laboratory diagnosis of, 372 hospital damage and, 1155­1156 education on, 598 long-term consequences of, in-kind donations, 1157 epidemiological and economics 374­375 lessons learned, 1159, 1161 research, 600 malnutrition and, 372 mobilization of resources, 1158­1159 epidemiological transitions and, 124 measles immunization and, 377 preparedness funding, 1158 estimated prevalence of, 591, 592t mortality rate and, 372, 374, 374f prevention and mitigation, foot problems and, 597 oral rehydration therapy and. See oral 1155­1156, 1158 health systems and operational rehydration therapy (ORT) funding for, 1159 research, 600 personal and domestic hygiene, 377 reconstruction funding, 1159 1366 | Index research and development agenda, focus on most cost-effective interventions dracunucliasis. See guinea worm 1159­1160 for developing countries, 165 eradication resources for emergency response, 1158 on suicides, 607 drinking water. See water supply search and rescue, 1157 on tuberculosis, 295­296 driving injuries. See traffic injuries shelters, 1157 district hospitals, 1211­1228 drowning. See unintentional injuries strategic approach to, 1160­1161 central financing of, 1220­1221 drug dependence, 907­931 water systems, repair of, 1156 changing disease spectrum, implications abstinence-oriented treatments, 910­911 disasters, 1147­1162 of, 1226 adolescents and, 1111 See also disaster relief clinical management of, 1223 amphetamines, 922­923 age and, 1149 clinical services provided by, 1214­1215 interventions, 923 classification of, 1147 cost-effectiveness of services, 1216, burden of disease associated with, 909, climatic disasters, 1151 1219, 1220t 914­916, 915t communicable diseases and, 1151, 1154 costs of care in, 1218­1219, 1218t cannabis, 917­919 earthquakes, 1149­1150, 1150t cross-cutting services and, 1215, interventions, 919 economic valuation of, 1151­1152 1222­1226 causes and health consequences of, environmental health and, 1154 defined, 1212 907­909 gender and, 1149 demand for services, 1221­1222 cocaine, 919­922 hazards, defined, 1148 economics of, 1216­1219 interventions, 921­922 health and economic losses from, economics of scale and, 1213, 1213f cost-effectiveness of interventions, 1148, 1148f efficiency and, 1213, 1217­1218, 1219t, 912­913, 915t hospitals and health installations, 1224­1225 comparing different interventions, damage to, 1153 equity issues and, 1214 913­914 hurricanes, 1151 factors influencing performance of, criminal justice interventions, 911­912 interventions, 1153­1156 1219­1222 detoxification, 910, 912, 913 long-term impact, 1151­1153, 1152t framework for delivery of services in developing countries, 916 mass casualties, treatment of, 1154 by, 1212f drug-free treatment, 912­913 poverty and, 1149 health information systems and, 1216 cost-effectiveness of, 913­914 as public health condition, 1147­1153 hospital-acquired diseases in, 1224 ecstasy (MDMA), 923­924 response and rehabilitation, 1153­1154 human resources and, 1223 heroin use risk integration with other local health-related antecedents of, 907­908 defined, 1148 services, 1215 health consequences of, 908­909 geographic distribution of, 1148­1149 levels of care, 1217 HIV infection and, 909 short-term health burden of, 1149­1151 national level and costs, 1217 mortality and morbidity, 909 surveillance in, 1008, 1154 private sector and, 1217 prevention of, 909­910 terminology for, 1147­1148 quality improvements and accreditation reducing heroin-related harm, 910 tsunamis, 1150 of, 1223­1224, 1224f, 1226 interventions, 909­916 volcanic eruptions, 1150 research and development agenda, 1226 methadone maintenance, 910, 911, 913 vulnerability services provided by, 1214 research and development agenda, 916 defined, 1148 supervision and, 1216 self-help groups, 911, 913 factors affecting, 1149 supply of services, 1222 drug resistance, 1031­1051, 1332­1333 water and sewage systems, damage to, surgery and, 5b, 27, 1249, 1253­1255, agenda for action and, 1045­1046, 1340 1153, 1156 1253t, 1254t, 1256f anthelmintic, 479 discount rate, 262, 278 technology, information, and integration antibiotic use and. See antibiotics, discrimination in, 1223, 1226 indiscriminate use of against HIV positive individuals, 109 training and, 1215­1216 containing spread of resistant against persons with disabilities, 264­265 DNA markers, study of. See genomics micro-organisms, 1041­1042, 1340 social exclusion and public health, Doctors without Borders' Drugs for diagnostic tests and, 1040 1072n1 Neglected Diseases, 1323 diarrheal diseases and, 378 Disease Control Priorities in Developing Doha Declaration, 151 disease burden and, 1033­1036, 1033t, Countries (DCP) (Jamison domestic violence, 48­49, 202, 756, 760 1034t, 1340 et al., 1993) See also interpersonal violence DOTS and. See directly observed therapy burden of disease estimates, 279 Dominican Republic short course (DOTS) estimates of deaths by causes, 28 contraceptive use in, 1081 drug quality and, 1040 intervention focus of, 16 sexual risk behaviors in, 312 drug treatment strategies and, 1037­1039 Disease Control Priorities Project (DCPP) DOTS. See directly observed therapy economic burden of, 1036­1037 death estimates for, 11b short course global coordination and, 1042­1043 on diarrheal diseases, 378­379 Down syndrome, 941­945, 944t gonorrhea and, 1036 Index | 1367 drug resistance (Continued) new product development, 141­144 cardiovascular disease in, 647, 649, helminth infections and, 479 drug resistance and, 1043 655t, 656t history of, 122 financing in developing countries, 150 development assistance for health HIV/AIDS, 358 R&D costs in developing countries, to, 245 humans and drug use, 314, 378, 142­143 diabetes in, 592 1031­1032 R&D costs in industrial countries, interpersonal violence in, 757 integrated management of childhood 141­142 lymphatic filariasis in, 436 illness and, 1044 patent system and, 150­152 mental disorders in, 618 interventions, 1037­1044, 1038t pharmacogenomics, 129, 129t neurological disorders in, 628 at patient level, 1040­1041 policies on, 1323­1324, 1345­1346 proposed guidelines for HIV/AIDS in, at provider level, 1039­1040 polypill. See polypill 348­349 lessons learned, 1044­1045 prescription and use, 1039­1040, skin diseases in, 715 malaria. See malaria 1329­1330 strokes in, 633, 634 multi-drug resistant TB (MDR-TB), 289, pricing, 1326­1327 vaccine-preventable diseases in, 397, 399, 293, 294, 296, 297, 299, 301, 305, 1333 affordability, ways to increase, 150 402, 408 new drug development and, 1043­1044 differential or "tiered" pricing, water supply, sanitation, and hygiene pneumococci and, 1033­1034, 1045­1046 146­147 promotion in, 787 prescribing patterns and, 1039­1040 orphan drug acts, 147 Eastern Europe reducing selection pressure, 1039 procurement of, 1326­1327, 1345­1346 See also specific countries research and development agenda, 1046 for psychiatric disorders, 605­606, 608, arsenic contamination of water in, 821 retail pharmacies and outlets and, 1040 609­610, 611, 612, 613, 618­619, 621 cardiovascular disease, 649, 649t risk factors and, 1031­1036 public education about, 98 diabetes in, 591 sexually transmitted infections, 314, 1036 quality assurance of, 1327 drug dependence in, 908, 916 shigella and, 1034­1036, 1045 regulation of, 92b, 1345­1346 hospitals in, 1348 surveillance and, 1042­1043 research and development agenda, 5b, kidney and urinary system diseases in, TB and, 294, 304­305, 1036, 141­144, 1335, 1336t 695, 698­699 1045­1046, 1333 for respiratory diseases of adults, 686­688 strokes in, 633 transmission of resistant pathogens for rheumatoid arthritis, 974­975 TB, spread of, 291 and, 1033 sales tax exemptions on, 220 EBM Toolbox, 121 drugs, 1323­1337 subsidies for medicines and medical Ebola, 1006 See also specific drugs and diseases supplies, 217 Economic Commission for Latin America adherence of patient to therapy, for TB, 304 and the Caribbean (ECLAC), 1330, 1346 WHO prequalification process for, 1151, 1152 for cardiovascular disease, 15­16, 1327­1328 economic growth and consequences 650­654, 656, 658, 856­863, 859t, Drugs for Neglected Diseases, 1323 adolescent interventions and, 1117 862t, 863t drunk driving, 893, 894t African trypanosomiasis and, 455 for cholesterol, 857 dysentery, 372, 378 air and water pollution interventions and, counterfeit, 1327 See also diarrheal diseases 827­828 dispensing, 1330 alcohol consumption interventions and, disposal of, 1329 E. coli, 371­372 899­900 distribution and storage, 1328­1329, See also diarrheal diseases contraception or family planning 1345­1346 Early Breast Cancer Trialists' Collaborative and, 1086 donated drugs, management of, 1329 Group, 580 CVD interventions and, 864 financing of, 1330, 1332t early childhood education, 221 dengue and, 455 future challenges for, 1331­1334 earthquakes, 1149­1150, 1150t disasters, effect on, 1151­1152 generic, 1326, 1328 East Africa helminth infection interventions and, for hemoglobinopathies, 672 See also specific countries 477­478, 478f for influenza, 683b, 684 development assistance for health to, 245 of high-technology medicine, 124­125, insulin. See insulin fluorosis of teeth in, 731 133­134, 133t for kidney and urinary system diseases, indoor air pollution and energy programs improved health, effect of, 4b, 7­8, 165 699­700, 703 in, 803, 809 of infectious disease outbreaks, 112b local production of, 1328 leishmaniasis in, 453 leishmaniasis and, 455 for malaria, 418­420 malaria in, 425 maternal and perinatal conditions and, management of, 1345­1346 water supply in, 773 522 Model List of Essential Drugs, 1324­1325 East Asia and Pacific quality of health care and, 1301­1302 for neurological disorders, 629, 631, 632, See also specific countries school health interventions and, 634­635, 637, 640, 641 alcohol consumption in, 898 1099­1101 1368 | Index stunting, wasting, and micronutrient defi- communication systems for, 1267 diarrheal diseases and, 373 ciency disorders, 562­563 costs and cost-effectiveness of in disasters, 1154 WB projections for 2000­15, 183 interventions, 49, 1265­1266, health systems' efficacy and, 1340 economic policies to encourage healthy 1266t, 1267t, 1268, 1270t indoor air pollution and, 797­798 behavior choices, 842­844 deaths from trauma, 1278n1 interventions for environmental control, ecstasy. See drug dependence defined, 1262 177­179 eczema, 708 description of, 1261, 1262t learning and developmental disabilities education disaster relief. See disaster relief and, 936b See also training of medical personnel documentation and quality mental disorders and, 605 adolescents and, 1115 assurance, 1272 posttraumatic stress disorder on alcohol abuse, 893 equipment and supplies, 1266­1267, (PTSD), 605 on diabetes, 598 1271­1272, 1275­1278t respiratory diseases of adults and, 690 early childhood education, 221 financing for, 1272 surveillance, 1006­1007 of girls, 1095 first aid, 1264 enzyme-linked immunosorbent assay on helminth infections, 472, 475 future challenges for, 1274 (ELISA) HIV/AIDS education campaigns, 344, health facility-based system, 1271­1272 African trypanosomiasis and, 462 344b, 1105 helicopter ambulances, 1268 leishmaniasis and, 457 on illegal drug use, 910 interventions, 1263­1272 EPI. See Expanded Program on information, education, and lay responders, 1264, 1267t, 1270t Immunization communication campaigns, 176 legislation on, 1272 Epidemic Intelligence Service (U.S.), 1002 life-skills and health education, 1115 lessons learned, 1272 epidemics on lifestyle changes for health, 837­838, malaria and, 1262 flu. See influenza pandemic, possibility of 842, 859 for newborns, 540­541 HIV/AIDS. See HIV/AIDS on malaria, 422 paramedical personnel, 1264­1265, 1270t surveillance for outbreaks, 1005­1006 malnutrition and school performance, 563 cost-effectiveness of combining with epidemiological trends, 6­7 on osteoarthritis, 972 lay responders, 1267t modern epidemiology, 120­124, 125 patient education, effect of, 98 defined, 1262 epidemiologists peer education programs, 1115 planning and implementation of, 1272 field epidemiologists providing evidence, referral hospitals' role in, prehospital care, 1264­1271 1001­1002 1233­1234, 1234t research and development agenda, training of, 1002 on risk factors, 121 1272­1274 epilepsy, 629­631 sex education, 344, 344b, 1121 response time, 1265 childbirth and, 948 Egypt training for hospital emergency workers, community-based treatment of, 110 diarrheal diseases in, 167, 176, 374, 378 1271, 1273b costs and cost-effectiveness of food subsidy programs in, 216 triage, defined, 1262 interventions, 47 health sector reform in, 1315 universal emergency care, 1271 in developing countries, 637, 638t kidney and urinary system diseases emphysema. See chronic obstructive faith healers and, 630 and, 697 pulmonary disease (COPD) and onchocerciasis and, 435 schistosomiasis in, 479 asthma patient compliance, 630 school health insurance program in, 187 EMRs (electronic medical records), 1300 prevalence, incidence rate, remission, and El Salvador end-of-life care for HIV/AIDS patients, mortality, 630 contracting for health care services 351­353 recommendations, 640 in, 189 end stage renal disease (ESRD). See kidney risk factors, 630 interpersonal violence in, 766 and urinary system diseases stigma of, 631 occupational health risks in, 1136b endangered species and traditional treatment gap for, 630, 641 elder abuse, 756 medicine, 1286 epistemology, defined, 110 See also interpersonal violence endemic treponematoses, 709 equalization funds, 228 electrification, 795, 803­804, 810 endoscopic surgery, 130 equity issues, 259­260 electronic medical records (EMRs), 1300 energy medicine, 1282 contraception and, 1086 ELISA. See enzyme-linked immunosorbent See also complementary and alternative cost-effectiveness analysis and, 272 assay medicine (CAM) CVD interventions and, 863 embryonic stem cells, 130 Energy Sector Management Assistance district hospitals and, 1214 emergency medical services, 1261­1279 Programme, 811 Down syndrome and, 944­945 absence of ambulance system and, 1273b England. See United Kingdom drug accessibility and, 1323 ambulances, 1267­1271, 1269t, 1270t environmental risk factors health improvements and, 4b, 7, 104b burden of disease, 1262­1263, air and water pollution, 817­832 helminth infection interventions and, 476 1263t, 1274f clinical epidemiology and, 120 IMCI and, 1184 Index | 1369 equity issues (Continued) Europe fecal-oral transmission, 317 justice and special concern for the worst See also specific countries See also diarrheal diseases off, 262­263 cardiovascular disease in, 649, 654, 655t, Fédération des Associations de Lutte primary care and, 1198 656t contre la Drépanocytose en referral hospitals and, 1239 cleft lip and palate in, 731 Afrique (Federation of Associations sexually transmitted infections and, 318, diarrheal diseases in, 378 to Control Sickle Cell Anemia in 318t, 319t food manufacture in, 845 Africa), 678 eradication of infectious diseases, hemoglobinopathies in, 670, 673 fee-for-service basis, 19 1163­1176 heroin use in, 908 female genital mutilation (FGM), 202 See also infectious and communicable dis- HIV/AIDS, proposed guidelines for, fertility and population growth, 1075­1076 eases; specific diseases 348­349 fetal alcohol syndrome, 936b ergonomics in workplace, 1137 kidney and urinary system diseases in, fetal monitoring during childbirth, 948 Eritrea, malaria in, 418, 421 695, 698, 700 fibromyalgia, 965 esophageal cancer, 570, 573 mental disorders in, 618, 619 financing, 16­21, 225­242 ESRD (end stage renal disease). See kidney occupational lung diseases in, 690 See also fiscal policies to promote health and urinary system diseases road traffic injuries in, 746 absorptive capacity of countries, Estonia salt fluoridation in, 726 232­233, 232t asthma in, 686 sexual risk behaviors in, 312 aid effectiveness, 23, 192­193, 233, ethics, 259­270 strokes in, 634 250­256 See also equity issues transportation choices in, 840 MDGs and, 185, 192­193 in choosing research priorities, 114, vaccine-preventable diseases in, 394, 399, reaching local communities, 255­256 265­269 400, 402, 408 challenges and transitional nature of, in cost-effectiveness analysis. See European Program for Intervention 239­240 cost-effectiveness analysis Epidemiology Training, 1002 composition of health financing by region current controversies in research ethics, European Union approval of and income level (2001), 230t 267­269 pharmaceutical products, 152­153 for contraception, 1332t discrimination against persons with evidence-based medicine, 121, 144, by country income level (2001), 18, 18t disabilities, 264­265 1001, 1026t development assistance. See development fair chances and best outcomes in terms excreta disposal. See sanitation assistance and health policy of, 264 improvements distributing and sourcing health genomics era and, 135 exercise. See physical activity interventions expenditures, 229 goals of ethical review of research, Expanded Program on Immunization district hospitals and, 1220­1221 266­267 (EPI), 44, 121, 383, 389, 397, 405b, for drugs, 1330, 1330t, 1332t in health resource allocation, 259­265 1177, 1186, 1324, 1347 for emergency medical services, 1272 hemoglobinopathies and, 670, 678 eye disease fiscal sustainability, 233 informed consent of human research sub- See also blindness; onchocerciasis; functions involved in, 225­227, 226f jects, 266­267, 269 trachoma for health information systems, placebo controls and, 268 diabetes and, 597 1018­1019 responsibility for health needs, 265 for health promotion and disease rights of host communities and, 268­269 fall-related injuries. See unintentional prevention, 211­223 standard of care and, 267­268 injuries in low-income countries, 5b, 20­21, state secrets of experimentation with family care leave, 221 211­212, 234­236 human subjects, 269n1 family planning, 1075­1090 community-based health insurance, in stem cell therapy, 130 See also contraception 231, 236 Ethiopia cost-effectiveness as intervention, 203 needs gap, 234 African trypanosomiasis in, 458 fertility behavior change, 510 new global alliances and funds, 234 child mortality in, 544b organization of programs, 1081 PRSPs, 235­236 dental treatment in, 729 social marketing of, 1081 macroeconomic consequences of aid, development assistance for health successful programs in, 167 23­24 to, 245 time interval between MDGs requirements, 185­193, 191b, 192t diarrheal diseases in, 378 pregnancies, 1079 in middle-income countries, 5b, 19­20, health care expenditures in, 245 vertical vs. integrated programs, 236­239 leprosy in, 444 1081, 1085 alternative risk pooling arrangements, malaria in, 414, 421, 426 family violence, 756 236­238 maternal deaths in, 544b See also interpersonal violence donor disengagement from, neonatal deaths in, 531 FDA. See U.S. Food and Drug 238­239 skin diseases in, 708, 709 Administration national health services, 238 1370 | Index private health insurance, 238 fluoridation, 972 kidney disease and, 696 single pool vs. virtual single pool, 238 dental caries. See dental caries mental disorders and, 605 sources of financing, 238 negative health effects of, 821 genetically modified crops, 131­132 universal coverage initiatives, 236, FOCUS on Young Adults, 1114, 1115t genetics 237­238, 237t Focusing Resources on Effective School diabetes and, 593 mobilizing government revenues, Health (FRESH), 784, 1096, 1096t, helminth infections and, 470 229­231 1097, 1098b, 1099, 1101, 1102­1103t hemoglobinopathies and, 673 models of, 226 folic acid, 836, 842, 939, 940­941, 948 insulin resistance syndrome and, 834b national health service (NHS) model, 226, folklore and alternative medicine, 1282 genital herpes, 312 238 food. See nutrition See also sexually transmitted for new product development, 146­150 food manufacture, 841­842, 856 infections (STIs) pharmaceutical firms in developing foot problems and diabetes, 597 genital mutilation, 761, 763 countries, 150 foreign direct investment and health genomics, 107­109, 126­130, 128b, 133, 135 private sector, 146­147 gains, 8 Genomics and World Health (WHO public-private partnerships, Framework Convention on Tobacco report), 127, 133, 135 147­150, 149t Control, 882 Germany public sector, 146 Framingham, Massachusetts, study of homeopathic products, use in, 1282 Poverty Reduction Strategy Papers males and heart disease, 120, 855 influenza in, 684 (PRSPs) and, 235­236 France transportation choices in, 840 for primary care, 1199­1200, 1200t homeopathic products, use in, 1282 universal coverage approach of, 237 private foundations, role of, 182, 253­255 skin diseases in, 714 GFHR. See Global Forum for Health See also specific foundations free trade agreements, 1141 Research private insurance model, 226, 238 FRESH. See Focusing Resources on Effective Ghana project support vs. budget support, 23 School Health breastfeeding programs in, 559, 561 purchasing, 226 fuel choice and air quality. See indoor air deworming programs in, 478 reallocation based on cost-effective pollution emergency transport to hospital in, 1264 criteria, 5b, 186 full income, 8­9, 9f food fortification in, 558 revenue sources, 226, 227, 231t, 240n2 defined, 4b, 8 health sector support program in, 251 risk pooling, 226, 227­232, 236­238 fungal infections of skin, 713­715 malaria in, 1092 social insurance model, 226 medical workers in, 1309 SWaps, 234­235 Gambia newborn survival rates in, 540 trends by country income level, malaria in, 418 organizational structure of health systems 231­232, 231f respiratory diseases of children in, 492 in, 93, 1345 for vaccines, 406­407, 1331, 1332t GATB. See Global Alliance for TB Drug road traffic injuries in, 746­747 Finland Development sexually transmitted infections in, CVD and lifestyle changes in, 837, 837b GAVI. See Global Alliance for Vaccines and 323, 325 diabetes and lifestyle changes in, 593 Immunization traditional medicine in, 1283­1284 rheumatoid arthritis in, 974 gender differences, 195­210 gingivitis, 729­730 tobacco taxes in, 882 See also women See also oral and craniofacial diseases and firearms and violence, 760, 763, 765 African trypanosomiasis and, 456 disorders See also interpersonal violence background, 195­197 Global Alliance for Improved Nutrition, first aid, 1264 burden of disease and, 197­198 182, 254 fiscal policies to promote health, disasters and, 1149 Global Alliance for TB Drug Development 211­223 hearing loss and, 958 (GATB), 140, 142 family care leave, 221 leishmaniasis and, 455 Global Alliance for Tuberculosis, 182 lessons in using, 222t mental disorders and, 605 Global Alliance for Vaccines and in low-income countries, 211­212 research agenda, 206­207 Immunization (GAVI), 24, 182, 234, maternity leave, 221 sex distinguished from gender, 196 244, 245, 253, 389, 397, 406, 491, sick leave, 221 sexually transmitted infections 492, 493 subsidies for health and health-related and, 318 Global Burden of Diseases 2000 study products, 212­218, 212t, vision impairment and, 955 (World Health Organization), 613 213­214t general primary care. See primary care Global Campaign against Epilepsy, 631 consumer subsidies, 212­218 generic drugs, 1326, 1328 Global Forum for Health Research producer subsidies, 218 genetic disorders (GFHR), 106, 114, 157, 161, 749 taxes. See taxation hemoglobinopathies and, 669­670, 671, Global Fund to Fight AIDS, Tuberculosis workplace health, 220­221 672, 678 and Malaria, 110, 147, 182, 234, 243, fiscal sustainability, 233 See also hemoglobinopathies 245, 305, 332, 425, 1333 Index | 1371 Global HIV Prevention Working Haemophilus influenzae type B (Hib). health systems Group, 332 See Hib disease prevention adolescents and, 1122 Global Immunization and Vision Haiti assessment of approaches to, 91­97 Strategy, 394 contracting for health care services in, capacity of Global Initiative on Children's 188, 1299 defined, 86n3 Environmental Health HIV/AIDS in, 358 low-capacity environments, solutions Indicators, 812 malaria in, 413 in, 98­99, 99t Global Polio Eradication Initiative, 1347 performance-based contracts with civil service reform and, 96 Global Programme against AIDS, 14 NGOs in, 252 constraints on delivery, 89­91, 89t, Global Programme for the Elimination of hand washing as intervention, 187, 90t, 1340 Lymphatic Filariasis, 439, 440 189­190, 377, 379, 556, 1041 cost-effective interventions aimed at, Global Strategy on Diet, Physical Activity, Harvard lifesaving study, 39b 27, 53­58 and Health, 219 HBV vaccine, 572­573 crisis in, 107 global surveillance networks, healers, use of in traditional medicine, 630, health care organizations, specificities 1011­1012, 1012f 1283, 1286­1287 of, 1341b global warming, 104b, 118 health care organizations, specificities history and current themes, 88­89, 107 Global Water Supply and Sanitation of, 1341b horizontal approach to. See horizontal Assessment 2000 Report (WHO health care workers. See medical workers approach to health systems and UNICEF), 772­773, 779 health education. See education human resources. See medical workers globalization, effect of, 3, 104, 104b, health gains, 5­6 IMCI and, 1178 1140­1141 See also specific diseases, regions, and improving quality, effect of, glomerulonephritis, 696­697 countries 188­189, 1178 See also kidney and urinary system effect on economic welfare, 4b, 7­9 inequities in. See equity issues diseases lessons of experience, 165­179 information and surveillance. See goiter, 554 linked to research, 160­161 information systems; surveillance gonorrhea. See sexually transmitted research methods on, 166 lack of information on, 99­100, 100b infections (STIs) successful programs for management. See management of gout. See musculoskeletal disorders behavioral change interventions, health services government reform and health care 175­177, 175b mass campaigns, 88 delivery, 90­91 cases selected for study, 166­169 occupational health and, 1141­1142 to meet MDGs, 185­190 combination or bundled organizational structures and financing Grand Challenges in Global Health, 115 interventions, 178 of, 92­94 Greece demand-side incentives, 175 outsourcing management of, 93­94, 93b dengue in, 452 environmental control interventions, primary care and, 1195­1196 hemoglobinopathies in, 669, 673 177­179 quality assessment and assurance, 96­97 malaria in, 421 intervention type, programmatic regulatory approach, 91­92, 92b, 96, 97 growth faltering. See stunting, wasting, and characteristics, and policies, 170­178 research and development agenda, micronutrient deficiency disorders product-intensive interventions, 99­100, 116 Guatemala 171­173, 171b service-level mechanisms, 98, 1340 field epidemiology in, 1002 service-intensive interventions, stewardship approach, 91­92 food fortification in, 558, 562 173­175, 173b strategic purchasing by, 1348­1349, 1349f indoor air quality and energy programs health information systems. strengthening of, 87­102, 1002 in, 190, 811 See information systems successes of, 88 interpersonal violence in, 766 health insurance surveillance. See surveillance occupational health risks in, 1136b community-based, 231, 236 sustainability of, 233 performance-based contracts with NGOs private, 231, 238 systems-level mechanisms, 97­98 in, 252, 252b universal coverage approaches, 236, vertical approach to. See vertical approach guinea worm eradication, 167, 169, 176, 237­238, 237t to health systems 779, 1173 health management information system weak, 99, 1340 Gulf Cooperation Council Group (HMIS), 1017 health workers. See medical workers Purchasing Program, 1327 Health Metrics Network (HMN), 1020b hearing loss guns and violence, 760, 763, 765 health policy acute ear infections in children, 484 See also interpersonal violence See also policy instruments aging and, 958 challenges for, 6­7, 1339­1340 causes and epidemiology of, 953, H. pylori infection, 569, 573, 574 cost-effectiveness and, 36b 957­960, 958t Haddon matrix, 739 IMCI and, 1177­1178 chronic middle ear infection, 959 1372 | Index cost-effectiveness of interventions, 954t global distribution and frequency of, hip replacement surgery, 973 gender differences and, 958 665­667, 665f, 666f, 666t, 667f historical achievements in medicine, hearing aids, 960 HIV/AIDS and, 670 119­120 interventions, 959­960 international and national support HIV/AIDS, 13­14, 331­369 mortality and, 958­959 groups, 678 See also specific countries and regions research and development agenda, 960 interventions, 45, 670­671, 677­678 action under uncertainty, 333­347 risk factors for, 958 normal hemoglobin, 663 adolescents and, 1112 years lived with disability and population genetics and dynamics, anemia in AIDS patients, 359­360 DALYs, 959 664­665 antiretroviral treatment for. helicobacter pylori. See h. pylori infection prevention, 669­670 See antiretroviral drugs (ART) helicopter ambulances, 1268 research priority of, 126 blood safety practices as prevention helminth infections, 467­482 screening and carrier detection, 671 of, 346 See also guinea worm eradication sickle cell diseases. See sickle cell anemia blood screening for, 346, 348 adolescents and, 1116 and related diseases blood transfusion and, 346­347 age-associated prevalence of, 469, 469f spectrum of inherited hemoglobin bloodborne transmission, 346­347 burden of, 470­471, 476­477 disorders, 663­664 breastfeeding and, 334, 336, 375 causes and characteristics of, 467­468 thalassemias. See thalassemias burden of, 333, 334t control of, 478­479, 783­784 hepatitis B chemoprophylaxis to treat, 349, 354 costs and cost-effectiveness of burden of, 396 circumcision as possible protection interventions, 46, 472­476, 475t cancer and, 569 against, 320, 322, 360b deworming (anthelmintic drug injecting drug users and, 315, 909 community health and nutrition treatment), 472­474 occupational risk of health care workers programs and, 1070­1071 school-based, 473, 474­475, 1092, for, 1130, 1130f concentrated epidemic, 348­349 1098b sexual transmission of, 315 condom use for prevention, 167, 170, economic benefits of interventions, vaccine for, 121­122, 244, 315, 389, 396, 176­177, 216, 345, 348b, 1080 477­478, 478f 406, 1324, 1325 low levels of use, 184 education and communication about, See also vaccine-preventable diseases costs and cost-effectiveness of 472, 475 hepatitis C interventions, 356t epidemiology of, 468­470 hemoglobinopathies and, 670 based on mode of transmission, excreta disposal and, 783­784 injecting drug users and, 909 344­345 global prevalence and distribution of, 467, occupational risk of health care workers prevention interventions, 336­347 468t for, 1130, 1130f school-based prevention programs, preschool children and, 473, 474 herbal products, use of, 1282­1283, 1283t 1105 research and development initiatives, heroin. See drug dependence successful prevention programs, 479­480 herpes simplex virus. See sexually 336­347, 339­343t risk factors for, 470 transmitted infections (STIs) development assistance in response to, sanitation improvements and, 472 H5N1 influenza. See avian influenza threat 244, 245 school-age children as high-risk Hib disease prevention, 167, 172, 244, 389, diagnostics development for, 144, 351 population for, 473, 474, 476, 396, 1324 discrimination against HIV positive 1092, 1098b See also vaccine-preventable diseases individuals, 109 selective treatment, 472 burden of disease, 396 drug resistance and, 358 soil-transmitted, 467­468 for children, 485­486, 491­492 drug therapy for, 1333 targeted treatment, 472, 476 high blood pressure. See cardiovascular economic effects of, 8, 9 universal treatment, 472, 476b disease (CVD), subheading: epidemic profiles to develop prevention water supply and, 779 hypertension guidelines, 347­349 hemoglobinopathies, 663­680, 669­672 high-income countries fungal infections and, 714 bone marrow transplantation, 669, 672, See also specific countries general interventions for transmission 675, 677, 678 breast cancer screening in, 578 of, 344­345 burden of, 667­669, 669t cancer diagnosis and treatment in, 559, generalized high-level epidemic, 349 clinical features, 664 582­583 generalized low-level epidemic, 349 costs and cost-effectiveness of diagnosis combination of high blood pressure, health challenge of, 4b, 7, 88 and management, 673­677, cholesterol, and obesity in, hemoglobinopathies and, 670 674t, 675t 859­860, 863 infected blood or blood products, thalassemias, 673­675, 674t, 675t morphine use for pain in, 984­985 exposure to, 335 ethical and social issues, 678 neonatal deaths in, 531 infection, determinants of, 333­336 future research needs, 678­679 smoking restrictions in, 876 infectivity, 334­335 Index | 1373 HIV/AIDS (Continued) second-line and subsequent referral hospitals. See referral hospitals injecting drug users and, 335, 346, 909, therapies, 358 small hospitals in developing countries 913, 917, 920 sex education and, 344, 344b and DALYS averted, 1255b interventions, 4b, 43, 344­345, sexual transmission of, 334­335, 334t surgery in. See surgery 350­360, 360b interventions to prevent, 345 households as producers and demanders of kidney disease and, 697 sexually transmitted infections and, care, 186­188 laboratory monitoring of immune 314, 345 HSV-1 and HSV-2, 315, 360b function as guide for therapy, skin diseases of, 707, 708, 709, 717 See also sexually transmitted 358­360 successful prevention programs, 332b, infections (STIs) leishmaniasis and, 453 336­347, 348b human capital approach. See value of a life expectancy, effect on, 6, 106 background data for, 336, 337­338t statistical life (VSL) low level epidemic, 347­348 theory and practice, 347­349, 347t Human Genome Project, 108, 126 malaria and, 418 symptom-based care, 352 Human Hookworm Vaccine Initiative, 480 monitoring of immune function to guide TB and, 291, 292­293, 294, 301, human papillomavirus and cancer, 569 therapy, 358­360 302, 304­305 human resources. See medical workers monitoring toxicity, 359­360 universal precautions, 346 human subjects research and ethics, mother-to-child transmission (MTCT), vertical transmission, 335­336 266, 269n1 332, 336, 339­343t, 345­346, voluntary counseling and testing (VCT), See also informed consent of human 345b, 376 332, 336, 344, 347, 348, 349 research subjects neurological problems in children women and, 195, 196 human T cell lymphotropic virus, 315 with, 935b workplace health and, 221 hurricanes, 1151 nutrition programs and food security, HMN (Health Metrics Network), 1020b hygiene. See hand washing as intervention; 352­353 home-based care, 1204, 1343­1344 sanitation improvements obstacles to control, 331­333 homeopathy, 1282­1283, 1283t, 1285, hyperpigmentation, 718 occupational risk of health care workers 1286, 1288 hypertension, 15 for, 1130, 1130f homicide. See interpersonal violence See also cardiovascular disease (CVD) opportunistic infections (OIs) and, Honduras hyperthyroidism, congenital, 353­355, 353f, 1339 cervical cancer screening in, 205 945­946, 949 oral manifestations of, 730 Chagas disease in, 438 hypothyroidism, 627 pain control for. See palliation community-based programs in, 1059, hypoxemia diagnosis, 488 peer-based intervention programs for, 1060b, 1067 344­345 interpersonal violence in, 766 IAEA. See International Atomic Energy perinatal transmission, 335­336 respiratory diseases of children in, 492 Agency prescribed therapy, adherence to, 358 Hong Kong IAVI. See International AIDS Vaccine prevention and management, 13­14, 43, avian influenza in, 682b Initiative 167, 348b drug dependence in, 916 The ICD-10 Classification of Mental and development assistance for, 254 influenza in, 684 Behavioral Disorders (ICD-10) lack of access to, 331­332 hookworm, 467, 471, 480, 783­784 (WHO 1992), 606t low levels of participation, 184 See also helminth infections ICPD. See International Conference on strategies for, 332b horizontal approach to health systems, Population and Development prevention-care synergy, 349­350, 350t 88, 99 IDSR (Integrated Disease Surveillance and prophylaxis for opportunistic management of health services and, Response), 1003­1004 infections, 355 1346­1347 IFF. See International Finance Facility (IFF) psychological support, 352 hormone replacement therapy, 967, proposal rape and, 758 971, 972 IHD. See ischemic heart disease referral hospitals and, 1239 Hospice Uganda, 990 illicit opiate abuse. See drug dependence research and development agenda, 114, hospitals ILO. See International Labour Organization 360­361, 360b accreditation of, 1296 (ILO) on occupational health risks respiratory disease associated with, changes in hospital practice, 130­131 ILO­WHO Joint Committee on 485, 489 defined, 1262 Occupational Health, 1131 response to therapy, 359 disaster damage to, 1153, 1155­1156 IMCI. See integrated management of rigorous evaluations, lack of, 332­333 district hospitals. See district hospitals childhood illness school-age children and, field hospitals for disaster relief, 1157 immunizations. See vaccine-preventable 1092­1093, 1093f governance of, 1348 diseases; vaccines; specific diseases cost-effectiveness of school-based management structure of, 94, 1342­1343 imprisonment for dependent illicit prevention programs, 1105 outsourcing management of, 93b opioid use, 911­912 1374 | Index improved health malaria in, 414, 418 fuel subsidy programs to control, 217 See also health gains morphine use in, 990­991 health impacts of, 795­796, 796t, 797t lessons of experience, 165­179 National Cancer Registry program, 1009 interventions and policy, 799, 800­801t value of, 158­160 neonatal deaths in, 531, 545, 545b kerosene and LPG as cleaner fuels, 803, incentive pay to medical providers, 4b, neurological disorders in, 628 805, 810, 811b 94­95, 95b, 170, 1299, 1310­1311, occupational risks of silica dust in, lack of lighting and, 797 1313­1320 1137, 1138b lessons learned, 808­811 community workers, 1063 performance-based contracts with NGOs levels of pollution and exposure, 795 income levels and health gains, 6 in, 253 MDGs on poverty reduction and, INDEPTH (International Network of Field quality of care in, 1298 799, 812 Sites with Continuous rotavirus vaccine in, 377 method used for determining attributable Demographic Evaluation of skin diseases in, 714, 715 disease burden, 796 Populations and Their Health in smallpox eradication in, 1163, 1170 research and development agenda, Developing Countries), 1021 strokes in, 633 811­812 India TB in, 253, 291 solid fuel use, 797 ambulances in, 1268 training for emergency medical infant care. See perinatal conditions arsenic contamination of water in, 821 personnel, 1273b infant mortality. See child mortality rates asthma in, 688 vaccine-preventable diseases in, 404b, 406 infectious and communicable diseases Bhopal catastrophe, 819, 819b, 1133, 1149 vital events registration in, 1021 See also specific diseases breast cancer in, 578, 584 water pollution in, 821, 825b, 828 burden of, 105 burn-related injuries in, 741 water supply in, 773­774 colonization efforts and control of, 120 cardiovascular disease in, 658, 659, women in, 195 control of 861, 863t Indonesia defined, 1164­1165 cataract surgery in, 1258b avian influenza in, 682b distinguished from eradication, 1165 child survival in, 1068 community-based programs in, 1064 cost-effective interventions for, 53 combination of high blood pressure, dengue in, 459 disasters and, 1151, 1154 cholesterol, and obesity in, 853­854, dental caries in, 725 elimination of, defined, 1165 861, 863t diarrheal disease in, 1303 eradication of, 1163­1176 community-based programs in, kidney and urinary system diseases See also specific diseases 1064, 1066 and, 697 certification process for, 1167 dengue in, 452 leprosy in, 441 defined, 1164 dental treatment in, 729 maternal deaths in, 533 distinguished from control, 1165 development assistance for health to, 245 mental disorders in, 621 economic considerations, 1167­1169 diabetes in, 597, 1333 neonatal deaths in, 533 frameworks for, 1165­1169 diarrheal diseases in, 372, 377, 378 polio in, 1170 geographic and environmental factors, district hospitals in, 1217 rice subsidy in, 216 1165­1166 drug addiction in, 908 skin diseases in, 708, 715 interventions to block transmission, emergency medical care in, 1273b smallpox in, 1163 1166­1167 epilepsy in, 630 TB in, 291, 295, 296 laboratory containment and, 1166 food fortification in, 558 tsunami in, 1149 local vs. international net benefits, food subsidies in, 216 indoor air pollution, 190, 793­815 1167­1168 hemoglobinopathies in, 664, 665, causes and burden of, 793­799 natural resistance to reinfection 668, 670 COPD and, 688, 690, 691­692, 804, 805 and, 1166 HIV/AIDS in, 254, 358 costs and cost-effectiveness of operational considerations, 1166­1167 immunization services in, 404b interventions, 799­808, 801b, 802b potential reservoirs and, 1166 indoor air quality and energy programs cost assumptions, 804 private vs. social net benefits, 1167 in, 190, 809, 809b cost-benefit analysis, 807­808, 808t scientific considerations, 1165­1166 Integrated Child Development Services cost-effectiveness analysis, 804­807, short-term vs. long-term net Program, 556 806t, 812 benefits, 1167 Kaposi's sarcoma in, 730 effectiveness assumptions, 804­805 surveillance and, 1166 kidney and urinary system diseases in, implementation period, 805 transmissibility and, 1166 695, 697, 700, 702, 703 DALYs and deaths due to, 797, 798t vertical vs. horizontal programs, leishmaniasis in, 452, 459, 461 economic effects of fuel collection, 797 1168­1169 leprosy in, 437, 445 electrification in rural areas, 795, extinction of, defined, 1165 lymphatic filariasis in, 436, 439, 440, 803­804, 810 gene therapy and, 127 442, 443 environmental consequences of, 797­798 kidney disease and, 697 Index | 1375 infectious and communicable sources outside of health sector, 1019t family and community practices and, diseases (Continued) strengthening of systems, 1023 1178, 1180, 1185­1186 new diseases since 1970, 112b, 122 surveillance. See surveillance future potential of, 1189­1190 partial control of, 121­122, 125 user fees to help support, 1019 health systems and, 1178, 1187­1188 research and development agenda, verbal autopsies, 1021 health workers and, 1178, 1186 106, 132 vital events monitoring, 1021 interventions, 1178­1180 skin diseases. See skin diseases information technology, potential of, 117 lessons learned, 1184­1188 water-related diseases, 775­776, 775t, 776t genomics field, 130 management process, 1179f infective endocarditis, 647 informed consent of human research Multi-Country Evaluation of IMCI See also rheumatic heart disease subjects, 266­267, 269 Effectiveness, Cost, and Impact INFECTOM (Information, Feedback, inherited disorders of hemoglobin. See (MCE), 1181­1184, 1183b Contracting with Providers to hemoglobinopathies newborn care and, 545b Adhere to Practice Guidelines, and Initiative for Vaccine Research Department policy shift to, 1177­1178, 1187 Ongoing Monitoring), 1303 (WHO), 407 research and development agenda, influenza and pneumonia, 484­485, Initiative on Public-Private Partnerships 1188­1189 681­684, 682­683b for Health, 148 integration of medical sciences in future, See also respiratory diseases of adults injuries, 16 132­133 diabetes and, 597 See also interpersonal violence; intellectual property and drugs, 150­152 influenza pandemic, possibility of, 4b, 7, unintentional injuries intentional injuries 106­107 adolescents and, 1111 costs and cost-effectiveness of See also avian influenza threat alcohol use and, 890 interventions, 48 information, education, and burden of, 107 Inter-Agency Coordinating Committee, 172 communication campaigns, 176, costs and cost-effectiveness of Inter-American Development Bank, 1149 178, 187, 475 interventions, 48­49 International AIDS Vaccine Initiative condom use, 1084 emergency care and, 1263 (IAVI), 182, 244, 254, 255b rheumatoid arthritis, 967 occupational injuries. See occupational International Atomic Energy Agency information, value of (VOI), 161­163, 162b health (IAEA) information systems, 1017­1030 prevention strategies, 48­49 African trypanosomiasis and, 460 benefits of improved information, surveillance, 1007­1008 International Centre for Diarrheal Disease 1023­1024 Injury Surveillance Guidelines Research, 374 collection, management, and analysis (WHO), 1007 International Classification of Functioning, systems, 1019­1022 insecticide-treated nets (ITNs) for malaria Disability, and Health (WHO), 937, costs and cost-effectiveness of improved prevention, 217, 421, 423, 424t, 425, 953, 957 information, 1023­1027, 1024t, 556, 1041 International Code of Marketing of 1025t, 1027t Institute of Medicine (IOM; U.S.) Breastmilk Substitutes, 539 data collection, 1019­1022 on global action on drug resistance, 1043 International Conference on Population direct expenditures for health on quality of medical care, 17, 17b, 599, and Development (ICPD), 499, information, 1018­1019 1293, 1294­1295, 1295b 513, 1095 dissemination and use of health insulin, 593, 596, 1328 International Council for the Control of information, 1023, 1297, 1305­1306 insulin resistance syndrome, 834b Iodine Deficiency Disorders, 559 district hospitals and, 1216 Integrated Disease Surveillance and International Diabetes Federation, 597 electronic medical records (EMRs), 1300 Response (IDSR), 1003­1004 International Dispensary Association, 1326 financing of, 1027­1028 integrated management and treatment International Energy Agency, 794, 810 health service statistics, 1021­1022 of TB, 295 International Finance Facility (IFF) information and communication integrated management of childhood proposal, 234, 250 technologies (ICT), 1022 illness (IMCI), 1177­1191 International Foundation of Internet access, 1022 antimicrobial drug provision and, 1185b Dermatology, 708 lessons learned, 1028 causes and burden of child mortality, 1177 International Labour Organization (ILO) national census, 1020 challenges when weak health systems, 99 on occupational health risks, 1128, national health account (NHA) CHNPs and, 1056 1129, 1130, 1135, 1140 framework, 1022 costs and cost-effectiveness of ILO­WHO Joint Committee on policy on information, 1019 interventions, 50, 1180­1184, 1181f Occupational Health, 1131 research and development agenda, on diarrheal diseases, 376, 384 WHO­ILO Joint Effort on Occupational 1028­1029 district hospitals and, 1215 Health and Safety in Africa, 1134 resource-tracking subsystem, 1022 drug resistance and, 1044 WISE (Work Improvement in Small sample surveys, 1020 equity issues and, 1184 Enterprises), 1133 1376 | Index International League against Epilepsy, 631 categories and pertinent policy long-term management of existing International Monetary Fund (IMF) instruments of, 59, 274b, 275 vascular disease, 651­652 on AIDS impact on full income, 9 child mortality and, 11­13 invasive interventions, 651 criticisms of structural adjustment community level of care. nonpharmacological programs and fiscal ceilings of, 233 See community-based programs and interventions, 652 Poverty Reduction Strategy Papers treatment pharmacological interventions, (PRSPs), 235­236 cost-effectiveness of. See cost-effective 651­652 International Network of Field Sites with interventions secondary prevention, 654 Continuous Demographic costs of ISDR (International Strategy for Disaster Evaluation of Populations and amount of health (by service or Reduction), 1147 Their Health in Developing intervention) US$1 million Islamic Republic of Iran Countries (INDEPTH), 1021 will buy, 25t earthquake in, 1149 International Partnership for determination of, 280­281 hemoglobinopathies in, 669, 670 Microbicides, 244 defined, 273­277 Isfahan Healthy Heart Program in, 842 International Society of Nephrology, distribution by cost-effectiveness ratio, 53f kidney and urinary system diseases in, 700, 703 district hospitals. See district hospitals 697, 698 International Strategy for Disaster drugs. See drugs ITNs. See insecticide-treated nets (ITNs) Reduction (ISDR), 1147 general primary care. See primary care for malaria prevention International Trachoma Initiative, 182 for MDGs, 183­185, 184t IUATLD (International Union against International Union against Tuberculosis packaging of, 49­50 Tuberculosis and Lung and Lung Disease (IUATLD), 691 personal. See personal interventions Disease), 691 International Zinc Nutrition Consultative population-based. See population-based IUDs, 1080 Group, 554, 565 interventions ivermectin Internet access, 1022 quality as determinant of distribution in Africa, 168, 172 interpersonal violence cost-effectiveness, 50 lymphatic filariasis and, 439 adolescents and, 1111 referral hospitals. See referral hospitals onchocerciasis and, 109, 110, 168, 172, burden and causes of, 756­759 intrauterine devices, 1080 440­441, 444, 957 collecting and managing data on, 765 intussusception and rotavirus vaccine, 376 cost-effectiveness of interventions, iodine deficiency Jamaica 764­765 in children, 554, 558­559 community-based programs in, 1065, cost of, 760 costs and cost-effectiveness of 1066, 1067 data on, 756­757 interventions, 561 hemoglobinopathies in, 664, 668, 676 cost-benefit data needed, 765 interventions, 558­559 quality of health care in, 1295­1296 deaths resulting from, 757, 757f, 757t, 758t lessons learned, 563 salt fluoridation in, 168, 173, 726 in developing countries, 758 mental retardation and, 562 Japan economic impact of, 760­761 during pregnancy, 936b cleft lip and palate in, 731 effects on public finance, 761 salt iodination program to combat, dengue in, 452 firearm registration, 765 168, 173 kidney and urinary system diseases future challenges for, 766­767 in school-age children, 1094 in, 695 interventions, 761­764 IOM. See Institute of Medicine strokes in, 633 legislation and shelter for abused women, IQ. See cognitive function water pollution in, 829 764­765 Iran. See Islamic Republic of Iran Japanese encephalitis parent training and home Ireland vaccine for, 406, 1325­1326 visitation, 765 development assistance for health See also vaccine-preventable diseases prevention from, 244 Johns Hopkins School of Public Health on implementation of, 765­766 fluoridation of water in, 725 development of pneumococcal primary, 761, 763­764, 766 iron deficiency. See anemia vaccines, 493 secondary and tertiary, 764 ischemic heart disease (IHD), 645­646 Joint Learning Initiative, 26b strategies, 761­764, 762­763t burden of, 649­650, 1278n2 Joint United Nations Programme on research and development agenda, cost-effectiveness of interventions, HIV/AIDS. See United Nations 765­766 650­654, 655t, 656t, 658 Programme on HIV/AIDS risk factors, 759, 759t, 760 epidemiological transition and, 647, 648t (UNAIDS) support services for victims, 766 linking costs and effectiveness justice and special concern for the worst youth intervention, 765 in developing countries, 653­654, off, 262 interventions 655t, 656t juvenile violence, 760 See also specific diseases and conditions in high-income countries, 658 See also interpersonal violence Index | 1377 kangaroo mother care, 540 knee replacement surgery, 973 water supply in, 771, 772, 773, 774, 776 Kaposi's sarcoma, 730 knowledge yellow fever in, 396­397 Katayama Syndrome, 468 See also information systems LDDs. See learning and developmental Kazakhstan diffusion of, 5b, 103, 117, 187 disabilities cardiovascular disease in, 649, 649t types of, 110 lead exposure, 820 Kenya Koch, Robert, 119 learning and developmental disabilities African trypanosomiasis in, 458 Korea, Republic of (LDDs), 933­951 drug dispensing in, 92b drug prescribing and dispensing attention deficit hyperactivity disorder fluorosis of teeth in, 731 in, 1346 (ADHD), 935b helminth infections in, 475 Industrial Safety and Health Act, 1140 autism, 935b HIV/AIDS in, 336 universal coverage approach of, 237 burden of, 934­937 hospitals in, 1217, 1218, 1219t, 1343, urban life in, 841b causes of, 934t 1345, 1348 Kuznets curve, 749b community-based rehabilitation and, IMCI program in, 1180 946­947 indoor air pollution in, 795, 802­803, 802b Lancet on need for new health systems congenital hyperthyroidism, 945­946 interpersonal violence in, 757 research specialty, 99 costs and cost-effectiveness of malaria in, 418, 426, 1092 Lao People's Democratic Republic interventions, 933 onchocerciasis in, 440 drug policy in, 1346 community-based rehabilitation, 937 pain control for cancer in, 986b water pollution in, 829 food fortification, 940­941 road traffic injuries in, 749b Latin America and Caribbean neonatal screening, 943, 945­946 traditional medicine in, 1283, 1285, 1286 See also specific countries Down syndrome, 941­945, 944t water supply in, 773 abortions in, 1078 education and work and, 933 Kenya Medical Research burn-related injuries in, 741 environmental exposures, 936b Institute­Wellcome Trust cardiovascular disease in, 650, 654, 655t, fetal alcohol syndrome, 936b Collaborative Research 656, 656t impairment, disability, and participation Programme, 92b Chagas disease in, 167, 436, 442, 446 levels of, 937 kidney and urinary system diseases, contraceptive use in, 1076, 1080, 1086 interventions, 933, 935­936b 695­706 dengue in, 452, 454­455, 459, 461 in low- and middle-income countries, acute renal failure, 698 development assistance for health to, 245 938­948, 939t benign prostatic hypertrophy, 698 diabetes in, 597 lifelong duration, 933 burden of, 695­696, 696t diarrheal diseases in, 373 neurological disabilities of causes of diseases of, 696­698 disasters, impact in, 1152, 1152t childhood, 935b diabetes and, 698 ecstasy (MDMA) use in, 924 prevalence, 933, 937­938 economic benefits of interventions, excreta disposal in, 780 prevention levels, 937 700­702, 701t food subsidy programs in, 216 research and development agenda, 948 future challenges for, 704 hurricanes in, 1151 unintended consequences of successful or genetic diseases, 696 indoor air pollution in, 803 partially successful interventions, 937 glomerulonephritis, 696­697 interpersonal violence in, 757 Lebanon HIV/AIDS and, 697 leishmaniasis in, 453 air pollution in, 828 hypertension and, 698 lymphatic filariasis in, 440 hemoglobinopathies in, 670 implementation of control strategies, malaria in, 414, 421 legislation 702­703 mental disorders in, 619, 622 cigarette and tobacco use warnings, infections, stones, and obstructive neonatal deaths in, 533 168, 876 uropathy, 697­698 occupational health risks in, 1128, 1135, drunk driving, 893 interventions to delay CKD (chronic 1136b first public health act in UK, 120 kidney disease), 699­700 onchocerciasis in, 441, 957 food advertising and, 842 prevention of kidney failure, 703 performance-based contracts with NGOs food manufacture and, 841 renal replacement therapy (RRT), 698, 703 in, 252 illicit opioid use, 911­912 research and development agenda, proposed guidelines for HIV/AIDS in, occupational health, 1131­1132 703­704 348­349 road safety, 747, 893 risk factors for kidney disease, 699 referral hospitals in, 1239 smoking bans in public places, 875­876 scabies and, 710 road injuries in, 1111 leishmaniasis schistosomiasis. See schistosomiasis salt fluoridation in, 726 burden of, 455­456 TB and, 697 sector adjustment loans to, 251 characteristics and transmission, 452­453 teaching and research centers, vaccine-preventable diseases in, 400, costs and cost-effectiveness of establishment of, 703 402, 408 interventions, 46, 459 1378 | Index economic impact of, 455 low-income countries natural history and course, 610 HIV/AIDS and, 453 See also specific countries policy and service implications, 619­621 management and control strategies of, burden of disease in, 31t public support for cost-effective 457­458, 461­462 cancer diagnosis and treatment in, intervention package, 621­622, 622t social impact of, 455­456 582­583 malaria vaccine development for, 462 causes of deaths in, 30t ACT and, 421, 424 leprosy combination of high blood pressure, affordability and scaling up, 425 burden of, 437 cholesterol, and obesity in, 860­863 anemia and, 417­418 characteristics and transmission, 435­436 diagnostics development in, 144 artemisinin to treat, 111, 1288 costs and cost-effectiveness of drug abuse, illicit, 916 See also artemisinin combination interventions, 444­445 financing health in, 5b, 20­21, 27, therapy (ACT) interventions and effectiveness of, 441 229­231, 234­236 benefit-cost ratios, 427 patterns of skin diseases in community See also financing burden of, 414­418, 414f, 415t and, 709 fiscal policy for health in, 211­212 causes and diagnosis of, 413­414 skin depigmentation in, 715 morphine use for pain in, 985 cerebral malaria and congenital leukemia, 315 neonatal deaths in, 531, 547b neurological disorders, 414, 935b levels of care new product development in, 142­143 chemoprophylaxis, 420 community level. See community-based private spending on health in, 231 child deaths due to. See child programs and treatment LPG Rural Energy Challenge, 810 mortality rates district hospitals. See district hospitals lung cancer children and, 417­418 general primary care. See primary care clinical epidemiology and, 120 civil engineering and, 422 referral hospitals. See referral hospitals indoor air quality and, 795 costs and cost-effectiveness of LF. See lymphatic filariasis mortality rates and, 570 interventions, 418­427, 422­427 life expectancy, increases in, 3, 4b, 6, occupational risk of, 1128, 1128t, switching first-line drugs, 423­424, 6t, 103 1129, 1130 424f, 424t See also mortality rates screening for, 574­575 deaths due to in low- and middle-income valuing of, 157­163 tobacco use and, 573, 871­872, 872f countries, 415­416, 416t of women, 5, 5f, 197, 201­202 workers' risk for asbestos and smoking development assistance and, 244 lifelong medical management, 4b, 27, 116 exposure, 1128, 1128t diagnosis and treatment of, 418­420 lifestyle lymphadenitis, 435 diagnostics for, 144 See also specific lifestyle choice lymphatic filariasis (LF) drug resistance and, 122, 414, 423­425, (e.g., alcohol use, tobacco use) burden of, 436 424t, 1036, 1046 cancer and, 569, 833 characteristics and transmission, 434 drug treatment of, 418­421 cardiovascular disease and, 647, 649, 652, costs and cost-effectiveness of economic benefits of control, 425­426 659, 833­836 interventions, 45­46, 442­443, 443t economics of interventions for, 422­427 chronic diseases and, 833­836, 834t interventions and effectiveness of, emergency medical services and, 1262 combination of high blood pressure, 439­440 genomics and mosquitoes, 108 cholesterol, and obesity and, 856 skin diseases and, 709 health education on, 422 diabetes and, 593, 1333 HIV/AIDS and, 418 helminth infections and, 470 macroeconomic consequences of aid, 23­24 home repellents and insecticide use neurological diseases and, 634 Madagascar for, 422 taxation to discourage high-risk breastfeeding programs in, 559, 561 indirect and comorbid risks, 417­418 behaviors, 177 cervical cancer screening in, 576, 577t indoor residual spraying (IRS), 421, 423, liquid petroleum gas, 217, 803, 805, contracting for health care services in, 188 424t, 1163 810, 811b influenza in, 684 innovative approaches for, 427 Live for Life (Johnson & Johnson), 838b leprosy in, 437 interventions, 45, 418­422 liver cancer, 569, 570, 572, 573, 574 malaria in, 414 combination or bundled local level. See community-based programs oral health and education program in, 726 interventions, 178 and treatment sexually transmitted infections in, 324 oral antimalarial treatment, 1092 Louisiana and free condom distribution skin diseases in, 708 IPT during pregnancy and infancy, 420, program, 216 major depressive disorders, 610­611 423, 424t low birthweight See also mental disorders ITNs for prevention. See insecticide- cardiovascular disease and, 121 cost-effectiveness methods and results, treated nets (ITNs) for malaria as cause of neonatal death, 534, 540 613­619, 614­618t prevention diabetes and, 121 epidemiology and burden, 610­611 larviciding and fogging, 421­422 malaria and, 418 interventions, 611 low birthweight and, 418 Index | 1379 malaria (Continued) maternal undernutrition, 510 birth preparedness, 511 macroeconomic impact of, 426 micronutrient deficiency, 1094 burden of, 508 maternal conditions and, 417­418 productivity loss and, 562­563 causes and conceptual frameworks, 505 mosquito repellents, 422 research and development agenda, comparison of alternative intervention patient management, 427 564­565 packages, 514­517, 515­516t, policy research on, 427­428 school-age children and, 1093 518­519t prevention research for, 427 short-term hunger of children, 1094 components of comprehensive safe moth- research and development agenda, vitamin B12 and dementia, 633 erhood strategy, 508­509, 509b 427­428 malpractice, 167, 1296 costs and cost-effectiveness of school-age children and, 1092, 1106 management of health services interventions, 46, 203, supervision and policy change, 420­421 See also integrated management of 513­522, 520t vector control, 421­422 childhood illness (IMCI) cost-effectiveness ratios, 517, 520t, vertical approach to, 88 ambulatory care, 1343 521­522 Malawi capital charging and, 1345 model assumptions, 525, 525t, 526t clinical officers in, 94 in community and social care settings, DALYs for, 508 diarrheal disease treatment in, 1303 1343­1344 defined, 500 emergency medical transport in, 1268 contracting for services. See outsourcing diabetes and, 597 HIV/AIDS in, 336 of health care services economic benefits of intervention, 522 household fuel use in, 793, 794f drugs and, 1345­1346 epidemiology of, 499­508 malaria in, 421, 426 hospitals, 1342­1343 fertility behavior change, 510 maternal and perinatal mortality in, 1313 human resources and. See medical fetal alcohol syndrome, 936b training of medical personnel in, 94, 1313 workers focus conditions and their risk factors, Malaysia intellectual resources and, 1346 501­504t anesthesia safety in, 1298 management, defined, 1340­1342 HIV/AIDS and, 43, 375 childbirth conditions in, 547b new public management, 92, 1341­1342 See also breastfeeding dengue in, 452 physical resources and, 1345 home-based care, 511 epidemiological transition in, 125 in primary care facilities, 1343 interventions, 508­513 Health Technology Assessment provider autonomy and, 1348 iodine deficiency and, 936b Unit in, 125 public vs. private provision, 1348 lessons learned, 522­524 hospital management in, 1345 quality of care and, 1342 malaria and, 417­418 occupational health in, 1134 relationships and, 1346­1347 treatment during pregnancy, 420 road traffic injuries in, 740, 744 strategic management, 1339­1352 male involvement, 511 skin diseases in, 715 strategic purchasing, 1348­1349, 1349f maternity leave from work, 221 males. See gender differences strengthening of, 1345 MDGs on, 499 Mali successful initiatives, 1344­1349 medical advances in, 123 childhood immunization programs in, Management Sciences for Health, 1326 nature and characteristics of, 500 1024 Managing Drug Supply (Quick et al.), 1326 neural tube defects during pregnancy, dental treatment in, 729 manipulative medicine, 1282 939­940 development assistance for health to, 245 See also complementary and alternative nutritional interventions, 510 diarrheal diseases in, 378 medicine (CAM) personal interventions, 510­513 skin diseases in, 708 marijuana. See drug dependence policy considerations and malnutrition mass campaigns, 88, 176 approaches, 513 See also stunting, wasting, and mass drug administration (MDA) population-based interventions, 510 micronutrient deficiency disorders lymphatic filariasis and, 171, 439, 442, postpartum care, 512 adolescents and, 1111 443, 446 primary-level care, 511­513, 515­516t childhood illnesses and, 563, 1093 onchocerciasis and, 171­172, 444, 446 quality of evidence on, 509­510 community-based interventions, 559, 561 successful programs utilizing, 171­172 research and development agenda, See also community health and massage 126, 524 nutrition programs (CHNPs) back pain and, 975 resource use and costs, 517 conceptual framework for causes manipulative medicine and, 1282 risks associated with, 1078­1079 of, 1057f Maternal and Child Insurance Program secondary-level care, 513, 515­516t diarrhea and, 372 (World Bank), 251 successful programs for improving, 168 human capital formation and, 562 maternal conditions, 499­529 maternal mortality learning and developmental disabilities See also abortions; childbirth conditions abortion and, 1078 and, 936b alcohol consumption and fetal alcohol causes and conceptual frameworks, 505 leishmaniasis and, 453 syndrome, 936b emergency care and, 1263, 1278n3 1380 | Index family planning and, 1078­1079, 1084 paramedical personnel, 1264­1265 major depressive disorders, 610­611 health challenge of, 4b, 203, 499 defined, 1262 mood disorders, 608­611 levels, trends, and differentials, 505­508, peer review of, 1296­1297, 1299 panic disorders, 612­613 506t, 507f performance-based professional policy and service implications, reducing in low-income countries, 547b recognition for, 1299­1300 619­621 Mauritius performance-based remuneration for, posttraumatic stress disorder (PTSD), combination of high blood pressure, 1299, 1348 605, 610 cholesterol, and obesity in, 856 primary care workers, 1200, 1201b public support for cost-effective food manufacture in, 841 remuneration of (generally), 1310, intervention package, tax rate reduction on alcohol, 900 1315­1317 621­622, 622t MCE. See Multi-Country Evaluation of optimal combination of compensation schizoaffective disorder, 609 IMCI Effectiveness, Cost, and Impact and incentives, 1319­1320 schizophrenia. See schizophrenia and MDA. See mass drug administration payment methods, 1317­1318 nonaffective psychoses (NAP) MDGs. See Millennium Development Goals research and development agenda, 1321 stigma of, 606­607, 641 MDMA (ecstasy). See drug dependence safety and effectiveness of new health pro- suicide, 607 MDT. See multidrug therapy fessions, 1313 mental retardation measles skills of, 1344, 1344t See also learning and developmental See also vaccine-preventable diseases surgery and, 1256 disabilities (LDDs) burden of, 395­396 traditional healers, 1283, 1286­1287 Down syndrome, 941­945 elimination programs, 98­99, 168, 170, training. See training of medical iodine deficiency and, 562 173, 1164 personnel Merck & Co. and ivermection, 110, 168, predisposition for diarrheal diseases, 377 unionization of, 1315 172, 440, 444 vaccine for, 98­99, 121, 395­396, 1325 workforce planning to meet policy goals, metabolic disorders, screening for, 937 vaccine gap and, 1326 1311­1312 metabolic syndrome, 834b Mectizan. See ivermectin medicinal plants, use of, 1286 methadone. See drug dependence Medicaid and Oregon Health Services See also complementary and alternative Mexico Commission, 263 medicine (CAM) air pollution reduction in, 824b medical workers, 1309­1322, 1310t medicine, history of, 119­120 arsenic contamination of water in, 821 advice for governments on, 1320 Medicines for Malaria Venture (MMV), cardiovascular disease in, 650 brain drain of, 94, 95b 140, 142, 182, 244 combination of high blood pressure, development assistance for health, need to Mediterranean region cholesterol, and obesity in, target, 245 hemoglobinopathies in, 665, 670, 673, 853­854 in district hospitals, 1223 674, 675 dengue in, 455 experience levels of, 5b melasma, 718 diarrheal diseases in, 374, 377 health sector reform and, 1311 meningitis, 292, 396, 486, 936b, 1334 drowning in, 741 IMCI programs and, 1178, 1186 See also vaccine-preventable diseases fall-related injuries in, 748­749 importance of, 94­96, 1344­1345 mental disorders, 16, 605­625, 606t, food fortification in, 558 incentive pay for, 4b, 94­95, 95b, 1299, 614­618t HIV/AIDS treatment in, 13, 358 1310­1311 See also specific disorders kidney and urinary system diseases incentives and motivation for, 1310­1311, adolescents and, 1111 in, 703 1313­1320, 1316­1317t amphetamines and, 922 National Crusade for Quality in Health context and external incentives, anxiety disorders, 611­613 Care in, 1299 1314­1315 bipolar disorders, 608­610 respiratory diseases of adults in, 691 group incentives, 1318 burden of, 107, 605, 606t, 1334 school enrollment in, 1101 optimal combination of compensation cannabis and, 918 skin diseases in, 708, 715 and incentives, 1319­1320 cocaine use and, 920 social welfare program (originally in practice, 1315 costs and cost-effectiveness of PROGRESA) and health sustainability and, 1318­1319 interventions, 47, 613­619, improvements in, 24, 167­168, typology of incentives, 621­622, 622t 169, 175, 177, 217­218, 1101 1313­1314, 1314t estimation of population-level costs, stomach cancer in, 573 MDGs, human resource needs to 613­619, 615­618t universal coverage approach of, 237 meet, 1310 estimation of population-level VAT exemptions for medicines, physician as needed resources, 1310, 1311t effectiveness of treatments, services, and some foods, 220 numbers and types of, 1312­1313 613, 614t water supply in, 774 occupational risk for HIV, hepatitis B, and interventions, 605­606, 1343 micronutrient supplementation, hepatitis C, 1130, 1130f effectiveness of, 614, 618 1058­1059 Index | 1381 Middle East and North Africa Marginal Budgeting for Bottlenecks mosquitoes See also specific countries Model, 191b See also vector control cardiovascular disease in, 650, 656t on maternal and perinatal lymphatic filariasis and, 434 contraceptive use in, 1076, 1080 conditions, 499 malaria and. See malaria development assistance for health MDG Needs Assessment Model, 191b yellow fever and, 396, 1163 to, 245 progress report (2004), 182­183 motor vehicle injuries. See traffic injuries diabetes in, 592, 1333 prospects for time remaining, 183 Mount St. Helens (U.S.), 1150 ecstasy (MDMA) use in, 924 scaling up for use of interventions, Mozambique epilepsy in, 631 183­185 government reform and health care hemoglobinopathies in, 665, 668 estimating cost of, 191b delivery in, 91 leishmaniasis in, 453, 462 stimuli external to health sector and, 183 leprosy in, 437 neurological diseases in, 631 strengthening of core public health neonatal deaths in, 531 proposed guidelines for HIV/AIDS in, functions, 189­190 skin diseases in, 708 347­348 on TB, 290, 305 Multi-Country AIDS Program in skin diseases in, 714 tracking progress and information needs Africa, 248 strokes in, 634 for, 1018 Multi-Country Evaluation of IMCI vaccine-preventable diseases in, 399, 400, on vaccine-preventable diseases, 389 Effectiveness, Cost, and Impact 402, 408 on water quality, 829 (MCE), 1181­1184, 1183b, 1185, middle-income countries mind-body medicine, 1282, 1286, 1185b, 1187, 1195 See also specific countries 1288­1289 multidrug therapy burden of disease in, 31t See also complementary and alternative See also polypill cancer in, 584 medicine (CAM) leprosy and, 441, 442, 445 causes of deaths in, 30t mineral fortification. See salt multisectoral determinants of health, 11b financing health in, 5b, 19­20, 27­28, minimally invasive surgery, 130­131 musculoskeletal disorders, 963­980 236­239 ministry of health, 231­232, 238 aging and, 963 See also financing miscarriages, 939 ankylosing spondylitis, 976 midwives, 174, 511, 545 See also abortions arthritis, types of, 965, 967 See also childbirth conditions MMV. See Medicines for Malaria Venture cost-effectiveness of interventions, 972, migration, effect of, 104b modeling, 281 973­975 brain drain of medical personnel, 94, 95b, molecular epidemiology, 107­109 disease-modifying antirheumatic 1309­1310 monogenic diseases, research priority drugs, 974­975 Millennium Development Goals (MDGs), of, 126 lessons learned, 976­977 24, 181­194 mood disorders, 608­611 primary interventions, 972 contracting out of health care services, See also mental disorders secondary interventions, 972­973 effect of, 188­189 bipolar disorders, 608­610 tertiary interventions, 973 cost of achieving, 190­191, 245 cost-effectiveness methods and results, back pain, 963, 966, 975­976 by countries, 190­191, 191b, 192t 613­619, 614­618t chiropractic manipulation and, 1288 globally, 190 major depressive disorders, 610­611 occupational risks for, development assistance and, 185, policy and service implications, 619­621 1136­1137, 1137f 192­193, 234, 244, 245­246 public support for cost-effective burden of, 963­965, 964t on diarrheal diseases, 384 intervention package, research needs, 977 effective interventions available for, 621­622, 622t causes and epidemiology of, 965­966 183­185, 184t Morocco combination of hypertension, cholesterol, underuse of, 184­185 contraceptive use in, 1083 and obesity and, 851 endorsement at UN Millennium respiratory diseases of adults in, 691 cost-effectiveness of interventions, Summit, 182 trachoma control program in, 168, 170, 968­972, 969­970t government and quality of policies, 174, 178 research needs, 978 effect of, 185­190 morphine use for pain. See palliation interventions, 206 government spending, mobilizing of, mortality displacement, 819 research needs, 978 192­193 mortality rates lessons learned, 976­977 health systems geared toward, 97 See also specific diseases and conditions osteoporosis, 963, 965, 967­968, 970 importance to all countries, 183 causes of in low- and middle-income lessons learned, 977 indoor air pollution and, 799, 812 countries, 30t primary interventions, 970­971 on infant mortality, 7, 531, 534, 547­548 in cost-effectiveness analysis, 260 secondary interventions, 971­972 list of health-related goals, 182b lowering of, 9­10 physical activity and, 970­971 Maquette for Multisectoral Analysis, 191b surveillance and, 1009 preventive strategies, 966 1382 | Index research and development agenda, Nepal neuropsychiatric diseases 977­978 arsenic contamination of water in, 821 research and development agenda, 109 risk factors for, 966, 966t leishmaniasis in, 452, 459 technological advances and, 131 screening for, 971 leprosy in, 437 new product development. See research and surgery for, 968 maternal mortality in, 510 development symptomatic treatments for, newborn care in, 535 new public management, 92, 1341­1342 966­967, 967t oral health program in, 733­734 New Zealand women and, 201, 206 reproductive health in, 474 capital charging and management of Myanmar respiratory diseases of adults in, 691 physical resources in, 1345 dengue in, 454 respiratory diseases of children in, 492 combination of high blood pressure, kidney and urinary system diseases sexually transmitted infections in, 323 cholesterol, and obesity in, 855, 859 and, 697 TB in, 295 fluoridation of water in, 725 as opium producer, 908 tobacco taxes in, 882 interpersonal violence in, 765 oral precancer and cancer in, 730 Netherlands suicide prevention for youth in, 1121 myocardial infarction. See acute myocardial breast cancer in, 584 universal coverage approach of, 237 infarction (AMI) COPD and asthma in, 686 newborns. See child mortality rates; development assistance for health perinatal conditions NAFTA (North American Free Trade from, 244, 254 disorders of. See congenital and Agreement), 1141 food manufacture to reduce fat developmental disorders NAP. See schizophrenia and nonaffective content, 845 NGOs. See nongovernmental organizations psychoses hospitals in, 1224­1225 Nicaragua Narcotics Anonymous, 911 transportation choices in, 840 community-based programs in, National Academy of Sciences (U.S.), 9 universal coverage approach of, 237 1066, 1067 National Center for Complementary and neural tube defects during pregnancy, injury surveillance system of, 168 Alternative Medicine 939­940 interpersonal violence in, 760, 766 (NCCAM), 1289 neurological disorders, 627­662 kidney and urinary system diseases National Center for Policy Analysis Alzheimer's disease (AD) and other in, 702 (Harvard University) study of dementias, 627­629 occupational health risks in, 1136b life-saving interventions, 39b burden of, 628, 629t, 632 performance-based remuneration for National Commission for the Protection of in children with HIV/AIDS, 935b health workers in, 1299 Human Subjects, 266, 267 costs and cost-effectiveness of pesticide poisoning prevention measures National Crime Prevention Council, 760 interventions, 47 in, 1139 National Evaluation of Pharmacother in developing countries, 635­640, 635t, nicotine. See tobacco use and control apies for Opioid Dependence 636t, 638t, 639t Niger Project, 913 epilepsy, 629­631 doctor visits for sick children in, 89 National Institute for Occupational Safety See also epilepsy IMCI program in, 1187 and Health (NIOSH; U.S.), 1134 learning disabilities and, 935­936b Nigeria National Institute of Allergy and Infectious See also learning and developmental hemoglobinopathies in, 664 Diseases, 114, 115 disabilities (LDDs) IMCI program in, 1180 National Institutes of Health (U.S.) malaria and. See malaria insecticide-treated nets (ITNs) for on complementary and alternative Parkinson's disease, 632­633 malaria prevention in, 217 medicine, 1281­1282 See also Parkinson's disease (PD) lymphatic filariasis in, 439 on COPD, 684 recommendations, 640 malaria in, 426 on eradication of infectious research and development agenda, neonatal deaths in, 531 diseases, 1164 640­641 neurological disorders in, 628 research fund allocation assessments, 161 stroke, 633­635 polio in, 254b, 1168 national statistics office, role of, 1018 burden of disease, 634 refugee populations and surveillance, 1008 natural disasters, 1147­1162 cost-effectiveness of interventions in nonaffective psychoses and schizophrenia. See also disasters developing countries, 638t, See schizophrenia and nonaffective NCCAM (National Center for 639t, 640 psychoses (NAP) Complementary and Alternative frequency of types, prevalence, noncommunicable diseases, 15­16 Medicine), 1289 incidence rates, mortality, and See also specific diseases and conditions neonatal conditions. See child mortality disability after stroke, 633 burden of, 105, 1331 rates; childbirth conditions; perinatal interventions, 634­635 cost-effective interventions for, 53 conditions recommendations, 640 increase in, 4b, 6 neoplasms. See cancer risk factors, 633­634 research priorities for, 115­116, 132 Index | 1383 nonfilarial lymphoderma, 709 obesity protective equipment research, 1143 nongovernmental organizations (NGOs) aggregate costs of, 846 research and development agenda, contracting out of medical services to, chronic diseases and, 833­835 1142­1143 93­94, 96b, 188 in combination with high blood pressure risks posed by exposure to asbestos and mental disorders and, 620­621 and cholesterol, 851­868 smoking, 1128, 1128t performance-based contracts with, See also cardiovascular disease (CVD) silicosis risks, 1137, 1138b 252­253, 252b, 1299 diabetes and, 593 smoking bans, 876 sanitation and, 782 food taxes to discourage, 219­220 state or government interventions for, North American Free Trade Agreement interventions to reduce, 857­858, 863 1131­1132 (NAFTA), 1141 low birthweight and, 121 surveillance and reporting of, 1133, 1143 nurses musculoskeletal disorders and, 966 technology and, 1143 See also medical workers risk factor for cancer, 574 workplace-based interventions, need for, 94 in school-age children, 1094 1132­1133 as primary care providers, 620 surveillance data on, 1009 occupational lung diseases, 689­690, 689f, nutrition obsessive-compulsive disorders, 612 1128, 1128t adolescents and, 1116 See also anxiety disorders See also respiratory diseases of adults breastfeeding supplementation, 555, 557b obstetric care. See childbirth conditions; silicosis, 1137, 1138b combination of high blood pressure, maternal conditions OCP. See Onchocerciasis Control Program cholesterol, and obesity and, 856 obstructive airways disease. See chronic OEPA. See Onchocerciasis Elimination community-based health and nutrition obstructive pulmonary disease Program for the Americas programs. See community health (COPD) and asthma Office of Human Research Protections and nutrition programs (CHNPs) occupational health, 1127­1145 (U.S.), 269n2 costs and cost-effectiveness of access to health care, 1133 onchocerciasis interventions, 46, 559­561, 560t, back pain risks, 1136­1137 burden of, 436­437 844­846, 845t capacity building and, 1134­1135 characteristics and transmission, 434­435 dental caries. See dental caries causes of conditions in developing world, costs and cost-effectiveness of development assistance and, 248b 1127­1128 interventions, 45­46, 443­444 diabetes and, 593 company health and wellness programs, interventions and effectiveness of, diarrheal diseases and, 376 838, 838b 440­441, 957 diet contributing to CVD, diabetes, and control of nonoccupational skin diseases and, 709 related conditions, 27, 633­634, 649, exposures, 1133 successful programs to control, 168, 652, 659, 699, 700, 833­836 costs and cost-effectiveness of 171­172 food fortification, 218, 254, 558, 842, interventions, 1135­1139 Onchocerciasis Control Program (OCP), 940­941 disease and injury research, 1143 168, 169, 178, 437, 440, 443, 444, 957 dental caries. See dental caries economic aspects of intervention, Onchocerciasis Elimination Program for folic acid. See folic acid 1135­1139 the Americas (OEPA), 440, 441, 443 food manufacture and. See food fiscal policies to promote, 220­221 Open Society Institute, 987 manufacture global burden of disease from opportunistic infections (OIs) and food subsidies, 212­216, 555, 842 occupational health risks, 1128­1131 HIV/AIDS, 353­355, 353f, 1339 genetically modified crops and, hearing loss and, 958 oral and craniofacial diseases and 131­132 helminth infections and, 470 disorders, 723­736 global school feeding campaign, 1098b implementation, 1139­1142 chronic gingivitis, 729­730 HIV/AIDS and, 352­353 implications for health system chronic periodontitis, 729­730 improvements from agricultural development, 1141­1142 cleft lip and palate, 731 productivity, 190, 220 improving working conditions for, common-risk-factor intervention infants and, 555­556 1131­1133 programs, 731­732, 732t lifestyle and, 835­836 incidence rates of nonfatal injuries, 1139f cost-effectiveness of oral health care, micronutrient supplementation, individual interventions, 1133 733­734, 733f 1058­1059 informal workforce and, 1127 dental caries. See dental caries pregnant women, 510 international interventions for, 1131 developmental disorders, 731 sales tax exemptions on healthy foods and interventions, 1131­1135 fluorosis of teeth, 731 medicines, 220 research and development HIV/AIDS and, 730 school programs. See school health agenda, 1142 noma (cancrum oris), 730­731 programs lessons learned, 1139­1140 oral precancer and cancer, 730 supplementary feeding, 1059 migrant workforce and, 1127­1128 periodontal diseases, 729­730 taxes on unhealthy foods, 219­220 preventive measures, 1135 research and future actions, 732 1384 | Index oral precancer and cancer, 572, 730 implementation of strategies to improve, Pasteur, Louis, 119, 1163 oral rehydration therapy (ORT), 374, 378, 989­991 patent system and drugs, 150­152 1059, 1180, 1303 interventions for pain relief, 982­987 pathophysiology, 122­124 cost-effectiveness of, 379, 383t legal controls on opioid drugs, 986­987 PCD. See Partnership for Child Oregon Health Services Commission, 263 measurement of pain, 981 Development Organisation for Economic Co-operation morphine use, 984­985, 986f, 987b, 1333 PD. See Parkinson's disease and Development (OECD) on pain in patients with cancer and AIDS, 982 pediculosis, 708 development assistance for health, research and development agenda, peer education programs, 344­345, 1115 243, 244 991­992 peer review in medical profession, orphan drug acts, 147 resources available to countries for 1296­1297, 1299 osteoarthritis. See musculoskeletal disorders developing national palliative care pelvic inflammatory disease, 312 osteopathic medicine, 1282, 1288 programs, 991b See also sexually transmitted osteoporosis. See musculoskeletal disorders three-step analgesic ladder, 984, 984f infections (STIs) out-of-school children and youth, 1098, Pan African Initiative, 462 penicillin 1121­1122 Pan American Health See also antibiotics, indiscriminate use of outpatient treatment Organization (PAHO) development of, 122 See also primary care campaign to eliminate polio, 168 resistance to, 122, 1033­1034 TB, quality of care and treatment of, 295 Chagas disease campaign. See Southern penicillin prophylaxis for newborns with outreach and referral programs, 174, 1186, Cone Initiative on Chagas disease sickle cell anemia, 670, 676­677 1194­1195 injury surveillance system of, 1007 performance-based professional recogni- outsourcing of health care services, 93­94, Regional Core Health Data Initiative tion for providers, 1299­1300 93b, 96b, 1320, 1347­1348 of, 1019 performance-based remuneration for quality of care, 188­189 revolving fund for drugs, 1327 providers, 1299, 1348 Safe Water System Initiative of, 216 perinatal conditions, 499­529 Pacific islands. See East Asia and Pacific Panama breastfeeding. See breastfeeding Packard Foundation, 182 scabies in, 710 burden of, 508 PAHO. See Pan American Health universal coverage approach of, 237 Chagas disease and, 434 Organization yellow fever in, 396­397 comparison of alternative intervention pain control. See palliation panic disorder, 47, 611, 612­613 packages, 514­517, 518­519t Pakistan Pap smear, 575 complementary feeding practices diarrheal diseases in, 217, 372, 378 Papua New Guinea and, 376 drug addiction in, 908 lymphatic filariasis in, 439 costs and cost-effectiveness of hemoglobinopathies in, 670 malaria in, 413, 414 interventions, 46, 513­522, 520t interpersonal violence in, 756 newborn survival rates in, 540 cost-effectiveness ratios, 517, 521­522 kidney and urinary system diseases in, quality of health care in, 1293 model assumptions, 525, 525t, 526t 697, 702, 703 skin diseases in, 715, 718 DALYs for, 508 leprosy in, 444 water supply in, 777 defined, 500 neonatal deaths in, 531 Paraguay economic benefits of intervention, 522 polio in, 254b Chagas disease in, 438 epidemiology of, 499­508 refugee populations and surveillance, 1009 out-of-school youth in, 1122 fetal alcohol syndrome, 936b respiratory diseases of children in, 493 paramedical personnel, 1262, 1264­1265 focus conditions and their risk factors, road traffic injuries in, 749b parasites. See helminth infections 501­504t TB in, 291 Parkinson's disease (PD), 632­633 interventions, 508­513, 536­537t PAL. See Practical Approach to Lung Health burden of, 632 lessons learned, 522­524 palliation, 981­993 costs and cost-effectiveness of levels, trends, and differentials of deaths, adequacy of and barriers to pain control interventions, 47, 637, 638t 505­508 in developing countries, 984­985 interventions, 632­633 MDGs on, 499 burden of pain from cancer and AIDS, personal intervention, 632­633 policy considerations and 981­982 prevalence, incidence rate, and approaches, 513 cancer and, 572, 581­582, 983­984t, 1333 mortality, 632 research and development agenda, 524 classification of pain, 982b recommendations, 640 resource use and costs, 517 costs and cost-effectiveness of morphine technological treatment of, 131 retinopathy, 954­955 and drugs, 987­989, 988t Partnership for Child Development (PCD), sexually transmitted infections of mother defined, 59 478, 1095 and, 315 effects of pain, 981 Partnership in Statistics for Development tetanus and, 394 HIV/AIDS and, 351­353, 981­982 in the 21st Century, 1028 transmission of HIV/AIDS, 335­336 Index | 1385 periodontal diseases, 729­730 physicians, 94 research, 109 See also oral and craniofacial diseases and See also medical workers summary of, 70­77t disorders pigmentary skin disorders, 708, surgery, 1248­1249 personal interventions 717­718 using measures other than $/DALY Alzheimer's disease, 629 PLACE (Priorities for Local AIDS Control averted, 83­85t "boutique medicine" resulting from Effort), 324­325, 326 population growth as concern, 104b, genomics, 108 placebo controls and ethics, 268 1075­1076, 1086 combination of high blood pressure, Planet Health program, 838b posttraumatic stress disorder (PTSD), 605, cholesterol, and obesity, Planned Parenthood Association of South 610, 621 856­860, 861t Africa, 203 See also anxiety disorders in developed countries, 859­860 pneumoconiosis. See occupational lung Poverty Reduction Strategy Papers in developing countries, 860­863 diseases (PRSPs), 235­236 CVD, 48, 856­858 pneumonia and HIV/AIDS, 354 poverty reduction support credits (PRSCs), defined, 59, 274b pneumonia and influenza, 484­485, 248, 250­251 hearing loss, 959­960 681­684, 682­683b Practical Approach to Lung Health (PAL), maternal conditions, 510­513 See also respiratory diseases of adults; 295, 691 Parkinson's disease, 632­633 respiratory diseases of children pregnancy. See abortions; family planning; population-based interventions vs., 50 drug resistance and, 1033­1034, 1045 maternal conditions; prenatal care stroke, 634­635 IMCI and, 1180 programs summary of, 60­69t improved quality of treatment for prehospital care, 1264­1271 using measures other than $/DALY children with, 1303 premature births. See childbirth conditions; averted, 78­82t Poland low birthweight pertussis. See diphtheria, pertussis, tetanus cardiovascular disease in, 649, 844b prenatal care programs Peru food price policies in, 219 See also maternal conditions collaborative improvement model tobacco-related legislation in, 168 community health and nutrition in, 1298 policy instruments programs, 1058 contracting for health care services in, 189 See also legislation costs and cost-effectiveness of, 203 household fuel use in, 793, 794f defined, 59 failure of women to participate in, 184 IMCI program in, 1181, 1186, 1187 ensuring use of resources for greatest medical advances in, 123 immunization in, 405b effect, 97­98 routine prenatal care, 511­512, maternal and child mortality in, 1225b polio 514­517 quality of health services in, 1225b burden of, 395 prescriptions. See drugs sexually transmitted infections in, 323 elimination, goal of, 121, 168, 172, 1164, prevention TB control program in, 168­169, 175 1170­1173 See also specific diseases and conditions pesticides, 821, 822 costs of, 1171­1173, 1172t prevention fatigue, 312 pharmaceuticals. See drugs IDA credit buy-downs for, 253, 254b primary. See primary prevention pharmacogenomics, 129, 129t vaccine for, 172, 395, 406, 1164, 1171, research and development in preventive Philippine National Epidemic Surveillance 1325, 1326 medicine, 107­109 System, 1004, 1004t See also vaccine-preventable diseases secondary. See secondary prevention Philippines pollution. See air and water pollution; price increases to deter high-risk childbirth in, 20, 21f environmental risk factors; indoor behaviors. See specific behavior community-based programs in, air pollution (e.g., tobacco use) 1064, 1066 polypill primary care, 1193­1209 diarrheal diseases in, 374, 378 cost-effectiveness analysis of, 275, 598 antiretroviral therapy for HIV/AIDS immunization in, 20, 21f preventing CVD, 598, 860­861, 863t and, 1204 kidney and urinary system diseases in, pooling. See risk pooling community-oriented, 1202, 1204 697, 702 population-based interventions comparison of proposed basic packages of quality of health care in, 1296 alcohol use and control, 896­899, 897t interventions, 1206­1207t surveillance in, 1004, 1004t Alzheimer's disease, 628­629 comprehensive vs. selective care, universal coverage approach of, 237 combination of high blood pressure, 1193­1194, 1195t volcanic eruptions in, 1150 cholesterol, and obesity, cost-effectiveness of interventions, water pollution in, 829 855­856, 859 1199, 1199t phobias, 612 CVD and, 47­48 defined, 1193 See also anxiety disorders defined, 59, 274b district health system and, 1202­1203 physical activity interventions, 652, 835, hearing loss, 959 effectiveness of, 1196­1198, 1198b 838, 841b, 842, 846, 847, 970, maternal conditions, 510 equity goals and, 1198 972­973, 1112 personal interventions vs., 50 expenditures for, 1199­1200, 1200t 1386 | Index "goodness of fit" within health system, public health Plan-Do-Study-Act cycle and, 1298 1195­1196 history of, 120­126 policy interventions to improve, health effects and, 1197 surveillance. See surveillance 1296­1300 home-based care, 1204 Public Health Schools without Walls, 1002 population-level considerations, 1295 human resources for, 1200, 1201b public knowledge. See knowledge process management of health services and, 1343 public-private partnerships for new comparative research on outcomes and maternal and perinatal conditions and, product development, process, 1305 511­513, 515­516t 147­150, 149t defined, 1294 nonhealth effects and, 1197­1198 pyoderma, 708, 709, 710­713, 712­713t, professional retraining for, 1297 nurses and, 620 718, 719t public-private provision of care and, 1297 outreach services of, 1194­1195 referral hospitals and, 1233 outsourcing management of, 93­94, 93b quality-adjusted life years (QALYs), 28, research and development agenda, 1304 priority setting for, 1202 260, 261, 265 social macroeconomic benefits of, private sector resources for, 1200­1202 quality of health care, 1293­1307 1301­1302 public sector resources for, 1204 administrative data and, 1300 standardized patients, use of, relationships with local communities, 1203 assessment and assurance, 96­97 1300­1301 research and development agenda, change of structural conditions and, structure, defined, 1294 1203­1204 1296­1298 TB and, 292­299 scaling up, 1199­1203 chart abstraction and, 1300 total quality management and, 1298 subsidies for, 217 clinical vignettes and, 1301 primary prevention collaborative improvement model race cancer and, 572­574 and, 1298 See also African Americans CVD and, 47­48 cost-effectiveness of improved process, dengue and, 454 research and development, 658 1303­1304 sexually transmitted infections and, 316 defined, 59 definition and framework, 1293­1295, radiation therapy for cancer, 581 interpersonal violence and, 761, 1294f, 1295b raloxifene, 971, 972 763­764, 766 as determinant of cost-effectiveness, 50 rape, 758 Priorities for Local AIDS Control Effort in developing countries, 1295­1296 See also interpersonal violence (PLACE), 324­325, 326 direct observation and recording of referral hospitals, 1229­1243 priorities for research, 5b, 58, 103­118, 126 visits and, 1300 access and equity issues, 1239 See also research and development dissemination of information advice and support to lower levels, 1233 priority setting and cost-effectiveness. on quality variation, 1305­1306 appropriateness of utilization of, See cost-effective interventions on technology, 1297 1238­1239 privacy and personal genetic district hospitals and, 1223­1224, clinical services of, 1231­1233 information, 109 1224f, 1225b context of health system and, 1237 private foundations, role of, 182, 253­255 drugs and vaccines, 1327 cost-effectiveness of, 1241­1242 See also specific foundations economic benefits of, 1301­1302, 1302f current balance of care and, 1237­1239 private sector economic costs of, 1302­1303 defined, 1230, 1230t district hospitals and, 1217 electronic medical record (EMR) determinants of appropriate balance of new product development, 146­147 and, 1300 referral-level care, 1236­1237 primary care and, 1200­1202 emergency medical services, 1272 education and training, 1233­1234, 1234t priority setting and, 38 high volume of care and, 1299 efficiency of, 1240­1242 quality of care and, 1297 internal enabling environment and, 1298 externalities and intangible benefits of, referral hospitals and, 1240­1241 interventions affecting provider practice 1235­1236 spending on health in low-income and, 1298­1300 functions and benefits of, 1230­1236, countries, 231 legal mandates and, 1296­1297 1231­1232t producers low use due to low quality, 89­90 health expenditure, share of, 1237­1238 subsidies for, 218 management of health services and, HIV/AIDS and, 1239 taxes on, 220 1342, 1344 indirect contribution to health product development priorities, 139­155 measurement of, 1300­1301 systems, 1235 See also research and development national and local clinical guidelines, 1297 negative impacts of, 1236 Project Accept, 360b organizational change and, 1297 private vs. public hospitals, 1240­1241 PRSCs. See poverty reduction support credits outcomes quality assurance and quality PRSPs (Poverty Reduction Strategy comparative research on outcomes and improvement, 1233 Papers), 235­236 process, 1305 research and development agenda, 1235 psychiatric disorders. See mental disorders defined, 1294 surgery and, 1249­1250 PTSD. See posttraumatic stress disorder peer review and, 1296­1297, 1299 unit costs of care, 1238, 1238t Index | 1387 referral programs. See outreach and referral emergency medical services, 1272­1274 emergency medical services, 1274 programs ethics and, 114, 265­269 inherent problems in setting, 112­113 refugee populations and surveillance, expansion of global research capacity, 117 need to be in global health agenda, 1008­1009 freedom of scientific inquiry for, 118 115­116 registration of births, deaths, marriages, funding for, 135, 254 participants and decision makers, etc., 533, 1021 gains in health linked to, 160­161 role in, 114­115 regulatory approach, 91­92, 92b, 96, 97, genomics, 107­109 product development, 139­155 176­177, 1131­1132 global health agendas, 106­107, 115 promising topics for global health rehabilitation valuing of, 157­163 agenda, 116 defined, 59 global research collaboration, 111­112, shift of paradigm for, 117 disasters, 1153­1154 112b, 117 systematic and evidence-based learning and developmental disabilities health systems, 99­100, 116 approach to, 113­114 (LDDs), 937, 946­947 hearing loss, 960 push funding, 254 leprosy and, 441 helminth infections, 479­480 quality of health care, 1304 religious customs HIV/AIDS, 114, 360­361, 360b reciprocity in, 110­111 alternative medicine and, 630, 1281 importance of, 135­136 recommendations for, 116­118 contraceptive use and, 1080 incentives needed for, 22 referral hospitals, 1235 reproductive health, 499, 1078­1079 indoor air pollution, 811­812 respiratory diseases of children, 493­494 adolescents and, 1112, 1113t, 1114­1116 infectious diseases, 106 rich country diseases vs. poor country sexually transmitted infections and, information systems, 1028­1029 diseases, 106, 157 315, 316 investment in, 116­117, 244 schistosomiasis, 479­480 research and development, 21­22, learning and developmental school health programs, 1105­1106 107­109 disabilities, 948 shigella vaccine, 378, 383 See also knowledge; technological progress lifestyle changes and chronic diseases, 846 subsidies for, 218 and health gains malaria, 427­428 surgery, 1255­1257 adolescents and, 1122­1123 malnutrition in childhood, 564­565 surveillance, 1012­1013 air pollution, 829­830 maternal and perinatal conditions, 524 TB, 303­304 alcohol use and control, 902­903 medical workers, 1321 10:90 issue, 106, 157 alternative medicine, 1289 molecular epidemiology, 107­109 traditional medicine, 1289 "appropriate science" for developing new product development, 139­155 tropical diseases, 115, 445­446, 446t world, 109­111 development cycles, 140 vaccines, 407­408, 1334­1335, 1334t "appropriate technology" for developing development institutions, 140­141 water pollution, 829­830 world, 111 diagnostics, 144­146 women's conditions, 206­207 "best buys," 113t ethics and, 114, 265­269 zinc deficiency, 564 cancer, 583­584, 1333 financing and institutional respiratory diseases of adults, 681­693 chronic diseases and lifestyle changes, 846 arrangements, 146­150, 254 acute diseases, 45, 681­684, 682­683b, combination of high blood pressure, food fortification, 558 683t, 786 cholesterol, and obesity, 864­865 patent law and, 150­152 avian influenza, 682b community health and nutrition regulatory and liability issues for, chronic respiratory diseases, 684­689, programs, 1071­1072 152­153 685f, 687t contraceptives and family planning, 1087 vaccines, 143­144 cocaine use and, 920 creation of global health architecture, newborn health and, 524, 547 COPD (chronic obstructive pulmonary 117­118 noncommunicable diseases, 115­116 disease) and asthma, 684­689, 685f CVD, 864­865 nonexclusivity and, 103 See also chronic obstructive definition of research, 103 nonrivalry and, 103 pulmonary disease (COPD) and diabetes, 600 occupational health, 1142­1143 asthma diarrheal diseases, 383­384 palliation, 991­992 future research needs, 691­692 disaster relief, 1159­1160 patent law and, 150­152 general approach to lowering risk of, drug dependence, 916 polypill to prevent CVD. See polypill 690­691 drug development, 5b, 141­143, population-based research, 109 HIV/AIDS, 485, 489 1335, 1336t preventive medicine, 107­109 occupational lung diseases, 689­690, 689f drug resistance, 1046 primary care, 1203­1204 pneumonia and influenza, 681­684, economic approaches to valuing, 157­163 priorities for research, 5b, 58, 682­683b, 683t methods to assess value, 161­163 103­118 respiratory diseases of children, economics of already in global health agenda, 115 483­497, 786 diagnostics, 144­146 approaches to, 112­115 acute ear infections, 484 vaccine development, 143­144 biomedical research, 126 acute pharyngitis, 483­484 1388 | Index air pollution and, 820 risk pooling, 226, 227­232, 236­238 See also water supply bronchiolitis, 485 river blindness. See onchocerciasis disaster damage to, 1153 case management of, 487­489, 490t, road traffic injuries. See traffic injuries effect on burden of diseases, 492­493, 494 Rockefeller Foundation 786­791, 1163 causes and burden of, 483­485 funding commitment for disease control epilepsy and, 631 cost-effectiveness of interventions, in developing countries, 147 excreta disposal, 779­784 489­491, 490t, 491t International AIDS Vaccine Initiative calculation of burden of diseases, HIV/AIDS and, 485, 489 (IAVI) support from, 255b 788­789, 788t hypoxemia diagnosis, 488 MDGs and, 182 costs of promotion, 782 influenza, 485 on sanitation and hygiene, 783 diarrheal disease and, 384, interventions, 485­489 stimulating growth of public-private part- 782­783, 782t lessons learned on control strategies, nerships, 147 direct health benefits, 782­784 491­493 vaccine-preventable diseases and, 397 effect on burden of diseases, lower respiratory tract infections on yellow fever eradication, 1163 786­791 (LRIs), 484­485 Roll Back Malaria Partnership, 182, 234, effect on other disease categories, newborns and resuscitation, 539­540, 418, 419b 783­784 539b, 545 Romania hygiene promotion and, 786 pneumonia, 484­485 cardiovascular disease in, 649, 649t levels of service, technologies, and diagnosis of, 487 pain control in, 987b costs, 779­780, 780f intramuscular antibiotics for, 488 Rotary International, 254b, 1171 policy implications, 781­782 oral treatment of, 488 rotavirus, 371, 376­377 reduction in diarrheal disease, 786­791, treatment guidelines for, 488­489 See also diarrheal diseases 786t, 787f, 789t, research and development agenda, rubella 790t, 791t 493­494 See also vaccine-preventable diseases social benefits, 780, 780t school-age children, 1093 congenital rubella, 935b, 937 water supply and, 783 upper respiratory tract infections vaccination, 406 willingness to pay, 781 (URIs), 483­484 vaccine gap and, 1326 helminth infections and, 470, 472 vaccine strategies, 491­492 Russian Federation promotion of hygiene, 784­786 retinopathy of prematurity, 954­955 cardiovascular disease in, 649, 1298 calculation of burden of diseases, rheumatic heart disease (RHD), 647 collaborative improvement model 788­789, 788t See also cardiovascular disease (CVD) in, 1298 costs of, 785 burden of, 650 interpersonal violence in, 766 effect on burden of diseases, 786­791 cost-effectiveness of interventions, 653, milk fluoridation in, 726 effect on diarrheal diseases, 785 656, 657t, 658 sexually transmitted infections in, 318 effect on respiratory infections, epidemiological transition and, 648t Rwanda 785­786 strokes and, 634 government reform and health care evidence and, 784 rheumatoid arthritis. See musculoskeletal delivery in, 91 reduction in diarrheal disease, 786­791, disorders 786t, 787f, 789t, rickets, 965 SAFE (surgery, antibiotics, face washing, 790t, 791t risk factors environmental change) strategy to sustainability of, 784­785 See also specific risks combat trachoma, 168, 170, 174, subsidy for, 217 (e.g., cholesterol levels, obesity) 178, 956 SARS (severe acute respiratory syndrome), behavioral risk factor surveillance system, Safe Motherhood Initiative, 499, 1215 112b, 122, 682­683, 683t defined, 998 Safe Routes to School program, 840b economic effect of, 1010 CVD and, 15, 851­868, 854t Safe Water System Initiative, 216 surveillance and, 997, 1005 development of concept of, 120 St. John's wort, 1283 Saudi Arabia diabetes and, 593, 851 sales tax exemptions on healthy foods and cardiovascular disease in, 650 drug resistance and, 1031­1036 medicines, 220 hemoglobinopathies in, 664 education about, 121 Salmonella, 371 kidney and urinary system diseases genomics and, 108, 108b See also diarrheal diseases and, 697 helminth infections and, 470 salt polio in, 1170 lifelong medical management of, 4b fluoridation, 168, 173 scabies, 708, 709­710, 711t, 719t multiple risks in combination, effect of, iodination, 168, 559 See also skin diseases 109, 851­862 reduction in food content of, 836, Scandinavia musculoskeletal disorders and, 845­846, 846t See also specific countries 966, 966t sanitation improvements, 187, alcohol use and control in, 893 TB and, 304 189­190, 377 dental treatment in, 729 Index | 1389 schistosomiasis on lifestyle necessary for health, 837­838 antibiotic use for, 314, 322 See also helminth infections malaria and, 1092, 1106 bacterial STIs and their sequelae, 314­315 age-associated prevalence of, 470 malnutrition and, 1093­1094, 1106 behavioral risk factors burden of, 471 out-of-school children and, 1098, associated with acquisition and causes and characteristics of, 468 1121­1122 transmission, 317 control programs, 479 pattern of disease as age specific, compensating behavior changes for, cost-effective interventions for, 472, 474 1092, 1092f 317, 322 epidemiology of, 468­470 research and development agenda, for exposure to infected sex partners, kidney and urinary system diseases 1105­1106 316­317 and, 697 sanitation, 783­784 burden of, 319­321 mass drug administration and, 444 sectoral roles DALYs gained from prevention and research and development initiatives, in implementation, 1101­1104 treatment, 319­320, 319t 479­480 policy and economic issues in impact of STIs on HIV, 320­321, 321t water supply and, 776, 779 defining, 1101 circumcision as possible protection schizoaffective disorder, 609 targeting food aid, 1106 against, 320, 322, 360b schizophrenia and nonaffective psychoses Scientific Working Group on Malaria, 427 coinfection with sexually transmitted (NAP), 605, 607­608 Scotland pathogens, 316 See also mental disorders pain control for cancer in, 986b community-level interventions, 322­323 costs and cost-effectiveness of transportation in, 840 condom use for prevention, 312, interventions, 47, 613­619, screening for diseases and conditions 1080­1081 614­618t breast cancer, 577­579, 578t, 579t, 580t control strategies, 203, 323 epidemiology and burden, 608 cancer screening, 123, 573, 574­575, 582 scaling up of, 324­325 genomics and, 131 cervical cancer, 205, 575­577, 577t core groups and bridge populations, interventions, 608 congenital hyperthyroidism, 945­946 313­314 natural history and course, 607­608 diabetes, 594, 598 costs and cost-effectiveness of interven- public support for cost-effective Down syndrome, 941, 942­945, 944t tion, 323­324 intervention package, liver cancer, 574 determinants of costs, 323­324, 324t 621­622, 622t lung cancer, 574­575 demographic and social risk markers, 316 school health programs, 13, 1091­1108 metabolic disorders, 937 drug resistance and, 314, 1036 acute respiratory infections and, 1093 musculoskeletal disorders, 971 duration of infectiousness, 317, 322 burden of disease and, 1094­1095 prenatal screening, 942­946, 944t, 948 effectiveness of interventions, 321­323 cash-transfer programs to increase school sickle cell disease, 675­677 epidemiology and control, 313­314 enrollment, 1101 thalassemias, 673­675 epidemiology and management (since cognitive improvements and, 1100 secondary prevention 1993), 312­313 cost-effectiveness of interventions, cancer, 572, 574­579 gonorrhea, 314 1098­1099, 1099t defined, 59 group-level interventions, 322 dental caries and. See dental caries secondhand smoke, 875­876 HIV/AIDS and, 312, 314 designing effective programs, self-help programs for substance abusers, See also HIV/AIDS 1104­1105 911, 913 income and inequality, 318, 318t, 319f development of programmatic approach, self-targeting, 216 individual-level interventions, 322 1095­1097, 1096t Senegal prevention of acquisition, 322 economic benefits of interventions, contracting for health care services in, prevention of transmission, 322 1099­1101 188, 1297 research and development agenda, FRESH framework. See Focusing HIV/AIDS in, 106, 358 325­326 Resources on Effective School Health mortality decline in, 10 societal determinants, 318 global school feeding campaign, 1098b skin diseases in, 708, 715 syndromic management, 313, 325 for helminth infections, 473, 474­475, sensory deficits. See blindness; hearing loss syphilis, 314 476, 1092, 1094, 1098b sentinel surveillance, 1002, 1021 vaccines for, 322 HIV/AIDS and, 1092­1093, 1093f, 1105 severe acute respiratory syndrome. viral STIs and their sequelae, 315 improved school attendance, benefits of, See SARS shigella, 371, 374, 378 1100­1101 sex distinguished from gender, 196 See also diarrheal diseases infectious disease and, 1092­1093 sexual health, 1087 drug resistance and, 1034­1036, 1045 insurance program, 187 adolescents and, 1112, 1113t, vaccine development needed for, interventions, 1095­1098, 1098b, 1099t 1114­1116 378, 383 key stakeholders in implementation, 1104, sexually transmitted infections (STIs), sick leave from work, 221 1105t 311­330 sickle cell anemia and related diseases lessons learned, 1101­1105 adolescents and, 1112 See also hemoglobinopathies 1390 | Index burden of, 667­669, 669t sleeping sickness. See African respiratory diseases of adults in, 691 clinical features of, 664 trypanosomiasis sales tax exemptions on healthy foods costs and cost-effectiveness of smallpox eradication, 121, 165, 168, 172, in, 220 interventions, 674t, 675­677, 676t 1163, 1169­1170, 1169t, 1170t sexually transmitted infections in, global distribution and frequency of, smoking. See tobacco use and control 323, 325 665­667, 665f, 666t, 667f smuggling of tobacco products, 877 tobacco-related legislation in, 168 interventions, 669­672, 677­678 soap subsidy, 217 South America. See Latin America and neonatal screening and prophylaxis, social anxiety disorder, 611, 612 Caribbean 675­677 See also anxiety disorders South Asia population genetics and dynamics, social epidemiology See also specific countries 664­665 research and development agenda, 109 alcohol consumption in, 898 Sierra Leone social marketing breastfeeding in, 539 development assistance for health to, 253 adolescents and, 1116, 1122 cardiovascular disease in, 649, 650, 652, neonatal deaths in, 531 of contraceptives, 1081 654, 655t, 656t Significant Caries (SiC) Index, 724, 724t Social Marketing for Adolescent Sexual cost-effective interventions for, 50, 51t silicosis. See occupational lung diseases Health Project, 1122 neglected low-cost opportunities and Sindh Instate of Urology and social security reforms, 237­238, 237t high-cost interventions in, 54­57t Transplantation, 703 soda consumption and obesity, 835, 836 dengue in, 454, 455, 457 Singapore soil-transmitted helminth infections (STH development assistance for health Fit and Trim Program, 842 infections). See helminth infections to, 245 hemoglobinopathies in, 670 solidarity funds, 228 diet in, 836 Kaposi's sarcoma in, 730 Somalia epilepsy in, 631 National Healthy Lifestyle Program, 843b African trypanosomiasis in, 458, 460 food manufacture in, 845 road traffic injuries in, 744 refugee populations and hygiene promotion in, 784­785 transportation choices in, 839 surveillance, 1009 leishmaniasis in, 455 Single Convention on Narcotic Drugs somatic cell nuclear transfer, 130 malaria in, 414, 420 (1961), 986 South Africa maternal and perinatal conditions in, skin diseases, 707­721 African trypanosomiasis in, 458, 460 520t, 521t See also leprosy cardiovascular disease in, 650, 656 mental disorders in, 614, 618, 619, 622 acne, 708 cervical cancer screening in, 205 neonatal deaths in, 531 bacterial skin infections, 709, 710­713, combination of high blood pressure, neurological diseases in, 628, 631 712­713t cholesterol, and obesity in, oral cancer in, 572 Buruli ulcers, 709 853­854, 862t performance-based contracts with NGOs current status of current community Community Health Intervention in, 253 control measures in dermatology, 720 Programme, 842 proposed guidelines for HIV/AIDS in, economic assessments and diseases in Coronary Risk Factor Study in, 843b 348­349 developing countries, 711­713t, dental treatment in, 729 road traffic injuries in, 746 716­717t, 718­720, 719t drunk-driving enforcement in, 901­902 strokes in, 634, 640 eczema or dermatitis, 708 excreta disposal in, 782 syphilis in, 314 effective therapies, 709, 709t HIV/AIDS treatment in, 376 tobacco use and control in, 880 endemic treponematoses, 709 hospitals in, 1218, 1231, 1296, 1347 vaccine-preventable diseases in, 397, 400, filarial lymphoedema, 709 interpersonal violence in, 760, 761, 402, 408 fungal infections, 713­715 765­766 vital events registration in, 1021 HIV-related, 707, 708, 709, 717 Kaposi's sarcoma in, 730 women in, 195 nonfilarial lymphoderma, 709 kidney and urinary system diseases Southeast Asia onchocerciasis and, 709 in, 703 See also specific countries patterns at community level, 708­709 malaria in, 420 avian influenza in, 682b pediculosis, 708 outsourcing of health care services ecstasy (MDMA) use in, 924 pigmentary disorders of, 708, 717­718 in, 93b hemoglobinopathies in, 665, 666 prevalence of, 708 Perinatal Education Programme mental disorders in, 620­621 pyoderma, 708, 709, 710­713, 712­713t, in, 545 rheumatic heart disease in, 647 718, 719t Perinatal Problem Identification strokes in, 633 scabies, 708, 709­710, 711t, 719t Programme in, 546b Southern Cone Initiative on Chagas Tinea capitis, 714­715, 716t, 718, 719t poisonings in, 747­748 disease, 167, 170, 177, 438 Tinea imbricata (Tokelau ringworm), 715, primary care in, 1201 Spain 717t public view that public clinics dilute drug-related deaths in, 909 tropical ulcers, 709, 715, 717, 718 drugs, 98 malaria in, 421 Index | 1391 Special Programme for Research and costs and cost-effectiveness of school health programs in, 1101 Training in Tropical Diseases interventions, 559­561, 560t skin diseases in, 707, 709 (WHO), 114, 115, 145, 148t, DALYs lost due to in children under age smallpox in, 1163 423, 446 five, 552, 553t strokes in, 634 spina bifida, 939­940 deaths due to in children under age five, syphilis in, 314 Sri Lanka 552, 552t TB in, 302 hemoglobinopathies in, 675 economic benefits of interventions, tobacco use and control in, 880 malaria in, 414, 418, 421 562­563 tsetse flies in, 454 neonatal death reduction in, 533, 547b interventions, 555­559 vaccine-preventable diseases in, 395, 397, prenatal care and childbirth in, 168, iodine deficiency, 554, 558­559 399, 400, 402, 408 169, 174 iron deficiency, 553­554, 558 vaccines and drugs, access to, 1346 primary care in, 1202 lessons learned, 563­564 zinc deficiency in, 554 water supply in, 774 research and development agenda, subsidies for health and health-related STATCAP, 1028 564­565 products, 212­218, 212t, 213­214t statistics, use of, 121, 1021­1022 school-age children and, 1093, 1093f consumer subsidies, 212­218 See also information systems vitamin A deficiency, 553, 556­558 diet, 843 stem cell therapy, 130 zinc deficiency, 554, 559 food subsidies for young children, 555 sterilization, 1080, 1082 Sub-Saharan Africa producer subsidies, 218 stewardship, 91­92, 1122, 1346, 1348 See also specific countries substance abuse. See alcohol use and STH (soil-transmitted helminth) African trypanosomiasis control in, control; drug dependence infections. See helminth infections 454, 456 Sudan stillbirths. See child mortality breastfeeding in, 539 African trypanosomiasis in, 456, 460 stomach cancer, 569, 570, 573, 574 cardiovascular disease in, 647, 649, 650, guinea worm eradication in, 169 stones and obstructive uropathy of 654, 655t, 656t leishmaniasis in, 453, 459 kidneys, 697­698 contraceptive use in, 1076, 1080 polio in, 1170­1171 See also kidney and urinary system cost-effective interventions for, 50, 52t refugee populations and diseases neglected low-cost opportunities and surveillance, 1009 Stop TB Partnership, 182, 289 high-cost interventions in, 54­57t suicide, 16, 607, 758 stoves, use of. See indoor air pollution diabetes in, 1333 See also interpersonal violence strategic management of clinical services, diarrheal diseases in, 373 adolescents and, 1111, 1121 1339­1352 drug treatment in, 1346 costs and cost-effectiveness of interven- strategic purchasing, 1348­1349, 1349f epilepsy in, 631 tions, 48 Strategies for Enhancing Access to guinea worm eradication in, 167, 169, 176 sulfadoxine-pyrimethamine, 45, 92 Medicines Project, 92b hemoglobinopathies in, 664, 665, 678 resistance to, 1032, 1036 stroke, 633­635 HIV/AIDS in, 106, 1070­1071 sulfonamides, 122 See also cardiovascular disease (CVD); hospital-acquired diseases in, 1224 surgery, 1245­1259 neurological disorders hospitals in, 1221b burden of surgical conditions, 1245­1248, acute management and secondary indoor air pollution in, 803 1247t prevention of, 48 influenza in, 684 by condition and region, 1248t burden of, 634 interpersonal violence in, 757 methods of, 1246 cost-effectiveness of interventions in iron deficiency in, 553­554 for cancer, 580­581, 584 developing countries, 637, 639t, 640 leishmaniasis in, 455 for cardiovascular disease, 651, 654, 658, frequency of types, prevalence, incidence malaria in, 413, 414, 415, 417­418, 421, 857­858 rates, mortality, and disability after 423, 426, 935b cataract surgery, 956 stroke, 633 maternal and perinatal conditions in, causes of surgical conditions, 1245­1248 interventions, 634­635, 635 520t, 521t, 522 community clinics and, 1249, 1253­1255, ischemic hear disease (IHD) and, 646 measles in, 396 1253t, 1254t, 1256f personal interventions, 634­635 mental disorders in, 614, 618, 619, coordinated model system for, 1251 recommendations, 640 621­622 costs and cost-effectiveness of risk factors, 633­634 neonatal deaths in, 531, 533 interventions, 49­50, 1251­1255, stunting, wasting, and micronutrient neurological diseases in, 631 1256, 1257b deficiency disorders, 551­567 noma (Cancrum Oris) in, 730­731 cost assumptions, 1253t adolescents and, 1111 poisonings in, 747­748 method for estimating costs and causes and burden of, 551­554, 552t proposed guidelines for HIV/AIDS DALYs, 1252 community-based health interventions, in, 349 at district hospitals, 5b, 27, 1249, 1068 provider autonomy in, 1348 1253­1255, 1253t, 1254t, 1256f 1392 | Index for epilepsy, 631, 637 informal networks as elements of, deworming programs in, 478 for hearing improvement, 959 1004­1005 DOTS development in, 111 high volume of care and quality injury surveillance, 1007­1008 drug shops, regulation in, 92b of, 1299 Integrated Disease Surveillance and HIV/AIDS in, 320, 331, 352 hip replacement surgery, 973 Response (IDSR), 1003­1004 hospitals in, 1218 human resources and, 1256 integrated surveillance, defined, 998 IMCI program in, 1181­1183, 1183b, interventions, 1248­1251 laboratory-based surveillance, 1184, 1184f, 1185b, 1186, 1187 for kidney disease, 701­702 1002­1003, 1003f kidney and urinary system diseases knee replacement surgery, 973 objectives of, 998­999, 999b in, 702 minimally invasive, 130­131 occupational health, 1133, 1143 leprosy in, 437 for musculoskeletal disorders, 968 as part of national public health lymphatic filariasis in, 440 for Parkinson's disease, 632­633, 637 systems, 1000 malaria in, 421, 425 population-based interventions, passive surveillance, defined, 998, 1021 malnutrition in, 1068 1248­1249 population-based surveys used management of health care in, research and development agenda, in, 1002 1345, 1348 1255­1257 principles and uses of, 999, 999t, 1021 mental disorders in, 620 resource requirements for services and refugee populations and, 1008­1009 Multi-Country Evaluation of IMCI level of care, 1250­1251t research agenda for, 1012­1013 Effectiveness, Cost, and Impact SAFE strategy to combat trachoma. routine health information system, (MCE) in, 1181­1183, 1183b See SAFE (surgery, antibiotics, face defined, 998 oral health and education program washing, environmental change) SARS and, 1005 in, 726 strategy to combat trachoma sentinel surveillance, 1002, 1021 primary care in, 1343, 1348 single-shot surgical services, successful strategies for, 1002­1005 regulation of health care in, 1296 programs involving, 174 syndromic surveillance, defined, road traffic injuries in, 749b at tertiary hospitals, 1249­1250 998, 1008 sexually transmitted infections in, surveillance, 997­1015 as tool to improve public health, 320, 322 active surveillance, defined, 998, 1021 1001­1002 skin diseases in, 708, 720 analysis and dissemination of surveillance training for, 117 strokes in, 633 data, 999­1000 SWaps (finance instruments), 234­235 training of medical staff on acute avian influenza and, 1005­1006 Sweden respiratory infections of children as basis for evidence-based development assistance for health in, 1297 decisions, 1001 from, 244 vital events registration in, 1021 behavioral risk factor surveillance system, diabetes and lifestyle changes in, 593 water supply in, 773 defined, 998 fluoridation programs in, 729 Tanzania Essential Health Intervention biologic terrorism and, 998, 1008 oral health programs in, 729, 733 Program (TEHIP), 1024, categorical surveillance, defined, 998 oral precancer and cancer in, 730 1026­1027, 1026b, 1202, 1208n2 chronic disease and, 1009­1010 Switzerland targeting of interventions. See specific disaster situations and, 1008, 1154 iodine fortification program in, 554 diseases and conditions drug resistance and, 1042­1043 salt fluoridation in, 725­726 taxation Ebola and, 1006 synovial fluid replacement, 973 on alcohol, 218, 893, 895t economics of, 1010­1011, 1010t syphilis. See sexually transmitted consumer taxes, 218­220, 219b emergency surveillance, 1008 infections (STIs) discouraging high-risk behaviors by, 177, environmental public health surveillance, 213­215t 1006­1007, 1006f T. vaginalis, 315 on drugs, 1327 eradication of infectious diseases tacit knowledge, 110 as method of revenue collection, and, 1166 Taiwan 227, 229 establishing and maintaining system for, arsenic contamination of water producer taxes, 220 999, 1000b in, 821 promoting health policies, 213­215t, field epidemiologists providing evidence, strokes in, 633 218­220 1001­1002 universal coverage approach of, 237 sales tax exemptions on healthy foods and future of, 1011 tamoxifen, 574, 580 medicines, 220 global surveillance networks, 1011­1012, Tanzania on tobacco, 177, 218, 874­875, 875f, 1012f adolescents in, 1116 879f, 882 health information and management African trypanosomiasis in, 458 on unhealthy foods, 219­220 system, defined, 998 community-based programs in, TDR. See Special Programme for Research indicator, defined, 997 1066, 1068 and Training in Tropical Diseases Index | 1393 technological progress and health gains, oral and craniofacial diseases and reducing noncommunicable disease, 10, 119­137 disorders in, 723 4b, 27 See also genomics road traffic injuries in, 744 reducing supply of tobacco, 877 "appropriate technology" for developing sexual risk behaviors in, 312 kidney disease and, 699, 700 world, 111 sexually transmitted infections in, 323 musculoskeletal disorders and, 966 array technology, 128 skin diseases in, 708 nicotine addiction, 873­874 community health and nutrition tobacco taxes in, 882 nicotine replacement therapies, programs and, 1063 vaccine-preventable diseases in, 406 876­877, 878 cost-effective interventions and, 53 water pollution in, 829 oral precancer and cancer and, 730 in developing countries, 133 Thalassemia International Federation, 678 periodontal disease and, 729 dissemination of information on, 1297 thalassemias, 127f price increases and, 878, 879, 882, 1116 high-technology hospital practice, 125, See also hemoglobinopathies respiratory diseases of adults, 685 1223, 1226 bone marrow transplantation, 675 smoking trends, 869­870, 870t history of 20th century for science, burden of, 667­669 smuggling of tobacco products, 877 technology, and medicine, 120­126 clinical features of, 664 strokes and, 633, 634 immunization technologies, 1335 control and treatment, 669­672 taxation to discourage use, 177, 218, information and communication cost and effectiveness of diagnosis and 874­875, 875f, 879f, 882 technologies (ICT), 1022 management, 673­675, 674t, 675t vision loss and, 954 newborn care, 547 global distribution and frequency of, warnings about, 168, 876 occupational health and, 1143 665­667, 666f, 666t, 667f workers' risk for asbestos and smoking potential of information technology, 117 options for control and management exposure, 1128, 1128t product development priorities, 139­155 of, 677 Togo, malaria in, 421 vaccine technologies, 1334 population genetics and dynamics, total quality management in health teeth. See dental caries; oral and craniofacial 664­665 care, 1298 diseases and disorders screening and counseling, 673­675 tourism TEHIP. See Tanzania Essential Health Tinea capitis, 714­715, 716t, 718, 719t dengue and, 455 Intervention Program See also skin diseases leishmaniasis and, 453 television watching, 835, 838b Tinea imbricata (Tokelau ringworm), malaria and, 426 terrorism, 104b 715, 717t tax rate reduction on alcohol in Mauritius attacks as disasters, 1161n1 See also skin diseases and, 900 biological, 107, 998, 1008 tobacco-related deaths, 15f, 120, 869, trachoma tertiary hospitals. See referral hospitals 870­871, 871t, 872f, 878­880, control program, 168, 170, 174, 178, tetanus. See diphtheria, pertussis, tetanus 879t, 882 956­957 Thailand tobacco use and control, 869­885 sanitation, prevention by, 784 avian influenza in, 682b, 1005­1006 adolescents and, 1111, 1116 water supply and, 779 child mortality in, 1068 advertising restrictions, 876 Trade-Related Aspects of Intellectual community-based programs in, 1059f, bans on smoking in public places, Property Rights. See TRIPS 1065, 1066, 1067 875­876 (Agreement on Trade-Related condom program in, 167, 170, 176 cancer and, 573, 833, 871­872 Aspects of Intellectual Property dengue in, 452, 455, 457, 459, 461 cardiovascular disease and, 647, 652, Rights) development assistance and health 659, 833 traditional medicine. See complementary policy, 248b cessation, 870, 872­873, 873f, 876­877 and alternative medicine (CAM) drug dependence in, 916 chronic diseases and, 833 traffic injuries, 16 epilepsy in, 630 cleft lip and palate and, 731 adolescents and, 1111 fall-related injuries in, 748­749 comprehensive control programs, 881 alcohol consumption as cause of, hemoglobinopathies in, 668, 670, 673, COPD and, 688, 690, 691­692 893, 894t 674­675 costs and cost-effectiveness of bicycle helmet legislation and HIV/AIDS in, 13, 106, 167, 348b, 352 interventions, 49, 877­880, 878t, enforcement, 747 incentive pay to doctors in, 95, 95b 880t, 881t burden of, 107 Kaposi's sarcoma in, 730 deaths. See tobacco-related deaths causes of, 739, 739t kidney and urinary system diseases in, diabetes and, 597 costs and cost-effectiveness of 697, 702, 703 health consequences of smoking, interventions, 48 medical workers in, 1315 870­873, 871t economic burden of, 737, 739 milk fluoridation in, 726 interventions implementation of prevention and nutrition investment in, 248b government interventions, 873­874 control strategies, 748­749, 749b occupational health in, 1132 reducing demand for tobacco, 874­877 interventions, 742­744, 742t 1394 | Index motorcycle helmet legislation and DALY estimates for, 437­438, 438t outpatient and community-based enforcement, 747 interventions and effectiveness of, 45­46, treatment, 295 risk factors for, 739­740 438­442, 460­463 private sector treatment, 295 safer roads and, 743­744 kidney disease and, 697 quality of care and treatment of, 27, speed and safety regulations, 745­746, 746t management and control strategies of, 292­299 speed bumps, 746­747, 747t 457­458, 463t research and development agenda, training of medical personnel research needs and priorities, 115, 303­304 alternative medicine and, 1286, 1287 445­446, 446t risk factors, 304 changes needed in medical education for skin diseases, 709 service delivery, 304 global perspective, 134­135 See also skin diseases silicosis and, 689 for childbirth and newborn care, 545 tropical ulcers, 709, 715, 717, 718 successful programs in controlling, 94­95, community workers, 1063 See also skin diseases 168­169, 174, 175 continuing skill development, 1312, 1344 tsetse flies, 453, 456, 458, 462, 464 targeting poor in allocation of health dissemination and use of health informa- tsunamis, 1150 resources to fight, 5b tion for, 1023 tubal ligation, 1080 tobacco use and, 289, 304 district hospitals' role in, 1215­1216 tuberculosis (TB), 289­309 vaccination, 292, 304 epidemiologists, 1002 alternative approaches to diagnosis and See also vaccine-preventable diseases for hospital emergency workers, treatment, 294­295 Turkey 1271, 1273b burden and trends, 291­292 ambulances in, 1268 nurses, 94 averted and avertable burden, 302, condom program in, 216 for occupational health, 1134­1135, 1142 302f, 303f epilepsy in, 630 professional retraining to improve quality children and, 293 kidney and urinary system diseases in, of care, 1297 complementary strategies, 304 698, 702, 703 for surgery, 1256­1257 cost-effectiveness of interventions, leishmaniasis in, 459 for surveillance purposes, 117 40­43, 295­297, 296t, 297t, mental disorders in, 621 traditional healers, 1286 301­303, 301b universal coverage approach of, 237 transplant surgery, 130, 131 managing endemic TB, 297­299, 298f typhoid, 775t, 776t, 777, 778 transportation managing TB outbreaks, 299­301, 300f See also traffic injuries cost per case prevented, 297­299 Uganda automobile use, 839, 843 cost per death prevented and DALY African trypanosomiasis in, 454, 456, 460 costs of unreliable systems, 187 gained, 299 Ebola in, 1006 for emergency medical services, 1264, deaths due to in low- and middle-income government reform and health care 1267­1271 countries, 290­291 delivery in, 91, 247, 1315 See also ambulances cost per death prevented and DALY HIV/AIDS in, 13, 106, 320, 348b, 354, 358 intersectoral action for improvements gained, 299 Hospice Uganda, 990 in, 189 development assistance and, 244 IMCI program in, 1181, 1185b, policy and environmental design, diagnostics development for, 144, 1186, 1187 839­840 145, 145t interpersonal violence in, 756, 757, 763 walking and bicycle use, 839­840, 840b, DOTS strategy for treatment. See directly malaria in, 418, 421 843­844 observed therapy short course medical workers in, 94, 1312 trauma care. See emergency medical services (DOTS) primary care and removal of fees for Trichomonas vaginalis, 315 drug resistance and, 294, 304­305, 1036, health services in, 97b Trinidad and Tobago 1045­1046 road traffic injuries in, 749b health service agency reform in, 93, drugs for, 304, 1326 sexually transmitted infections in, 1318­1319 economic benefits of interventions, 312, 320 TRIPS (Agreement on Trade-Related 302­303 training of medical personnel in, 94 Aspects of Intellectual Property HIV/AIDS and, 291, 292­293, 294, 301, water supply in, 773 Rights), 150­152, 351, 1324 302, 304­305 U.K. National Institute for Clinical tropical diseases, 433­466 impact and targets, 304 Excellence, 1297 See also specific types infection, disease, and death, 290­291 UNAIDS. See United Nations Programme burden of, 436­438, 454­456 integrated management of, 295 on HIV/AIDS characteristics and transmission, 433­436, interventions, 292­294 UNESCO (United Nations Education, 451­454, 454t kidney and urinary system diseases Scientific, and Cultural costs and cost-effectiveness of and, 697 Organization), 1095 interventions, 442­445, 445t, latent infection, treatment of, 292­293 UNICEF. See United Nations Children's 459­460 MDGs and, 290, 305 Fund Index | 1395 unintentional injuries, 737­753 neonatal deaths in, 533 cardiovascular disease in, 647, 650, burden of, 737, 738f, 738t organizational structure of health systems 656, 658 burn-related injuries in, 92 cigarette smoking in, 875, 876, 881 indoor fires and, 797­799 Public Health Act (1848), 120 cleft lip and palate in, 731 interventions, 743t, 744­745 skin diseases in, 714 cocaine use in, 920 risk factors for, 741 strokes in, 633 combination of high blood pressure, causes of, 739­741, 739t universal coverage approach of, 237 cholesterol, and obesity in, 859 costs, cost-effectiveness, and economic United Nations COPD and asthma in, 686 benefits of interventions, 745­748 development assistance for health dengue in, 455 driving injuries. See traffic injuries from, 244 development assistance for health from, drowning International Conference on Population 244, 254 interventions, 743t, 745 and Development (1994), 196 diabetes in, 593, 598 risk factors for, 741 Millennium Development Goals. See diarrheal diseases in, 377 economic burden of, 737, 739 Millennium Development Goals drug development and approval fall-related injuries (MDGs) process, 153 interventions, 743t, 744, 748­749 Millennium Summit (2001), 182 See also U.S. Food and Drug risk factors for, 740­741 United Nations Children's Fund (UNICEF) Administration future challenges for, 750 Alma Ata Declaration. See Alma Ata ecstasy (MDMA) use in, 924 implementation of prevention and Declaration on Primary Health Care failure of "war on cancer" in, 112 control strategies, 748­749, 749b (1978) fluoridation programs in, 725, interventions, 742­743t, 742­745 Child-Friendly Schools framework of, 726, 729 occupational injuries, 1130, 1143 1095­1096 food fortification in, 558, 842 See also occupational health on development assistance for health, 243 hemoglobinopathies in, 664, 670, poisonings on diarrheal diseases, 374, 378 675, 676 childproof paraffin containers, on drug procurement, 1326 heroin use in, 908 747­748 guinea worm control program, 167 HIV/AIDS in, 354 interventions, 743t, 744 IMCI program. See integrated homeopathic products, use in, 1282 risk factors for, 740 management of childhood illicit drug use in, 908, 911 research and development agenda, illness (IMCI) interpersonal violence in, 764, 749­750 "marginal budgeting for bottlenecks" tool 765, 766 United Kingdom of, 541 iodine fortification program in, 554 air pollution causing deaths in, 819 on nutrition and child survival, 1057, kidney and urinary system diseases in, alternative medicine in, 1285 1057f, 1070 695­696, 698­699, 701, 702 capital charging and management of on soldiers under age of 18, 1111 occupational lung diseases in, 690 physical resources in, 1345 on vaccine management by Occupational Safety and Health Act cigarette smoking in, 872­873, 873f, 875, countries, 1329 (1970), 1134 876, 881 on vaccine-preventable diseases, 394 oral and craniofacial diseases and class differences in morbidity and vaccines from, 1327 disorders in, 723 mortality, 124 on water supply, 773­774 oral health programs in, 726, 729, 733 COPD and asthma in, 686 United Nations Development Programme, periodontal disease in, 730 dental treatment in, 729 247, 811, 1095 preterm birth in, 947 development assistance for health United Nations Education, Scientific, primary care in, 1197 from, 244 and Cultural Organization sexual risk behaviors in, 312 epilepsy in, 630 (UNESCO), 1095 smallpox eradication efforts of, fluoridation programs in, 725, 726 United Nations Foundation, 254b 1170, 1170t Health and Safety Executive, 1135 United Nations Industrial Development urbanization hemoglobinopathies in, 668 Organization, 825b, 1328 air pollution and, 819­820 heroin use in, 908, 911 United Nations Population Fund, lifestyle and, 840, 841b high-technology hospital practice in, 125 243, 1095 urinary system diseases. See kidney and homeopathic products, use in, 1282 United Nations Programme on HIV/AIDS urinary system diseases hospital throughput in, 125, 130 (UNAIDS), 182, 244, 245, 332, 333, Uruguay hospitals in, 1224 335, 347, 989 Chagas disease in, 436, 438, 442 International Finance Facility (IFF) United States U.S. Agency for International proposal by, 234 air pollution in, 828, 829t Development, 1297 London tax on cars driven into, 839 alternative medicine in, 1282, 1283, 1287 U.S. Agency of Healthcare Research and milk fluoridation in, 726 arsenic contamination of water in, 821 Quality, 1297 1396 | Index U.S. Centers for Disease Control and increasing coverage for traditional EPI, TB in, 295 Prevention 401­402, 401f, 401t, 404b, 405b traditional medicine in, 1285 on cost of drug resistance, 1037 new antigens, 402­406, 403t Vigisus project in Brazil, 1021 on drinking water quality, 216 distribution and storage, 1328­1329 violence free software for information from, 1022 economics of discovery and development, See also interpersonal violence guinea worm control program, 167 143­144 adolescents and, 1111 on measles elimination, 1164 financing of, 1331, 1332t costs and cost-effectiveness of surveillance systems of, 1000 future challenges for, 1331­1334 interventions, 48 training of epidemiologists, 1002 immunization technologies, 1335 Violence against Women Act on violence, 755, 760 inequities in access to (vaccine gap), 1326 (U.S 1994), 764 U.S. Department of Health and Human management of, 1329 viruses, control of, 122 Services, 346 policies on, 1324 See also specific type of virus U.S. Environmental Protection procurement of, 1327 vision loss. See blindness Agency, 828 quality assurance of, 1327 vital events registration, 1021 U.S. Food and Drug Administration, research and development agenda, neonatal death information from, 533 141­142, 152­153, 268, 841, 407­408, 1334­1335, 1334t vitamin A deficiency 1287, 1327 selection of, 1325­1326 blindness due to, 553 U.S. President's Emergency Plan for AIDS successful programs of, 172­173 childhood illness and mortality associated Relief, 332, 1333 WHO prequalification process for, with, 553 U.S. Public Health Service Panel on 1327­1328 costs and cost-effectiveness of cost-effectiveness, 262 value of a statistical life (VSL), 8, 9, 280 interventions, 561 user fees, 19, 97b, 187, 227, 1330 value of information (VOI), 161­163, 162b diarrheal diseases and, 376 valvulitis, 647 IMCI and, 1180 Vaccine Fund, 397, 491, 492 See also rheumatic heart disease interventions, 556­558 vaccine-preventable diseases, 389­411 vasectomy, 1080 lessons learned, 563, 565 See also vaccines; specific diseases and vector control mass drug administration and, 444 conditions African trypanosomiasis and, 458, pregnant women and, 510 burden of, 394, 397 462­463 school-age children and, 1094 DALYs lost from, 397, 399t Chagas disease and, 438, 442 vitamin B12 deficiency, 627 deaths averted and avertable dengue and, 457, 461 vitamin D supplementation, 967, 971­972 burden, 398t leishmaniasis and, 458, 462 VOI. See value of information causes and epidemiology of, 389­397, lymphatic filariasis and, 439­440 volcanic eruptions, 1150 390­393t malaria and, 421­422 voluntary counseling and testing (VCT) Expanded Program on Immunization. See onchocerciasis and, 440 for HIV/AIDS, 332, 336, 344, 347, 348, 349 Expanded Program on Venezuela volunteer community-based Immunization (EPI) Chagas disease in, 438 programs, 1064 financial sustainability of immunization dengue in, 455 VSL. See value of a statistical life programs, 406­407 food fortification in, 558 inequities in access to vaccines (vaccine verbal autopsies, 1021 war gap), 1326 vertical approach to health systems, 88, 174 burden of injuries from, 107 influenza, 684, 690, 692 development assistance and, 234 costs and cost-effectiveness of research and development agenda, management of health services and, interventions, 48 407­408, 1334­1335, 1334t 1346­1347 soldiers under age of 18, 1111 respiratory diseases of children, 485­487 primary care and, 1194 Warmi Project in Bolivia, 535 vaccines, 1323­1337 Viagra, 1327 wasting. See stunting, wasting, and See also vaccine-preventable diseases; Vietnam micronutrient deficiency disorders specific diseases avian influenza in, 682­683b water supply, 771­779 childhood diseases, 44­45 cervical cancer screening in, 205 calculation of burden of diseases, respiratory diseases, 485­487, 491­492 cholera vaccine in, 377 788­789, 788t community health and nutrition contraception, access to, 1086 classification and burden of water-related programs, 1059 dengue in, 452, 454, 457 diseases, 775­776, 775t, 776t cost-effectiveness analysis of, 44, diarrheal diseases in, 377 conditions for health effect, 777­778 276­277b health care reform in, 247 dental health. See dental caries costs and cost-effectiveness of programs malaria in, 418, 420, 421, 425 diarrheal disease and, 377, 776, 777, 777t, existing programs, 397­400, 400t sexually transmitted infections in, 323 786­791, 786t, 787f, 788­791, 789t, improving programs, 407 silicosis prevention measures in, 1139 790t, 791t Index | 1397 water supply (Continued) community health and nutrition polio eradication efforts and, 254b direct health effects, 775­779, 775t programs for, 1058 Poverty Reduction Strategy Papers disaster damage to, 1153, 1156 contraception. See contraception (PRSPs), 235­236 effect on other disease categories, cost-effectiveness of interventions, sector adjustment loans from, 251 778­779, 778t 203­206, 204­205t on skin diseases, 709 epidemiological questions and problems, conditions affecting women on surveillance, 1000 776­777 disproportionately, 205­206 on user fees, 227 excreta disposal and, 384, 783 conditions specific to women, 203­205 on water supply, 773 helminth infections and, 470 death in childbirth. See maternal World Bank Development Indicators, 379 hygiene promotion and, 786 mortality World Conference on Education for All levels of services and costs, 771­773, 772f depression and, 107, 196, 201, 202, 206 (1990), 1095 pollution. See air and water pollution discrimination against as issue, 104b World Development Report 1993: Investing purification subsidy programs, 216 gender-based conditions, 202­203 in Health (World Bank), 13­14, 28, quality of, 216, 771, 774 helminth infections and, 474 251, 271, 1026b, 1180, 1194, 1197 quantity of, 772 HIV/AIDS and, 195, 196 World Development Report 2004: Making reasonable access to, 187, 772 indoor air pollution and, 797, 803 Services Work for Poor People time-saving benefit, 773­775 inequity in health services for. See gender (World Bank), 171, 179n2 WaterAid, 782 differences World Food Program, 352 weak health systems, challenge of, 99, 1340 leishmaniasis and, 455 World Health Organization (WHO) weight. See low birthweight; obesity life expectancy of, 5, 5f, 197, 201­202 See also names of programs starting with Weil, Andrew, 1287 malaria and, 413, 414, 417, 418, 422 "WHO" Wellcome Trust, 134 milestones and influential works in Ad Hoc Committee on Research Relating WERC (Village Education Resource international women's health to Future Intervention Options, 106, Center), 782 (1980­2003), 208­209t 111, 113, 114 West Africa pregnancy. See maternal conditions; on adolescent burden of disease, 1110 lymphatic filariasis in, 436 prenatal care programs on African trypanosomiasis, 451, 453, onchocerciasis in, 172, 178, 435, 437, research priorities for conditions creating 454, 456, 458, 459, 462 440, 443 excess burden for, 206­207 on air pollution, 817 yellow fever in, 396­397 risks associated with unwanted on alcohol-use disorders, 902 Western Pacific pregnancies, 1077­1078, 1078t Alma Ata Declaration. See Alma Ata diarrheal diseases in, 373 sexually transmitted infections and, 315 Declaration on Primary Health Care neonatal deaths in, 533 underutilization of health services by, 197 (1978) rheumatic heart disease in, 647 vision impairment and, 955 on antimicrobial supplements in animal What Works Working Group, 117, 165 workers' health. See occupational health feed, 1041 WHO. See World Health Organization workplace violence, 760 on artemisinin derivatives, 1046 WHO Alliance for the Global Elimination See also interpersonal violence on asthma, 688 of Trachoma, 957 World Bank on blindness in childhood, 957 WHO Global Network of Collaborating approach to ensure resources go to where burden of disease estimates of, 279 Centers in Occupational Health, most needed, 97 on cancer, 572 1134, 1140 Child Development projects, 1064, 1068 on cardiovascular disease, 650 WHO Global Strategy for Containment of development assistance for health from, on cataract surgery, 1258b Antimicrobial Resistance, 1043 244, 245, 247 on Chagas disease, 436, 438 WHO-ILO Joint Effort on Occupational on development assistance for health Choosing Interventions That Are Health and Safety in Africa, 1134 needs, 245, 246 Cost-Effective initiative of, 860 WHO Priority Medicines Project, 1335 on drug procurement, 1326 Commission on Macroeconomics and WHO-UNICEF Effective Cold Store on economic growth for 2000­15, 183 Health. See Commission on Management Initiative, 1329 on excreta disposal, 779, 782 Macroeconmics and Health willingness-to-pay valuation, 280 International Development Association on contraceptive use, 1080 women (IDA) grants and credits, 243, on COPD, 684 See also gender differences 244, 254b on cost-effective analyses, 261 abortions. See abortions on kidney and urinary system diseases, on data collection, 1020 African trypanosomiasis and, 456 702, 703, 704 on defluoridation of water, 731 burden of disease and, 196, 197­199, on local production of drugs, 1328 on dengue, 451, 452, 454, 455, 457, 461 199t, 200t Maternal and Child Insurance on development assistance for health, 243 conditions specific to women, 199­201, Program, 251 on diabetes, 592 201t on MDGs financial requirements, 190 on diarrheal diseases, 375, 378 cancer and, 203, 205, 570 on pain control, 1333 on diphtheria, 394 1398 | Index on drug accessibility, 1323, 1333, on palliative care and pain relief, 989­991, years of life lost, 28 1345­1346 990b yellow fever drug approval process and, 153, on polio eradication, 121, 1164, 1171 control of, 1163 1327­1328 prequalification of drugs and vaccines by, vaccine for, 121, 396­397, 1324, 1325 on drug research priorities, 1335 1327­1328 See also vaccine-preventable diseases on epilepsy, 631 Project Focus: Ensuring Immunization Yemen on eradication of infectious diseases, 1174 Safety, 1132 onchocerciasis in, 436, 957 Expanded Program on Immunization. See Regional Office in Africa and polio in, 1170 Expanded Program on integrated disease population young adults. See adolescents and young Immunization (EPI) surveillance, 1178 adults Framework Convention on Tobacco on respiratory diseases of adults, 691 youth violence, 763­764, 765 Control, 573, 882 on respiratory diseases of children, Yugoslavia, malaria in, 421 on gender differences, 196 488­489, 493­494 Global Health Initiative of, 1095 on salt intake, 836 Zambia guidelines for national drug policies, on sexually transmitted infections, 319 African trypanosomiasis in, 458, 460 1324­1325 on skin diseases, 707 dental treatment in, 729 guinea worm control program, 167 smallpox eradication efforts of, 168, health service agency reform in, 93, on hearing loss from excessive noise, 959 1163­1164 1315, 1318 on helminth infections, 470, 471, 474b on sugar intake, 836 hospitals in, 1343 on hemoglobinopathies, 665, 672 on surveillance, 1000, 1007, 1009, 1011, IMCI program in, 1187 on HIV/AIDS, 332, 346, 355, 1333 1042 Kaposi's sarcoma in, 730 on homeopathy, 1286, 1287 on syphilis, 314 malaria in, 421 human reproduction and tropical disease on TB, 36b, 289 primary care in, 1343 research programs, 106 "3 x 5" program of, 1333 quality of health care in, 1344 ILO-WHO Joint Committee on training of epidemiologists, 1002 sexually transmitted infections in, 323 Occupational Health, 1131 on Trichomonas vaginalis, 315 Zanzibar IMCI program. See integrated manage- on unintentional injuries, 737, 741, 742 African trypanosomiasis in, 458, 460 ment of childhood illness (IMCI) on user fees, 1330 Zimbabwe on influenza, 684 on vaccine-preventable diseases, 394, African trypanosomiasis in, 458, 460 on interpersonal violence, 755, 756, 766 401, 1324 cervical cancer screening in, 205, on kidney and urinary system diseases, on vaccines, 1324, 1325t, 1327 576, 577t 697, 702, 703, 704 on violence, 765, 766 community-based programs in, on leishmaniasis, 451, 454, 455 on water pollution, 829 1066, 1067 on leprosy, 437, 444 WHO-ILO Joint Effort on Occupational contracting for health care services in, on lymphatic filariasis, 436, 439 Health and Safety in Africa, 1134 189, 1347­1348 on malaria, 88, 414­415, 1046, 1163 on yellow fever, 397 excreta disposal in, 782 on measles, 395 World Health Organization-International hygiene promotion in, 785 on mental disorders, 613, Union against Tuberculosis and infant mortality in, 1068 619­620, 1343 Lung Disease (WHO­IUATLD) Kaposi's sarcoma in, 730 Model List of Essential Drugs of, Global Project, 1043 malnutrition in, 1068 1324­1325 World Health Report 2000, 1313 oral health and education program on mortality by risk factors, 11b World Health Report 2002: Reducing Risks, in, 726 mother-baby package of interventions Promoting Healthy Life, 121, 132, primary health care nurses in, 1313 of, 514 788, 820, 1148, 1194, 1246 quality of health care in, 1297 multidrug therapy developed by, 441 World Medical Association. See Declaration regulation of health care in, 1296 on musculoskeletal disorders, 963 of Helsinki road traffic injuries in, 749b on neonatal deaths, 531­532, 539 World Report on Violence and Health sexually transmitted infections in, 318 on neurological disorders, 627 (2002), 765, 766 taxes on alcohol in, 218 on newborn care, 535­536 World Summit for Children (1990), traditional medicine in, 1283, 1284 on occupational health risks, 1128, 1129, 558, 559 zinc deficiency 1130, 1140 World Trade Organization (WTO) costs and cost-effectiveness of on onchocerciasis, 437 Agreement on Agriculture, 220 interventions, 561, 565 on oral and craniofacial diseases and Agreement on Technical Barriers to diarrheal disease and, 378 disorders, 723­724 Trade, 220 human capital formation and, 562 oral health and education program of, TRIPS and, 150­152 IMCI and, 1180 726, 731­732, 734 worms. See guinea worm eradication; mortality and, 554 on oral rehydration therapy, 378 helminth infections research on, 564 Index | 1399 Credits Cover and Interior Design Unlimited/Getty Images; chapter 37: CDC/Joe Miller/Reed and Naylor Design, Washington, D.C. Crnrick Pharmaceuticals; chapter 38: Richard Lord, www. rlordphoto.com; chapter 39: Richard Lord, www.rlordphoto. Part One Photographs com; chapter 40: Gabe Palmer/Corbis; chapter 41: Richard Chapter 1: Curt Carnemark/The World Bank; chapter 2: Lord, www.rlordphoto.com; chapter 42: Richard Lord, www. Sebastian Szyd/The World Bank; chapter 3: Richard Lord, rlordphoto.com; chapter 43: Richard Lord, www.rlordphoto. www.rlordphoto.com; chapter 4: Richard Lord, www.rlordpho- com; chapter 44: Richard Lord, www.rlordphoto.com; chap- to.com; chapter 5: Micheal Simpson/Getty Images; chapter 6: ter 45: Richard Lord, www.rlordphoto.com; chapter 46: Richard Lord, www.rlordphoto.com; chapter 7: Royalty- Richard Lord, www.rlordphoto.com; chapter 47: Royalty-Free/ Free/Corbis; chapter 8: Richard Lord, www.rlordphoto.com; Corbis; chapter 48: Royalty-Free/Corbis; chapter 49: Richard chapter 9: Richard Lord, www.rlordphoto.com; chapter 10: Lord, www.rlordphoto.com; chapter 50: Ray Witlin/The World Shehzad Noorani/The World Bank; chapter 11: Royalty-Free/ Bank; chapter 51: Richard Lord, www.rlordphoto.com; chap- Corbis; chapter 12: Royalty-Free/Corbis; chapter 13: Richard ter 52: Royalty-Free/Corbis Lord, www.rlordphoto.com; chapter 14: Royalty-Free/Corbis; chapter 15: Naylor Design Inc. Part Three Photographs Chapter 53: Richard Lord, www.rlordphoto.com; chapter 54: Part Two Photographs Richard Lord, www.rlordphoto.com; chapter 55: Shehzad Chapter 16: CDC; chapter 17: Richard Lord, www.rlordphoto. Noorani/The World Bank; chapter 56: Richard Lord, www. com; chapter 18: Pep Bonet/PANOS; chapter 19: Richard Lord, rlordphoto.com; chapter 57: Richard Lord, www.rlordphoto. www.rlordphoto.com; chapter 20: Richard Lord, www. com; chapter 58: Richard Lord, www.rlordphoto.com; chap- rlordphoto.com; chapter 21: CDC/PHIL/Corbis; chapter 22: ter 59: Richard Lord, www.rlordphoto.com; chapter 60: Pan American Health Organization; chapter 23: Dr. Dennis Richard Lord, www.rlordphoto.com; chapter 61: Richard Lord, Kunkel/Getty Images; chapter 24: Richard Lord, www. www.rlordphoto.com; chapter 62: Ray Witlin/The World Bank; rlordphoto.com; chapter 25: Richard Lord, www.rlordphoto. chapter 63: Gideon Mendel/Corbis; chapter 64: Richard Lord, com; chapter 26: Richard Lord, www.rlordphoto.com; chap- www.rlordphoto.com; chapter 65: Richard Lord, www. ter 27: Richard Lord, www.rlordphoto.com; chapter 28: rlordphoto.com; chapter 66: Richard Lord, www.rlordphoto. Richard Lord, www.rlordphoto.com; chapter 29: David com; chapter 67: Richard Lord, www.rlordphoto.com; chap- Becker/Getty Images; chapter 30: Royalty-Free/Corbis; chap- ter 68: Curt Carnemark/The World Bank; chapter 69: Royalty- ter 31: Royalty-Free/Corbis; chapter 32: 3D4Medical.com/ Free/Corbis; chapter 70: Richard Lord, www.rlordphoto.com; Getty Images; chapter 33: Richard Lord, www.rlordphoto.com; chapter 71: Richard Lord, www.rlordphoto.com; chapter 72: chapter 34: Spike Walker/Getty Images; chapter 35: Richard Richard Lord, www.rlordphoto.com; chapter 73: Richard Lord, Lord, www.rlordphoto.com; chapter 36: SIU/Visuals www.rlordphoto.com. 1401 ISBN 0-8213-6179-1